Contact 2016

Anghel Chiru · Nicolae Ispas Editors
CONAT 2016
International Congress
of Automotive and
Transport Engineering
CONAT 2016 International Congress of Automotive
and Transport Engineering
Anghel Chiru Nicolae Ispas
•
Editors
CONAT 2016 International

Congress of Automotive
and Transport Engineering
123
Editors
Anghel Chiru Nicolae Ispas
Department of Automotive Department of Automotive
and Transportation Engineering and Transportation Engineering
Transilvania University of Brasov Transilvania University of Brasov
Brasov Brasov
Romania Romania
ISBN 978-3-319-45446-7 ISBN 978-3-319-45447-4 (eBook)

DOI 10.1007/978-3-319-45447-4
Library of Congress Control Number: 2016953012
© Springer International Publishing Switzerland 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG Switzerland
Preface
In the field of creation, manufacturing and testing of modern automobiles are mani-
fested demands and requirements, sometimes contradictory, related to: security, aero-
dynamics, design, energy efficiency, comfort, noise, vibrations, entertainment, respect
for the environment, reliability, preservation of resources, technological optimization,
costs and recycling. The automotive industry is characterized, on the other hand, by a
strong competition in inventiveness and high technology products. Rapid and large-
scale evolution is dependent on the capacity of the research and design departments
of the manufacturing companies to assimilate and work with innovative engineering
methods. Numeric simulation represents one of the effective environment that support
the activities related to:
– New materials engineering
– Product–process interdependency
– Development of future automotive concepts: hybrids, autonomous and
interconnected
– Virtual reality: simulation, optimization of process management
– Topological and technological optimization
– Cost efficiency of research and development.
Introduction of RDE (Real Driving Emissions) measure starting from September
2017 for Euro 6c (new models) and Euro 7 (starting with January 2020) is a new
challenge for large automotive manufacturers. The advancement in internal combustion
engines and electronic systems used in control and analysis constitutes guaranties for
future solutions for treating and filtering the noxious and particulate emissions.
To respond to the major challenges faced by the automotive world, the authors
of the papers presented in the technical sections of CONAT 2016 International Con-
gress offer their vision and solutions in the following areas:
– Advanced Engineering Methods
– Innovative Solution for Motor Vehicles
– Heavy and Special Vehicles
– Automobile and Environment
– New Materials, Manufacturing Technologies and Logistics
– Advanced Transport Systems and Road Traffic
– Accident Research and Analysis.
The presented papers were selected from 222 proposals submitted. All papers were
developed by 218 specialists from 14 countries. Assessment and systematization of
presentations were performed by an international scientific committee. They were
analyzed for the novelty and scientific content, as well as for the layout of each paper,
following the requirements imposed by FISITA, SIAR, SAE International and Springer
Publishing.
VI Preface
The initiation, implementation and completion of activities for preparing, collecting

and finalizing of the volume containing the Proceedings of CONAT 2016 Congress
were possible, thanks to the involvement and professionalism of Dr. Dinu Covaciu, the
Coordinator of the Technical Programme of Congress. We thank him for all his efforts.
We also express our gratitude to our colleagues from Transilvania University, and to
the members of the scientific committee and reviewers, to organizers and partners from
the companies Schaeffler, Renault Romania, Atlassib, Horiba, AVL, Tedata and
Brahms, for their support and involvement in all activities of organization of the XIIth
edition of the International Congress of Automotive and Transport Engineering
CONAT 2016.
Anghel Chiru
President of the CONAT 2016 Scientific Committee
Nicolae Ispas
President of the CONAT 2016 Organizing Committee
Scientific Committee
President of the Scientific Committee

Anghel Chiru Transilvania University of Brasov, Romania
Coordinators of the Congress Sections

Ion Preda Advanced Engineering Methods
Transilvania University of Brasov, Romania
Gheorghe Al. Radu Innovative Solution for Motor Vehicles
Mircea Năstăsoiu Heavy and Special Vehicles
Corneliu Cofaru Automobile and Environment
Gheorghe Mogan New Materials, Technologies and Logistics
Daniela Florea Advanced Transport Systems and Road Traffic
Nicolae Ispas Accident Research and Analysis
Scientific Committee Members

Cristian Andreescu University Politehnica Bucharest, Romania
Michael Butsch University of Applied Sciences Konstanz, Germany
Kwang Min Chun Yonsei University, Korea
Giovanni Cipolla Politecnico di Torino, Italy
Michael Elicker Schaeffler Technologies AG & Co. KG, Germany
Nicolae Filip Technical University of Cluj-Napoca, Romania
Eduard Schaeffler Technologies AG & Co. KG, Germany
Golovatai-Schmidt
Paul Guempel University of Applied Science Konstanz, Germany
Peter Kuchar University of Applied Science Konstanz, Germany
Jan Leuridan Siemens Industry Software, Belgium
Abdellatif Miraoui Cadi Ayyad University, Morocco
Eugen Mihai Negrus University Politehnica Bucharest, Romania
Hariton Poparad Transilvania University of Brasov, Romania
Serban Raicu University Politehnica Bucharest, Romania
Adrian Rienaecker University of Kassel, Germany
Corina Sandu Virginia Tech, USA
VIII Scientific Committee
Adolfo Senatore Federico II University of Naples, Italy

Roger Sierens Ghent University, Belgium
Cornel Stan West Saxon University of Zwickau, Germany
Ion Tabacu University of Pitesti, Romania
Jan Unarski Institute of Forensic Research Krakow, Poland
Nicolae Vasiliu University Politehnica Bucharest, Romania
Ion Visa Transilvania University of Brasov, Romania
Beside the members of the Scientific Committee listed here, the papers were reviewed
by a total number of 73 reviewers from academic institution, research centers and
industry.
Definitions, Acronyms, Abbreviations

Al Aluminum
CAD Computer-aided design
CAE Computer-aided analysis
CG Center of gravity
CGAP Gap element connection
CI Cast iron
CM Cross member
CMVR Central Motor Vehicle Rules
CVD Complete vehicle development
FEA Finite element analysis
Fig. Figure
FOS Factor of safety
FUPD Front under run protection device
Govt. Government
GVW Gross vehicle weight
HD Heavy duty
LH Left hand
NINC Number of increments
NLPARM Nonlinear parameter
PDD Product design and development
PGAP Gap element property specification
RH Right hand
RLDA Road load data acquisition
RPM Revolution per minute
RUPD Rear under run protection device
SPC Single point constraint
T Ton
TM Transmission housing
UPD Under run protection device
VECV Volvo Eicher Commercial Vehicle
Contents
Advanced Engineering Methods
Novel Simulation Methodology for the Estimation of the Flow Mark Risk
Associated with the Injection of Mass Colored Plastic Materials . . . . . . . . . . 3
Alexandru Ioan Ciobanas
Numerical and Experimental Investigations of a Twin Sheet Thermoplastic

Structure with Rectangular Frusta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Stefan Tabacu, Claudiu Diaconescu, and Alexandru Oltean
Expanded Polypropylene Finite Elements Analysis in Transport

Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Alexander Schwarz, Catalin-Andrei Runcianu, and Andreas Gebhardt
Functional Description and Design Aspects of Hybrid Transmissions. . . . . . . 36

Martin Vornehm
Comparative Study on the Performances of Aerodynamic Devices Used

in Decreasing of the Automobiles Lift Force . . . . . . . . . . . . . . . . . . . . . . . 48
Angel Huminic, Gabriela Huminic, Marius Carauleanu,
and Constantin Ciolofan
Analysis of the Influence of the Drive Force Distribution Between Axles

on an Automobile Stability in Its Curvilinear Motion. . . . . . . . . . . . . . . . . . 55
Andrzej Reński
About the Steering Kinematics of the Road Trains . . . . . . . . . . . . . . . . . . . 64

Ion Preda and Dinu Covaciu
Evaluation by Experimental Methods of the Parameters that Influence

the Behavior of Various Passenger Cars Classes in the Braking Process. . . . . 75
Adrian Todoruţ, Nicolae Cordoș, István Barabás, and Sorin Miheț
Evaluation of the Behavior in Cornering for Different Classes of Passenger

Cars by Numerical Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Nicolae Cordoș, Adrian Todoruţ, István Barabás,
and Radu Daniel Mureșan
Structural Dynamic Applications Using Principal Component Analysis

Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Mircea Niculescu, Cristi Irimia, Ioan Calin Rosca, Mihail Grovu,
and Maria Violeta Guiman
X Contents
Weight Optimization of Existing Transmission Housing for Heavy Duty

Commercial Vehicles Using Correlated Simulation and Experimental Tools . . . 100
Kandreegula Suresh Kumar and Dheeraj Kumar Singh
New Approaches in Designing a Race Car Chassis for an Engineering

Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Doru Ionut Furnea and Silviu Butnariu
Presentation of Tests with a Dynamometer Related to the Determination

of Fuel Consumption in the NEDC and WLTC Driving Cycles and as Well
Related to the Dynamic Behavior of Vehicles . . . . . . . . . . . . . . . . . . . . . . . 120
Michael Butsch and Valerio Rescigno
Validation Procedure for Worldwide Harmonized Light Vehicles Test

Cycle via Hardware in the Loop - Real Time Testing . . . . . . . . . . . . . . . . . 128
Bogdan Ovidiu Varga, Nicolae Burnete, and Calin Iclodean
Experimental Investigation of a Vehicle Behavior Using Different Complex

Data Acquisition Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Ilie Dumitru, Racila Laurentiu, Tutunea Dragos, Matei Lucian,
Dima Alexandru, and Oprica Alexandru
Validation of Kinematic Simulation of Sprocket Contacts of Chain Links

by Experiments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Karim Shalaby, Florin Corciova, and Simona Lache
Belt Losses Evaluation for a Push-Belt CVT . . . . . . . . . . . . . . . . . . . . . . . 152

Valerian Croitorescu, Codrin Cuciurean, and Mircea Oprean
Kinematic and Dynamic Study of a Mechanism for a Vehicle Front

and Rear Stabilizer Bars. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Nicolae Dumitru, Cristian Copilusi, Ionut Geonea,
and Alexandru Margine
Methods for Modeling an Elastic System with Permanent Contour Coupling

Deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Hariton Poparad
Reinvestigation of the Electromagnetic Valve Train (EMVT) Technology

via Multidomain Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Christian Trebehs and Doru Groza
Study on Mixture Formation at an Experimental Spark Ignition Engine . . . . . 187

Dan Mihai Dogariu, Anghel Chiru, Stelian Țârulescu, Marius Lazăr,
and Vlad Ștefan Stancu
Contents XI
Aspects Concerning Modeling of Combustion for Compression Ignition

Engine with Injection Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
Dan Mihai Dogariu, Anghel Chiru, Stelian Țârulescu, and Marius Lazăr
Comparative Study on the Performances of the One-Way Hydraulic Valves . . . 203

Marius Lazăr, Anghel Chiru, and Dan Mihai Dogariu
Estimation of Boiling Points of Brake Fluids . . . . . . . . . . . . . . . . . . . . . . . 209

István Barabás, Adrian Todoruț, and Nicolae Cordoș
Modelling and Optimization of Constructive Form for GO-NOGO Exterior

and Interior Diameter Gauges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Iulian Alexandru Orzan and Constantin Buzatu
Systems Engineering Requires Digital Twins of Machine Elements . . . . . . . . 227

Willi Theodor Gruender
Collaborative Design and Use of Interactive Simulations: Boost

the Learning Environment in Road Vehicle Dynamics Curriculum . . . . . . . . 234
Moein Mehrtash and Dan Centea
Experimental Researches on the Magneto-Rheological Dampers Response

to the Control Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Alexandru Dobre, Nicolae Vasiliu, and Cristian Nicolae Andreescu
Innovative Solution for Motor Vehicles
Contributions in Experimental Research Concerning Diesel Fuel Supply

and Lubrication in the Case of Comparative Study Between Euro V and IV
Common Rail Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Doru Baldean, Mihai Aurel Crișan, and Adela Ioana Borzan
Experimental Research Regarding the Possibility of Biofuel Fumigation

Supply Method on a Single Cylinder Compression Ignited Engine . . . . . . . . 264
Doru Baldean, Nicolae Vlad Burnete, Dan Moldovanu, Ferenc Gaspar,
Adela Ioana Borzan, and Mihai Aurel Crișan
Study on a Stratified Charge Spark Ignition Automotive Engine . . . . . . . . . . 274

Edward Rakosi, Sorinel Talif, Gheorghe Manolache,
and Adrian Sachelarie
Researches on Cooling Air Flow Control Devices Using on Cars

with Internal Combustion Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
Spiridon-Cristian-Dan Dăscălescu, Marius Receanu,
Laurenţiu Dimitriu, and Liliana Vornicu-Albu
Modeling the Energy Evaluation for an Electric Machine . . . . . . . . . . . . . . . 295

Valerian Croitorescu and Alexandru Manta
XII Contents
Balancing of a Single Stage Reciprocating Compressor with Elastic

Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Doru Groza, Ioan Călin Roșca, and Gheorghe Alexandru Radu
The Calculation Algorithm for the Determination of the Temperature

at the End of the Intake Stroke for GDI Engines . . . . . . . . . . . . . . . . . . . . . 311
Tudor Mitran, Vasile Blaga, Sorin Moca, George Dragomir,
and Dinu Fodor
Diagnosing the Operation of a Locomotive Diesel Engine Based on the

Analysis of Used Oil in the Period Between Two Technical Revisions . . . . . 319
Cătălin V. Zaharia, Rodica Niculescu, Victor Iorga, Cătălin Ducu,
Adrian Clenci, and Bogdan Aron
Heavy and Special Vehicles
Simultaneous Influences of Tyre Pressure and Steering Geometry upon the

Theoretical Speed Ratio in the Running Gear System of Four-Wheel Drive
Tractors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Mircea Nastasoiu and Nicolae Ispas
Analysis of the Comfort of the Seats Used in Light Military Vehicles . . . . . . 339
Marian Mitroi and Anghel Chiru
Aspects Regarding the Kinematic Optimization of a Tracked Military

Vehicle’s Transmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
Octavian Alexa, Radu Vilău, Ovidiu-Constantin Ilie, Daniela Voicu,
and Marin Marinescu
Automobile and Environment
Research on Gasoline Homogenous Charge Compression Ignition (HCCI)

Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
Corneliu Cofaru
Study of the Diesel Engine Cycle Variability at LPG Fuelling . . . . . . . . . . . 371

Liviu Nemoianu, Alexandru Cernat, Constantin Pana,
Niculae Negurescu, and Cristian Nutu
Direct Injection of Diesel and Ethanol in a Diesel Engine - A Numerical

Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
Nicolae Vlad Burnete, Călin Iclodean, Dan Moldovanu,
and Nicolae Filip
Contents XIII
The Effects of a Diesel-Ethanol Blend Without Additives on Diesel Engine

Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Nicolae Vlad Burnete, Călin Iclodean, George Popescu,
and Nicolae Filip
The Influence of Biodiesel on Engine Combustion and CO2 Emissions . . . . . 395

Bogdan Cornel Benea
Modeling the Performances of a Vehicle Provided with a Spark Ignition

Engine Using a Double Energy Source, Gasoline and Liquefied
Petroleum Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
Valentin Nişulescu, Florian Ivan, Gheorghe Fratila,
and Gheorghe Banca
Modeling the Performances of a Vehicle Provided with a Hybrid Electric

Diesel Propulsion System (HEVD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
Gheorghe Bancă, Florian Ivan, Gheorghe Frățilă,
and Valentin Nișulescu
Aspects of Experimental Research on Hydrogen Fuelled Automotive Diesel

Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
Ionel Mirica, Alexandru Cernat, Constantin Pana, Niculae Negurescu,
and Cristian Nutu
Researches on Combustion Quality for a Single Cylinder Diesel Engine . . . . 435

Stelian Tarulescu and Radu Tarulescu
Urban Transportation Solutions for the CO2 Emissions Reduction

Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
Stelian Tarulescu and Radu Tarulescu
Mission CO2 Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451

Roland Seebacher
Experimental Research Concerning the Influence of the Electrochemical

Reactions Induced by Products of an Electrolytic Cell Over the Emission
Pollutants in ICE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
Marius Nicolae Moldoveanu, Cheng Lin Tang, and Cornel Aramă
Energy-Based Approach to Model a Hybrid Electric Vehicle and Design

Its Powertrain Controller and Energy Management Strategy . . . . . . . . . . . . . 471
Nicole Salloum, Serge Francis, and Charbel Mansour
Dynamic Control of an Electric Vehicle with Traction Induction Motor . . . . . 482

Marius Constantin Georgescu
XIV Contents
A Propulsion System for Means of Transport with Non-autonomous

Electrical Traction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
Tudor Mitran, Nicolae Chioreanu, Catalin Chioreanu, Alexandru Rus,
and Horia Beles
Validation of a Human-and-Hardware-in-the-Loop Control Algorithm

Using Real Time Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
Ionut Alexandru Stoica, Marius Valentin Bataus,
and Ioan Mircea Oprean
Electronic Control Systems of E-Smart Vehicle. . . . . . . . . . . . . . . . . . . . . . 509

Radu Tarulescu and Stelian Tarulescu
Battery Management System of E-Smart Vehicle . . . . . . . . . . . . . . . . . . . . 517

Radu Tarulescu and Stelian Tarulescu
New Materials, Manufacturing Technologies and Logistics
Worldwide Diversity Management of “Global Access” Renault/Dacia’s

Vehicle Range by Using a Modern PLM System . . . . . . . . . . . . . . . . . . . . 527
Bogdan Bădescu and Pascal Candau
Potential Capabilities of Shape Memory Driven Automotive Devices. . . . . . . 536

Joachim Strittmatter, Paul Gümpel, and Oskar Lattner
Applicability of Composite Forming Technologies for Automotive

Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
Lucian Rad and Anghel Chiru
Constructive Optimization of Composite Materials Automotive

Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 556
Anghel Chiru and Lucian Rad
The Evaluation of Rheological Properties of Composites Reinforced

with Hemp, Subjected to Photo and Thermal Degradation . . . . . . . . . . . . . . 565
Mariana Domnica Stanciu, Ioan Curtu, Mihai Groza, and Adriana Savin
Optimization of the Plastic Injection Molding Parameters by Using the

Taguchi Method to Prevent Jammed Components in the Assembly Process
of Mechatronic Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
Ion Cristian Braga and Anisor Nedelcu
Key Characteristics of the World Class Manufacturing Concept

in the Production of Chassis for Buses Industry . . . . . . . . . . . . . . . . . . . . . 585
Luminiţa Ştirbu, Cosmin Știrbu, Nadia Belu, and Viorel Nicolae
Contents XV
Advanced Automotive Assembly Line Trends as Tools in Optimizing

Production Line Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 592
Hariton Poparad
Cleaning Methods for Flux Pollution Measurement in Automotive Coolant

Loop Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599
Hervé Bourgeais and Claire Demarcq
Design and Development of Chassis Along with Novel Under Run

Protection Devices for Commercial Vehicles with the Use of an Innovative
Approach Resulting in Lighter Weight of the Vehicle . . . . . . . . . . . . . . . . . 610
Ashesh Anil Shah and Himanshu Arora
The Quality Management Principles and Their Incidence Within ISO 9001:
2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 620
Laurentiu Aurel Mihail
Product Quality Control Optimization for Selection of Measurement

and Control Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 629
Constantin Buzatu and Iulian Alexandru Orzan
Improvement of the 8D Analysis Through a System Based on the “Internet

of Things” Concept Applied in Automotive Industry . . . . . . . . . . . . . . . . . . 635
Viorel Nicolae, Laurenţiu Mihai Ionescu, Nadia Belu,
and Știrbu Luminița Elena
A Case Study Regarding the Implementation of Six Sigma in an Assembly

Process for the Automotive Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
Dorin-Ion Dumitrascu, Adela-Eliza Dumitrascu, and Anghel Chiru
Advanced Transport Systems and Road Traffic
Analysis of the Influence of One-Way Streets on the Urban Road Networks

Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 653
Daniela Florea, Dinu Covaciu, and Janos Timar
Cooperative Smart Intersection as an Enabler of Advanced Traffic

Management Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
Laura Coconea, Gonzalo Alcaraz, and Renato Augusto Lira de Andrade
Bus Routing Safety for the Transportation of Children to School . . . . . . . . . 676

Eleni Chalkia, Jose Maria Salanova Grau, Evangelos Bekiaris,
Georgia Ayfandopoulou, Chiara Ferarini, and Evangelos Mitsakis
Improving the Road Traffic Regulation in the Area of Roundabout

Intersections Function of the Traffic Streams Size . . . . . . . . . . . . . . . . . . . . 684
Elena Neagu, Andrei Alexandru Boroiu, Ionel Vieru,
and Alexandru Boroiu
XVI Contents
Traffic Optimization in Urban Area - Roundabout Versus Lights Case

Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694
Nicolae Filip and Claudiu Golgot
Mathematical Algorithm for Calculating the Total Traffic Lights Cycle

in Junctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702
Lucian Matei, Ilie Dumitru, and Laurentiu Racila
Study on the Influence of Intersections with Forest Roads upon the Traffic
Flows on Highways. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 710
Constantin Alexandru Bratu and Dinu Covaciu
Smart Solar Electric Tricycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 721

Dahbi Mohamed, Said Doubabi, and Ahmed Rachid
Programming an Autonomous Mini-vehicle in Narrow Environments . . . . . . 729

Emanuel Giurgiu, Teodora Girbacia, and Gheorghe Leonte Mogan
About Uncontrollable Reactions of the Driver Due to Skin Mechanical

Stimuli. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 737
Alexandru Bogdan and Silviu Butnariu
Development of an Advanced Driver Assistance System Using RGB-D

Camera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746
Alin Pantea, Florin Girbacia, and Teodora Girbacia
A Method for Transforming Electric Vehicles to Become Autonomous

Vehicles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 752
Mihai Kocsis, Andreas Schultz, Raoul Zöllner,
and Gheorghe Leonte Mogan
Accident Research and Analysis
Cars Crashes with Cars of Same or Different Generation – Occupant’s

Loads, Timings and Acceleration’s Effects . . . . . . . . . . . . . . . . . . . . . . . . . 765
Nicolae Ispas and Mircea Nastasoiu
Crash Tests Data Acquisition and Processing . . . . . . . . . . . . . . . . . . . . . . . 782

Dinu Covaciu and Dragoş Sorin Dima
Research Regarding Occupant’s Movement in the Case of Frontal Collision

Using High-Speed Video Recording . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 790
Alexandru Ionut Radu, Corneliu Cofaru, Bogdan Tolea,
Daniel Dragos Trusca, and Horia Beles
Contents XVII
Research Regarding the Influence of Vehicle’s Safety Restraint Systems

on Its Occupants in Case of Rear-End Collision . . . . . . . . . . . . . . . . . . . . . 798
Alexandru Ionut Radu, Corneliu Cofaru, Bogdan Tolea,
Daniel Dragos Trusca, and George Toganel
The Assessment of the Head Injury of a Pedestrian in Comparison

with a Cyclist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
Bogdan Tolea, Stelian Tarulescu, Daniel Dragos Trusca,
George Toganel, and Alexandru Ionut Radu
Determination of Kinematic and Dynamic Behavior in the Driver’s Skull

upon the Impact with the Steering Wheel . . . . . . . . . . . . . . . . . . . . . . . . . . 812
Oana Victoria Oţăt, Nicolae Dumitru, Victor Oţăt, and Ilie Dumitru
Pedestrian Dynamics During the Contact Phase with the Vehicle . . . . . . . . . 820
Adrian Șoica
Example of a High-Speed, Side-Impact, Car Crash Reconstruction

Using a Planar Multibody Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 829
Petru A. Simionescu and Ion Preda
Is the Virtual Homologation for Pedestrian Protection Viable? . . . . . . . . . . . 836

Silvia Santos, Carolina Álvarez-Caldas, and José L. San Román
Comparative Analysis of Kinematics and Dynamics Parameters

that Characterize Vehicle-Pedestrian Collision. . . . . . . . . . . . . . . . . . . . . . . 844
Alexandru Petrescu and Liviu Petrescu
The Importance of Consumed Energy for Deformation

in Study of Collision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 856
Mihai Clinciu, Anghel Chiru, and Stefanita Ciunel
Aspects Regarding the Reconstruction of Traffic Events. . . . . . . . . . . . . . . . 865

Mihai Clinciu, Anghel Chiru, and Stefanita Ciunel
Vehicle Driver Drowsiness Monitoring and Warning System . . . . . . . . . . . . 873

Tiberiu Vesselenyi, Alexandru Rus, Tudor Mitran, Bogdan Tataru,
and Ovidiu Moldovan
Research Regarding Pedestrian Visibility During Night-Time

Using Photo Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 881
Bogdan Tolea, Daniel Dragos Trusca, Csaba Antonya,
Alexandru Ionut Radu, and Dragos Sorin DIMA
Estimating the Costs Caused by Road Traffic Accidents in Romania . . . . . . . 889

Alin Drosu and Corneliu Cofaru
Advanced Engineering Methods
Novel Simulation Methodology
for the Estimation of the Flow Mark Risk
Associated with the Injection of Mass Colored
Plastic Materials
Alexandru Ioan Ciobanas(&)
Renault Technologie Roumanie, Bucharest, Romania

alexandru-ioan.ciobanas@renault.com
Abstract. Due to their desired metallic effect, the so called “special effect”
plastic materials are increasingly used in the automotive industry. Aluminum
flakes fillers are commonly used in the automotive industry, especially for mid
and large exterior trim plastic parts. However, an important drawback of these
materials is the high risk of generating undesirable flow marks on the visible
side of the part. In this context, this study focuses on developing a novel
simulation methodology in order to predict the flow marks produced by the
differential orientation of the Al flakes. In a first step, a physical reasoning that
explains the link between the presence of technical features on the part skin (e.g.
clips, ribs) and the flow perturbations at the source of Al flake orientations is
proposed. This reveals the importance of the 3D fountain flow at the flow front
for the transport of oriented particles at the part surface. In a second step, a new
simulation methodology was developed and implemented in Moldflow injection
molding software. The novelty of the simulation method consists in adapting the
fiber orientation model to a simple midplane mesh approach, with the aim of
predicting the flow mark zones. Even if the 3D fountain flow is not modeled for
a midplane approach, one can still predict the flow mark generation zones by
analyzing the flow perturbation generated at the core zone. This is one important
feature of the present simulation methodology and is based on the physical
reasoning proposed in the first step. Finally this new simulation methodology
was applied to the injection of an exterior trim part developed by Renault and a
correlation between the simulation results and the injection trails was under-
taken. The risk zones revealed in the simulations correlates well with the flow
mark zones obtained on physical parts, validating thus the present approach.
Keywords: Mass colored plastic material Aluminum flakes filler Flow

marks Simulation Moldflow software
1 Introduction
During the last decades, the plastic producers and fillers developers were increasingly
focusing on developing “special effects” metallic pigment like aluminum pigments,
gold-bronze pigments or pearlescent pigments [1, 2]. Metallic pigments are increas-
ingly used in a variety of applications ranging from mass colored injected parts to

A. Chiru and N. Ispas (eds.), CONAT 2016 International Congress of Automotive
and Transport Engineering, DOI 10.1007/978-3-319-45447-4_1
4 A.I. Ciobanas
special coatings paintings. The focus here will be on analyzing the effect of using
aluminum pigments for the production of injected plastic parts. Indeed, due to their
desired metallic effect, aluminum flakes fillers are the most commonly used in the
automotive industry, especially for mid and large exterior trim plastic parts.
The advantage of using this type of fillers is twofold. The metallic aspect plastic
parts are colored in mass, hence less prone to scratching defects, while, on the other
hand, there is an important economic advantage because the expensive painting process
which was initially used for obtaining the desired surface effect is not needed anymore.
However, important drawbacks of using metallic fillers can be identified. On one
hand, any weld lines will be highly marked on the parts surface. On the other hand, and
most importantly, the risk of generating flow marks (e.g. darker zones or darker
plumes, Fig. 1) on the part surface is high, a consequence of the flow front perturba-
tions during the filling of the cavity.
Fig. 1. (a) The sudden change in the skin slope determines a local flow perturbation and a flow
mark appear in the wake of this zone; (b) The tiger stripes appearing when injecting with a
different mass colored material clearly reveal the flow perturbation created by the sudden change
in the skin shape
While the weld lines can be quite easily predicted by analyzing the part topology
(or by using a simple fill simulation analysis) the prediction of flow marks which are
produced by local flow perturbation during cavity filling is much more difficult. Often,
change in the skin slopes or even small technical features on the back of the part can
produce important flow marks on the part skin.
In this context, the developing of a simulation methodology that could predict the
flow marks risk zones would be extremely useful during the product design process. In
the following we will firstly propose a physical mechanism explaining the main cause
of flow marks. In a second step a novel simulation methodology is developed in order
to predict the flow marks produced by the differential orientation of the Aluminum
flakes. The simulation methodology was implemented in Moldflow injection molding
software where the fiber orientation feature was used for the prediction of Al flakes
orientation.
Novel Simulation Methodology for the Estimation of the Flow 5
2 Aluminum Pigments and the Metallic Effect
The Al pigments used in the automotive industry are mainly platelet-shaped particles
(flakes) with a diameter ranging from 2–200 µm and a thickness of 0.05 – 2 µm [3]
(Fig. 2). The main manufacturing process is the milling of high purity aluminum
powder in the presence of a solvent (usually white spirit).
Fig. 2. SEM images of Aloxal [4] showing the plated-shaped Al particles (flakes)
The metallic effect is achieved by a combination of reflection and scattering of light

[1]. On one hand if the flakes are aligned with the part surface (Fig. 3) the reflection at
the surface of the flakes will be important and the metallic effect amplified. On the other
hand, if the light is scattered at the edges of the flakes a local glittering effect is obtained.
The particle size significantly influences the brilliance or the darkness of the part: a
satin finish for small diameters (5–25 lm), a brilliant effect for average diameters (15–
70 lm), while for large sized particles (30–200 lm) a sparkle finish is obtained. Notice
as well that the particle size distribution is also important: the tighter the particle
diameter distribution, the more brilliant the part will be.
Due to the nature of the non-Newtonian fluid, the velocity profile close to the cavity
walls (Fig. 5a) is characterized by high gradients. Hence high shear forces are gener-
ated at the cavity walls [5]. Furthermore, in a flow exhibiting high shear forces, the
flakes tend to align themselves in order to reduce their resistance to flow [6]. In this
context, in the proximity of the part surface, the Al flakes will have the natural ten-
dency to orient parallel to the cavity/part surface. However, flow perturbation may arise
during cavity filling, leading to zones where the flakes will be oriented more or less
Fig. 3. The metallic effect is obtained if the flakes found in the exterior layers are oriented
parallel to the part surface
6 A.I. Ciobanas
perpendicular to the part surface. In the latter case, the reflection of the light in these
zones is modified and the surface aspect will be significantly less bright compared with
the zones where Al flakes are parallel to the part surface. One usually calls this flow
marks (Fig. 1). In the next section a physical mechanism is proposed, explaining the
strong relation between the flow perturbations appearing during the filling of the cavity
and the development of flow marks.
3 Local Flow Perturbations at the Source of Flow Marks
Let us consider the filling of a constant thickness part (Fig. 4) and that the filling is
done a constant flow front velocity (the injection profile is thus adapted to the evolution
in time of the melt front surface). If one looks at the velocity profile in a thickness
section, two flow zones can be identified: a fountain flow close to the melt front and a
developed flow behind the melt front. While for the developed flow a quasi-2D fluid
flow can be assumed (velocities along z direction are negligible, vz 0), for the
fountain zone, the fluid flow is fully 3D with non-negligible vz velocities.
Fig. 4. During the filling of a constant thickness cavity two flow zones can be identified: a
developed flow zone (*2D flow) and the “Fountain flow” zone (3D flow)
3.1 Developed Flow Zone – Wall Proximity

Let us focus on the velocity field close to the cavity walls. Because the filling of the
cavity is usually very fast (to prevent plastic solidification during filling), one could
assimilate the plastic melt as a power law non-Newtonian fluid for which the viscosity,
l; is a power function of the shear rate, c: l / cn1 . For most thermoplastics, the
exponent is n 0:2 0:3: For such a fluid, the steady state velocity profile along the
part thickness has a shape as shown in Fig. 5a with high velocity gradients close to the
walls and consequently high shear rate rxy in the zone close to the cavity walls [4],
Fig. 5b.
In this context, one clearly expects that the highly sheared fluid will strongly
interact with the Al flakes found at this zone: the strong shear forces will act on the two
faces of the flakes creating a rotation momentum that will align the flakes parallel with
the xy plane, thus parallel to the part surface (Fig. 5b). Hence, in the absence of other
Fig. 5. (a) The velocity profile within the developed zone, characterized by high velocity
gradients at the walls of the mold; (b) An xy plane section close to the cavity wall reveals the
high rxy shear rate and consequently the natural tendency of the flakes to orient parallel to the
part surface
flow perturbations, one expects that, at the developed flow zone and close to the cavity
walls, the Al flakes will align parallel to the part surface, amplifying the metallic finish
of the part surface.
3.2 Developed Flow Zone – Core Zone

At the core zone, due to the symmetry of flow with respect to xy plane, the shear
stresses rxy are negligible. Hence, at the core of the flow, the Al flakes have no
tendency to orient in the xy plane/parallel to the surface.
If now one focus on the velocity field in the xy plane, at a section taken at the core
zone, one can see that, if no flow perturbation (e.g. ribs, high curvature of the skin) are
present along the flow path, the velocity field at the core zone will be uniform with
small gradients along y direction (Fig. 6a, the flow field upwind of the rib). On the
other hand if such a perturbation element exists along the flow path, the velocity field at
the core zone will develop significant gradients along y direction. Sectioning the flow at
Fig. 6. (a) In the wake of a perturbation element (rib) the velocity field at the core of the plastic
part is modified significantly; (b) In the wake of the rib important shear rxz forces are developed,
orienting the flakes perpendicular to the part surface.
8 A.I. Ciobanas
this perturbed zone with an xz plane will reveal important rxz shear forces in the wake
of the perturbation element (Fig. 6b). Similar with the previous case, the fluid will
interact with the Al flakes found at the core zone, flow characterized with high rxz
shearing forces. Important friction forces will act on the two faces of the flakes creating
a rotation momentum having the tendency to align the flakes parallel with the xz plane,
hence perpendicular to the part surface.
It follows thus that flow perturbation elements can indeed orient Al flakes per-
pendicular to the part surface, however these perturbation will act mainly at the core of
the flow.
On the other hand, the flow marks are zones where the particles oriented perpen-
dicular to the part surface are found close to the part surface and not at the core zone. In
this context, an additional mechanism must be accounted in order to explain the
presence of flow marks in the wake of the perturbation elements.
3.3 Fountain Flow Zone Role for the Generation of Flow Marks
Let us consider the configuration shown in Fig. 7 where a perturbation element (a rib) is
placed along the flow path during the filling of the cavity. Upwind of the rib, the
advancing front is not-perturbed and, as discussed earlier, the high shear rate rxy at the
cavity walls will determine a strong alignment of the Al flakes: close to the cavity walls,
the particles will be mostly parallel to the surface and a metallic effect will be obtained.
Fig. 7. The physical mechanism at the base of creating flow marks in the wake of perturbation
elements: the flow perturbation downwind of the rib will orient the flakes perpendicular to the
surface at the core zone; these are transported to the fountain zone where the transversal flow
(vz 6¼ 0) will further transport the particles to the part surface, creating a flow mark
Due to the filling of the rib, downwind of the rib a perturbation of the core flow is
generated. Significant shear stresses rxz will develop at the core zone and consequently
the flakes will align perpendicular with the part surface in the wake of a perturbation
element.
On the other hand, the hot core zone is feeding the “Fountain flow” zone found just
ahead (Fig. 7). This way, the core oriented flakes are rapidly transported into the
Fountain flow zone where the transversal flow existing there (vz 6¼ 0) will displace
further the oriented particles at the cavity surface. A flow mark is thus generated in the
wake of the rib.
4 Simulation Model and Correlation
In the previous section a physical framework explaining the main causes for the
generation of flow marks was proposed. The local flow perturbations appearing during
the filling of the cavity are at the source of flow marks. In this context, the use of a
simulation model that could predict these flow perturbations and the associated particle
orientation would be extremely useful. In the following a novel simulation method
proposed for estimating the flow mark risk is presented, together with a correlation
between simulation and experimental results.
4.1 The Model

The simulation methodology was implemented in Moldflow injection molding soft-
ware. The fiber orientation feature included in Moldflow was used for the prediction of
Al flakes orientation. The novelty of the simulation method consists in adapting the
fiber orientation model to a simple midplane mesh approach, with the aim of predicting
the flow mark zones.
The fiber orientation model used has at its base the Folgar-Tucker orientation
equations [7]. A derivation of the original Jeffrey model [8], the Folgar-Tucker model
additionally uses a phenomenological term that accounts for fiber interactions. In this
process two semi-empirical coefficients are introduced: Cl , a phenomenological
parameter that must be correlated with the experiment, and Dz , a coefficient which is
introduced for midplane modelling only. There are some studies [9, 10] providing
guidelines for choosing Cl and Dz values with respect to the aspect ratio of the fibers
and the thickness of the part, but, providing the empirical nature of the coefficients, a
correlation with the experiment is always necessary.
4.2 The Study Case

A number of injection trials were available at Renault for a series of mass colored parts
including Al pigments. In this context, the present study focuses on modeling one of
these parts, namely the “Rear Bumper Add On” from Sandero project (Fig. 8).
The part is sequentially injected by using three tab gates (one central and two
laterals tab gates). Notice that the two laterals gates are only used for packing (are
opened at the beginning of the packing phase, after the cavity is filled), hence the part is
filled solely by the central gate. Because in the study we focus on estimating the flow
marks, a phenomena developing during the filling phase only, we simplify the
Moldlflow model by considering only the central gate.
The part was injected using the Lyondell Basell TRC 277P polymer filled with a
precise fraction of Al flakes. Notice that the original data sheet (udb file) for TRC 277P
contains the material data for the plastic matrix only, not including the Al filler. The
10 A.I. Ciobanas
Fig. 8. The front and rear views of the “Rear Bumper Add On” part. The green arrows indicate
the injection gates positions. The main flow marks are marked on the visible side. On the core
side the pin clips (perturbation elements) at the source of the flow marks are also marked.
mass concentration, the diameter distribution as well as the aspect ratio of the Al flakes
were provided by the material supplier to Renault and were manually added to the udb
file (at Renault and material supplier request, all the filler parameters are kept confi-
dential in this article). With respect to these parameters the range for the two empirical
coefficients Cl and Dz were established [9, 10].
On one hand, Dz is a factor that affects the orientations of fiber along the thickness
of the part (1 for a fully 3D problem and 0 for a 2D problem). Given the large thickness
of the analyzed part (3.5 mm) the general approach is to set Dz ¼ 1. Notice that it was
found that the influence of this parameter on the present simulation results was neg-
ligible. In the Folgar-Tucker model the coefficient Cl accounts for the fiber-fiber
interaction and is highly depended on Al flakes aspect ratio. Given a flake aspect ratio
(diameter/thickness = D/h) of the order of 50–100, the order of magnitude for this
coefficient is 103 [9, 10]. A results sensitivity analysis at the variation of the Cl
coefficient has been carried out and the value Cl ¼ 0:003 provided results that corre-
lated best with the injection trial results.
Finally, the process parameters used in the simulation are similar with the one in the
injection trials. The main injection parameters are synthetized in Table 1.
Table 1. The main process parameters used for the rheology simulations
Process parameters Value Process parameters Value
Mold surface T 40 °C Fill time 1.8
Melt injection T 250 °C Injection profile Constant 375 cm3/s
IMM manufacturer Krauss Maffei V/P switchover 99 %
IMM type KM 1300 T Packing time 19 s
IMM screw diameter 120 mm Packing profile Constant 40 MPa
IMM max clamping force 1300 T Cooling time 13 s
The nominal thickness of the part is 3.5 mm. Midplane extraction of the original
3D part was obtained using the casting method available in ANSA preprocessing
software. It was observed that the level of refinement of the mesh and thickness
gradients is very important to model correctly the flake orientations and the flow
perturbations. Therefore, at the perturbation zones, a finer mesh was used and special
care was taken to model correctly the thickness gradients at the evolution from the
3:5 mm of the part skin to the smaller thickness of the pin clips (Fig. 9). The total mesh
has approximately 150 000 tria elements.
Fig. 9. The midplane mesh was refined at the perturbation zones in order to better capture the
flow perturbation effects. The color range → thickness of the part.
Similar with the real injection tool, the hot runner (Fig. 10) was modeled using a
hot runner channel, a hot sprue (both having D ¼ 20 mm) and a Runipsys™ hot nozzle
type B, (manifold:D ¼ 20 mm, gate: D ¼ 6 mm). The U shaped cold runner was
modeled with beams, while the lateral tab central gate was modeled in midplane mesh
with a thickness variation similar with the real case (Fig. 10).
Fig. 10. The feeding system details: hot runner; hot nozzle, cold runner and cold gate
4.3 Simulation Results and Correlation with Experiment

Let us focus now on the flow marks zones found on the injected parts. The main flow
mark zones on the right hand side of the part are shown in Fig. 11. Each of this flow
mark is produced by a perturbation element (pin clips or sharp edges of the part
surface). The source for flow marks F1 and F2 are found at the flow perturbations
induced by the pin clips P1 and P2, while the flow mark F3 is a consequence of the
flow perturbation at the sharp edge shown in Fig. 11b (P3 zone).
12 A.I. Ciobanas
Fig. 11. (a) The flow marks identified on the right hand side of the part; (b) The perturbations
elements at the source of the flow marks (pin clips for P1 and P2 and sharp edge for P3); Al
flakes orientation tensor (1st direction) at the core zone: front (c) and rear view (d)
In Moldflow the orientation of the particles is defined by a second order tensor (6

components, symmetric, trace = 1). The three principal directions of the orientation
tensor are quantifying the main orientations of the particles at a given point and their
values are between 0 and 1 (from low to high probability of fiber alignment). The first
principal direction (largest value) represents the direction along which the most fibers
are aligned, while the third principal direction (smallest value) represents the direction
along which the fewest fibers are aligned.
In the analysis of Al flake orientation it was essential to account for the limitation of
the mathematical model behind the midplane approach: the Hele-Shaw approximation.
In this approximation the “Fountain flow” zone is not modeled and the oriented par-
ticles generated at the core zone due to flow perturbations cannot be transported to the
part surface and solidified in the frozen layer (creating thus the surface flow marks).
Despite this limitation, one can still predict the flow mark generation zones by ana-
lyzing the particle orientation tensor (1st direction) at the core zone, where strong shear
stresses rxz will develop. This is an essential feature of the present simulation
methodology and is based on the physical reasoning proposed in Sect. 3.
In Fig. 11c, d, the particle orientation tensor (1st direction) at the core zone and at
the end of filling are presented for the right hand side of the part. As one can observe the
model successfully manage to reproduce the flow mark risk zones. Even if the thickness
of the flow marks does not correspond well to the physical ones, qualitatively, the flow
marks are well reproduced. One reason for this non-agreement resides in limitation of
the midplane model: the oriented particles generated at the core zone remain within the
core zone and will diffuse rapidly in the wake of the perturbation, hence the
non-agreement of the flow marks thickness. One other reason for this non agreement
reside in the fiber orientation model used by Moldflow (Folgar-Tucker model) that has
limitations at high fillers concentrations. Despite all these drawbacks and simplifica-
tions, the present model successfully manages to predict the flow marks risk zones.
In Fig. 12 the relation between the perturbation elements and the flow marks
appearing on the left hand side of the part is presented. Similarly with the right hand
side, the flow marks F5 and F6 are a consequence of flow perturbation at the pin clip P5
Fig. 12. (a) The flow marks identified on the left hand side of the part; (b) The perturbations
elements at the source of the flow marks (pin clips for P4 and P5 and sharp edge for P3); (c) Al
flakes orientation tensor (1st direction) at the core zone and at the end of filling
and sharp edge P6 respectively. It is worthwhile to notice that pin clip P4 does not
produce a visible flow mark on the part surface. In agreement with the trial result, the
present model does not reproduce a significant orientation zone corresponding to P4,
Fig. 12c. This comes to support even further the correctness of the present approach.
The difference between the effects produced by P1 and P4 pin clips comes from the
asymmetry of their positioning (Fig. 13). The relative position of the perturbation
elements with respect to the flow direction influences significantly the level of local
flow perturbation and consequently the risk to develop a flow mark.
At the lateral zones the model also reproduces two additional zones with significant
Al flake orientation (Fig. 13), having as source the pin clips P7 and P8. Interestingly,
the physical parts are not showing significant visual defects at these zones (some flow
marks can be observed but they have a small intensity). One reason for this
none-agreement between the experiment and the simulation might reside in the fact that
Fig. 13. The asymmetry between the positioning of P1 and P4 pin clips. Pin clips P7 and P8 are
at the source of the two orientation zones (lateral red circles).
14 A.I. Ciobanas
these two zones are filled last, just before the start of packing. Most probably, the
oriented Al flakes generated at the core zone don’t have the time to be transported at the
fountain zone and then reach the part surface to create visible flow marks.
As discussed in the previous section, Al flake orientations are highly correlated
with the shear stresses developed along the flow lines: (e.g. rxy in the close proximity
of the cavity surfaces; rxz at the core zone, in the wake of the perturbation element). By
using Folgar-Tucker formulation for the orientation model, the solver must compute at
each node the components of these shear stresses. Unfortunately, in the set of
post-processing variables, Moldlfow software do not include the components of the
local shear stresses. This makes it difficult to validate directly the correlation between
shear stresses and flake orientation, as discussed in Sect. 3.
One can still evidence, though, the flow perturbations produced by the technical
features by analyzing the velocity magnitude evolution at the core zone, during the
filling of the clips. In Fig. 14b the velocity magnitude in the axial xy plane (0 nor-
malized thickness) is plotted at 0.9 s after the start of injection, time which corresponds
with the filling of P1 clip. One can observe the strong perturbation of the velocity field
and a clear evidence of high rxz shear stresses zones at the core zone (highlighted zones
in Fig. 14b). As discussed, these strong rxz shear stresses are at the source of the
orientation of Al flakes perpendicular to the part surface.
Fig. 14. (a) The flow marks produced by the P1 clip; (b) The velocity magnitude at the core
zone (the axial plane, 0 normalized thickness) showing the significant flow perturbations
produced by the filling of the P1 clip.
5 Conclusions
For plastic materials using metallic fillers (e.g. Al flakes) there is an important risk of
generating flow marks (e.g. darker zones or darker plumes) on the part surface. The
flow marks are a consequence of the flow front perturbations during the filling of the
cavity. In this context, the present study focus on developing a simulation methodology
that could predict the flow marks risk zones.
First, a physical mechanism that explains the source of flow marks is proposed. On
one hand, close to the cavity walls, the high rxy shear forces will align the Al flakes
parallel with the part surface, amplifying the metallic effect. On the other hand, if a
perturbation element (e.g. rib, clip, sharp edge) is placed along the flow lines, a flow
perturbation is generated at the core of the flow. High rxz shear forces will develop at
the core zone in the wake of the perturbation element, orienting the Al flakes
perpendicular with the part surface. Hence the Al flakes oriented perpendicular to the
part surface are generated at the core of the flow. An additional mechanism must be
accounted to explain the presence of these particles close to the cavity walls and in the
wake of the perturbation elements (i.e. the flow marks). It is found that the hot core
zone is feeding the “Fountain flow” zone found just ahead of the perturbation element.
The core oriented flakes are thus rapidly transported into the Fountain flow zone where
the transversal flow will further displace the particles at the cavity surface. A flow mark
is thus generated in the wake of the rib.
In a second step, a new simulation methodology was developed and implemented in
Moldflow injection molding software. The novelty of the simulation method consists in
adapting the fiber orientation model to a simple midplane mesh approach, with the aim of
predicting the flow mark zones. Even if the 3D fountain flow is not modeled for a
midplane approach, one can still predict the flow mark generation zones by analyzing the
flow perturbation generated at the core zone. This is one essential feature of the present
simulation methodology and is based on the physical reasoning proposed in Sect. 3.
Finally, the new methodology is applied to the modeling of the “Rear Bumper Add
On” part from Sandero project. The injection trails available at Renault are subse-
quently used to correlate the simulation and experimental results.
The simulation results presenting the Al flakes orientation distribution at the core
zone show a very good correlation with the flow marks zones identified on real parts. In
this context, the present simulation methodology can be used as a reliable tool for
predicting the flow mark risk zones for parts injected with plastic materials containing
Al flakes fillers.
Acknowledgments. The present study would not have been possible without the fruitful col-
laboration between Renault DE-VD department, Delta Invest injection plant and Lyondell Basell
material supplier. The author wants to thank particularly MaxPol Moreau for the help on material
technical information, Philippe Madelaine for his precious help during the injection trials and
Marian Codreanu for his useful suggestions.
References
1. Schaal, M.: Special effects pigments for plastics. In: The 2nd International Conference on
Fillers for Polymers, Cologne, Germany, Paper 11 (2006)
2. Carl Schlenk Aktiengesellschaft entreprise. http://www.schlenk.com/index.php?id=35&L=1
3. Marrion, A.: The Chemistry and Physics of Coatings. Royal Society of Chemistry,
Cambridge (2004)
4. Wißling, P.: Metallic Effect Pigments: Fundamentals and Applications. Vincentz Network
GmbH & Co KG, Hannover (2006)
5. Kazmer, D.O.: Injection Mold Design Engineering. Hanser Publisher, Munich (2007).
Chapter “5.3.5 Power Law Model”, pp. 99–101
6. Wheeler, I.: Metallic Pigments in Polymers. Rapra Tehnology Ltd., Shawbury, United
Kingdom (1999). Chapter “7.4 Orientation”, p. 71
7. Folgar, F., Charles, T.: Orientation behavior of fibers in concentrated suspensions. J. Reinf.
Plast. Compos. 3(2), 98–119 (1984)
16 A.I. Ciobanas
8. Jeffery, G.: The motion of ellipsoidal particles immersed in a viscous fluid. Proc. Roy. Soc.
Lond. Math. Phys. Eng. Sci. 102(715), 161–179 (1922). The Royal Society
9. Moldflow Documentation: Analysis sequence > Fiber orientation analysis > Moldflow’s
fiber orientation models
10. Eberle, A., et al.: Fiber orientation kinetics of a concentrated short glass fiber suspension in
startup of simple shear flow. J. Nonnewton. Fluid Mech. 165(3), 110–119 (2010)
Numerical and Experimental Investigations
of a Twin Sheet Thermoplastic Structure
with Rectangular Frusta
Stefan Tabacu1,2(&), Claudiu Diaconescu2, and Alexandru Oltean2

1
University of Pitesti, Pitesti, Romania
stefan.tabacu@upit.ro
2
ALSECA Engineering, Bucharest, Romania
{claudiu.diaconescu,alexandru.oltean,
stefan.tabacu}@alseca.com
Abstract. Thin walled structures with rectangular frusta discussed by this study
are manufactured using vacuum formed thermoplastic sheets that finally define a
unitary structure that is a highly efficient solution providing outstanding per-
formance by means of using small amounts of materials and high-tech, low
energy consumption equipment. Sample parts formed by 2 × 2 cells with a cell
dimension of 10 × 10 mm and a total height of 18.3 mm were tested in com-
pression. Average crushing force and energy absorbed were calculated.
Numerical models of the structure were developed and solved using Ls-Dyna
explicit and implicit capabilities. Generally, both implicit and explicit solver
provided good results in terms of crushing force and energy absorbed, while the
failure pattern was better reproduced by the implicit solver. The implicit solver
requires longer runtimes and this may add some constrains, thus for the explicit
solution the simulation time was investigated in order to setup a parametric
analysis. It was found that by using smaller cells the total energy absorption can
be increased with respect to the available design volume, yet a limitation is
added by the manufacturing process regarding the depth of the shape, draft angle
and thickness of the part.
Keywords: Cellular structures Twin-sheet forming Axial crushing Explicit

and implicit simulations Parametric analysis
1 Introduction
Cellular or patterned structures are of a particular interest due to high energy absorption
performances during crushing while maintain a low specific weight. This way the
structures may be a candidate for demanding industry activities like the automotive
field [1, 2] where keeping a low vehicle weight is a major concern. The particular shape
of the finished parts gives the required strength (crushing force) and energy absorption
performances. The work of Bartl et al. [3] addressed a composite with a porous mineral
granulate embedded in polyamide, Kathiresan et al. [4] a composite with laminated
fibers, Zou et al. [5] a honeycomb structure and Sashikumar et al. [6] an aluminum
egg-box structures. A specialized structure that uses the SKYDEX® material was

18 S. Tabacu et al.
investigated by Zhu et al. [7]. Akisanaya and Fleck [8], Gupta [9] and Mamalis et al.
[10] investigated the deformation process and structural performances of conical frusta
structures. The egg-box structures are closely related and prove to be a reliable solution
for energy absorption devices as presented by Lam et al. [11] for thermoplastic com-
posites and by Nowpada et al. [12] for a metallic material. Twin sheet thermoplastic
structures are highly efficient solutions that provide outstanding performance by means
of using small amounts of materials (shaped in rather complex geometry) and
high-tech, low energy consumption equipment. The structures are manufactured using
vacuum formed sheets that finally define a unitary structure. The manufacturing process
is presented in Fig. 1.
Fig. 1. Twin sheet structures. Manufacturing process
The thermoplastic sheets are heated up to the process temperature (glass transition
phase) and then inserted in the forming machine. Using vacuum the sheet are indi-
vidually shaped according to the design specification and finally the upper and lower
molds are closed, in order to fuse the top and bottom sections, giving the unitary
structure of the part.
2 Experimental Investigation of the Twin Sheet Structures
Experimental tests were performed on existing structures (http://www.hombach-

kunststofftechnik.de). Figure 2(a) presents a photographic image of the part and Fig. 2
(b) presents the main dimensions associated to the structure.
The total height of the structure is 18.6 mm and the walls have a thickness of
0.31 mm. The draft angle of the rectangular frustum is of 9.4°. The frustum joins the
flat surface by a filleted section with a radius of 0.30 mm.
The cellular structure sets of four cells (disposed in a 2 × 2 cells configuration)
were cut and used for the compression test, performed using an universal testing
machine.
Numerical and Experimental Investigations of a Twin Sheet 19
Fig. 2. Cellular structure (a) cellular structure (scale in mm); (b) main dimensions.
The speed of the moving plateau was set to 2.5 mm/min and the total travel (tool
displacement) was limited to 9 mm in order to avoid the increase of the force due to
heavily compaction of the sample. Figure 3 presents the crushing process.
Fig. 3. Experimental testing (a) initial configuration; (b),(c) stage I - deformation of the flanges;
(d) stage II - collapse of the columns; (e) compaction.
Figure 4 shows a section of the collapsed structures.
Fig. 4. Collapsed structure (a) folds created on top; (b) folds created on bottom.
20 S. Tabacu et al.
The first stage of the crushing process consists in the deformation of the flanges of
the cellular structure. Once buckling occurs, folds are created on the wall of the frusta
sections. The crushing force was recorded during the experiments (Fig. 5).
Crushing force vs. tool displacement.

550
500
450
400
350
Force [N]
300
250
200
150
100
Sample A
50
Sample B
0
0 1 2 3 4 5 6 7 8 9
Tool displacement [mm]
Fig. 5. Measured crushing force
The average crushing forces Pm;expA and Pm;expB were computed using available
experimental data. For the first case the average value is of 326 N while for the second
case the average value is of 338 N. Results are listed in Table 1.
Table 1. Experimental data.

Parameter Sample A Sample B Notes
Height [mm] 18.6 18.6 Geometrical
Wall thickness [mm] 0.31 0.31 features of the parts
Draft angle [º] 9.4º 9.4º
Configuration 2 × 2 cells 2 × 2 cells
Average force [N] 326 338 Tool travel speed:
Maximum force [N] 440 457 2.5 mm/min
Energy [J] 2.83 2.94
Experimental data were used also to identify the polypropylene material parameters
of the formed material (www.matweb.com). The mean crush force is dependent of the
value of yield stress (Abromowicz and Jones [13], Jones [14], Wierzbicki et al. [15])
while the failure strain gives the amount of internal energy accumulated by the
structure. Equation (1) gives the average (mean) crushing force.
pffiffiffiffiffiffiffiffiffiffiffiffi
Pm ¼ 4 M0 9:6 3 ci1 =t þ p tan h ð1Þ
where M0 ¼ 1=4 r0 t2 is the fully plastic moment per unit length; r0 is the flow
stress for the material; t is the thickness of the part and ci1 is the sum of half width and
half breadth defined as ci1 ¼ l þ2 w.
Using experimental results Eq. (2) was used to identify the value of flow stress of
the material used to manufacture the samples.

Pm;expA þ Pm;expB pffiffiffiffiffiffiffiffiffiffiffiffi
¼ 4 M0 9:6 3 ci1 =t þ p tan h ð2Þ
2
where Pm;expA and Pm;expB are the average values of the experimental crushing forces
listed in Table 1. A value of 34:5MPa was assigned for the flow stress and further used
for the numerical simulations considering a bilinear material model with a null hard-
ening modulus. A simple material model (isotropic with kinematic hardening -
*MAT_003) was selected and by fitting experimental data, a value of 1500 MPa was
assigned to the Young’s modulus and 1.5 (150 %) for the strain at failure following the
experimental behave of the structures. The strain rate effect was neglected for the
selected material models.
3 Numerical Investigation of Twin Sheet Thermoplastic

Structures
The structures with rectangular frusta were investigated using numerical methods.
A Matlab [16] code was developed in order to define a parameterized numerical model
for the analysis used both for material identification phase and parametric study. Based
on the extensive capabilities of Ls-Dyna [17] both explicit and implicit simulation were
performed. For the simulation cases mentioned before (implicit and explicit solvers),
the computed crushing force obtained using the numerical models are presented in
Fig. 6.
Crushing force vs. tool displacement. Crushing force vs. tool displacement.
550 550
500 500
450 450
400 400
350 350
Force [N]
Force [N]
300 300
250 250
200 200
150 150
100 Sample A 100 Sample A

Sample B Sample B
50 50
Simulation - Explicit Simulation - Implicit A
0 0
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
Tool displacement [mm] Tool displacement [mm]
a) b)
Fig. 6. Experimental and numerical crushing force (a) Explicit analysis; (b) Implicit analysis
22 S. Tabacu et al.
The total simulation time is closely related to the required or computed timesteps used
during the simulation. Time unit is constrained by the set of consistent units assigned to
the numerical model, defined for the proposed model as kg, mm, ms. The simulation time
was set, by a number of trial and error simulations, to 40 ms. For the implicit analysis the
maximum time step (parameter DTMAX) was set to 0.05 ms (Implicit A Cases) for a
short runtime and to 0.01 ms (Implicit B Cases) for results improvements. Based on the
definition of the numerical model, the timestep for the explicit analysis was defined by the
solver without triggering the mass scaling procedure. The numerical model was devel-
oped using shell elements of type Belytschko-Tsay. A number of 32 elements were
defined along the edges of the cell and 36 elements along the edge of the rectangular
frustum. Contacts were defined at the interface of the cellular structure with the stationary
wall (ground) and with the flat moving wall. A self-contact was defined for the cellular
structure. The coefficient of friction was set to 0.30 for the single surface contact and to
0.7 for the contact between the part and the walls.
For the analysis of the results the reference value for the average force and energy
was calculated as the median of the results obtained from the experiments. Results are
summarized in Table 2 (Fig. 6).
The results obtained using the numerical simulation are similar to the experimental
data. For the average crushing force the maximum value was obtained for the Implicit
analysis B. The computed internal energy accumulated by the structure is similar for all
simulation cases with the experimental results. The current mesh size (32 × 36 ele-
ments) was defined by multiple runs in order to provide reliable results for the solver
Table 2. Numerical results. Comparison with experimental data.

Average force [N] Error [%] Energy [J] Error [%]
Experiment A 326 – 2.83 –
Experiment B 338 – 2.94 –
Experiments A, B 332 Reference 2.89 Reference
Simulation – Expl. 329 –0.75 2.82 –2.25
Simulation – Imp. A 331 –0.15 2.89 0.17
Simulation – Imp. B 349 5.28 2.83 –1.19
Fig. 7. Mesh density (cell x frustum edge) (a) 32 × 36 elements; (b) 22 × 24 elements; (c)
16 × 18 elements.
types and parameters used for the numerical analysis. The influence of the mesh density
was investigated in order to identify a reliable solution for the numerical analysis.
Implicit analysis (Implicit A) with a larger maximum timestep (DTMAX = 0.05
ms) required the shortest simulation times and for the numerical model with a dense
mesh (32 × 36 elements) it provided good results in terms of average crush force and
total deformation energy. Plots of the crushing force obtained from the simulations are
presented in Fig. 8.
Crush force. Mesh size. Explicit analysis. Crush force. Mesh size. Implicit analysis. DTMAX=0.05 ms. Crush force. Mesh size. Implicit analysis. DTMAX=0.01 ms.
600 550 550
500 500
500 450 450
400 400
400
350 350
Force [N]
Force [N]
Force [N]
300 300
300
250 250
200 200
200
150 150
100
Mesh 32x36 100 Mesh 32x36 100 Mesh 32x36
Mesh 24x27 Mesh 24x27 Mesh 24x27
50 50
Mesh 16x18 Mesh 16x18 Mesh 16x18
0 0 0
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
Tool displacement [mm] Tool displacement [mm] Tool displacement [mm]
Fig. 8. Crushing force. Models mesh size. (a) Explicit analysis; (b) Implicit analysis A;
(c) Implicit analysis B
The force plots obtained for the model with a mesh size of 16 × 18 show the it
failed to match the deformation pattern for Implicit cases A.
4 Parametric Analysis
Based on the results and findings presented by now a parametric analysis was per-
formed (on a pack of 2 × 2 structures) and the following parameters were investigated:
cell size and sheet thickness (of the formed structure, mainly the thickness of the
rectangular frustum). The length of the cell ranges from 10 to 25 mm in steps of 5 mm,
the draft angle is set to 9.4°, while the thickness of the formed part ranges from 0.3 to
0.6 mm in steps of 0.1 mm. The travel of the moving wall was limited to half of the
cell’s height (Fig. 2b)).
Figure 9 presents the results recorded during the computer simulation using the
explicit solver in order to speed up the process (results show that the average force is
well estimated). For the preliminary process of structure dimensioning the analytical
solution may provide the starting point while the numerical analysis will complete and
improve the design.
The energy absorption increases while decreasing the cell’s dimensions. For a cell
with edge’s length of 10 mm and a thickness of 0.5 mm the energy absorbed is of
2.5 J. For a cell with edge’s length of 20 mm and a thickness of 0.5 mm the energy
absorbed is of 4.0 J. This show that for the same available space using smaller cell the
energy absorbed will be higher due to the increased efficiency of the structure. Fur-
thermore the absorbed energy can be increased with the wall’s thickness.
Overall, the methods of improving the structural response can be derived from the
analysis of the mean crushing force. The only limitation is related to the dimensions of
24 S. Tabacu et al.
Average crushing force [N] E nergy [J]

12 0
0
4 .5 5
3.
12 00
4
14
3
00
0.55 10 0 12 12 00 0.55
0 4 .5
80
00 5
0
10
00
10 00 4
80 0
2. 3.
thickm ess [m m ]
0.5 0.5
thickness [m m ]
10 0 10 00
5 5 4 .5
80 0 3
80 0 4
60 0
0.45 0 80 0 10 0 0.45 3 .5
0
80 0 2. 3 3 .5
0.4 60 0 0.4 5
60 0 80 0 2
60 0 3
60 0 2 .5
0.35 0.35
4 04 0 0
6
00
Computed avergate force 2 2. 5
60 0 1 .5
40 0 Analytical solution 2
0.3 0.3
10 15 20 25 10 15 20 25
length: L 0 [mm] length: L 0 [mm]
a) b)
Fig. 9. Energy absorbed and Specific energy absorption. (a) average crushing force; (b) energy
absorption.
the structure as the height of the cell and draft angle give the thickness of the part and
manufacturing process requirements.
5 Conclusions
Twin sheet thermoplastic structures were investigated using experiments and numerical
simulation. Sample parts were tested in compression in order to determine the average
crushing force and the energy absorbed. The data were used also for a parameter
identification of the material model used for the numerical simulations.
Results were evaluated in terms of average crushing forces, energy absorbed and
failure pattern. For dense mesh definitions, results obtained using the implicit solver are
more accurate, although one major shortcoming resides in the long runtime required for
the model to provide a solution. When the maximum timestep was increased the
implicit solver failed to provide an accurate solution. Therefore the performances of
the explicit solver were evaluated. As the material model was not strain rate sensitive
the influence of simulation time was investigated. Varying the simulation time from
10 ms to 50 ms the outputs were in a close range. As a consequence based on the
explicit solver a parametric analysis was performed. The length of the cell ranged from
10 mm to 25 mm while the thickness of the part varies from 0.3 mm to 0.6 mm.
Average crushing force was computed and compared with the analytical solution. It
was found that for the preliminary investigations of such structures the analytical
equation provides a good start point. Also the parametric analysis pointed that
whenever possible the energy absorption performances can be increased by using
structures with small cell. One limitation may be added by the geometry of the part
(depth and draft angle) and the manufacturing process.
References
1. Tarlochan, F., Samer, F., Hamouda, A.M.S., Ramesh, S., Khalid, K.: Design of thin wall
structures for energy absorption applications: enhancement of crashworthiness due to axial
and oblique impact forces. Thin Walled Struct. 71, 7–17 (2013)
2. Tanlak, N., Sonmez, F.O.: Optimal shape design of thin-walled tubes under high-velocity
axial impact loads. Thin Walled Struc. 84, 302–312 (2014)
3. Bartl, F., Klaus, H., Dallner, R., Huber, O.: Material behaviour of a cellular composite
undergoing large deformations. Int. J. Impact Eng 36, 667–679 (2008)
4. Kathiresan, M., Manisekar, K., Manikandan, V.: Performance analysis of fibre laminated
thin conical frusta under axial compression. Compos. Struct. 94, 3510–3519 (2012)
5. Zou, Z., Reid, S.R., Tan, P.J., Li, S., Harrigna, J.J.: Dynamic crushing of honeycombs and
features of shock fronts. Int. J. Impact Eng. 36, 165–176 (2009)
6. Sashikumar, S., Chirwa, E.C., Myler, P., Qian, P., Matsika, E.: Numerical investigation of
the collapse behaviour of an aluminium egg-box under quasi-static loading. Int. J. Crash-
worthines 17, 582–590 (2012)
7. Zhu, F., Dong, L., Ma, H., Chou, C.C., Yang, K.H.: Parametrized optimal design of a novel
cellular energy absorber. Int. J. Mech. Sci. 86, 60–68 (2014)
8. Akisanya, A.R., Fleck, N.A.: Plastic collapse of thin-walled frusta and egg-box material
under shear and normal loading. Int. J. Mech. Sci. 48, 799–808 (2006)
9. Gupta, P.K.: A study on mode of collapse of varying wall thickness metallic frusta subjected
to axial compression. Thin. Wall. Struct. 46, 561–571 (2008)
10. Mamalis, A.G., Manolakos, D.E., Ioannidis, M.B., Kostazos, P.K., Hassiotis, G.: Finite
element simulation of the axial collapse of thin-wall square frusta. Int. J. Crashworthines 6,
155–164 (2001)
11. Lam, S.W., Tao, X.M., Yu, T.X.: Comparison of different thermoplastic cellular textiles
composites on their energy absorption capacity. Compos. Sci. Tech. 13–14, 2177–2184
(2004)
12. Nowpada, S., Chirwa, E.C., Myler, P., Chinnaswamy, G.K., Matsika, E.: Egg-box panel for
commercial vehicle front-compressive loading test. Int. J. Crashworthines 15, 59–70 (2010)
13. Abramowicz, W., Jones, N.: Dynamic axial crushing of square tube. Int. J. Impact Eng 2,
179–208 (1984)
14. Abramowicz, W., Wierzbicki, T.: Axial crushing of multi corner sheet metal columns.
J. Appl. Mech. - T. ASME 56, 113–120 (1989)
15. Wierzbicki, T., Recke, L., Abramowicz, W., Gholam, T.: Stress profile in thin -walled
prismatic columns subjected to crush loading - I Compression. Comput. & Struct. 51,
611–623 (1994)
16. The MathWorks Inc. MATLAB. Natick, Massachusetts, United States (2009)
17. Hallquist, J.O.: LS-DYNA Keyword User’s Manual. Version 970. Livermore Software and
Technology Corporation, Livermore CA (2003)
Expanded Polypropylene Finite Elements
Analysis in Transport Application
Alexander Schwarz1, Catalin-Andrei Runcianu2(&),

and Andreas Gebhardt1
1
FH Aachen University of Applied Sciences, Aachen, Germany
{a.schwarz,gebhardt}@fh-aachen.de
2
Transilvania University of Brasov, Brasov, Romania
runcicta@gmail.com
Abstract. The current technology has reached a maturity degree which requires
a higher complexity of the analyses performed in the development of a product.
Most of the technological innovations involve recently discovered materials that
exhibit an intricate behavior. For this reason, in the development of a new
product, it is no longer satisfactory to perform static linear finite elements
analyses and nonlinear analyses are required. The analyses which imply the use
of nonlinear materials present certain difficulties and the risk of misleading
results if not performed correctly. This paper intends to be an overview of a
finite elements analysis of a transportation box made from expanded
polypropylene. It highlights the steps that are required to reduce the inaccuracy
of the results and minimize the gap between simulation and reality and also the
problems which can occur along the process.
Keywords: Expanded Polypropylene Finite elements analysis Material

testing Product development
1 Introduction
The current paper focuses on finite elements analysis for a relatively new material
called expanded polypropylene (EPP) which is a polymeric foam. The purpose of it is
to describe the steps that have been taken in order to develop a new transportation box
made from expanded polypropylene.
The main steps necessary to perform finite elements analysis for nonlinear materials
are: material testing, curve fitting and validation of mathematical model and simulation.
This paper can serve as a guideline for finite elements analysis not only for expanded
polypropylene but for any hyperelastic material, materials very common in the auto-
motive industry (rubber used in tires, different blends of plastics from the vehicle
interior, etc.).

Expanded Polypropylene Finite Elements Analysis 27
2 Expanded Polypropylene
Expanded polypropylene (EPP) was produced for the first time in the 1970´s in the
effort to develop different forms of polypropylene. It was at first designed for auto-
motive components in order to produce parts with a smaller weight while maintaining
the necessary mechanical properties to satisfy the requirements of the application.
2.1 Applications
EPP has managed to make its way in different domains of manufacturing. It initially
started to be used in automotive engineering because of the energy absorption prop-
erties. In the automotive sector, this foam is mostly used for interior components such
as stowage systems, door panels, floor levelers, pillars, parcel shaves, toolkits, head
rests and sun shades. Some of the reasons for using EPP for vehicle interiors are:
energy absorption, good elastic behavior, in combination with metals it absorbs more
energy, shape flexibility in molding, good surface aspect, and the possibility to easily
obtain a multitude of surface finishes.
Worth mentioning is the fact that in some scientific research papers [1], expanded
polypropylene was successfully used to filter runoff water which contains road
chemical residuals from tires, crank-case oil or other constituents and thus resulting in
pollution reduction.
Although there are many important areas in which this type of foam plays an
important part, the largest industry which revolves around polymeric foams is the
packaging one.
2.2 Properties and Behavior

EPP is a closed-cell foam which means that the gas is held within the cells. While in an
open-cell structure the gas is able to flow between the cells and the foam is able to
compress more, EPP is not able to compress as much due to its closed structure. Since
the gas in a foam has the least mechanical strength, EPP has better mechanical strength
properties in comparison with open-cell foams like polyurethane foams, this being one
of the reasons for the increasing popularity of EPP in comparison with other expanded
polymers.
EPP is used in different applications, having multiple roles such as comfort prov-
idence, components protection, and energy absorption. This foam is not crushable and
not totally recoverable. It has a hyperelastic behavior but also shows some inelastic
properties like residual strain and stress softening. Hysteresis is a phenomenon that can
be observed at EPP. Although a hysteretic effect can be observed during testing, EPP
rarely returns to its original size due to the presence of residual strain. The residual
strain present after loading and unloading is reversible and gradually with time it
decreases and after long periods of annealing time it can disappear [2].
28 A. Schwarz et al.
3 Material Testing
In order to perform a finite element analysis for a product, it is necessary to obtain the
hyperelastic properties of the material. In this case, several different density material
probes have been tested in order to ensure a satisfactory database. The density range
tested was from 59 kg/m3 up to 270 kg/m3. The specimens used had the density of
59 kg/m3, 80 kg/m3, 127 kg/m3, 152 kg/m3 and 270 kg/m3. The annotations used in
the paper (e.g. EPP59) can be decoded as expanded polypropylene with ρ = 59 kg/m3.
3.1 Test Set-up and Results

The machine model used was a Zwick/Roell ZmartPro. The speed of the machine was
set at 20 mm/min and the specimens were tempered for several hours at the laboratory
temperature of 22° C. A total of 25 specimens were analyzed to obtain accurate and
stable results (5 specimens/density) (Fig. 1).
In Fig. 2, the average values obtained in the tests are represented, values which
highlight the behavior of expanded polypropylene. It can be observed that the ultimate
tensile strength increases as the density of the EPP is higher, which translates into the fact
that the increase of the mechanical strength of the foam is accompanied by a decrease of
the elasticity. The values obtained prove that density is a characteristic that has a major
impact on the elastic properties of EPP and although expanded polypropylene is
considered a very elastic material, it tends to lose some of those elastic properties with the
increase of the mass to volume ratio.
Taking as reference the average value of EPP59, it can be observed the fact that for
EPP80 the strain level increased by 15 % and the strain level decreased by 30 % while
for EPP127 the stress level increased by 116 % with a decrease of the strain of 23 %.
For the EPP152, the level of the ultimate tensile strength increased by 169 % while the
strain decreased with 34 %. The highest density EPP of 270 kg/m3 exhibited the
highest increase of the value of the ultimate tensile strength of 408 % with a decrease
of the strain level of 59 % (Fig. 3). The lack of a clear trend in the strain level is a cause
of material imperfections appeared in the process of machining of the specimens.
Fig. 1. Tensile stress-strain curves (left) and tensile specimens (right)

Fig. 2. Average values of ultimate tensile strength and strain
Fig. 3. Stress and strain percentual evolution reported to EPP59 values
4 Curve Fitting and Validation of Material Model
After obtaining the stress-strain curves from the material testing, it was proceeded with
a curve fitting process to model and validate the material’s stress-strain characteristic.
In modern finite elements analysis software, there is a wide range of hyperelastic
mathematical models available, each one having its particularities. Depending on the
elastic properties of the material, certain models are more appropriate than others in
reproducing the behavior of polymeric foams and several papers that compare
hyperelastic models are available [3, 4].
The accuracy of the analysis depends on the model obtained by curve fitting,
especially for hyperelastic materials which present a highly nonlinear stress-strain
curve and large deformation ability. This step is very important because it allows early
detection of poor quality stress-strain curves that do not allow a convergence of the
solution.
4.1 Curve Fitting of Material Model

The curve fitting of the model was done using the ANSYS Workbench “Curve Fitting”
tool which allows the user to introduce the stress-strain data obtained in the tensile test
and to compute the model parameters. The engineering stress-strain data was used to
calculate the model parameters and to generate a computed stress-strain curve that is
described by a mathematical equation. Curve fitting was done for each model to
establish the best model for the analysis.
The “Neo-Hookean” [5], “Arruda-Boyce” [6], “Gent” [7] and “Blatz-Ko” [8]
models produced a linear fitting that is inadequate to implement in the analysis. The
“Ogden 3rd Order” model [9] fitting resulted in a curve similar to compression testing
while the “Yeoh 3rd Order” model [10, 11] was more accurate but with a high gap that
could lead to analysis inaccuracies (Fig. 4).
The best fit was obtained with the “Mooney-Rivlin 5 Parameter” [12, 13] and
“Polynomial 2nd Order” [14] model (Fig. 5). It is important to specify the fact that the
“Mooney-Rivlin 9 Parameter” model presented some convergence problems because of
a high number of coefficients (Table 1), reason for which the “Mooney-Rivlin 5
Parameter” model was used.
Fig. 4. “Yeoh 3rd Order” model fitting (left) and “Ogden 3rd Order” model fitting (right)
Fig. 5. “Mooney-Rivlin 5 Parameter” model fitting (left) and “Polynomial 2nd Order” model
fitting (right)
Table 1. Material constants for “Mooney-Rivlin” model

“Mooney-Rivlin 5 Parameter” material constants
C10 [MPa] C01 [MPa] C20 [MPa] C11 [MPa] C02 [MPa]
−80.06794103 83.6836129 7007.799701 −15883.48274 9091.217841
“Mooney-Rivlin 9 Parameter” material constants
C10 [MPa] C01 [MPa] C20 [MPa] C11 [MPa] C02 [MPa]
−101.5888166 105.4101781 −334637.9565 669310.0379 −334374.4965
C30 [MPa] C21 [MPa] C12 [MPa] C03 [MPa]
9188538.891 −30466113.39 33791889.23 −12482791.23
4.2 Validation of Material Model

After the definition of the material using the experimental data, the model was vali-
dated. The validation process represented a simulation of the tensile test. The output of
the validation was the force-elongation curve which was compared with the
force-elongation curve from the experimental test.
The force-elongation curve resulted from the simulation, validated the
“Mooney-Rivlin 5 Parameter” model and it highlighted the fact that it is an accurate
model even if it did not reach the same elongation as in the experimental test. For the
EPP59, in the experimental phase, the maximum elongation of 12.19 mm was obtained
while in the simulation the elongation was 11.4 mm, resulting in a difference of
0.79 mm. For the EPP270, the maximum elongation of 4.63 mm was achieved in the
test while in the simulation the elongation was 5.38 mm, resulting in a difference of
0.75 mm. Worth mentioning is the fact that for high stress-low strain materials, the
model is not as good as for low stress-high strain (Fig. 6) due to a big curve offset,
although in both cases there is a difference between the real and simulated maximum
elongation.
Fig. 6. EPP59 validation (left) and EPP270 validation (right)

5 Finite Elements Analysis
For the analysis with finite elements, a box was designed with the intent to be used in
courier transportation and to be handled by one person. This type of boxes made from
expanded polypropylene is gaining popularity in merchandise distribution and small
goods delivery. The dimensions of the box used in the simulation are
580 × 380 × 305 mm (L × W × H) without the sealing top and 40 mm wall thickness.
These dimensions permit easy handling of the box by a single person.
5.1 Material Properties, Meshing and Boundary Conditions

In what concerns the material, the EPP with the lowest density was considered as
reference in the simulation due to the fact that it has the lowest cost. Also, EPP59 has
the lowest mechanical strength highlighted in the tensile test, with a yield strength
estimated at 0.2 MPa and an average ultimate tensile strength of 0.67 MPa.
The mesh used was a hex-dominant mesh which provides accurate results with a
reasonable amount of elements that leads to an efficient computational effort and time.
The meshing and analysis were performed in ANSYS Workbench 15. The analysis was
set as static structural with “Large deflections” option activated. Very important for
such materials is the meshing because of the large deflections that can occur and cause
convergence issues. To avoid such problems, the meshing was done using an
“Aggressive mechanical” quality control algorithm to obtain better quality elements.
For the boundary conditions, the case in which the box is handled by one person
was considered. The load used is the maximum allowed load for handling by men of
15 kg, accordingly to the Romanian legislations. In order to simulate the contact patch
of the hand on the base of the box, the base face was split to create two rectangular
areas in which a fixed support was set (Fig. 7).
Fig. 7. Boundary conditions
5.2 Results
The simulation highlighted the fact that the maximum principal stress also known as
tensile stress and the minimum principal stress also knows as compression stress are
located in the geometrical line determined by the handlers fingers (Fig. 8).
Fig. 8. Tensile stress (left- top view) and compression stress (right- bottom view) results
From the values of the stress, it can be mentioned the fact that the box made from
the lowest density EPP is suitable for the application and it only experiences elastic
deformation without a risk of failure considering that the yield point was determined at
0.2 MPa and the tensile stress has a maximum value of 0.011 MPa and the com-
pression stress has a maximum value of 0.028 MPa.
In order to compare the impact of density on product development, a simulation
with EPP270 was also performed and the result showed the fact that with EPP270 the
deformation was 0.029 mm while the EPP59 experienced a deformation of 0.16 mm,
EPP270 being able to reduce the deformation with 82 % (Fig. 9).
Fig. 9. EPP59 total deformation (left) and EPP270 total deformation (right)
6 Conclusions
Expanded polypropylene is a material that is gaining popularity due to its good

mechanical properties and low weight. In the future, it is expected to increase its market
share, especially in the transportation industry where it is used for transportation boxes.
Regarding the application of the finite element method, EPP is a material that is
difficult to model with finite elements due to the fact that it exhibits a nonlinear
behavior and presents large deflections. It can be mentioned the fact that for modeling
with finite elements, the “Mooney-Rivlin” is an adequate mathematical model that
gives accurate results with the observation that it is very good for low-density EPP
while for high-density EPP it exhibits some differences between experimental and
simulation data. Also, the “Polynomial” hyperelastic model has proven its suitability
for EPP but has an additional parameter in comparison with the “Mooney-Rivlin”, a
parameter that can significantly increase the computational time. Although the literature
presents some material models classification using as classification criteria the accuracy
of the fit, the classifications should be used only as a source of information and the
decision regarding the best model should be taken based on the fitting of the experi-
mental data obtained.
Another aspect is that, in order to perform a simulation with nonlinear materials, the
experimental data is very important as it represents the starting point of the simulation
and in order to obtain good results it is necessary to pay extreme attention to the tensile
tests and to the implementation of the experimental data into the mathematical model
and also to perform the material model validation to early detect convergence problems
caused by improper test data.
As the results of the simulation have showed, expanded polypropylene with a
density of 59 kg/m3 represents a material that is strong enough for the application and
is suitable for the manufacturing of a transportation box because it also is lightweight
and can be easily manipulated and also can be easily produced using a steam molding
process based on beads.
Despite of a reduced deflection of the box in the case in which EPP with a density
of 270 kg/m3 was used, a higher density is not justified in low loads transportation and
would only represent over-engineering and increase the costs of the product.
High-density polymeric foams should be taken into consideration when the require-
ments of the application involve high loads and manipulation shocks like those present
in forklift manipulation. Nevertheless, the simulation using a higher density material
has revealed what are the benefits of a more dense and expensive EPP from the
deformational point of view.
The present paper certifies the fact that expanded polypropylene is an adequate
material for designing transportation boxes and not only, and also highlights the role of
material density in product design and how it can impact the behavior of a product.
References
1. Kang, S.W., Kim, S.K., Yun, S.L., Ban, H.J., Lee, W.T., Kwak, P.J.: Field application of
particle filtration with EPP (expanded polypropylene) engineered media of road runoff.
Mater. Sci. Forum 695, 614–617 (2011)
2. Zhang, X., Andrieux, F., Sun, D.: Pseudo-elastic Description of Polymeric Foams at Fnite
Deformation with Stress Softening and Residual Strain Effects. Elsevier, Amsterdam (2010)
3. Bergström, J.: Mechanics of Solid Polymers Theory and Computational Modeling, pp. 236–
275. Elsevier Inc, San Diego (2015)
4. Marckmann, G., Verron, E.: Comparison of hyperelastic models for rubber-like materials.
Rubber Chem. Technol. 79(5), 835–858 (2006)
5. Treloar, L.R.G.: The elasticity of a network of long chain molecules (I and II). Trans.
Faraday Soc. 39(36–64), 241–246 (1943)
6. Arruda, E.M., Boyce, M.C.: A three-dimensional constitutive model for the large stretch
behavior of rubber elastic materials. J. Mech. Phys. Solids 41, 389–412 (1993)
7. Gent, A.N.: A new constitutive relation for rubber. Rubber Chem. Technol. 69, 59–61
(1996)
8. Blatz, P.J., Ko, W.L.: Application of finite elastic theory to the deformation of rubber
materials. Trans. Soc. Rheol. 6, 223–251 (1962)
9. Ogden, R.W.: Large deformation isotropic elasticity—on the correlation of theory and
experiment for incompressible rubberlike solids. Proc. R. Soc. Lond. A Math. Phys. Sci.
326, 565–584 (1972)
10. Yeoh, O.H., Fleming, P.D.: A new attempt to reconcile the statistical and phenomenological
theories of rubber elasticity. J. Polym. Sci. Polym. Phys. 35, 1919–1931 (1997)
11. Yeoh, O.H.: Characterization of elastic properties of carbon-black-filled rubber vulcanizates.
Rubber Chem. Technol. 63, 792–805 (1990)
12. Mooney, M.: A theory of large elastic deformation. J. Appl. Phys. 11(9), 582–592 (1940)
13. Rivlin, R.S.: Large elastic deformations of isotropic materials. IV. further developments of
the general theory. Philos. Trans. R. Soc. Lond. Ser. A Math. Phys. Sci. 241(835), 379–397
(1948)
14. Rivlin, R.S., Saunders, D.W.: Large elastic deformations of isotropic materials VII.
experiments on the deformation of rubber. Philos. Trans. R. Soc. Lond. Ser. A 243(865),
251–288 (1951)
Functional Description and Design Aspects
of Hybrid Transmissions
Martin Vornehm(&)
LuK GmbH & Co. KG, Buehl, Germany

martin.vornehm@schaeffler.com
Abstract. This paper describes a method used at Schaeffler to analyze inte-

grated hybrid transmission architectures with one electric motor. Both the
properties of their modes and aspects of mode transitions become clear using
special visualization techniques. The method was implemented in MATLAB,
and several transmission examples will be discussed. This leads to design
aspects and requirements concerning the overall transmission layout and the
components. Examples of beneficial components from Schaeffler are discussed.
Small shifting elements allow more space for the electric motor. Clutches,
brakes and CVT subsystems that are the right size for hybrid transmissions
provide the required comfort. Low-friction bearings and on-demand actuators
contribute to increased range of electric driving.
Keywords: Hybrid Transmission Architecture Modes Mode transitions

CVT Dimensioning Clutch Actuator Bearing PHEV MATLAB DHT
1 Introduction
Hybridization and electrification are important trends in the automotive industry that
have been driven by CO2 or other emission targets and upcoming changes like auto-
mated driving. The expected increase in market share from 2 % hybrid drivetrains to
20 % in 2020 will partly come from plugin hybrid electric vehicles (PHEV) [1]. These
incorporate a large battery and a powerful electric machine, but also represent a
paradigm change: the transmission is the key to the engines of the future and not vice
versa [2]. Among the many hybridization options, some solutions are preferred, and
suppliers can provide hybrid components and systems for these solutions.
As shown in Fig. 1, preferred solutions cluster around local Pareto optima. Beyond
each optimum, adding more effort (components or electrical power) will only slightly
increase the benefit, until a new driving functionality or drivetrain simplification
becomes available. A first optimum is 48 V technology near 15 kW, which allows CO2
reduction and basic electric maneuvering. Beyond 40 kW, high voltage P2 systems
represent another preferred solution, since they also provide the required performance
for electric city driving and improved vehicle dynamics. A third, still unclear, optimum
at even higher power is expected when a reduced number of gears becomes viable.
These powerful electric motors can be placed closer to the transmission output, where
the classification of Fig. 2 indicates several variants like P2.5 and P3. The complexity
increases even more, because multi-mode mixtures of these variants are also possible.

Functional Description and Design Aspects of Hybrid Transmissions 37
Fig. 1. Pareto diagram of hybrid drivetrains
Fig. 2. Positions P0 – P4 in the drivetrain of an electric motor, Clutch types K0 – K2 of several

hybrid drivetrains. The icons on the right represent some variants of the transmission-integrated
P2.5 or P3 configurations that contribute specific functional modes each
Recent overviews [3] have begun to collect and classify the increasing number of
proposed transmission architectures and defined the term “dedicated hybrid transmis-
sion” (DHT) for a subclass of hybrid transmissions. A DHT’s functionality depends on
the electric machine and without the electric machine, some requirements are not
fulfilled.
The remaining challenge of hybrid transmissions is that gears – conventionally
characterized by their ratios, i.e. by one number each – must become operating modes
(with several characteristic properties per mode). A holistic description in particular is
required for the new types of operating modes “continuous speed modes,” “charging/
sailing modes,” “pure E-modes” and “pure ICE-modes” as well as various types of mode
changes.
38 M. Vornehm
2 Generalization from Non-hybrid to Hybrid Modes
A non-hybrid drivetrain connects an internal combustion engine (ICE) and an

output-shaft (OUT), which means a two dimensional state-space of rotation speeds.
Each gear is a line in this state-space and can be characterized by its slope or its ratio.
Thus, each gear can be depicted by one point on a ratio scale, and several gears by
several points. This is shown on the left in Fig. 3 with a nonlinear vertical axis.
The state-space of torques could be depicted in the same way and can similarly be
described by the torque-ratio as a single characteristic number per gear. From these two
characteristic values (speed-ratio and torque-ratio), one characteristic in a last step can
be omitted when the fundamental law of power conservation is applied (neglecting
losses). Engineers are thus accustomed to identifying a gear using primarily a single
number: its ratio.
A hybrid drivetrain, combining an ICE, an electric motor (EM) and an OUT to
drive a vehicle, offers for each mode a subspace of the state-space spanned by these
three speeds plus the three torques of ICE, EM and OUT. As in the non-hybrid case,
both speed and torque space have a linear symmetry, meaning that whenever the
drivetrain can work in some operating mode at speeds of ICE, EM and OUT, it will
mathematically work as well with any multiple of this speed triplet.
The speed-operating modes can thus be represented by their orientations, i.e. by
points on a unit sphere. A suitable projection from this sphere onto a flat map is shown
in Fig. 4. This projection always yields straight lines for grids and for operating modes.
The axes, however, are nonlinearly scaled. Two gears are shown in Fig. 4, with each
forming a dotted beam with a characteristic orientation. Each of these creates one
characteristic spot on the projection plane, which is the two dimensional diagram seen
on the right in Fig. 3.
Fig. 3. Characteristics of non-hybrid transmissions, and expansion to the hybrid case (C = pure
combustion engine mode with electric motor disconnected, V = variable e-CVT mode,
S = sailing/charging/cranking mode with ICE and EM functionally connected but transmission
output functionally disconnected, E = pure electrical mode with ICE functionally disconnected)
Fig. 4. State-space of the speeds in hybrid transmissions and the projection plane for its modes
Again, a similar idea can be applied to the torques, leading to a torque-sphere and a
torque-map. In this way, the six dimensions (three speeds and three torques) can be
reduced to four projected characteristics (two for speed ratios and two for torque
ratios). In a last step, the power conservation law helps reduce the number of char-
acteristic dimensions from four to three.
When looking at only two-dimensional diagrams, this last step requires attention:
any operating mode of a hybrid transmission has one degree of freedom. If the oper-
ating mode has fixed speed ratios (depicted as dots in Fig. 3), then this freedom exists
in torque (power). Alternatively the operating mode can have continuously variable
speed ratios (the line-shaped modes shown on the right in Fig. 3). In this case the
torques have to obey some constraint, e.g. the torque-balance-law on a planetary gear
set or a zero-torque-condition for a disconnected EM, OUT or ICE.
3 Examples of Graphical Representations for Operating

Modes
Some examples will clarify the difference between the various P-types and their
mixtures. The pictures shown here are MATLAB figures and were computed by an
analysis program using symbolic mathematics. The program input defines the trans-
mission structure by a syntax-like (not a matrix-like) description of the transmission’s
graph. Based on this structural definition, the linear equation systems for speeds and
torques are formed and solved of all modes. The characteristics of these solutions are
then displayed or numerically evaluated.
3.1 P2-Double-Clutch-Transmission (DCT)

Remembering Fig. 2, a P2-DCT with six speeds has the map of operating modes
shown in Fig. 5. The fixed ratios indicated by black dots and numbered 1-6/R are in the
X position = 1, meaning the same speeds for EM and ICE. These operating modes
allow continuous torque-ratios for load-point shifting of the ICE to improve fuel
efficiency.
40 M. Vornehm
Fig. 5. Hybrid mode diagram of a P2 – double-clutch-transmission
Disconnecting the ICE by disengaging a K0 clutch (see small icon) creates pure
electric driving modes (depicted by slanted lines). These modes are lines, because the
disconnected ICE can actually run at any speed including shut-down. The relevant
property of these purely electric drive-modes is their slope: the more straight the line,
the higher the ratio of the E-mode and thus the electric drive force.
This type of transmission also offers a charging/sailing-mode, which is depicted by
the vertical line at X = 1. In this mode, only the ICE and the EM are connected via K0, and
the rest of the drivetrain and gears are in neutral. Consequently, there is power exchange
only between ICE and EM, for example to charge the battery or to crank the ICE.
As is conventional for diagrams like Fig. 5, the ratios and values used do not
include the final-drive ratio.
3.2 DCT with Integrated Hybrid

A widely known alternative to the P2-DCT is to place the motor on one of the DCT
input shafts – either on the odd gears or on the even gears [4]. Both are often called
P2.5-DCT. Additional classifications are design-based with coaxial or non-coaxial
electric machine [5]. The diagrams corresponding to these two types of transmissions
are shown in Fig. 6. There are pure electric modes (slanted lines), pure ICE modes
(horizontal lines), the sailing/charging mode (vertical line) and a number of discrete
Fig. 6. Schematic operating modes of the two types of P2.5 – DCT

combinations of ICE and EM-speeds (dots) where load-point shifting is possible. Gear
numbers like “4/3” represent fourth gear for the ICE, but third gear for the EM.
Compared to the P2–DCT, a benefit is that the K0 clutch (Fig. 2) and its actuator can
be eliminated. Although the double-clutch allows many comfortable mode-changes,
some of them require more complex control, like preselect or intermediate modes.
3.3 P3-DCT
With the electric motor on the transmission output shaft, the six speeds show up as
numbered points, but now the motor runs at different speeds compared to the ICE
except for fifth gear. The slanted line in Fig. 7 corresponds to the only one purely
electric drive mode.
With a “small” motor, this simple P3 has a known target conflict, namely, the
electric mode would require a high ratio and a more straight line in Fig. 7 for sufficient
wheel torque. However, this would also mean very high speeds of the electric motor
when driving fast. Consequently, there is a tendency to use either a powerful electric
motor for P3, or more than just one electric gear.
Also, without additional disconnect devices, there is no charging/sailing mode
(vertical line).
Fig. 7. P3 – DCT transmission operating modes
3.4 Integrated AT-Hybrid with Several Modes and e-CVT

Using planetary gears and placing the motor anywhere inside the transmission adds
modes to the diagram. To explain the effects, a 9-speed transmission was chosen and
the electric motor was virtually placed in parallel to one of the existing transmission
brakes. With that brake released, a purely electric mode is created that allows
launching. Also, a new type of operating mode appears, often called e-CVT. In Fig. 8
these are the three other diagonal lines that connect some of the dots. It is noteworthy
that there is a difference between real CVTs and torque-balanced e-CVTs. Real CVTs
allow any power flow at any speed-ratio through the CVT, whereas e-CVTs only allow
restricted stationary power-flow-types at given speed-ratios. There are zones in which
the comfort-relevant power-flow to the transmission output is limited by this.
42 M. Vornehm
Fig. 8. Planetary-AT with integrated hybrid and several additional modes
4 Operating Mode Transitions
An intuitive understanding of how to use and change operating modes is to “jump” or

“walk up” in the diagram during vehicle acceleration. The new aspect in hybrid
transmissions is that sometimes fixed ratio operating modes (dots) are connected by
other operating modes (lines).
A vehicle launch is an example of this kind of transition. In Fig. 8, a straight line
(electric mode) allows the vehicle to be electrically launched without using the ICE or a
launch clutch. When a sufficiently high vehicle speed has been reached, the combustion
engine can be started. After adjusting its speed to that of the first gear, the shifting
element can engage the first gear (dot). Subsequently, all control freedom of a parallel
gear mode can be applied such as boost, recuperation battery recharging. In some cases
like in Fig. 8, both first and second gears are on the E-mode line. Depending on the
battery state of charge or the vehicle speed already reached, there are thus several
options for how and when to activate the ICE as support.
Another standard type of shift-path detail can be explained in Fig. 6 on the right:
gears 1/2, 2 and 3/2 (the variants of gears 1, 2 and 3 with electric ratio staying in second
gear) are connected by the line of a purely electric driving mode, which, by means of a
sufficiently powerful motor, lets friction torque handover be avoided. The same is true
for gears 3/4, 4, 5/4, but with a different electric mode (line). Depending on the next
gear change expected, gear 3/2 or 3/4 has to be preselected, which can be done without
torque interruption by using the horizontal purely combustion-engine-mode and its
torque.
If quantitative criteria on traction force and power limits are fulfilled, also e-CVT
modes can be helpful for transitions from one “dot” to the other “dot”. This can allow
friction-clutches/brakes to be replaced by dog-clutches/brakes. Wherever no such line
connects two “dots”, friction-elements have to perform torque-handover and
speed-synchronization for comfortable gear changes.
The only transmission design in which changing ratios is automatically comfortable
is the hybrid CVT like [6]. As CVTs are very smooth, they match nicely with electric
driving, which is also very comfortable. Due to this, any remaining mode change
requires augmented focus and control effort. In multi-mode hybrid CVT, the variability
of the CVT helps to optimize the efficiency of both the combustion engine and the
electric system.
5 Design Aspects
Packaging an electric motor with more than 80 kW into a transmission requires

designers to focus on the following aspects for both inline and transverse layouts:
• The minimum length of all components
• The minimum power losses of all components, including actuators
In addition or as a consequence, a disruptive change in transverse transmission
layout is an option, as shown in Fig. 9. With the given engine-interface and differential
position, hybridization inside the bell housing is straightforward, but limited in length.
An increase in electrical power beyond that limit requires one of these things: (1) the
combustion engine is shrunk or moves (Fig. 9, left path); (2) the transmission gears and
motor swap positions (middle path); (3) the electric motor moves off-axis (right path);
or (4) addition of an electric axle instead of an integrated hybridization.
All of these variants have been published as designs [4], or are already on the
market [7].
Fig. 9. Fundamental design options for transverse hybrid drivetrains when increasing the
electrical power. Integrated hybrid transmissions represent the three left options
6 Clutches/Brakes and Their Dimensioning
Besides the potential of replacing friction clutches/brakes by synchronized shifting

elements or even by dog clutches, the capacity of the friction elements requires a more
detailed assessment in hybrid transmissions.
A simplified version of today’s procedure for the torque dimensioning of a clutch or
brake is shown in Fig. 10 on the left. For all gears the required torque capacity (either
absolute or relative to the ICE torque) is calculated, and the maximum is selected as the
basis for the geometrical design (here: 65 %, due to gears 1 or R).
In the hybrid case (shown on the right in Fig. 10), this maximum must be taken
from a matrix of gears (lines) and electric torque supports (columns). In some of the
gears like gear 1, the torque from the electric motor has no influence on the required
44 M. Vornehm
Fig. 10. Current clutch/brake dimensioning process and for integrated hybrid transmissions
clutch capacity. In other gears like gear 7 or reverse, there is such an influence. The
result will usually be a higher required capacity of the clutch or brake (here: 82 % due
to gear R).
For hybrid transmission development to be successful, knowledge of the system
and calculations beyond the standard must be combined with detailed optimization of
the friction elements, like the groove pattern, knowledge of the facing material or
drag-torque minimization by separation assisting designs.
6.1 Short Shifting Components

Every millimeter of length is worth many hundred Watt of electrical motor-power.
Short shifting elements can thus make the difference between the “fun of electric drive”
and “lame driving behavior.” Recent Schaeffler designs and prototypes [8] have
reduced the length of layshaft gear shifting systems from 38 mm to less than 30 mm
(Fig. 11), improvements that can be considered enablers for integrated hybridization.
Fig. 11. Synchronization unit with reduced length for manual, automated and hybrid
transmissions
6.2 Bearings
The importance of friction reduction in bearings is often underestimated, but this kind
of reduction is always linked to fuel consumption, lubrication and cooling. For electric
driving, friction reduction is also related to the capacity and cost of the battery.
For a vehicle intended to drive 1 h electrically (say 30 miles), each Watt of drag
power means spending money with the weight and space for 1 Wh of ineffective
battery capacity. The identification of the optimum bearing design in combination with
detailed optimization, like the bearing-raceways for instance [9], is therefore a
worthwhile endeavor, since this can reduce the losses in the entire transmission by
100 W or more.
A surprisingly high friction reduction can be achieved by using short needle roller
bearings in place of planetary thrust washers. The needles in these bearings (Fig. 12)
have a diameter of only 1 mm. As reported in [10], this can reduce CO2 emission by
1 g/km or save about 40 W.
Fig. 12. Short axial needle bearing integrated in a planetary set
6.3 On-demand Actuation

For friction elements and shift elements, electromechanical actuation systems show
impressive performance and energy efficiency with as low as 20 W average power
consumption for the entire actuation of a DCT [11]. Local control units such as those in
the applications shown in Fig. 13 are beneficial for hybrid transmissions because they
can contribute to distributed safety designs. Optimized system combinations of actuator
and actuated component are one key factor for any kind of future transmission.
46 M. Vornehm
Fig. 13. Schaeffler shift actuators, clutch actuators and clutches in mass produced DCT
7 Summary
Hybrid transmissions represent a rewarding development direction for modern drive-

trains to meet current and future demands. In addition, hybrid transmissions are also
challenging in terms of complexity, system development and component expertise.
Entering this field requires a supplier and partner, who combines production and
development competencies at all system levels. This includes the abstract architecture
assessment, an understanding of the overall layout, subsystem design as well as opti-
mized components and their intricate details.
References
1. Wagner, U.; Faust, H.; Scheidt, M.; Kraus, M.: Die Vielfalt im Antriebsstrang der Zukunft
managen (Managing the diversity of future drivetrains), MTZ 76(12), 18–24. Springer
Vieweg, Wiesbaden (2015)
2. Gutzmer, P.: The transmission is the key to the engines of the future. In: 36th International
Engine Symposium, Vienna, 7 May 2015
3. Fischer, R.: Dedicated hybrid transmissions – a new category of transmissions. In:
Proceedings of the 14th International CTI-Symposium, Berlin, 7 December 2015
4. Blessing, U., Meissner, J., Schweiher, M., Hoffmeister, T.: Scalable hybrid dual-clutch
transmission. ATZ worldwide 116(12), 4–9 (2014)
5. Fischer, R., Schneider, G., The XSG family – dry clutches and electric motors as core elements
of the future automated gearbox. In: 7th LuK Symposium Baden-Baden, pp. 173–188, 11 April
2002
6. Lauinger, C., Götz, A., Walter, B., Finsterbusch, M., Simon, E.: CVT hybrid – low emission,
high performance. In: 16th Congress Drivetrain in Vehicles, Friedrichshafen, VDI-Berichte
2276, pp. 827–845, 22 June 2016
7. Jelden, H.; Philipp, K.; Weiss, N.; Kessler, A.: Der plug-in hybrid des modularen
querbaukastens von volkswagen (The plug-in hybrid of volkswagen’s modular transversal
platform), MTZ 75(4), 40–47, Springer Vieweg, Wiesbaden (2014)
8. Utz, J., Lichtenwald, V., Hirt, G.: Short synchro (STSS) – reduced shift travel for MT’s and
DCT’s. In: 16th Congress Drivetrain in Vehicles, Friedrichshafen, VDI-Berichte 2276,
pp. 231–242, 22 June 2016
9. Petery, G.; Zintl, G.: Optimized transmission solutions with innovative products. In: 3rd CTI
Symposium, Suzhou, 17 December 2014
10. Gegner, M.; Beeck, F.: Turning new directions: surprising potentials in planetary
transmissions. In: 14th Congress Drivetrain for Vehicles, Friedrichshafen, 24 June 2014
11. Goeckler, M.: Transmission actuation – further development of existing systems and new
application areas. In: 2nd VDI Conference on Control Solutions for Transmissions,
Friedrichshafen, VDI-Berichte 2256, p. 771ff, 16 June 2015
Comparative Study on the Performances
of Aerodynamic Devices Used in Decreasing
of the Automobiles Lift Force
Angel Huminic1,2(&), Gabriela Huminic1, Marius Carauleanu1,

and Constantin Ciolofan2
1
Aerodynamics Laboratory, Transilvania University of Brasov,
Brașov, Romania
{angel.h,gabi.p}@unitbv.ro
2
R&D Department, INAS Craiova, Craiova, Romania
{angel.h,ciolofan}@inas.ro
Abstract. Being a continuous subject of research, this study presents new

results concerning devices used to generate downforce and their influence in
road vehicle aerodynamics. Thus, using a mass production car model, the study
examines the effects of a rear wing and a short underbody diffuser on the main
aerodynamic characteristics, lift and drag. The results show that the lift of the
studied car decreases due to both diffuser and rear wing, more significantly due
to the later one. Concerning the drag, it decreases only in the case of underbody
diffuser.
Keywords: Automobile Aerodynamic devices Lift Drag
1 Introduction
In order to improve the performances of the modern cars regarding fuel consumption,
stability and handling, there are used various aerodynamic devices. Most used in this
field are the rear wing and the underbody diffuser, which have the potential to reduce
significantly the lift, also the drag in some cases.
These devices have been extensively studied as separate component, using either
generic car bodies or real car models, and the obtained results were published in many
articles and research studies.
Concerning the flow through the underbody diffusers, previous reported works have
been performed by Buchheim et al. [1], George [2], Howell [3], Cooper et al. [4, 5],
Senior and Zang [6], Zhang et al. [7] and Huminic et al. [8, 9]. All of these studies are
emphasizing the influence of the underbody diffusers in the mechanism of generation
and increasing of downforce. Some analytical models were also proposed in order to
facilitate the selection of optimum geometry [5, 9].
Regarding the effect of the rear wings on road vehicles aerodynamics, several
studies were published by Katz and Dykstra [10–12], and Diasinos and Gatto [13]. The
relevance of the wings in this field is revealed also by a large number of patents in the
international databases.

Comparative Study on the Performances of Aerodynamic Devices 49
Thus, being a continuous subject of research, this study examines experimentally

the effects of a rear wing and a short underbody diffuser on the main aerodynamic
characteristics of a mass production car.
2 Experimental Setup
The experimental study was performed using a 1:5 scale model of an automobile
shown in Fig. 1 (side view), with the following main dimensional characteristics:
length l ¼ 0:7 m, width w ¼ 0:27 m and the reference aria Aref ¼ 0:065 m2.
In the first stage were analyzed the influence of two short underbody diffusers on
main aerodynamic characteristic, drag and lift coefficients, relative to the base
configuration:
– simple plane underbody diffuser having the angle ad ¼ 20
– plane underbody diffuser filleted to the surface of car body with the radius
r ¼ 10 mm.
The ration between diffuser length (ld ) and length of the body (l) was ðld =lÞ ¼ 0:15,
which corresponds to a relative short underbody diffuser.
Fig. 1. Studied diffusers: (a) simple plane underbody, (b) plane underbody diffuser filleted to the
surface of car body
In the second stage, a rear wing was added to the model. It was placed at a distance h
relative to the body, as depicted in Fig. 2. The section of the aileron was a circular arc of
radius R ¼ 55 mm with the chord c ¼ 55 mm. The span of the wing was b ¼ 200 mm.
Relative to the base configuration of car, there were performed experiments for
three values of the ratio h=l:
50 A. Huminic et al.
Fig. 2. Studied car with rear wing
Fig. 3. Wind tunnel view and detailed of the tested car model within test chamber
h 15 h 25 h 35
¼ ; ¼ and ¼ :
l 700 l 700 l 700
The tests were conducted using the infrastructure of the Aerodynamics Laboratory,
Transilvania University of Brasov, within the wind tunnel (see Fig. 3) having a closed
test section of 1.2 m × 0.6 m × 1.2 m, maximum velocity of 40 m/s and turbulence
lower than 0.5 pct. It has a four component strain-gauge balance with a PC-based
system for data acquisition and a moving belt device for ground effect simulation.
Certified metrological equipments are used for various measurements concerning road
vehicle aerodynamics (1:5 scale models), small wind turbines, aviation structures and
active flow control using Coandă effect.
The current experiments were performed for a reference air velocity vair ¼ 120
km/h (vair ¼ 33:3 m/s), which correspond to a Reynolds number Re ¼ 1:45 106
computed with the length of model l ¼ 0:7 m.
3 Results and Discussions
The results concerning the influence of the underbody diffusers on the main aerody-
namic characteristics, drag coefficient (CD ) and lift coefficient (CD ), Eq. (1), are
recorded in the Table 1.
D L
CD ¼ q 2 ½; CL ¼ q 2 ½; ð1Þ
air vair air vair
2 Aref 2 Aref
where D and L are the drag and lift measured on the car model, and qair is the
density of the air during measurements.
Table 1. The underbody diffusers influence on drag and lift

Base Car with simple Car with
configuration ofplane underbody underbody
the car diffuser diffuser filleted
to the surface of
the car
CD CL CD CL CD CL
0.3961 −0.0705 0.3892 −0.0974 0.3863 −0.1032
Having as reference the values of the base configuration, CDref ¼ 0:3961 and
CLref ¼ 0:0705, were recorded the following relative variations of the aerodynamic
coefficients, Eq. (2), as presented in Table 2.
Table 2. Relative variations of drag and lift concerning the underbody diffuser
Car with simple Car with
plane underbody underbody
diffuser diffuser filleted
to the surface of
the car
DCx ½% DCz ½% DCx ½% DCz ½%
−1.74 −38.15 −2.47 −46.38
CD CDref CL CL
DCD ¼ 100½%; DCL ¼ ref 100½%: ð2Þ
CDref CLref
Thus, the results show that it is possible to decrease the aerodynamic characteristics
of vehicles using relatively short underbody diffusers. The decreasing is more signif-
icant for the lift coefficient, about 40 %. This value could be increased more if the
underbody is filleted to the body of car.
The results concerning the influence of the rear wing on drag and lift are recorded
in the Table 3.
Table 3. The rear wing influence on drag and lift

Base Base configuration of the car, with rear wing
configuration of h ¼ 15
l ¼ 700 l ¼ 700
h 25 h 35
l 700
the car, without
rear wing
CD CL CD CL CD CL CD CL
0.3961 −0.0705 0.4363 −0.0846 0.4602 −0.0909 0.4696 −0.0928
As shown, the lift is decreasing due to the rear wing in a similar way as for the
underbody diffuser, but in this case this positive effect is accompanied by an increase in
drag, which leads to an increase of the fuel consumption, too. Also, the downforce and
drag due to the rear wing are increasing with the value of ratio h=l.
Having as reference the values of the base configuration, were recorded the fol-
lowing relative variations of the aerodynamic coefficients, DCD and DCL , as shown in
Table 4.
Table 4. Relative variations of drag and lift concerning the rear wing
h
l ¼ 700
15 h
l ¼ 700
25 h
l ¼ 700
35
DCD ½% DCL ½% DCD ½% DCL ½% DCD ½% DCL ½%
10.15 −20.00 16.18 −28.94 18.55 −31.63
Hence, from aerodynamic point of view, the results show that the configuration of
car with underbody diffuser is more favorable than the solution with rear wing, because
the later one also leads to an increase of the drag.
Figures 4 and 5 depict the flow pattern in the wake of car without and with rear
wing. As shown, in the first situation there is a large vortex behind the car, which flows
toward the ground and dissipates faster than the second situation, where an additional
smaller vortex can be seen due to the presence of the wing, which also contributes to
the increase of drag.
Fig. 4. Flow pattern for car without rear wing
Fig. 5. Flow pattern for car with rear wing, h=l ¼ 35=700
4 Conclusions
The study examines experimentally the effects of a rear wing and a short underbody
diffuser on the main aerodynamic characteristics of a mass production car.
The results show that it is possible to decrease the lift of vehicles using both
(relatively short) plane underbody diffuser and rear wing. The efficiency of the diffuser
could be increased significantly if the underbody is filleted to the body of car.
The underbody diffuser also leads to the decrease of drag, unlike the rear wing,
which causes an increase in drag, and consequently an increase of the fuel consump-
tion. Both solutions can be used simultaneously if it is necessary to increase signifi-
cantly the downforce and stability, and the drag reduction is not the main aerodynamic
goal.
References
1. Buchheim, R., Deutenbach K.R., Laxckoff, H.J.: Necessity and premises for reducing the
aerodynamic drag of future passenger cars. SAE Technical Paper 810185 (1981). doi:10.
4271/810185
2. George, A.R.: Aerodynamic effects on shape chamber, pitch, and ground proximity on
idealized ground vehicle body. ASME J. Fluids Eng. 103, 631–638 (1981)
3. Howell, J.P.: The influence of a vehicle underbody on aerodynamics of a simple car shapes
with an underfloor diffuser. In: Vehicle Aerodynamics, R.Ae.S. Conference, Loughborough,
pp. 36.1–36.11 (1994)
4. Cooper, K.R., Bertenyi, T., Dutil, G., Syms, J., Sovran, G.: The aerodynamic performance of
automotive underbody diffusers. In: SAE Technical Paper 980030 (1998). doi:10.4271/
980030
5. Cooper, K.R., Sovran, G., Syms, J.: Selecting automotive diffusers to maximise underbody
downforce. In: SAE Technical Paper 2000-01-0354 (2000). doi:10.4271/2000-01-0354
6. Senior, A.E., Zhang, X.: The Force and pressure of a diffuser-equipped bluff body in ground
effect. ASME J. Fluids Eng. 123, 105–111 (2001)
7. Zhang, X., Senior, A., Ruhrmann, A.: Vortices behind a bluff body with an upswept aft
section in ground effect. IJ Heat Fluid Flow 25, 1–9 (2004)
8. Huminic, A., Huminic, G.: Computational study of flow in the underbody diffuser for a
simplified car model. In: SAE Technical Paper 2010-01-0119 (2010). doi:10.4271/2010-01-
0119
9. Huminic, A., Huminic, G., Soica, A.: Study of aerodynamics for a simplified car model with
the underbody shaped as a Venturi nozzle. IJ Veh. Des. 58(1), 15–32 (2012)
10. Katz, J., Dykstra, L.: Study of an open-wheel racing-car’s rear-wing aerodynamics. In: SAE
Technical Paper 890600 (1989). doi:10.4271/890600
11. Katz, J., Dykstra, L.: Effect of wing/body interaction on the aerodynamics of two generic
racing cars. In: SAE Technical Paper 920349 (1992). doi:10.4271/920349
12. Katz, J., Dykstra, L.: Effect of passenger car’s rear deck geometry on its aerodyanamic
coefficients. ASME J. Fluids Eng. 114(2), 186–190 (1992)
13. Diasinos, S., Gatto, A.: Experimental investigation into wing span and angle-of-attack effects
on sub-scale race car wing/wheel interaction aerodynamics. J. Exp. Fluids 45(3), 537–546
(2008)
Analysis of the Influence of the Drive Force
Distribution Between Axles on an Automobile
Stability in Its Curvilinear Motion
Andrzej Reński(&)
Institute of Vehicles, Warsaw University of Technology, Warsaw, Poland

arenski@simr.pw.edu.pl
Abstract. In the case of vehicles with four-wheel drive, it is important how the
driving torque is distributed between wheels of the front and rear axles. The
target of the work was to develop the method of control of driving forces
distribution in order to increase the vehicle directional stability, particularly in
critical states of its motion.
The paper presents the results of calculations conducted according to ISO
standard 4138 “Steady-state circular driving behavior”. Results of calculation
showed that in particular riding conditions distribution of drive forces can have
significant influence on a stability characteristics of the vehicle, namely its
characteristics can be changed from understeer to oversteer or vice-versa. This
effect can be practically utilized in a control of vehicle directional stability.
Keywords: Automobile directional stability Driving forces distribution Tire

slip characteristics Understeer Oversteer
1 Introduction
Longitudinal forces (driving forces or braking force) in addition to lateral forces can
have a significant effect on the automobile stability in its curvilinear motion. Thus, in
the case of vehicles with four-wheel drive, it is important how the driving torque is
distributed between wheels of the front and rear axles [1]. In automobiles with com-
bustion engines it is nowadays realized by more and more advanced differentials. It can
be solved much easier in the case of electric driven vehicles equipped with two engines,
which drive each of axles separately. The target of the work was to develop the method
of control of driving force distribution in order to increase the vehicle directional
stability, particularly in critical states if its motion.
The paper presents the results of calculations conducted according to ISO standard
4138 “Steady-state circular driving behavior” [6]. Tire slip characteristics demon-
strating influences of lateral and longitudinal forces on tire side slip angles were used in
the calculations. Particularly in the calculation it was taken into consideration that
longitudinal force acting on the wheel causes the increase of tire side slip angle. Results
of calculation showed that in particular riding conditions drive forces distribution can
have significant influence on a stability characteristics of the vehicle, namely the
characteristics can be changed from understeer to oversteer or vice-versa.

56 A. Reński
This effect can be practically utilized in a control of vehicle directional stability.

The calculation shows that proposed method of stabilization of vehicle motion can be
effective particularly in critical states when tires adhesion reaches its limit values.
2 Vehicle Model
For mathematical description of vehicle curvilinear motion different models can be

used from very simple linear two-degree-of-freedom to very complicated models. In
this paper simple two-degree-of-freedom vehicle model was applied, which is often
named as bicycle model (Fig. 1). Its formal description can be found in publications on
vehicle dynamics, e.g. [2, 4] and others. If constant longitudinal velocity is assumed,
the vehicle motion can be described by two coordinates: displacement along y-axis
(with respect to the vehicle axis system) and yaw angle ψ (rotation about z-axis) or by
their two derivatives – lateral velocity y_ and yaw velocity w_ respectively.
Forces acting on the vehicle have been marked as follows: Y1, Y2 – lateral forces
acting on one front or rear tire in tire-road contact point, X1, X2 – longitudinal forces
acting on one front or rear tire in tire-road contact point, positive by driving, negative
by braking.
Fig. 1. Vehicle model: x, y - vehicle axis system, CG – center of gravity, l12 – wheelbase, l1, l2 –
front and rear axle to CG, R – cornering radius, v – velocity, v1, v2 – velocities of front and rear
axles, δ - steer angle, α1, α2 – slip angles of front and rear wheels, w_ - yaw velocity.
Analysis of the Influence of the Drive Force Distribution 57
As the angles δ, α1, α2 are relatively small, it can be assumed for simplicity that:
sinα ≈ α, tgα ≈ α, cosα ≈ 1. With this assumption, according to Fig. 1, the following
geometrical relation can be obtained:
l12
¼ d a1 þ a 2 : ð1Þ
R
The quotient l12/R is named Ackerman angle δA, i.e. the steer angle, which should
be applied when cornering on the circle with radius R, if tires slip angles would not be
taken into consideration. Therefore
dA ¼ d a1 þ a2 ð2Þ
Steer angle δ depends on steering wheel angle δsw
dsw
d ¼ ; ð3Þ
is
where is is overall steering ratio. Finally
dsw
dA ¼ a1 a2 ð4Þ
is
The difference of slip angles on the right side of the equation depends on centrifugal
force acting on the vehicle, therefore on the lateral acceleration ay and on vehicle
parameters namely: center of gravity position and cornering stiffness of front and rear
wheel tires. On the basis of this slip angles difference the simplest definition of under-
and oversteer is formulated. An automobile understeers when the slip angles of front
tires are bigger than the slip angles of rear tire and it oversteers in the opposite case.
3 Under- and Oversteer According to ISO Standard
According to the ISO Standard 4138 “Steady-state circular driving behavior”, the
automobile handling characteristics is determined not only on the basis of slip angles
difference but also on how the steering wheel angle changes according to changes of
lateral acceleration. The standard introduces the concept of understeer gradient GS
1 ddsw ddA
GS ¼ ð5Þ
is day day
If it can be assumed that the steering ratio is is constant
dd ddA
GS ¼ ð6Þ
day day
58 A. Reński
Positive value of GS indicates that with increasing value of lateral acceleration the
steer angle δ increases faster than the theoretical Ackerman angle, and the vehicle has
understeer handling characteristic. The negative value of GS indicates however the
oversteer characteristic. This method of evaluation of the handling characteristic is
more compatible with the feelings of the driver, which evaluates the behavior of the
automobile on the basis of the relationship between the change in steering angle, and
the subsequent change of the lateral acceleration.
ISO Standard 4138 describes the test commonly applied for assessment of handling
characteristic of an automobile. The test consists on cornering on the circle with
constant radius R with different constant velocities v. The results are presented in the
form of plots of the steering wheel angle δsw as a function of the lateral acceleration ay
(Fig. 2). The lateral acceleration depends on velocity v as follows
v2
ay ¼ ð7Þ
R
or
ay ¼ v w_ ¼ R w_ 2 ð8Þ
where w_ is the vehicle yaw velocity. Exemplary test results obtained for vehicles with
different handling characteristics are presented on Fig. 2.
Fig. 2. Assessment of the handling characteristic on the basis of understeer gradient GS:
Cornering on the circle with constant radius R with different constant velocities v (1 – understeer
vehicle, 2 – neutral, 3 - oversteer, 4 – vehicle with variable characteristic;
GS ¼ 0; GS [ 0; GS\0)
4 Tires Characteristics
Slip angles of given tire depend on forces acting on it: longitudinal X, lateral Y and
vertical Z. Respectively, for front and rear tire
a1 ¼ a1 ðX; Y; Z Þ a2 ¼ a2 ðX; Y; Z Þ ð9Þ
Generally above dependences are nonlinear. They are presented in a form of dia-
grams, which are known as tire cornering characteristics. In Fig. 3 there is presented an
exemplary diagram of dependence between side slip angle α, lateral force Y and vertical
load Z, while in Fig. 4 the dependence between slip angles and lateral and longitudinal
forces Y and X for given value of vertical load. From that last diagram, the influence of
longitudinal force on the side slip angle can be read. The envelope of curves in Fig. 4
represents the limit of tire-to-road adhesion. When isotropic friction properties of tire
can be assumed this limit can be presented in a form of circle with radius µmZ, where
µm is the maximum value of tire-to-road friction coefficient. Hence the tire adhesion
limit can be defined in form of inequality
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X 2 þ Y 2 lm Z ð10Þ
Accurate dependence for specific tires should be obtained from road or bench tests.
For calculation, so called tire models are often used, e.g. Dugoff’s Model [3], Pacejka’s
Magic Formula [5] and others, which present tire characteristics in form of equations
approximating real dependences. These models, however, require to deliver a large
number of parameters describing properties of given tire, which often are difficult to
obtain.
Fig. 3. Example of tire cornering char- Fig. 4. Example of dependence between lateral
acteristic [7] force Y and longitudinal X for different values of
slip angles α [7]
60 A. Reński
For this study, curves from Fig. 3 have been approximated by the equation
describing a family of parabolas
Y ðaÞ ¼ A a2 þ B a ð11Þ
According to values of parameters A and B, different parabolas corresponding to

different values of vertical load can be obtained.
Characteristics from Fig. 4 can be approximated however with a family of ellipses
described by the equation
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

Y ¼ Y ð aÞ 1 X 2 Xmax 2 ð12Þ
where Y(α) is the value of lateral force calculated from equation of parabola. The force
Xmax is the maximum value of tire-to-road adhesion force for given vertical load Z and
adhesion coefficient μm: Xmax = Z μm.
5 Investigation of Influence of Drive Force Distribution

on Vehicle Handling
Approximated tire characteristic according to Eqs. (11) and (12) were used in calcu-
lations, which were applied in investigation of influence of drive force distribution on
vehicle handling. The investigation consisted on calculation of steer angle as a function
of lateral acceleration done for an automobile cornering on the circle of constant radius
R with different constant velocities v. According to the plain single track (bicycle)
vehicle model used here several simplifying assumptions were adopted in the calcu-
lations, namely: The body roll was not taken into consideration. Therefore the differ-
ences between forces acting on the inner and outer wheel of given axle are not taken
into account. Furthermore, by steady-state cornering with big radius, small steer angle
and constant speed can be assumed, that the only lateral force acting on the vehicle is
centrifugal force. For these assumptions tangential forces X and Y were calculated in
the following way: Lateral forces Y acting between tires and road surface counteract the
centrifugal force
m v2
Fr ¼ ð13Þ
R
The values of lateral forces acting on one wheel of front or rear axle depends on
their static distribution
m v2 l2 m v2 l 1
Y1 ¼ 0:5 ; Y2 ¼ 0:5 ð14Þ
R l12 R l12
Longitudinal forces X counteract however the vehicle resistance forces, namely

aerodynamic drag force
Fp ¼ 0:5 q S cx v2 ð15Þ
(where: ρ - air density, S - vehicle frontal area, cx - aerodynamic drag coefficient, v -

vehicle velocity) and rolling resistance force on horizontal grounds
Ft ¼ m g f0 ð1 þ k v2 Þ ð16Þ
(where: m - vehicle mass, f0 - rolling resistance coefficient at small speed, k -

proportionality coefficient, v - velocity). If the ratio of drive force distribution is κ, then
the drive forces acting on one front or rear wheel (assuming that traction forces of the
inner and outer wheels are equal) can be calculated as follows:
1 j 1 1
X1 ¼ ðFp þ Ft Þ; X2 ¼ ðFp þ Ft Þ ð17Þ
2 jþ1 2 jþ1
For assumed value of radius R and for different values of vehicle velocity v, the
forces X1, Y1 and X2, Y2 were calculated on the basis of Eqs. (14) and (17). Then, after
using Eq. (12) and after solving of quadratic Eq. (11), the values of slip angles α1 and
α2 have been obtained. Finally, with use of transformed Eq. (1)
l12
d ¼ þ a1 a2 ð18Þ
R
steer angles δ as a function of vehicle velocity were calculated. After taking into
account the dependence between velocity v, the circle radius R and the lateral accel-
eration ay (Eq. (7)), the diagrams according to standard ISO 4138 were done.
The calculations were done for vehicles with different center of gravity position,
with different drive forces distribution and for different values of tire-to-road adhesion
coefficient. Exemplary diagrams are presented on Figs. 5 and 6.
1,2 1,2
1 1
Steer angle [deg]
Steer angle [deg]
0,8 0,8
0,6 0,6
deltaA deltaA
0,4 delta1 0,4 delta 4
delta2 delta 5
0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
a Lateral acceleration [m/s2] b Lateral acceleration [m/s2]
Fig. 5. Dependence of steer angle from lateral acceleration for vehicle with the CG in the
middle of the wheel base (a) and for vehicle with the CG moved forward (b) and with different
drive forces distribution: front wheel drive (delta1, delta4), front to rear drive force distribution
κ = 0.67 (delta2, delta5), rear wheel drive (delta3, delta6); Ackerman angle – delta A. Tire–to–
road friction coefficient µm = 0.8
62 A. Reński
1,2 1,2
1 1
Steer angle [deg]
Steer angle [deg]

0,8 0,8
0,6 0,6
deltaA deltaA
delta2 delta 5
0 0
0 0,5 1 1,5 2 2,5 0 0,5 1 1,5 2 2,5
a Lateral acceleration [m/s2] b Lateral acceleration [m/s2]
Fig. 6. Dependence of steer angle from lateral acceleration for the same vehicles as on Fig. 5,
but in the case of tire–to–road friction coefficient µm = 0.2
Results of calculation presented on Figs. 5 and 6 showed that vehicles with front
wheel drive, both when the center of gravity is in the middle of the wheelbase, and
when it is moved forward, have understeer characteristic for all values of lateral
acceleration. However, the vehicles with the center of gravity in the middle of
wheelbase and with grater rear wheels drive force or the rear wheel drive vehicles
present the tendency to oversteer. All vehicles with the center of gravity moved to front
have tendency to understeer at small lateral accelerations. A special case is a vehicle
with rear wheel drive which initially understeers, but it changes its characteristic to
oversteer at grater values of lateral acceleration (from ca. 4 m/s2 for µm = 0.8 and from
ca. 0.7 m/s2 for µm = 0.2).
6 Conclusions
The realized calculations confirm that lateral forces and their distribution between axles
have influence on the vehicle handling. However, this impact is not significant for the
average motion conditions and it can be observed mainly by cornering upon the
adhesion limit. This is due to the fact that, beyond these limit situations, the usually
drive forces are much smaller than the lateral forces. Therefore, the use of changes in
drive forces distribution for the enhancing of vehicle handling can be effective manly
during critical maneuvers.
References
1. Brukalski, M.: Introduction to dynamic square method. Proc. Inst. Veh. 3(103), 29–35 (2015)
2. Dixon, J.C.: Tires, Suspension and Handling. Society of Automotive Engineers, Warrendale
(1996)
3. Dugoff, H., Fancher, P., Segel, L.: An analysis of tire traction properties and their influence on
vehicle dynamic performance, SAE Technical Paper 700377 (1970). doi:10.4271/700377
4. Gillespie, T.D.: Fundamentals of Vehicle Dynamics. Society of Automotive Engineers,

Warrendale (1992)
5. Pacejka, H.B.: Tyre and Vehicle Dynamics, 2nd edn. SAE International, Warrendale (2006)
6. Passenger cars. Steady-state circular driving behaviour - Open-loop test methods. ISO
4138:2012
7. Zomotor, A.: Fahrwerktechnik: Fahrverhalten (Chassis Technology: Road Behaviour).
Vogel-Buchverlag, Würzburg (1987)
About the Steering Kinematics of the Road
Trains
Ion Preda(&) and Dinu Covaciu

{pion,dinu.covaciu}@unitbv.ro
Abstract. These paper studies the road trains turning process in the hypothesis
of Ackerman steering. Both tractor semitrailer and tractor full-trailer combina-
tions are analyzed. Using planar models and geometric relations, the temporal
differential equations of the combination are derived considering that the speed
and pivoting angle of the steerable wheels are known as time functions.
If the tractor-semitrailer and tractor-dolly-trailer are easier to model, the case
of the combination of a tractor with a trailer having steering mechanism is more
complicated.
The kinematic model can be useful in many driving situations as parking
manoeuvers, passing through round-about or following a certain path. From the
mathematic model, the road train off-tracking can be also obtained.
Keywords: Road trains Tractor-trailer combination Dolly Articulated

vehicle Steering kinematics Ackermann steering Tractrix equations
Clothoid Swept path Off-tracking
1 Vehicle Steering
It is not easy to find a good definition for the vehicle steering (even everybody
“knows” what that means). However, the authors will take the risk to try one enun-
ciation. The steering is the process that allows to a vehicle that moves in a direction
close to its longitudinal direction to generate centripetal acceleration. This seems to be
an acceptable definition for all the vehicles types, for example airplanes, boats on
waves or ground vehicles on rough terrain.
Now, for clarity, it is necessary to be given some others definitions:
• the longitudinal direction is the direction from rear to front (or vice versa); nor-
mally, the driver (or pilot) looks almost in that direction;
• the trajectory of a vehicle is the path (geometrical locus) described by its center of
gravity during movement; of course, each point of a vehicle will have its own
trajectory;
• the osculating circle at a given point of the vehicle’s spatial trajectory is the circle
passing through that point and representing the best approximation of the curve in
the considered point [1, 2]; the center and radius of the osculating circle at a given
point are called, respectively, turning center (center of curvature) and turning
radius (radius of curvature) of the trajectory at that point;

About the Steering Kinematics of the Road Trains 65
• the centripetal acceleration is the component of the vehicle’s acceleration acting

on the inner normal to the trajectory on that point, from the vehicle’s center of
gravity to the turning center [1, 3]; while the longitudinal acceleration change the
magnitude of the velocity (the speed), the centripetal acceleration change the
direction of the velocity.
For the steering process, it is important that the vehicle moves almost longitudi-
nally: a vehicle moving only in lateral direction isn’t steering, but sliding. It is also
important for the driver or pilot to see that the vehicle course is close to its heading.
Reducing the area of discussion to the ground vehicles, a vehicle is steering only if
its wheels or tracks remains (at least partial) in contact with the ground, i.e. the vehicle
is not rolling over (not overturns).
If the ground surface is not very bumpy (is almost flat), the vehicle movement can
be considered planar (at least on small sectors), with a planar trajectory, and that
simplifies a lot the study of the steering process. In that case, the sum of all the external
forces acting on the vehicle may have two components acting in that plane: a com-
ponent tangent to the trajectory, providing thrust (drive) or braking, and another
component normal to the trajectory, realizing steering. If exists, a third component,
perpendicular to the plane, will induce bounce.
An extremely benefic characteristic of the wheel is that the force needed to make
that roll is many times smaller (about 100 times for modern wheels on good-state
roads) as the force needed to make the wheel slide. That means a wheel (even more for
an elastic one) has the tendency to move in a direction contained by its central plane
(perpendicular to its spinning axis).
This valuable property is used in the most cases to realize the vehicle steering.
Others modalities to steer ground vehicles consist in changing the camber angle of
elastic wheels (for bikes or others slanted vehicles) or to ensure different peripheral
speeds for the left and right wheels or tracks.
The actual study presents aspects about the road trains steering in the hypothesis of
negligible lateral slip of the wheels or tracks. It is possible to consider that only if the
actual grip-forces at the tires or tracks are small in comparison with theirs maximal
values. Such conditions may be obtained only on grounds which are flat and present
small lateral inclination, which is normally the case with the roads.
The planar shape of the vehicle’s turning trajectory may be described in particular
cases by a tractrix curve [4]. The word for word translation of the German name of
that curve is “the dog curve” [4, 5], because the movement of a “well behaving” dog in
leash which is following its master moving in straight line has this shape (the authors
emphasize in purpose the two particularities). Similarly, most of the towed vehicles
(trailers) will describe a tractrix (or tractory) curve when are following theirs tractors
(tugs) that are moving in straight line or in circle (the mentioned particular cases).
But, if the dog has unsubordination tendencies, pulling aside, its movement is not
anymore a tractrix. Continuing the similitude between a road train and a master with its
undisciplined leashed dog, the trajectory of the trailer equipped with a steering system
will be not a tractrix.
This paper will present an algorithm which permits to describe the planar move-
ment of road trains of different types, with trailers with or without steering possibilities,
66 I. Preda and D. Covaciu
with any steering lows and speed imposed by the driver, under two aspects of dis-
placement: position (trajectory) and velocity.
2 Importance of the Research Theme
At present, it is a large interest in the studying of road trains steering, because a very
large long-distance transport is based on the road trains. Due to the historical influences
and infrastructure development, there are different commercial vehicle concepts in the
geographical areas of the world (Europe [6] – up to 40 t overall mass and up to 25 m
overall length; USA [7]; Japan; Australia). These categories of vehicles present many
important similarities, but also have specific characteristics that are reflected in theirs
kinematics and dynamics.
The back-up and parking maneuvers are studied to make easier the driver’s task,
but also for a better design of parking and docking places. For various types of urban or
road-network development, the understanding of steering ability is also necessary for
the analysis of vehicle access and egress at junctions and accesses [8]. The swept path
study is used to demonstrate that the geometry of the road intersections or highway
connections is sufficient for the vehicle passing or to help the designer to modify the
junction design.
For the engineers, the analysis of the unwanted interferences between road trains
components or the behavior of the trailers steering-mechanisms are mandatory stages in
the vehicle design [9].
The influence of the driver maneuvers over the vehicle stability presents also a
major interest [10, 11]: active steering, rollover warning and avoidance [11], active
stabilization of car and caravan-trailer combination [12] or even autonomous driving
need detailed study of the road train steering.
3 Road Trains
A road train is an assembly (a combination) of at least one automotive vehicle and at

least one trailer. With respect to theirs steering abilities, the component vehicles may be
of different types (acronyms appertain to the authors):
• S2GA – steerable vehicle with two groups of axles (one front group and one rear
group); these are equipped with a steering system acting on at least one steerable
group of axle; this may be a motor vehicle or a full-trailer with steerable wheels;
• A1GA – articulable vehicle with one group of axles; normally, this group contains
one or two (driving or non-driving) non-steerable axles; that type of vehicle is
attached to another articulable vehicle, obtaining so an articulated vehicle (with two
semi-frames); the two vehicles in the combination may be positioned at a variable
angle by rotating them against a pivot;
• D1GA – dolly-type vehicle, a carriage having one group of axle (with non-steerable
or steerable wheels) and a system that permit its attachment to another component
of the road train;
• DB – draw-bar; this isn’t a true vehicle, but is a very important component of some
road train types and its position may alter the kinematics of the towed vehicles.
Table 1. Typical road train configurations

Joint type Component vehicles
Articulated vehicle 1 DoF A1GA + A1GA
Tractor & semitrailer 2 DoF S2GA + D1GA
Tractor & steerable trailer 3 DoF S2GA + D1GA
Tractor & dolly & trailer 3 DoF & 2 DoF S2GA + D1GA + D1GA
Tractor & drawbar & steerable trailer 3 DoF & 3 DoF S2GA + DB + S2GA
Combining these elementary vehicles, a huge number of road trains may be

imagined. In the Table 1, there are presented some of the most spread road trains types.
Nevertheless, if different shapes and dimensions of these combined vehicles are con-
sidered, the number of variants becomes extremely big.
For a good understanding of the reasoning presented in this paper, the most
important points for the component vehicles of the road trains will be noted further in
the same way (Fig. 1) and will keep the same meaning: O – origin (conventional);
G – center of mass; F and R – equivalent steering center for the front and rear wheels
(groups); A – point to be attached to another vehicle (in the front); H – point to attach
(hang) another vehicle (in the rear); Q – intersection point between the longitudinal
symmetry axis of the vehicle (RF) and the perpendicular from C on that axis; P – any
point of interest; C – center of turn.
Fig. 1. Names of the points attached to a vehicle
The points F, O, R, and Q are all disposed on the longitudinal symmetry axis of the
vehicle.
In the case of the vehicle schematized in the right side of the Fig. 1, which has two
non-steerable rear axles, the distance RF represents its equivalent wheelbase. If one
allows to the third axle to become self-steerable at low speeds (as happens on some
modern truck-tractors), the point R will superpose over the actual point O.
To realize the road train, the point A of the first trailer it is attached to the hanging
point H of the tug (tractor vehicle) and so on if others trailers are attached. It is also
possible that an automotive vehicle to push a trailer: in this case, the point A of the
pushing vehicle will be attached to one of the points H or A of the trailer. For example,
the superposition of the points A of one agricultural tractor and one trailer is adopted by
many (Romanian) farmers to maneuver more easily the trailer in narrow spaces.
4 Steering Kinematics
The steering kinematics of a vehicle (singular or in combination) is treated in many

important works for the vehicle dynamics, as [13–15]. A slight different approach will
be presented further.
On a solid body, the Euler’s theorem for the velocities distribution [1] states that:
v¼!
~ ~ ~
v0 þ x d ð1Þ
where ! v0 is the translation velocity of a reference (origin) point, ~

v – the translation
velocity of a certain point, x~ – the instantaneous rotational velocity of the body and
~
d – the position vector of the individual point with respect to the reference point (the
oriented distance between the two points). The velocities and the distance vectors have
both magnitude and orientation.
Particularizing the Eq. (1) for the planar movement of a vehicle [16], that takes the
scalar form:
v ¼ v0 þ x d ð2Þ
where v0, v and ω are the magnitudes (speeds) of the previous firsts vectors and d is the
distance (as magnitude) between the reference and individual points.
In the Fig. 2 it is represented schematically an all-wheel steering vehicle (upper
view) at a certain moment of time. At that moment, its planar movement is a simple
rotation with the speed ω against the point C (named turning center and having the
instantaneous translational speed equal with zero: v0 = 0). The translational speed vP of
any vehicle’s point P is then:
vP ¼ x RP ð3Þ
where RP is now the distance between the point C (the turning center) and the
respective point P.
The left scheme in the Fig. 2 is obtained from the one in the right by replacing the
two wheels of an axle with an equivalent one placed in the middle of each axle (points
F and R for the front and rear wheels, respectively).
The vehicle tendency to roll, opposing to skid (lateral sliding), means that, for small
to moderate lateral forces, a good steering system ensures that all wheel’s spinning-axes
Fig. 2. Schematized upper view of an all-wheel steering vehicle: right – four-wheel planar
model; left – simplified “bicycle” model
projections on the ground surface to be oriented through the turning center C. This
geometrical condition is known as the Ackermann condition [3, 9, 16].
Studying the Fig. 2 and realizing simple geometrical calculations [17] it results that
the distance CQ (on a line perpendicular to the vehicle’s longitudinal direction RF) is:
Wb
CQ ¼ RQ ¼ ð4Þ
tan dF þ tan dR
where Wb is the wheelbase and δF and δR are the mean (equivalent) steering angles of
the front and rear wheels (identical with the directions of the velocities !
vF and !
vR of the
points F and R). That means all the geometrical elements in Fig. 2 can be derived easily
if a distance between two points (for example the wheelbase RF = Wb) and the mean
steering angles δF and δR are known. And further, the speed in any vehicle’s point of
interest P can be calculated if the speed vX of any other point X is supplementary
known.
For example, the vehicle’s side slip angle β is:
QG QG
b ¼ tan1 ¼ tan1 ð5Þ
CQ RQ
and the speed of the point Q is
vQ ¼ vcosb ð6Þ
where v = vG is the vehicle’s speed (the speed of its center of gravity G).
Considering Eq. (4), the next step consists in the calculation of the yawing speed ω:
vQ v vQ
x ¼ ¼ ¼ ðtan dF þ tan dR Þ ð7Þ
RQ R Wb
where vQ and v are the speeds of the points Q and G (the center of gravity), RQ and R –
the distances from the point C (the center of turn) to the points Q and G; the distance R
is named the vehicle’s turning radius.
Little discussion it is necessary with respect to the model presented in Fig. 2. This
may be used for an all-wheel steering vehicle, at which the driver controls the steering
of the front wheels, while the mean steering angle of the rear wheels is imposed:
• through a static function δR = f(δF) (normally in a mechanical way);
• through an active function δR, depending on instantaneous driving conditions, as is
the case of some high-class modern cars.
Also, this model will correspond to an automotive front-wheel steering vehicle, if
one imposes δR = 0, or to a rear-wheel steering vehicle, if one imposes δF = 0. In these
two cases, the driver controls respectively the angles δF or δR. Furthermore, this model
is applicable to full-trailers equipped with steerable front wheels, rear wheels or even
all wheels [1]. In this case, the control parameter isn’t the driver’s action, but the
difference ψjk between the heading angles ψj of the vehicle or the drawbar in front of it
and ψk of the considered trailer (Fig. 3).
Fig. 3. Steering scheme of a road train obtained by a trailer k with one steerable axle attached to
an all-wheel steering tractor j
Considering a fixed to ground coordinate system (with a horizontal axis x and a

vertical axis y), in the Fig. 3 there are presented schematically the steering parameters
of a road train formed by an AWS tractor (type S2GA) and a steerable (full or semi)
trailer with one group of axles (D1GA). The indices j and k indicate respectively the
towing and towed vehicles.
The trajectories of the points of the tractor j are obtained starting for the scheme in
that figure. Accordingly, the trajectory of the point Fj is given by the equations:
Z Z
t t
xFj ¼ xF0j þ vF cos wj þ dFj dt yFj ¼ yF0j þ vF sin wj þ dFj dt ð8Þ
0 0
and the trajectory of the point Rj

xRj ¼ xFj Wb cos wj yRj ¼ yFj Wb sin wj ð9Þ
As can be seen in the Fig. 3, the scheme of the trailer (vehicle k) is almost similar
with the scheme of the tractor j (or, more general, the towing vehicle). The differences
consist in the way to calculate the two equivalent steering angles δFk (at the point
Fk ≡ Ak ≡ Hj) and δRk (at the point Rk).
The angle δFk at the point Fk isn’t changed directly by the driver’s action. Instead, it
will depend either:
• on the relative position angle ψjk = ψj – ψk between the component vehicles and,
• through the angle δHj (negative on the Fig. 3), on the turning radius Rj and on the
position of the hanging point Hj of the towing vehicle.
In this case, the steering angle δFk is:
dFk ¼ wjk þ dHj ¼ wj wk þ dHj ð10Þ
That means the angle δFk depends on the history of the vehicle movement.
The angle δRk at the trailer’s wheels is normally zero (δRk = 0), but there are
vehicles, as some articulated buses or trolley buses, at which that angle is a static
function δRk = f(ψjk) of relative position angle ψjk = ψj – ψk between the vehicles.
Such complicated trailer’s steering system are used to reduce the road train swept path
(steering corridor), necessary to accommodate the buses with some narrow streets of
towns. Another possibility consists in an active change of the angle δRk in order to
stabilize the road train.
To obtain the trailer’s heading angle ψk and the yawing speed ωk in the general
case, it is necessary to use the Eq. (7) for the towed vehicle k and to remember that the
yawing speed is the time derivative of the heading angle
dwk
¼ xk ðt; wk Þ ð11Þ
dt
This ordinary differential equation with initial condition (ψk = ψk0) may be solved
relatively easy by approximate methods.
With this specification, the velocities for all the points of the towed vehicle k may
be determined, and, starting from the position and velocity of the point Hk, the process
may continue for another trailer.
5 Algorithm Implementation and Results
The geometrical data of the combined vehicles is stored in different worksheets in

Microsoft Excel. Other tabulated data is stored in Excel in order to describe, by linear
interpolation, the variation laws of the driver’s steering actions and the tractor’s speed
(in the center of gravity).
The Excel data is imported in PTC Mathcad, the general calculation software
adopted for the implementation of the presented algorithm.
Starting from the data stored in Excel, simple vehicle sketches are automatically
realized in Autodesk Autocad and also different representations of the results obtained
in Mathcad. Two examples of such graphics are presented in Figs. 4 and 5.
Fig. 4. Almost clothoidal trajectory of a road train, followed by stabilization on circle
Fig. 5. Successive positions of a dolly on the trajectory for a straight line movement of the
tractor
In the Fig. 4 it is represented the path of a road-train passing from a straight motion
to a circular one. Starting with the third represented position, the steering angle of the
tractor’s front wheels increases with one degree per position, up to a maximum of 16º.
Once reached this value, the vehicle moves on a circle with the radius of 14.6 m.
Another example consists of a dolly having its initial position perpendicular to the
tractor, at left and at right. The successive positions of the dolly presented in the Fig. 5
correspond to a movement from the left to the right with 1 m per position. For a better
understanding of the trailer’s movement, the tractor wasn’t represented.
6 Conclusions
In this paper, the steering kinematics of the road train was studied. The presented
equations are applicable with minor changes to different vehicle types into the com-
bination, including trailers with steering systems and articulated vehicles. The algo-
rithm admits variable speed (including the reverse moving) and different steering lows
for the towing or pushing vehicle.
Also, obtaining the trajectories of different points of interest, the off-tracking of the
road train can be calculated in concrete driving situations.
Partial automation of the data input and results representation was realized through
the combination of three software types (Excel, Mathcad and Autocad), so valorizing
theirs best possibilities.
References
1. Olariu, V., Sima, P., Achiriloaie, V.: Mecanica tehnică (Technical Mechanics). Editura
Tehnică, Bucharest (1982)
2. https://en.wikipedia.org/wiki/Osculating_curve
3. Gillespie, T.D.: Fundamentals of Vehicle Dynamics. SAE, Warrendale (1992)
4. Steyn, S.: Neither a tractor, nor a matrix but a tractrix! (2008)
5. Weisstein, E.W.: Tractrix. From MathWorld-A Wolfram Web Resource. http://mathworld.
wolfram.com/Tractrix.html. Accessed May 2016
6. Fries, M., Sinning, M., Lienkamp, M.: Virtual combination of commercial vehicle modules
(virtual truck) for characterization of future concepts. IEEE (2015)
7. Bortoni-Anzures, L., Cruz-Victoria, J., Herrera-Rivas, H., González-Sánchez, D.I.,
Díaz-Romero, D., Huixtlaca-Cuautecatl, I.: Full vehicle instrumentation for dynamic
analysis. Ingeniería Mecánica, Tecnología Y Desarrollo 3(2), 60–65 (2009)
8. http://www.sweptpathanalysis.com/index.htm. Accessed May 2016
9. Alexandru, P., Dudiţă, F., Jula, A., Benche, V.: Mecanismele Direcţiei Autovehiculelor
(Motor Vehicles’ Steering Mechanisms). Editura Tehnica, Bucharest (1977)
10. Enache, V., Preda, I., Vulpe, V., Dogariu, M.: Theoretical and experimental research for
vehicle handling improvement. In: International Conference CONAT 2004, Brasov,
Romania, paper 1063 (2004)
11. Chen, B.C., Peng, H.: Rollover warning for articulated heavy vehicles based on a
time-to-rollover metric. ASME J. Dyn. Syst. Meas. Control 127, 406–414 (2005)
12. Fischer, G., Heyken, R., Trächtler, R.: Aktive gespannstabilisierung beim BMW X5 eine
weiterentwicklung der fahrdynamikregelung DSC. ATZ Automobiltechnische Zeitschrift.
2002(04), 330–339 (2002). (Active trailer stabilization of the BMW X5 - Next generation of
DSC Dynamic Stability Control)
13. Mitschke, M., Wallentowitz, H.: Dynamik der Kraftfahrzeuge (Dynamics of Motor
Vehicles). Springer, Berlin (2004)
14. Rajamani, R.: Vehicle Dynamics and Control. Springer, New York (2012)
15. Pauwelussen, J.: Essentials of Vehicle Dynamics. Elsevier, Boston (2015)
16. Preda, I., Ciolan, G.: Vehicle mathematical model for the study of cornering. Ann. Univ.
Oradea, Fascicle Manage. Technol. Eng. XI(XXI), 22–32 (2012)
17. Preda, I., Covaciu, D., Ciolan, G.: Design approach for the vehicle’s steering linkage fitted
on rigid axle. In: 3rd International Congress Science and Management of Automotive and
Transportation Engineering SMAT, University of Craiova, Romania (2014)
Evaluation by Experimental Methods
of the Parameters that Influence the Behavior
of Various Passenger Cars Classes
in the Braking Process
Adrian Todoruţ(&), Nicolae Cordoș, István Barabás, and Sorin Miheț
Faculty of Mechanical Engineering, Department of Automotive Engineering

and Transports, Technical University of Cluj-Napoca, Cluj-Napoca, Romania
{adrian.todorut,Nicolae.Cordos,
istvan.barabas}@auto.utcluj.ro, sorin_mihet@yahoo.com
Abstract. This paper presents a study - by experimental methods - on the

behavior of different classes of passenger cars in the braking process. In the
experimental determinations, the parameter values were chosen to characterize the
real studies case, commonly encountered, of the passenger cars use in the braking
process. Thus, it is considered the experimental determination of the braking
ability parameters of the studied passenger cars depending on tire pressure,
loading on axles, nature and condition of the rolling surface, tire type (summer or
winter) and type of braking system. The way in which behave the passenger cars
equipped with winter tires in a situation where the weather requires summer tires at
different inflation pressures, also show interest. For the frequency assessment of
the brake using in urban areas at different traffic densities it has been used Android
applications. The results were represented graphically in order to facilitate theirs
comparison.
Keywords: Passenger cars Android applications Braking Tire pressure

Axle load Rolling surface
1 Introduction
From the available literature in the field and from the specialized books and also from
the scientific papers, result the continue concerns of the researchers and opening up
innovative lines of research and development. Also, in order to reduce the costs of the
research emerges the development tendency, and the use of computer simulation
environments, that capture the possible variables in exploiting of vehicles, and to allow
a pragmatic interpretation of the results. In the experimental research there is tendency
of using the advanced technologies of the data acquisition and processing.
In the specialized literature [1–5] are deals aspects concerning the description of
physical phenomena in the braking of the vehicles which may underlie the develop-
ment of working models that capture a large range of situations commonly encountered
in the operation of the vehicles and to provide results on road safety, with emphasizing
the influence of various operating parameters of the vehicles in their braking behavior.

76 A. Todoruţ et al.
Due to continuous concerns of the researchers in the area of vehicle dynamics and
safety in their use, at present there are results [6–8] obtained through computer simu-
lation, experimental or mathematical modeling related to issues mentioned above,
through capturing concomitant as many variables that are likely to be encountered during
braking vehicles. The study of Sivaramakrishnan et al. [6] provides an experimental
analysis on how tire design parameters can influence tire characteristics such as peak
grip, braking stiffness and ultimately affect ABS performance for a passenger sedan and
high performance vehicle. Roth et al. [7] indicate through simulation and experimental
data that the relaxation length of a tire affects ABS performance only under certain
criteria. Some studies have shown the influence of longitudinal force-slip curve of the tire
on ABS performance and how every vehicle reacts differently to the same tire [8].
The performed experimental determinations are designed to reflect the behavior of
the car in the braking process, the tests aimed the evaluation of parameters that
influence the braking capacity of the car under study.
2 The Experimental Determination of the Parameters

that Influence the Behavior of the Cars in the Braking
Process
The experimental methods try to render the commonly situations encountered in traffic
and the behavior of the cars in braking process. In this sense it was followed:
• the experimental determination, using Android applications, for the frequency of
using the service brake in the case of travelling in urban traffic at different densities;
• the experimental determination of the axle loads for each of the tested cars,
according to the additional masses which they were loaded using the weighing
tester APF 110 Space, controlled by control unit Space PFC 750;
• the experimental determination of the braking distances of the cars for different
parameters for the various loadings situations and various condition of the road;
• determining the braking forces needed for each car axle, using the roller brake tester
NFP Space 750, for different additional loading cases of the cars, considered in the
test for determining the braking distance.
The loads on the car axle, depending on the additional masses (for the loading on
the car axles have been use bags filled with sand, each with a weight of approx. 50 kg)
wherewith were loaded the tested passenger cars (Citroën C4 1.6 HDI, Smart Forfour
1.3i, Seat Ibiza 1.4 16 V), were determined using the weighing suspension tester
Space APF 110 type, controlled by the control unit Space PFC 750 type, from the
endowment of the Research Center for Automotive Engineering and Transportation
respectively its related research structure - Research Lab for Analysis and Diagnostics
of Motor Vehicles, at Automotive and Transports Department at the Faculty of
Mechanical Engineering from the Technical University of Cluj-Napoca.
To determine the required braking force of the tested passenger cars, according to their
various additional loading and to determine the proper functioning of the braking system,
cars were tested using the tester with balance and braking rolls Space PFN 750 type.
Evaluation by Experimental Methods 77
Thus, have been determined the braking forces required for each axle at different
additional loading cases of the passenger cars, loading conditions identical to those in
the experimental test for determining the braking distance. Table 1 captured some data
of the studied passenger cars parameters in unladen state.
Thus, the additional load was introduced in the trunk of the car influencing the load
on the rear axle (Fig. 1) and thus creating approximately an equal distribution of
braking force on the car axles.
Table 1. Data relating to the parameters of the studied cars

Car brand Unloaded mass of Tire Tread Type of tyre/inflation The type of
the car, [kg] sizes depth, pressure, [MPa] braking system
[mm]
Seat Ibiza 1140 185/60 4 Summer/0.22 Disc/Drum
1.4 16 V R14
Smart 1053 175/65 5 Summer/0.22 Disc/Disc
Forfour R14
1.3i
Citroën C4 1304 205/55 4 Summer/0.22 Disc/Disc
1.6 HDI R16
The experimental measurements of the braking process were conducted both in

urban, on frequently circulated routes and on the test track.
The data on the achieved route, the position, the travel speeds and the tests dura-
tions to were obtained using the Android application Sports Tracker that uses the
smartphone’s GPS sensor.
In order to determine the utilization rates of the braking system and the number of
braking in urban areas, have been conducted a series of measurements using Android
applications. To measure decelerations in order to establish the number of braking
made by car in an urban environment, have been used application Accelerometer
Monitor that uses the smartphone accelerometer sensor to record and display these
Fig. 1. The correspondence between the loading mode and the mass distributed between the
axles (Seat Ibiza and Smart Forfour - small class car, B; Citroën C4 - compact class car, C). L1 -
Unloaded; L2 - Loaded (100 kg in the trunk); L3 - Loaded (200 kg in the trunk).
variations. The data are stored and on their basis it can be achieved variation graphics
of the acceleration in time.
The urban track were the measurements have been carried out, in length of 5.8 km
has been intensely circulated (Fig. 2). The recording device was fixed on a support
mounted inside of the tested car (Fig. 3). The resulting data is displayed on the mobile
terminal (Fig. 4) and can be saved for later analysis.
The measurements were performed twice for each of their testing times (7:50; 16:00;
22:00) in the month of May 2015. In choosing of hours for taking the measurements it
was considered to obtain data for different levels of traffic and thus the determining the
frequency of use the service brake in conditions similar to the real ones. Figures 5 and 6
are showing the case of the Citroën C4 1.6 HDI car. Determination of the braking
number was achieved considering the graphs obtained by the software for measuring the
accelerations, where the positive accelerations are specific for the car starting and
negative accelerations results are specific to the braking (see Fig. 6).
Fig. 2. The route of the car Fig. 3. The decelerations Fig. 4. The display data
recording device
65.4
53.4 48.5
41
14.32 20.28 9.07 24.2 18.1 33.1 29 17
Time, [min] Maximum speed, Medium Speed , Brake numbers

[km/h] [km/h]
hour 7:50 hour 16:00 hour 22:00
Fig. 5. The test data conducted in urban areas concerning the variations of different parameters
according to time of the test, along a route length of 5.8 km, for Citroën C4 car.
Data collected during the measurements were the basis for determining the
parameters of the braking testing.
Choosing the parameters and the route. Based on the obtained data during the tests
in urban areas have been set the test speeds for obtaining the braking space. Thus, the
speeds used in the test were: 20 km/h, 30 km/h, 40 km/h, 50 km/h, 60 km/h. The route
for carrying out the tests was chosen to conduct so as to provide the maximum speed of
60 km/h and the braking distance required to stop completely at this speed. Consid-
ering these conditions, the location was chosen at the shopping center Polus Center
Cluj-Napoca with agreement of the Mall management.
Fig. 6. The variations of the longitudinal accelerations, in the case of Citroën C4 car
The way to perform the tests and the measurements. The period of the testing was
05-06.05.2015, with an ambient temperature of 18–19°C. The preparation of the route
involved establishing the dimensions of a test tracks with total length of 80 m,
delimited with the reflective markings.
Marking the start of the braking was set at a distance of 50 m from the starting line,
effectively stopping space being measured from the marking start braking until the
front of the stopped car. Parameters depending on which was taken into account in
determining the braking distance required to stop: the running surface on
concrete-asphalt dry/wet; tire pressure (nominal pressure 0.22 MPa; low pressure
0.15 MPa); Used tire type (summer, winter); loading state of the vehicle (unloaded,
loaded with 100 kg and 200 kg, see Fig. 1); influence of the ABS (using ABS; ABS
disabled); the speed the car (20–60 km/h).
The obtained results at this stage have a graphical interpretation and refer to:
• testing the braking ability of the car (see Table 1) and the influence of a 100 kg
additional mass on the braking distance (Figs. 7, 8, 9) - the results can be seen that
with increasing the additional mass increases the distance needed to stop the car, so
the braking system is less effective;
• influence of tire pressure on the braking distance (Fig. 8) - the tests were conducted
for unloaded Citroën C4 car (compact);
• influence of the tire type on the braking distance (Fig. 9) - the tests were carried out
for unloaded Citroen C4, for summer tires and winter tires with the state of wear of
30 %, the size and inflation pressure thereof is based on Table 1;
• influence of the vehicle loading state respectively the running surface state on the
braking distance for different types of tires (Figs. 9, 10, 11);
• influence of the Anti-Lock Braking System (ABS) on the braking distance
(Fig. 12), ABS deactivation was achieved by removing the fuse from the fuse box
system of the car to which the tests were conducted. Besides the increasing of
the breaking distance, deactivating of the ABS led to loss of longitudinal stability of
the car. From the dynamic conditions, the load on the front axle is greater, and if the
braking forces are evenly distributed, rear axle wheels lock first or remain locked
and the car tends to skid.
To determine the required braking forces for the tested cars according to different
loads to which they were subjected, as well as to determine the proper functioning of
the braking system, the cars have been tested in the laboratory on the support brake
testing car using the tester with scales and braking rolls.
Using the tester Space PFC 750 type were determined for each axle loads of the
tested cars, according to the additional masses wherewith were loaded the cars. Using
the brake tester Space NFP 750 type were determined the braking forces required for
each axle, for different additional loading cases of the cars, loading conditions identical
to those used in the tests to determine the braking distance.
Following the measurements test in the laboratory were obtained results relating to:
• distribution of the braking forces for the studied cars, depending on the additional
load (Fig. 13);
• distribution of the necessary braking force on the axles, depending on the tire
pressure (Fig. 14).
Fig. 7. Additional loading influence on the Fig. 8. The stopping distance variation
stopping distance of the cars, fitted with according to the inflation pressure, for unloaded
summer tires, dry rolling surface. C4 Citroën car, equipped with summer tires,
dry rolling surface.
Fig. 9. Variation of the braking distance Fig. 10. Variation of the braking distance
depending on car loading, fitted with summer depending on car loading, fitted with winter
tires, on a dry and wet rolling surface (Citroën tires, on a dry and wet rolling surface (Citroën
C4 car). C4 car).
Fig. 11. Variation of the braking distances Fig. 12. Influence of the ABS on the braking
depending on the rolling surface condition distance, for unloaded car fitted with summer
for summer/winter tires (Citroën C4 car). tires (Citroën C4).
Fig. 13. Braking force depending on the Fig. 14. Axles braking force depending on tire
loading cars axles. L1 - Unloaded; L2 - pressure, for Citroën C4 car.
Loaded (100 kg in the trunk).
3 Conclusions
Studies, research and the conducted experimental tests in order to develop this work,
allowed the possibility of issuing final conclusions on the subject proposed as follows:
• in urban areas, the frequency of using the service brake is proportional to the
density of the road’s route;
• any additional vehicle weight increases the braking distance;
• the braking distance is directly influenced by the type of braking system;
• decreasing of the tire pressure increases the braking distance;
• using inappropriate season tires (summer tires - winter or winter tires - summer)
increased the required vehicle stopping distance, so less efficient braking system;
• the braking distance increases with increasing the additional mass of the car,
regardless of the used tire type or the condition of the rolling surface;
• wet rolling surface determines the decrease of the adhesion coefficient and thus
considerably increases the braking distance;
• disabling of the ABS system, lead to an increase of the braking distance and losing
the longitudinal stability of the car during of the braking process;
• the distribution of the braking force for each axle is depending on axle loads,
leading to a greater distribution of braking force on the rear axle, which negatively
influences the behavior of the car during braking.
References
1. Abe, M.: Vehicle Handling Dynamics, Theory and Application. Butterworth-Heinemann,
Published by Elsevier Ltd., Oxford (2009)
2. Blundell, M., Harty, D.: The Multibody Systems Approach to Vehicle Dynamics. Elsevier
Butterworth-Heinemann, Oxford (2004)
3. Crolla, D.: Automotive Engineering – Powertrain, Chassis System and Vehicle Body.
Elsevier Butterworth-Heinemann, Oxford (2009)
4. Pacejka, H.B.: Tyre and Vehicle Dynamics, 2nd edn. Elsevier Ltd., Delft (2006)
5. Wong, J.Y.: Theory of Ground Vehicles, 3rd edn. Editura John Wiley & Sons, INC., Ottawa
(2001)
6. Sivaramakrishnan, S., Singh, K.B., Lee, P.: Experimental investigation of the influence of tire
design parameters on anti-lock braking system (abs) performance. SAE Int. J. Passeng. Cars
Mech. Syst. 8, 647–658 (2015). doi:10.4271/2015-01-1511
7. Roth, V., Xie, K.J., Kidney J., Turner, J.: Investigation of tire longitudinal relaxation length
and its effect on dry stopping performance. In: 178th Technical Meeting of the Rubber
Division of the American Chemical Society, Inc. (2010)
8. Kidney, J., Mani, N., Roth, V., Turner, J., Branca, T.: Experimental and computational studies
of contact mechanics for tire longitudinal response. In: Proceedings of the 30th Tire Society
Conference, Akron, OH (2011)
Evaluation of the Behavior in Cornering
for Different Classes of Passenger Cars
by Numerical Modeling
Nicolae Cordoș(&), Adrian Todoruţ, István Barabás,

and Radu Daniel Mureșan
Technical University of Cluj-Napoca, Cluj-Napoca, Romania

{Nicolae.Cordos,adrian.todorut,
istvan.barabas}@auto.utcluj.ro,
dani_muresan77@yahoo.com
Abstract. This paper presents a study, using modeling methods, on the

behavior of different classes of passenger cars in cornering. Taking into account
the factors that influence the movement and tendency of the cars in cornering,
the parameter values in the mathematical model were chosen so as to charac-
terize the real situations encountered frequently in cars’ exploitation in the
cornering process. In order to have relevant results for the chosen cars, different
experimental loads (for each axle) were determined for different loading cases.
The analytical model, developed in Mathcad, takes into account the geometrical
parameters of the passenger cars and helps to determine the understeering or
oversteering tendencies. The results were represented graphically in order to
facilitate their comparison.
Keywords: Passenger car Cornering Drift angles Understeer Oversteer

Numerical modeling
1 Introduction
Specialized literature [1–6] describes the movement process of the vehicles and the
parameters influencing this process. Based on this, new working models can be
developed to capture more diverse situations encountered in the behavior of the motor
vehicles in cornering.
Due to the continuing concerns of researchers in the field of vehicle dynamics
[7–13] currently there are results obtained experimentally or by mathematical modeling
addressing the issues mentioned above.
In support of the research came a number of regulations in the field that are
captured in the works mentioned at [14–17].
Analytical models for calculation in MathCad were developed to assess the
cornering trends of the tested passenger cars in the case of their movement in a circle
with a constant radius.

84 N. Cordoș et al.
Considering the main parameters that influence the maneuverability of the pas-
senger cars, the paper aims to deliver results on the average value of the drift angles,
depending on:
– the speed – in the case of different axles loads of the car, on circular trajectories of
different radius;
– the lateral acceleration – in the case of different axles loads of the car;
– the coefficient of resistance to lateral deflection of the tyre;
– the position of the center of gravity.
2 Determination of the Distributed Masses on the Car Axles

for Different Loads
In order to have relevant results for the chosen cars, different experimental loads (for
each axle) were determined for different loading cases. These measurements were
carried out at the Department of Automotive Engineering and Transports at the Faculty
of Mechanical Engineering from the Technical University of Cluj-Napoca. The loads
on the car axle, depending on the additional masses with which the tested passenger
cars were loaded, were determined using the suspension tester Space APF 110 type
(Fig. 1) from the Research Center for Automotive Engineering and Transportation at
the Department of Automotive Engineering and Transports from Cluj-Napoca.
The different loads on the axle were obtained by the use of sandbags weighing
50 ± 1 kg, and in some cases passengers. Figure 2 captures the loading mode of the
passenger cars.
The average angle of the front wheels steering can be obtained by dividing the
steering wheel rotation angle by the transmission ratio of the steering system.
3 The Numerical Evaluation Method
The character of a vehicle within terms of maneuverability can also be studied ana-
lytically. In order to obtain the differences between several exploitation situations, the
loads on the passenger car axles have been determined, taking into account the fact that
the value of the longitudinal inclination angle of the road α is zero. To determine the
loads on the car sides and the reactions between the tire and the rolling surface at each
wheel, for each speed, it was also considered that the value of the transversal angle β is
null [4, 5].
The loads calculation on the sides of the car is influenced by the centrifugal force
that largely depends on the cornering radius R, on the mass of the car and on the
driving speed [5]:
ma v2
Fcf ¼ : ð1Þ
R
Evaluation of the Behavior in Cornering for Different Classes of Passenger Cars 85
The determination of the lateral forces at the car wheels is based on the value of the
reaction forces, according to the relation (2), where φ is the coefficient of longitudinal
adherence. The ratio between them and the coefficient of resistance to lateral deflection
of the tire k, helps to obtain the values of the drift angle (relation 3) [5].
Fig. 1. Mass determination on the front axle Fig. 2. The correspondence between the load-
for the studied passenger cars. ing mode and the mass distributed between the
axles (Seat Ibiza - small class car, B; Citroën
C4 - compact class car, C). L1 - Unloaded; L2 -
Loaded (150 kg in the trunk); L3 - Loaded
(passenger on the right in the front + passenger
on the right in the rear + 50 kg in the trunk).
Y1;2L;R ¼ 0:8 u Z1;2L;R ; ð2Þ
Y1;2L;R
d1;2L;R ¼ ; ð3Þ
k
where: Y1;2L;R - lateral forces, (1 - front axle, 2 - rear axle, L - left side, R - right side);
Z1;2L;R - normal reactions (1 - front axle, 2 - rear axle, L - left side, R - right side).
Lateral acceleration was calculated based on the relation:
v2
ay ¼ ; in g; ð4Þ
gR
where: R - radius of the circle on which the vehicle is moving, in m; g - gravitational

acceleration, in m/s2; v - speed, in m/s.
Following the mathematical model, using MathCad software, there were obtained
the variations of the average drift angles at the front and rear axles for the studied
passenger cars. This was done according to their different loads and movement on
circular trajectories with different radius.
4 Obtained Results
Figures 3 and 4 present the average value of the drift angles of the front wheels for
Citroën C4, depending on the speed (Fig. 3) and on the lateral acceleration (Fig. 4) in
different load cases. The tendency of understeering in the case of the unloaded vehicle is
stronger than when it is loaded with 150 kg in the trunk. The smaller the differences
between the values of the drift angles at front axle wheels, respectively rear axle, the
bigger the tendency of the passenger car towards a neutral steering. The same tendency
of drift angles variation is observed in the case of Seat Ibiza (see Fig. 5), which is driven
on a circular trajectory of 16 m radius, on a concrete dry road (three loading cases).
Fig. 3. The variation of the average value of Fig. 4. The variation of the average value of
the drift angles for Citroën C4 depending on the drift angles for Citroën C4 depending on the
the speed in the different loading cases. Load lateral acceleration in the different loading
1: front = 823 kg, rear = 481 kg, load 2: cases. Load 1: front = 823 kg, rear = 481 kg,
front = 808 kg, rear = 641 kg. load 2: front = 808 kg, rear = 641 kg.
Fig. 5. The variation of the average value of Fig. 6. The variation of the average value of
the drift angles for Seat Ibiza depending on the drift angles for Citroën C4 (unloaded),
the speed in the different loading cases. Load depending on the speed, on circular trajectories
1: front = 691 kg, rear = 449 kg, load 2: (different radii).
front = 675 kg, rear = 609 kg, load 3: front =
685 kg, rear = 555 kg.
In Fig. 6 it is shown the behavior of the unloaded Citroën C4 (see Fig. 2) on a dry
concrete road, on circular trajectories with different radii (as it can be seen in relation 1
an increased radius leads to a decrease of the inertial force).
The value of the resistance coefficient to the lateral deflection of the tire
(k) influences the value of the lateral deflection angle of the tire. This is illustrated in
Fig. 7, where it is considered that Citroën C4 moves unloaded on a circular trajectory
with a radius of 16 m, with a speed of 8.33 m/s. The coefficient of resistance to lateral
deflection of the tire [4, 5] is increasing from a minimum (300 N/degree) to a maxi-
mum value (700 N/degree).
In the analytical calculation, the position of the center of mass for the passenger cars
varies between (0.45…0.55)L, (L - wheelbase) in rapport to the front axle [3–5].
Figure 8 highlights the influence of the weight center position on the maneuverability
degree in the case when Citroën C4 is driven unloaded on a concrete dry road, following
a circular trajectory with a radius equal to 16 m, at a constant speed of 11.11 m/s. It is
noted that the position of the mass center determines the understeering, oversteering or
neutral trend in cornering of the passenger cars. The intersection of the two lines that
indicates the drift angles of the front wheels, respectively of the rear axle wheels, marks
the weight center for which the car will be characterized by neutral steering.
Fig. 7. The variation of the drift angle Fig. 8. The variation of the drift angle depend-
depending on the coefficient of resistance to ing on the weight center position for Citroën C4.
lateral deflection of the tire for Citroën C4.
Based on the numerical model and the graphical results interpretation, the following
can be stated:
– MathCad software generates complex calculations and useful charts;
– the behavior of the vehicle can also be analytically appreciated for different
exploitation situations;
– drift angle value increases exponentially with the speed of the vehicle (see Figs. 3, 5
and 6);
– an increased value of the resistance coefficient at the lateral deflection of the tire
determines a decreased drift angle (see Fig. 7);
– uniform mass distribution of the passenger car on the axle leads to a neutral
cornering tendency (see Fig. 8);
– starting from the basics about the stability of the passenger cars in cornering,
mathematical models can be developed for the study of the forces that occur in the
case of the particular movements.
5 Conclusions
Based on the studies undertaken in order to develop this paper, the following were
found:
– interest in the stability and maneuverability of the cars was shown, as a consequence
of the current high travel speeds;
– taking into account the ellipse grip concept, if the vehicle is driven on a rectilinear
trajectory and the longitudinal force (on a wheel) is maximum, then any modifi-
cation on the trajectory will determine the appearance of lateral forces and in order
to maintain the maneuverability of the vehicle, the value of the longitudinal force
will decrease;
– in the specialized literature, there is data indicating the trend of cars movement in
cornering depending on how they are organized;
– regarding the cornering trend of the cars, they are designed to be less sensitive to
minor changes such as the mass redistribution on the axles;
– it is essential that a car possesses a high maneuverability even without the inter-
vention of the electronic stability systems;
– the mathematical model suggests the tendency of the vehicles, but the values
obtained cannot be attributed to a real pattern (due to multiple simplifications).
Taking into account the results obtained using the mathematical model, the fol-
lowing directions of research can be considered:
– the extension of the theoretical research on: how the behavior of the cars is influ-
enced by other parameters than those analyzed in this paper (e.g.: race suspension
springs, the values of toe angle of the wheels, the caster angle); methods to correct
the understeering and oversteering trends of the cars; structure and composition of
the tires in order to obtain reduced values of the lateral deflection angles;
– increase the complexity of the mathematical model used for this research paper in
order to obtain results similar to the real ones and also analyze the vehicle behavior
for others exploitation situations.
References
1. Rajamani, R.: Vehicle Dynamics and Control. Springers, New York (2005)
2. Reza, N.J.: Vehicle Dynamics: Theory and Applications. Springers, New York (2008)
3. Tabacu, Ș., Tabacu, I., Macarie, T., Neagu, E.: Dinamica autovehiculelor: Îndrumar de
proiectare (Vehicle dynamics: Design Guidelines). Pitești, Editura Universităţii din Piteşti
(2004)
4. Todoruţ, A.: Bazele dinamicii autovehiculelor: Algoritmi de calcul, teste, aplicaţii (Vehicle
dynamics Fundamentals: Calculus algorithms, tests, applications). Editura Sincron,
Cluj-Napoca (2005)
5. Untaru, M., Câmpian, V., Ionescu, E., Pereş, Gh., Ciolan, Gh., Todor, I., Filip, Natalia,
Câmpian, O.: Dinamica autovehiculelor (Vehicle Dynamics). Universitatea Transilvania din
Braşov, Sectorul Reprografie U02, Braşov (1988)
6. Wong, J.Y.: Theory of Ground Vehicles, 3rd edn. Wiley, Ottawa (2001)
7. Li, L., Yang, K., Jia, G., Ran, X., Song, J., Han, Z.Q.: Comprehensive tire–road friction
coefficient estimation based on signal fusion method under complex maneuvering
operations. Mech. Syst. Signal Process. 56, 259–276 (2015). ISSN 0888-3270. http://
www.sciencedirect.com/science/article/pii/S0888327014004002
8. Moustapha, D., Sename, O., Dugard, L., Martinez-Molina, J.J., Gaspar, P., Szabo, Z.:
Integrated vehicle dynamics control via coordination of active front steering and rear
braking. Euro. J. Control 19(2), 121–143 (2013). ISSN 0947-3580. http://dx.doiorg/10.1016/
j.ejcon.2013.03.004
9. Nogowczyk, P.: Vehicle steerability evaluation on the basis of random steering input test. In:
10th Mini Conference on Vehicle System Dynamics, Identification and Anomalies, Paper
presented at: VSDIA 2006, 6–8 November 2006, Budapest, Hungary (2006)
10. Panke, D.A., Ambhore, N.H., Marathe, R.N.: Review on Handling characteristics of road
vehicles. J. Eng. Res. Appl. 4(7, Version 4), 178–182 (2014). ISSN: 2248-9622
11. Wang, B., Victorino, A.C., Charara, A.: State observers applied to vehicle lateral dynamics
estimation: a comparison between Extended Kalman filter and Particle filter. In: 39th Annual
Conference of the IEEE Industrial Electronics Society, Paper presented at: IECON 2013,
Viena, Austria (2013)
12. Wang, W., Xi, J., Chen, H.: Modeling and recognizing driver behavior based on driving
data: a survey. Math. Prob. Eng., p. 20, Article ID 245641 (2014). doi:10.1155/2014/245641
13. Wang, Y., Yang, J., Liu, H., Chen, Y., Gruteser, M., Martin, R.P.: Sensing vehicle dynamics
for determining driver phone use. In: The 11th International Conference on Mobile Systems,
Applications, and Services, 25th–28th June 2013, Taipei, Taiwan (2013)
14. ***Atestări inspectori ITP – Reglementările RNTR 1 (aprobate prin OMTCT nr.
2.133/2005, cu modificările și completările ulterioare) (Inspectors credentials ITP -
regulations RNTR 1 (approved by OMTCT no. 2133/2005, as amended and supplemented))
(2005)
15. ***ISO 4138:2004, Passenger cars - Steady-state circular driving behavior - Open-loop test
methods. Revised by ISO 4138:2012 (2012)
16. ***SAE J2181_201109, Steady-State Circular Test Procedure for Trucks and Buses (2011)
17. ***SAE J266, Steady-State Directional Control Test Procedures for Passenger Cars and
Light Trucks. SAE Surface Vehicle Recommended Practice (1996)
Structural Dynamic Applications Using
Principal Component Analysis Method
Mircea Niculescu1, Cristi Irimia1, Ioan Calin Rosca2(&),

Mihail Grovu1, and Maria Violeta Guiman2
1
Siemens Industry Software, Brasov, Romania
{mircea.niculescu,cristi.irimia,
mihail.grovu.ext}@siemens.com
2
Transilvania University, Brasov, Romania
{icrosca,violeta.guiman}@unitbv.ro
Abstract. Principal component analysis (PCA) is a statistical procedure that

uses an orthogonal transformation to convert a set of observations of possibly
correlated variables into a set of values of linearly uncorrelated variables called
principal components. The PCA method enables to reduce the study of a
complex noise and vibration problem, with multiple partially correlated refer-
ences, to the study of independent, uncorrelated problems. This paper describes
systematic processes for road noise improvement along with measurement and
analysis process. The noise sources are identified by using a source decompo-
sition method. In the next step the main noise paths are identified by using a
transfer path analysis method (TPA). Based on obtained results, the design
modification of body panels is suggested for road noise reduction by using a
panel contribution analysis. Finally the method will be applied to road noise
reduction process for a new vehicle.
Keywords: Noise reduction Transfer Path Analysis (TPA) Principal

Component Analysis (PCA) Panel contribution
1 Introduction
In the past decade the needs of the automotive industry have changed due to economic
and ecological reasons. The development cycles have become shorter and more cost
sensitive. The interior noise and vibration characteristics (NVH) of vehicles have
become important factors in influencing customers’ options. Many engineering tools
have been developed for analyzing and optimizing vehicle NVH. An important group
is represented by the transfer path analysis methods (TPA). The purpose of these
methods is to analyze the noise transfer between noise sources and a receiver [1]. For
example, the engine in a car is a noise and vibration source and the driver or another
passenger is the receiver. Knowing the noise source, there are many connections and
possible paths through which noise and vibration energy can arrive at the receiver.
Using TPA the critical loads and propagation paths can be identified.
The solutions for improving the road induced interior noise problems on driving
also are based on TPA. The source of road noise comes from contact between tire and

Structural Dynamic Applications Using Principal 91
road surface, but the most important transfer paths depend on components and coupling
elements. In the first step this paper first identifies contribution of structure borne noise
to interior noise from tire. Secondly, the contribution of the body panels to interior
noise is characterized. Finally, the body panel improvements are discussed. These steps
are applied to a new passenger car development process.
2 Deterministic Versus Random Loads
Deterministic Data. It is said about these data that are deterministic when both the
excitations and the responses are described by deterministic frequency spectra. A rep-
resentative example of deterministic data is the engine noise, Fig. 1. If it is considered
that the car is excited only by the engine, the excitation/response relation is related to a
single phenomenon, the engine turning. This means that measured crosspower data can
be converted into frequency spectra reported to the engine signal. One reference will
clarify all the responses. The Transfer Path Analysis will use the reference spectra as
input loads.
Random or Stochastic Data. Atypical example it is represented by the vibrations
generated in the car by the road inputs, Fig. 2. The crosspowers and autopowers
contain information regarding the correlation between excitation and response signal
with respect to other excitations/responses and itself. The input signals are correlated
and one needs a randomization approach to estimate the degree of correlation between
them. This indicates that it needs more than one independent source to explain all the
responses.
Fig. 1. Engine noise and vibration as Fig. 2. Random vibration caused by the road
deterministic excitation input
Multiple References without Correlation. When multiple phenomena are acting, it is

needed to consider more references to create the referenced spectra of the excitations
for each reference. If there are two references, ref 1 and ref 2, and an excitation
signal, excit1, than the crosspower matrix is given by:
92 M. Niculescu et al.
2 3
ref 1 ref 1 ref 1 ref 2 ref 1 excit 1
4 ref 2 ref 1 ref 2 ref 2 ref 2 excit 1 5 ð1Þ
excit 1 ref 1 excit 1 ref 2 excit 1 excit 1
The values ref 2 ref 1 and ref 1 ref 2 evaluate the coherence between the ref-
erence signals. If they are zero, there is no correlation between them and it is sufficient
to create the referenced spectra for excit1, with each reference. The forced response is
calculated with each of the spectra and sums the two responses to obtain the total
response. The addition is carried out using an energetic formulation as shown below:
ðoutputÞ2 ¼ ðoutput 1Þ2 þ ðoutput 2Þ2 ð2Þ
Multiple References with Partial Correlation. When the references are partially
correlated, it is necessary to calculate multiple referenced spectra. Following the pro-
cedure mentioned above to compute the responses would result in overestimating the
output, because each of the two referenced spectra does not only contain the infor-
mation related to one reference. As the reference is correlated to another reference, it
results the non-zero values ref 2 ref 1 and ref 1 ref 2 for the matrix shown
above. The Principal Component Analysis can solve this.
3 Road Noise Principal Component Analysis (PCA)
The Principal Component Analysis (PCA) allows reducing the study of a complex
noise and vibration problem with multiple partially correlated references, to the study
of independent, uncorrelated problems. The PCA splits the non-coherent vibrations into
coherent sets of vibrations and assigns the latter to the ‘virtual’ sources that can be
treated in a similar way to coherent sources. Singular Value Decomposition
(SVD) techniques are applied on a multi-reference crosspower matrix. The crosspower
matrix is decomposed into its principal components, which gives as many single ref-
erence sets of crosspowers as there are references. These new sets of references are
called principal components. The PCA process is shown in Fig. 3.
The references are the measurement locations, representative for the phenomena
that make the structure’s dynamic behaviour. The responses, or slave measurements,
are the measurement locations which are not used as references. They describe the
dynamic response as it is observed by the operator of the vehicle under test. The
reference signals are in general partially correlated to each other. An uncorrelated set of
principal component spectra can be obtained by using a principal component decom-
position of this reference set.
Virtual Crosspower Spectra. With the multi-reference crosspower averaging, data is
condensed into the following two matrices:

– the autopower matrix SX ;X of the references;

– the crosspower matrix SY ;X of the responses with respect to these references.
“slave” measurements
Y2
Y2
Y1 X2
Principal component X’2
Y1
X1 decomposition X’1
Y3
Y3
Partially coherent Orthogonal

reference set reference set
[SY,X]: multi-reference [SY,X’]=[SY,X][U] [SY,X’]: single reference
crosspower matrix crosspower matrix
[SX,X]: autopower matrix of [SX’,X’]= [U][SX,X][U]H [SX’,X’]: principal component
references autopower matrix
Fig. 3. Principal component analysis process

The reference autopower matrix SX ;X is decomposed by the eigenvalue decom-
position into its orthogonal, independent or principal component autopower matrix
SX 0 ;X 0 , the so-called virtual reference autopower:

SX ;X ¼ ½U SX 0 ;X 0 ½U H ð3Þ
The matrix ½U is the unitary eigenvector matrix, and ½U H is its Hermitian trans-
pose. The principal component autopower matrix SX 0 ;X 0 is a diagonal eigenvalue
matrix, the principal components being orthogonal with respect to each other. It rep-
resents the new orthogonal reference set, or the set of virtual independent references.
The elements of this diagonal areSii ði ¼ 1; nref Þ.
The crosspower matrix SY ;X can be transformed to a set of crosspowers with
respect to this new orthogonal virtual reference set. This gives:

SY ;X 0 ¼ SY ;X ½U ð4Þ

The matrix SY ;X 0 contains the virtual crosspower spectra, which are ‘‘single
reference’’ crosspowers, with respect to the virtual orthogonal references (principal
components). The elements of this matrix contain for a response j and a virtual ref-
erence i’, Sji0 .
Referenced Virtual Spectra. From the virtual crosspowers, the referenced virtual
spectra can be derived. This is finished by dividing the single reference virtual
crosspowers by the square root of the autopowers of each of the principal components.
These referenced spectra are to be used as input data to the Transfer Path Analysis
(TPA). For a response j and a virtual reference i’, the referenced virtual spectrum is
.pffiffiffiffiffiffi
ffi
calculated as: Sji0 Si0 i0 . The outcome of PCA consists of n sets of the referenced
virtual spectra, n being the number of references chosen in the PCA.
Referring to the principal component autopowers, the following observations can

be made:
– the singular values ri are the autopowers of the virtual references or the principal
component autopowers;
– the virtual references X’ are uncorrelated (diagonal matrix);
– principal component autopowers are sorted: r1 [ r2 [ . . . [ rn and offer an
indication of number of independent active phenomena that build up the structure’s
dynamic behaviour.
The referenced virtual spectra are used as the operational data for the transfer path
analysis model. The different virtual spectra related to the different independent sources
are considered as individual load cases, which can be processed in sequence.
4 Road Noise Reduction Using Source Decomposition

and Panel Contribution Analysis
4.1 Road Load Source Decomposition Process

The interior noise in automobiles is generally classified into airborne and structure
borne noise. The structure borne noise is influenced by the mount characteristics and
vehicle structure dynamics. The powertrain inputs and road inputs are the major
contributors to the structure borne sound problem [2–6].
First application of this study includes two purposes [7]:
– To find out the principal components (virtual references) those contribute to the
complex noise and vibration problems in order to reduce them;
– To compute the non-coherent steering wheel vibrations caused by the road load
inputs in the time domain, which are applied at four tire patch excitation points.
RL
FL RR RL
FR
FL=front left RL=rear left

FR=front right RR=rear right
Load – time history [N]
Fig. 4. Car model, tire patches and load time history
The load application points are the tire patches, where the large masses are
attached, Fig. 4. The interface points are the subframe and steering system points and
the response points are the steering wheel points.
The history loads contain the measured road load time data. These are derived from
a road profile measurement, which contains two random signals, one for the front-left
(FL) tire and one for the front-right (FR) tire in the Z-direction. For the rear tires the
loads in the time domain are the same, but they are delayed with some time. The forces
are high in amplitude as they were pre-processed to apply to a model with large masses
attached to the tire patches. These road loads are coming from a test database.
In the next step a crosspower set of the road loads is calculated. The crosspower set
handles crosspowers and autopowers. Crosspowers describe the spectral contents of the
part of a response signal that is correlated to a reference signal, whereas principal
components that relation between a response signal and a virtual reference signal. The
crosspower set will be used for road noise principal analysis case.
In Fig. 5 it is shown the total displacement of one point of the steering wheel as an
output response to the virtual reference spectra.
Fig. 5. Total displacement of the steering wheel point generated by the virtual spectra
The total displacement spectrum is shown for all virtual references (principal
components) together and the displacements (responses) for each virtual reference
separately over the complete frequency range. The responses (vibrations) for all
principal components (virtual references) are added by using the RSS averaging option,
because the energetic sum must be used as these responses are referenced towards the
orthogonal (uncorrelated) references.
4.2 Panel Contribution Analysis Process

Identification of problematic area in early stages of vehicle development is considered
as the most efficient way to reduce time and cost of design. Panel contribution analysis
is introduced as one of the most effective diagnostic method to identify the source of
interior acoustic problem. In this paper identification of panel contribution is realised
using measured road loads derived from a road profile measurement.
The vibro-acoustic performance of the car body is investigated by using finite
element analysis. The analysis process is shown in Fig. 6.
Structural FE Model
- Correlated random loads Acoustic FE Model
- Modal results
Projected on the
Acoustic model Pressure
response
Fig. 6. Vibro-acoustic analysis process
At first structural model of the car is made. Modal analysis of the structure allows
identification of the normal modes and frequencies. The structure is excited with virtual
loads (principal components) at the points of contact of the tires with road. Inside the
cabin is meshed yielding of the acoustic mesh. The cab structural mesh is projected to the
interior acoustic mesh. Sound pressures near the driver and passenger ears are analyzed.
Contributions of thirteen dominant panels are calculated for excitation forces.
Individual and total pressure responses spectra in the driver ears depending on the
virtual references are shown in the Fig. 7.
Fig. 7. Individual and total pressure responses spectra in the driver ears
Panel contributions at different frequencies in the driver left ear acoustic pressure
are presented in Fig. 8.
Acoustic pressure in the

driver left ear location
Fig. 8. Panel contributions in the driver left ear acoustic pressure
Major panel contributions in the driver left ear acoustic pressure for specified
values of frequency are shown in Fig. 9.
Front right door at 29Hz
Rear seat floor at 66 Hz
Rear seat floor at 53Hz

Wind shield at 88Hz
Fig. 9. Major panel contributions in the driver left ear
5 Conclusions
Coherence and Principal Component Analysis methods are commonly applied to

analyse the physical interrelations between stationary multichannel test data. Non-
coherent vibrations decomposing into a set of coherent vibrations (main components)
avoid overloading responses to such input loads. Figure 10 shows the RMS averaging
of the acoustic pressure spectra in the analyzed target responses (left, right ears of the
Fig. 10. Acoustic pressure spectra responses in the target locations for correlated vs. coherent
input loads
driver and passenger car) using correlated (red color) vs. coherent loads (blue color).
The response overloading is estimated as being 5 dB.
Panel contribution analysis ranks a set of panels with respect to their sound pressure
level contribution at a target position. The predicted contributions from all panels under
study are combined to find their sum.
The major panel contributions in the driver left ear acoustic pressure are the fol-
lowing: front right door at 29 Hz, rear seat floor at 53 Hz, rear seat floor at 66 Hz and
windshield at 88 Hz. An example of structural modification proposal for a new model
car is presented in Fig. 11.
Front doors at 29Hz
Structural doors proposal:

window frame needs
to be added
Fig. 11. Structural modification proposal for the front doors
Based on structural analyses one can conclude that for a diminishing of the road
noise generated by the front doors at 29 Hz, the windows frame needs to be added.
References
1. Van der Auweraer, H., Mas, P., Dom, S., Vecchio, A., Janssens, K., Van de Ponseele, P.:
Transfer path analysis in the critical path of vehicle refinement: the role of fast, hybrid and
operational path analysis. In: Proceedings of the SAE N&V Conference, Traverse City
(USA), SAE paper 2007-0102352 (2007)
2. Donders, S., Hermans, L., Nauwelaerts, E., and Chojin, S.: CAE technologies for efficient
vibro-acoustic vehicle design modification and optimization. In: International Conference on
Noise and Vibration Engineering, ISMA 2008, Lueven, Belgium, pp. 4127– 4142 (2008)
3. Koners, G., Panel noise contribution analysis: an experimental method for determining the
noise contributions of panels to an interior noise. SAE (2003)
4. Wolff, O.: Fast panel noise contribution analysis using large PU sensor arrays. In: Proceedings
of Internoise 2007, Istanbul, Turkey (2007)
5. Castillo, M. S., Schuhmacher, A.: Optimization of Sound Packages for Commercial Vehicles
Using Panel Contribution Analysis. SAE (2009)
6. Yoo, B. K., Chang, K.: Road Noise Reduction Using a Source Decomposition and Noise Path
Analysis. SAE (2005)
7. Virtual Lab User Guide: Virtual Lab 12. SL1, User Manuals
Weight Optimization of Existing Transmission
Housing for Heavy Duty Commercial
Vehicles Using Correlated Simulation
and Experimental Tools
Kandreegula Suresh Kumar(&) and Dheeraj Kumar Singh
VE Commercial Vehicles Ltd., Pithampur, India

{skkandreegula,dksingh}@vecv.in
Abstract. In the current market trends there is always a need to arrive for light
weight designs. For commercial vehicles, an attempt is made to re-design of
existing cast iron transmission housing-considering its structural strength and
stiffness. Advanced non-linear topology optimization methods have been
addressed as the most promising techniques for light weight and performance
design of powertrain housings.
Instead of conventional approach, a new finite element simulation approach
has been adopted to align the project time plan (design and development time) to
predict the stiffness and static factor of safety under defined VE commercial
vehicles limited durability load cases.
For this, it was used non-linear topology optimization technique and it was
experimentally verified. With this study, 20 % weight reduction was achieved in
transmission housing with the same overall structural stiffness – hence, same
methodology can be implemented in other structural housing.
Keywords: Transmission housing Topology optimization Powertrain HD

commercial vehicles Gap elements Non-linear RLDA
1 Introduction
The nature of commercial vehicle operation has always led the manufacturers to look
for ways to reduce operational costs in order to maximize the profits for vehicle
operators. The key factor that directly affects the profitability of commercial vehicle
operation is fuel efficiency of a vehicle as it accounts for majority of the operational
costs. The optimization of fuel efficiency can be achieved by weight reduction of
vehicle components and sub-systems. Weight reduction also provides additional
advantage as it is beneficial in loaded vehicle condition and unloaded condition. With
the advent of FEA techniques and computer aided analysis it has become possible for
the industry to explore complex component designs without exponential increase in
developmental costs.
The powertrain consists of the engine and drive train (i.e., transmission, clutch,
drive shafts, and drive axles). In HD commercial vehicles, the powertrain accounts for
about 40 % of the total weight. By increasing specific power density and optimization

Weight Optimization of Existing Transmission 101
of designs, vehicle weight can be reduced. Opportunities for weight reduction in the
drive train include lighter weight engines, transmissions and clutch housing [1].
For the HD commercial vehicles Powertrain component i.e. Transmission Housing,
an attempt is made to replace heavy transmission assembly with an optimized design in
order to take higher payloads and possess higher overall efficiency. For HD commercial
vehicles, the observed engine torque range is around 600 Nm * 900 Nm; hence
existing cast iron transmission housing has a 7 mm thick wall thickness with cross ribs
around the housing.
In the current market trends there is always a need to arrive at designs that lower in
weight & carry more pay load without compromising its strength bearing all the worst
gear loads. Recent developments in material technologies have made aluminum a
suitable replacement for light weight designs. However, the stiffness of aluminum is
considerable lower than the cast iron material, this requires base thickness of aluminum
design has to be increased by at least 30 %. This implies existing design base thickness
need to be increased from 7 mm to 9 mm, which will not at all commercially manu-
facturable for aluminum designs. And also, if we go for customized process for alu-
minum, the tooling cost increases by 40 % * 50 %. This is not viable, as it will
directly impact on the vehicle cost. Additionally due to low volume sales in this
platform, the time for return on investment is considerably high – which is not
favorable to the customer.
Hence, it is decided that the existing cast iron design has to be optimized using
CAE optimization techniques – with minimum changes in the existing tooling, without
changes in the existing design envelop and compatibility of other mating components
in the same vehicle platform.
To achieve same existing/proven cast iron housing stiffness for new housing
stiffness with defined worst gear loading conditions, the existing cast iron housing has
been analyzed first for defined worst gear loading conditions to know the stiffness at
different known bearing locations. The idea was to come with new housing with the
same stiffness arrived at the same locations as existing cast iron housing. Total 6 load
cases have been analyzed to counter all gear loads with 36 responses using Altair
OptiStruct software to meet the existing/proven cast iron design stiffness at the bearing
locations. The stiffness values for the new optimized transmission housing are equiv-
alent to that of existing proven cast iron transmission.
2 Topology Optimization
It is a mathematical approach that optimizes material layout within a given design

space, for a given set of loads and boundary conditions such that the resulting layout
meets a prescribed set of performance targets. Using topology optimization, engineers
can find the best concept design that meets the design requirements [2].
Topology optimization has been implemented through the use of finite element
methods for the analysis, and optimization techniques based on the method of moving
asymptotes, genetic algorithms, optimality criteria method, level sets, and topological
derivatives.
102 K.S. Kumar and D.K. Singh
Topology optimization is used at the concept level of the design process to arrive at
a conceptual design proposal that is then fine-tuned for performance and manufac-
turability. This replaces time consuming and costly design iterations and hence reduces
design development time and overall cost while improving design performance [3].
In some cases, proposals from a topology optimization, although optimal, may be
expensive or infeasible to manufacture. These challenges can be overcome through the
use of manufacturing constraints in the topology optimization problem formulation.
Using manufacturing constraints, the optimization yields engineering designs that
would satisfy practical manufacturing requirements. In some cases, additive manu-
facturing technologies are used to manufacture complex optimized shapes that would
otherwise need manufacturing constraints.
3 Process Implementation
First to make a design space without changing mating/interface parts and applied loads
& boundary conditions in the concept design. Then stiffness of concept design has to
compare to the existing/proven TM housing stiffness. This process continues till
concept/optimized design stiffness is correlated with existing design stiffness using
topology and size optimization. Final optimized design has to check for DFM (Design
for manufacturing) and DFA (Design for assembly). This process of topology opti-
mization is shown in the Fig. 1.
Fig. 1. Optimization – process implementation

4 FE Modelling
The geometry of the transmission housing is modeled in Pro-E, where the complete
wireframe generated. The data is then translated to Initial Graphics Exchange Speci-
fication (IGES/STP) format and read into Hyper Mesh™, where the transmission
housing is FE modeled by 3D Second order Tet mesh generation - using Tetra10 solid
elements, i.e. 3D (2nd Order) tetrahedral elements with 10 nodes.
After meshing transmission housing, all bearings (2 counter shaft bearings, 2 main
shaft bearings and 2 idler shaft bearings are meshed with the same element sizes and
connected all bearings to the transmission housing with the help of PGAP elements,
which will represents contact between the mating components, as shown in the Fig. 2.
Transmission Housing consists of 6 bearings, i.e. Input shaft (B1), Counter shafts
(B2 & B3), Main shaft (B4) and Idler Shaft Bearings (B5 & B6).
Fig. 2. Transmission housing - FE model
5 PGAP Elements
Any model with non-linear CGAP (gap) elements uses the non-linear solution
sequence. The element can be used for both static analysis and optimization; however,
the assumptions of small displacement and small strain still apply.
Gap elements (CGAP) are defined node-to-node and require a property card and an
orientation vector to be assigned to them. Gap element behavior is defined through the
PGAP property card. Before creating a gap element, it is recommended to create the
PGAP property card and then assign this property to the gap elements in the gaps panel
(or other panels) in Hyper Mesh. Most of the panels that create 1-D elements (for
example linear 1-D) can also be used to create gap elements by changing element
configurations [4].
On the PGAP card (PGAP – Gap element property), the initial gap opening is
defined by U0. The gap element offers very low stiffness until the gap is closed.
Friction can also be specified, and gets activated once the gap is closed.
6 Activating Non-linear Solution (NLPARM)
OptiStruct uses its linear solution sequence by default even if gap elements are present.
A non-linear solution is activated for a sub case by including the NLPARM parameter.
If NLPARM is not present in the sub case definition, the gap element has linear
behavior. In this situation, the gap status is determined once at the beginning of the
solution, and does not change as the solution progresses [4].
In Hyper Mesh, the NLPARM parameter can be defined by creating a load collector
with card image “NLPARM.” The NINC parameter on this card represents the number
of equal subdivisions that the total load in a given sub case will be divided into.
If NINC is blank, the entire load for a given sub case is applied at once. In most cases,
this is the only parameter that needs to be defined. This NLPARM load collector needs
to be included for the sub case along with static loads and SPCs.
7 Structural Analysis
According to the applied loadings originating from different categories of mechanics,

this non-linear elastic analytical procedure could further be divided into six load steps by
means of the superposition principle for simulating various operating processes of the
VE Commercial vehicles Ltd – CAE standard powertrain load cases, a shown below:
(1) 5.5 g Vertical load + First gear loads
(2) 5.5 g Vertical load + Second gear loads
(3) 5.5 g Vertical load + Third gear loads
(4) 5.5 g Vertical load + First gear loads
(5) 5.5g Vertical load + OD gear loads
(6) 5.5 g Vertical load + Reverse gear loads
Transmission housing assembly consists of all internal & external parts for CAE
structural analysis (housing, upper case, gears, shafts, bearings, etc.), approx. 250 kg.
The transmission housing is constrained at the front side and all gear loads are
applied at the Input shaft (B1 @ front), Counter shafts (B2 @ front & B3 @ rear), Main
shaft (B4 @ rear) and Idler Shaft Bearings (B5 @ front & B6 @ rear) locations, as
shown in the Fig. 3.
Fig. 3. Boundary conditions

First, the existing cast iron housing has been analyzed for defined worst gear
loading conditions to know about the stiffness at different known bearing locations, i.e.
B1 to B6. The idea was to come with new housing design with the same stiffness
arrived at the same locations as existing on the cast iron housing.
Total 6 load cases has been analyzed to counter all gearings with 36 responses
using Altair OptiStruct to meet the existing/proven cast iron design stiffness at the
bearing locations i.e. B1 to B6.
Figure 4 shows the design space for optimization. The wall thickness for design
space is taken as 10 mm because existing design has a 7 mm thick with cross-over ribs.
The design iterations from Optimization of new transmission housing are shown in
the Fig. 5a to d. After 8 iterations, the stiffness values are matched with existing design
stiffness values at all 6 bearing locations.
Finally, we have checked the strength of Iteration 9 design, that is safe under all
peak gear loading conditions and static factor of safety is more than acceptable
(FOS > 1.6) and is comparable with the existing transmission housing.
Fig. 4. Design space for optimization
8 Static Analysis Results
Figure 6 shows the stress plots for both existing design and Optimized design under
worst load case i.e. 5.5 g Vertical Load + Reverse Gear Loads. Both static results are
almost same with 15 kg weight saving.
Table 1 shows, the comparative (Existing design and Optimized design) results
summary for all 6 load cases. The static FOS are lower than the acceptable FOS 1.0.
Fig. 5. a. Iteration 1 Optimization results, b. Iteration 3 Optimization results, c. Iteration 5

Optimization results (all deflections values are in mm), d. Iteration 9 Optimization results (all
deflections values are in mm)
Fig. 6. Static analysis results
Table 1. Static analysis results summary
9 Experimental Verification
As per VE Commercial Vehicles Ltd, standard durability rig has been setup as shown
in the Fig. 7 and then tested the Optimized transmission assembly in which TM
housing is the test component. TM housing at each gear for maximum torque condition
has been evaluated with defined duty cycles (Torque 900 Nm, 2000 RPM) as shown in
the below Table 2. This Optimized TM design has completed the entire duty cycles
without any failures and passed the durability testing cycle.
Fig. 7. TM housing-rig setup
Table 2. TM housing - duty cycles
As per VE Commercial Vehicles Ltd, standard endurance test cycle has been
carried out on 3 vehicles for 180000 km each. Therefore a cumulative of 540000 km
has been covered on the vehicle without any failure. Hence vehicle validation has been
concluded and certified as OK for implementation on regular vehicles [5].
10 Conclusions
Using the proposed cast iron transmission housing in our organization –

• It is estimated that cost savings is of up to 10 %
• Reduction in weight is by 15 kg (20 %)
Almost 90 % correlation achieved with Strain test data and CAE simulation tool.
Topology Optimization process can be implemented for all existing costlier CI parts
– To reduce cost & overall weight of the vehicle.
An accurate FE simulation methodology is established to reduce the product
development time & cost of the design - To align the project time plan and to ensure
First Time Right at the development phase.
Time saving in product development with respect of new transmission housing
design – Which improves the product performance, reliability and durability.
Acknowledgments. The authors would like to express their gratitude to Mr. Rajinder S.
Sachdeva, Executive Vice-President, Technology Development, VE Commercial Vehicles Ltd.,
for granting permission to publish this work. Special thanks to Mr. Vimal Mulchandani, Vice
President, CVD and Mr. Sanjay Pakale, Dy. General Manager, Powertrain and Mr. Umashanker
Gupta, Dy. General Manager, CAE for their guidance and encouragement to develop present
work. Also, the contributions and support of all CAD, CAE engineers and respective testing
engineers are highly appreciated.
References
1. Trucks and Heavy-Duty Vehicles Technical Requirements and Gaps for Lightweight and
Propulsion Materials, U.S Department of Energy, Energy Efficiency & Renewable Energy,
Vehicle Technologies Office, February 2013
2. Kandreegula, S., Sukumar, N., Endugu, S., Gupta, U.: Structural non-linear topology
optimization of transmission housing and its experimental verification. In: SAE Technical
Paper 2015-01-0098 (2015). doi:10.4271/2015-01-0098
3. Ogata, Y., Suzuki, S., Hiraoka, M.: Development of topology optimization method for
reduction of transmission housing weight. In: SAE Technical Paper 2005-01-1699 (2005).
doi:10.4271/2005-01-1699
4. Altair – OptiStruct Help
5. Yadav, V., Londhe, A., Khandait, S.: Correlation of test with CAE of dynamic strains on
transmission housing for 4WD automotive powertrain. In: SAE Technical Paper
2010-01-0497 (2010). doi:10.4271/2010-01-0497
New Approaches in Designing a Race Car
Chassis for an Engineering Competition
Doru Ionut Furnea(&) and Silviu Butnariu
Transilvania University of Braşov, Braşov, Romania

ionutfurnea@yahoo.com, butnariu@unitbv.ro
Abstract. This paper presents the design and verification procedure of a race
car chassis, built respecting the rules of an international competition dedicated to
young researchers from all universities of the world. We used CAD/CAE
methods. The designed chassis has been subjected to various analyzes, corre-
sponding to the different stresses during operation: braking, cornering, asym-
metric loads, frontal impact, side impact, roll-over, torsional stiffness, buckling,
bending, modal analysis and transient analysis. A total of 3 versions of chassis
have been analyzed to extract the best results. After analysis, we could choose
materials, pipe size, and chassis structure. This project demonstrated that the
proposed chassis design meets the needs of the engineering competition
regulations.
Keywords: FEM analysis Design Car chassis Ansys Solidworks
1 Introduction
Motorsport, mainly formula type cars were always the starting point of all innovations
and progress which can be seen on the road cars daily. It is a highly competitive
environment in which high end research is achieved. There are more competitions
which gives researches from around the world a glimpse of how the motorsport
industry operates. These competitions are designed to challenge young researchers
from universities all over the world to plan, design, build and compete with formula
style vehicles.
One of these engineering competitions, run by the Institution of Mechanical
Engineers, which is held annually in countries around the world, causes various teams
from different universities to design, build, test and race a small single-seat, open
wheeled race car [2]. Typical for racing competitions, the teams must follow a strict set
of rules. To support the innovations of the teams, regulations are set to a minimum.
Points are awarded on a variety of events to the teams. By gaining the most points
the team wins the competition. The judges are industry specialists and high profile
engineers who award points in order to win the competition [3]. Teams are challenged
to design and build race cars in order to compete in static and dynamic events to
demonstrate their understanding of the engineering behind the car and also to test the
performance of the vehicle. In present paper, the aim is to design and test the car
chassis. This component may have different types of design including: ladder frames,
space frame, monocoque and backbone style.
New Approaches in Designing a Race Car Chassis 111
The proposed chassis is a space frame style design, using Computed Aided Design
(CAD), it has been created in Solidworks. After design, using finite element analysis
(FEA) techniques, different loading conditions have been simulated in order to assess
the build quality and design and how the chassis would react in different situations.
Optimizations helped very much due to the visualization of the critical areas and
also over-build areas of the chassis.
After the FEM analysis, it was found that under different load, the chassis per-
formed as intended and it was easily able to withstand stresses that were created under
specific conditions.
2 Problem Definition
This type of chassis is a development of the earlier ladder frame (Fig. 1). In both the
space frame and the tube-frame chassis, the main parts of the vehicle are attached to a
skeletal frame of tubes, and the body panels do not have a structural function.
Fig. 1. Chassis diagram
The advantage of the space frame rather than the open channel sections is that they
resist to higher torsional forces. The first space frame chassis were produced in 1930s
by designers who took them as models from architecture or aircraft design.
A perfect designed space frame would contain tubular sections arranged so that the
only forces acting on them are either tension compression. This is a great advantage as
materials have a better resistance to compressive and tensile loads rather than bending
loads. Since the space frame chassis were introduced in car racing in the 1960s, many
high end sports cars have been designed with this type of chassis due to the advantage
it gives in weight reduction while keeping its rigidity.
Until now, the team has been building the chassis with circular tubes made of
material SAE 4130. It has an excellent strength to weight ratio and a good weldability.
The 25CrMo4 is an alloyed heat treatable steel with a tensile strength of 750–
950 N/mm2. It is used mainly in automotive and aviation industry for components with
high toughness as axles, turbine parts, turbine rotors, crank shafts. Problems are met
when trying to attach the suspension system and the steering system to the structure.
Additional material may be required to have the required stiffness and strength
demands in specific areas. The necessary addition of tubes increases the complexity of
112 D.I. Furnea and S. Butnariu
manufacturability and the vehicle mass and center of gravity height reduces the han-
dling performances.
The crash impact system is located mainly at the front of the car. Regulations
determine the maximum decelerations and dimensioning. As an impact absorbent
material, foam has been used along with a mandatory 4 mm thick aluminum
anti-intrusion plate bolted to the chassis to meet the standards. Due to relatively low
speeds in the competition, aerodynamics are less important for this paper. Underbody
effects have been used to improve the cars performance. These are outside of this
project and are researched separately to improve future designs.
To summarize the above, the challenge becomes: design a lightweight, formula
type, race car space frame chassis considering ergonomic and impact properties,
meeting FSAE strength and stiffness regulations while creating an easy to manufacture
and low cost production chassis.
In order to achieve a high quality design, the following inputs are used: the loads
(it have been analyzed previously and are used to investigate the location of different
loads that act on the chassis); the structural limitations and possibilities with respect to
the feasible materials are investigated; safety and ergonomics are assessed; the engi-
neering aspects regarding the weight versus stiffness and manufacturability are
presented.
The main function of chassis are the following: to support all the loads attached to
it and load of driver; to withstand torsional deformations caused by the movement of
the vehicle; to withstand the forces generated by cornering, acceleration and braking of
the vehicle; to control the vibrations which are caused by the movement of the vehicle
on the road; to withstand frontal and side impact forces in order to protect the driver.
The principal loads on the chassis are from the suspension system and reaction
forces of the driver and engine (main masses of the race car) to the lateral and vertical
accelerations.
The chassis will be subjected to different types of loads that will simulate how it
would react under different situations. These are braking, cornering, asymmetric loads,
frontal impact, side impact, roll-over, torsional stiffness, buckling, bending, modal
analysis and transient analysis.
3 Methods and Materials
The main property of a race car is its performance. In the present design hypothesis, the
biggest advantage is achieved only through exceptional handling of the vehicle.
Besides the mass of the car, other important parameters can influence the cars handling.
Every competition has a set of rules to provide the fundamental safety and to avoid the
competition from becoming unfair. For designing the chassis, a number of very strict
rules has been taken into consideration, mainly concerning safety issues. In order to
achieve the objectives, we used the following actions: assessment performed before the
implementation of project ideas, to compare proposals; designing of race car chassis
based on FSAE requirements, design constraints and ergonomic constraints (tools:
Solidworks); checking the weight and influence of dynamic factors on the dynamic
behavior of the chassis (tools: Excel, MathCad); verification using FEM (tools: Ansys).
For this chapter of the report, data calculated will be used to simulate loads
encountered by the chassis of the car during braking, cornering, asymmetric and tor-
sional forces using finite element method. Using this data, displacements and areas of
high stress can be found and optimized in order to reduce the stress concentrations and,
if possible, the number of tubes in the frame and in doing so, reducing the overall
weight of the car and the manufacturability complexness [5].
For each of the simulations, boundary conditions have been added to make the
simulations as accurate as possible replicating actual racing conditions [6].
Forces have been added to nodes on the chassis that reflect the points where the
suspension is attached to the frame. Figure 2 shows 4 different tube sizes being used in
the chassis.
Fig. 2. BS16 chassis with tube sizes
One of the biggest influence factors in the competition is the driver. The chassis
can support the driver in his performance by ensuring that all controls can be easily
reached, by having a comfortable seating position and a good visibility over the front of
the chassis. The factors used in seeking a good seating position for the driver are:
horizontal and vertical position and angle of the seat with respect to the horizontal
plane; reclined or upright position; position of the steering wheel; horizontal and
vertical position of the pedal assembly; height and position of dashboard and front
hoop.
The cost is very important for the team to be able to achieve good quality and
performance from the overall race car, but even more so the technical skills that are
required to construct it. The material and tooling costs must be kept low. Space frame
chassis have some disadvantages due to enormously time consuming and costly to
build. Also production cannot be done using robots due to the structure’s complexity.
Therefore, the complexity must be kept limited to facilitate the team to build the
chassis themselves, reducing manufacturing costs for the project.
Apart from the structural requirements regarding the crash safety, there are some
tight regulations regarding the cockpit to keep the driver as safe as possible (Fig. 3).
This involves dimensional demands for the cockpit to allow a quick exit – or driver
egress- in case of accidents. For these dimensional demands, templates are being used.
The principal loads on the chassis are from the suspension system and reaction forces
of the driver and engine (main masses of the race car) to the lateral and vertical
accelerations. The suspension geometry is shown, along with the positions of the driver
and the engine. Connecting rod forces are simply push and pull forces. The chassis will
be subjected to four different types of loads that will simulate how it would react under
different situations. These are braking, cornering, asymmetric loads and bending.
Fig. 3. Geometric rules considering the front and main roll hoops [4]
The suspension system provides a smooth ride even when the wheels traverse
uneven terrain. This is achieved by damping the vibration that get transmitted to the
chassis through the wheels. The double wishbone suspension is used commonly in high
performance vehicles due to its superior kinematic response.
These are loads experienced by the car during deceleration. The loads result from
the contact between the contact patch of the tire with the tarmac. These forces are
transferred through the suspension system onto the chassis through the mounting
points.
Cornering loads are the loads produced by the vehicle tire during cornering. During
cornering maneuvers, loads act on the wheels, mainly on the outside of the corner. The
greater the lateral acceleration the greater the centrifugal force and the greater the tire
side forces we take place. The vehicle tends to roll due to a lateral acceleration. This
roll will vary directly with the vehicle’s roll movement and the mass of the vehicle. The
part of the car that is going to roll is the sprung mass. Centrifugal force, being an
acceleration, will act through the center of gravity of the sprung mass.
4 Results
4.1 Linear Braking
This simulation scenario shows how the chassis is supporting all the loads coming from
the wheels of the car. Forces were applied to the front and rear suspension joints
(Fig. 4). The front had a force of 1275 N applied to each wheel while the rear had a
force of 930 N applied rear.
Fig. 4. Forces and boundaries on the chassis
Fig. 5. Total deformation - Linear Braking
This was done to simulate the weight transfer to the front of the car, and by this, to
simulate the grater braking at the front of the chassis. This simulation is reflected as a
greater displacement at the front of the chassis as can be seen in the figure below.
It can be seen in Fig. 5 that approximately 50 % of the members of the chassis
carry stress during braking. The maximum stress is resulted in the node between the
Main Hoop and the back of the chassis at only 2 points to only 18.92 MPa. The yield
strength of the 25CrMo4 being at 460 MPa, this allows a large safety factor (Fig. 6).
Fig. 6. Maximum combined stress - Linear Braking
The large safety factor suggests that reduction of frame numbers or reduction of
diameter could be done. This is not possible because already the number of members
and diameters of the tubes are at the border of the FSAE rules. This results in reduced
options available for optimization. Also the displacement of the chassis is limited at
only 0.11 mm. This shows a good stiffness of the chassis and also a good handling
performance as the chassis has a very limited movement during braking.
Additionally, to simulate also the inertia forces which the car would support, the
main components meaning the driver and the engine were loaded into the simulation.
The driver forces had the origin by the points in which the belts are held by the chassis.
The engine forces resulted from the engine mounts.
Torsional stiffness has been found to be the highest priority as the connection
between the chassis and other components to be as rigid as possible. This is where the
largest stresses and deformations occur. The areas that require reinforcement are the
A-arm and suspension rocker mounting points, as these nodes transfer the loads from
the tires through the suspension and then to the chassis.
4.2 Optimization
In the same manner, we made analysis for cornering loads, bending, asymmetric loads,
frontal impact, main and front roll hoop, side impact and modal analysis. A few results
are shown in Table 1.
A total of 3 versions of chassis have been analyzed to extract the best results
(Figs. 7 and 8). This includes:
(I) with the white tubes. First case has been modeled to fit all components of the
racecar, but also to have the exact coordinates of the suspension mounting
points. From this basis, further improvement was possible.
(II) with improved tubing. From the first case, this model has an improvement of
17.72 % of the torsional stiffness displacement. Total difference in weight:
+1,475 kg from Ist case.
(III) with solutions from first and second variant. From the second case, this model
has an improvement of 5.93 % of the torsional stiffness displacement and
improvement of 19.30 % between first and third variant. Total difference in
weight from 2nd model: +1,995 kg.
Overall, from first case to the final design, a total of 23.65 % increase in torsional
stiffness displacement has been achieved with an increase of only 8.75 % in weight.
Table 1. Analysis and results [1]
Forces and boundaries - Total displacement - Tor- Maximum Combined Stress -

Torsional stiffness sional stiffness Torsional Stiffness
Forces and boundaries on the Maximum combined stress - Total Deformation - Linear
chassis Linear Braking Braking
Forces and boundaries on the Total deformation of the Maximum Combined Stress -
chassis - cornering chassis - cornering cornering
Forces and boundaries - Total Deformation - asym- Maximum Combined Stress -

asymmetric loading metric loading asymmetric loading
Forces and boundaries - Total Deformation - Frontal Maximum Combined Stress -

Frontal Impact impact Frontal Impact
Forces and Boundaries - Total Deformation - Main Maximum Combined Stress -

Main roll hoop Roll Hoop Main Roll Hoop
Fig. 7. Deformation optimization of chassis
Fig. 8. Stress optimization of chassis
5 Conclusions
The objective of this project is to design a chassis suitable for a mechanical engineering
competition. This report describes the design of a lightweight, high stiffness space
frame chassis, combining ergonomic, esthetic, structural and crash properties. The
chassis is the central element of a race car which bonds all systems together to work as
designed, with maximum efficiency and thus with great performance. All components
are attached to the chassis and subjected to different loads. The weight of the chassis
and of the major systems, and the manner in which it transfers these loads play a critical
role in its performance. Considering the performance requirements of the chassis,
meticulous virtual modelling and analysis validate a design and reduce the risk of
failure. It is important to remember that a race car means the sum off all of its parts,
meaning that only when all these systems work together, the overall performance will
be secured.
Furthermore, the tube dimensions have been chosen carefully, on the basis of
lightweight though stiff, results being evaluated using extensive Finite Element Analysis
techniques. The overall goals of this project ware taken into account and each goal was
achieved after continued research throughout the project. The chassis resulted was
subjected to all loads demanded by FSAE rules and with good results.
For the future, a recommendation would be to reevaluate the suspension system in
parallel with designing a new chassis. With good interaction between the 2 major
systems of the racecar, a lot of difficult engineering problems could be avoided by
making these two systems more suitable to one another, instead of working around a
pre-design suspension system. In this project, the design and manufacturing timings are
limited. In order to compete with top teams at the competition, time has to be divided in
tasks and daily evaluation must be performed. With this project timing, the margins of
error will be minimized, resulting also in overall improved quality and a limited
number of failed composite components.
Acknowledgements. This article contains the results of research conducted in preparation of

license’s thesis of the first author and was realized within the Partnership Programme in priority
domains - PN-II, which runs with the financial support of MEN-UEFISCDI Romania, Project no.
227/2014, System for Diagnosis and Therapy of Spine Diseases (SPINE).
References
1. Furnea, D.-I.: Designing a race car chassis for formula student engineering competition,
License thesis, Transilvania University of Brasov (2016)
2. Institution of Mechanical Engineers, Rules for Formula Student. http://events.imeche.org/
formula-student/about-us. Accessed 03 May 2016
3. Riley, W., George, A.: Design, analysis and testing of a formula SAE car chassis, SAE
Technical Paper 2002-01-3300 (2002). doi:10.4271/2002-01-3300
4. SAE International: FSAE Rules 2016, 25 March 2016 (2016). http://students.sae.org/cds/
formulaseries/rules/2015-16_fsae_rules.pdf
5. Mogan, G., Butnariu, S.: Modelarea şi Analiza cu Metoda Elementelor Finite în Inginerie.
Teorie şi Aplicaţii practice (sistem integrat) (Modeling and Analysis with the Finite Element
Method in Engineering. Theory and Applications), Ediţia I, Editura Universitatii Transilvania
din Brasov (2014). ISBN: 978-606-19-0311-5
6. Butnariu, S., Mogan, G.: Analiza cu elemente finite în ingineria mecanica. Aplicatii practice
in ANSYS (Finite Element Analysis in Mechanical Engineering. Practical Applications in
ANSYS), Ed. Universităţii Transilvania (2014). ISBN 978-606-19-0474-7
Presentation of Tests with a Dynamometer
Related to the Determination of Fuel
Consumption in the NEDC and WLTC Driving
Cycles and as Well Related to the Dynamic
Behavior of Vehicles
Michael Butsch(&) and Valerio Rescigno
University of Applied Sciences Konstanz, Konstanz, Germany

{butsch,varescig}@htwg-konstanz.de
Abstract. Several possibilities of tests under load on a chassis dynamometer

are presented. Consumption measurements according standard driving cycles as
the New European Drive Cycle (NEDC) and Worldwide harmonized light duty
test procedure/cycle (WLTP/WLTC) make special attention to the observance of
the regulations necessary. The rotational masses of inertia and the load
depending on velocity have to match the required values.
Load tests as well allow the determination of the maximum acceleration in the
current gear and the slippage of the driven wheels.
Keywords: WLTC WLTP NEDC Chassis dynamometer Fuel

consumption Acceleration Slippage
1 Introduction
At the University of Applied Sciences Konstanz, a modern electronically controlled

dynamometer and several cars and motorcycles for tests are available, with which
numerous studies have already been carried out and some of the main results will be
presented. The paper is intended to explain different tests with load. The emphasis is
placed on the driving cycles NEDC (New European Driving Cycle) and WLTC/WLTP
(Worldwide harmonized Light vehicles Test Cycle/Test Procedure) and to the
requirements for proper conduct of investigations on a dynamometer. In a first step, the
two driving cycles NEDC and WLTC are presented in detail with their regulations for
tests on a dynamometer. These specific rules to conduct the experiments in the labo-
ratory will be discussed. The mentioned dynamometer and its equipment will be
presented. The various possibilities for operation and for carrying out various experi-
ments are shown. An important study with the view to increasing demands in terms of
reducing CO2 emissions is the consumption of a vehicle. Results are presented for
consumption measurement of a GM Vauxhall. The determination of the exact fuel
consumption is not based on a car computer but on a modified fuel system of the car
including fuel supply from an external gas can. By weighing the amount of used up
fuel, the consumption of the car can be measured exactly.

Presentation of Tests with a Dynamometer Related to the Determination of Fuel 121
Other important figures that can be determined on a dynamometer are the wheel
power, the power losses and eventually the engine performance. These values are
allowing the calculation of the brake specific power of the engine. From the charac-
teristic power curve, the full load characteristic curve of torque of the engine can be
determined.
Further interesting possibilities for analysis result of these measurements such as
the determination of the maximum gradeability and the maximum acceleration
capacity. As well the slippage related to load can be measured. In this paper the
possibilities of the determination of consumption data and driving dynamics data of
vehicles are presented. Also the requirements for the implementation of tests in the
NEDC and WLTC cycle will be explained.
2 Presentation of the Dynamometer
A detailed presentation of the dynamometer was already given in the context of

comparative measurements of a motorcycle with an electric drive and a motorcycle
with a combustion engine [1]. In this paper the specific ways to perform a test with a
predetermined load are presented.
Fig. 1. Test vehicle on the dynamometer
Figure 1 shows on the left side the front axle on the rollers of the dynamometer and
the optical speed measurement device. In the middle one can see the fixation of the
vehicle in the rear and as well the exhaust-gas ventilation. On the right side, the fixation
of the vehicle in the front and the fan are showed.
The dynamometer Autodyn 30 is a single-axis test rig of the company Super-Flow.
Due to the structural design of the dynamometer, the driven wheels have better contact
with the roller than having a dynamometer with two rollers. The dynamometer has an
eddy current brake, by means of which a braking torque can be generated. Thus, it is
possible to simulate in the test real driving resistances quite well [2].
A large tire contact area results due to the large roller diameter of 76.2 cm (30 ̋),
which provides good traction and thus a great power transmission capability. But
because of the curvature of the roller, the rolling friction resistance is still increased by
about 30 % related to the flat road (see Fig. 2) [3].
122 M. Butsch and V. Rescigno
roller diameter in in r = 0.3 m radius of the wheel

R radius of the roller
FXR/
FR
r/R
Fig. 2. Relationship between the rolling resistance on the chassis dynamometer FxR to the road
Fx (Source: Paulweber, Lebert [3])
The dynamometer offers three options of a driving cycle in which the eddy current
brake is active [2]:
Speed-Controlled Operation. In this mode it is possible to predetermine the velocity
profile and thus to approach the desired speeds during a driving cycle. This would be
for the implementation of the operating cycles based on the vehicle speed, a great
advantage since the driver doesn´t need to follow the speed with the accelerator pedal.
The software controls the speed.
The driver, for example having reached the speed of 50 km/h, would have no
possibility to accelerate any more. A disadvantage of this mode is that the driver has no
direct influence on the braking torque. So it is possible to drive at a desired speed with
both, partial or full load. However, the target of the driving cycles is to drive at a certain
speed with the power necessary on a flat road in real driving situations, as is imposed
by the regulations.
Torque-Controlled Operation. The torque profile of the test can be specified but
there is no possibility of a specification of torque related to speed.
Road Load Mode. In the road load mode the air resistance of the vehicle is simulated.
The speed-dependent determination of the air resistance is performed automatically. In
this mode only the product “FaxCd” of the front surface area and the drag coefficient of
the vehicle have to be fed:
FaxCd ¼ front surface area ðAÞ drag coefficient ðcw Þ ð1Þ
Using the FaxCd value, the eddy current brake simulates the air resistance related to
the speed of the vehicle. The road load mode corresponds best to the method which is
specified in the regulations. The chosen FaxCd value has to be crosschecked according
the regulation No. 83, Appendix 1 of the UN/ECE [4]. The total load F absorbed by the
chassis dynamometer has to be in a certain range defined as below:
“The load absorbed by the brake and the chassis dynamometer internal frictional
effects, between the speeds of 0 and 120 km/h is as follows:

F ¼ a þ b v2 0; 1 F80 ðwithout being negativeÞ ð2Þ
Where:
F = total load absorbed by the chassis dynamometer (N),
a = value equivalent to rolling resistance (N),
b = value equivalent to coefficient of air resistance (N/(km/h)2),
v = speed (km/h),
F80 = load at 80 km/h (N).”
The values of a, b and F80 are specified in the regulation no. 83.
The measured load F and the calculated values have to be compared. The tolerance
for the deviation of these values lies at ± 5 % at the speeds 120, 100, 80, 60 and
40 km/h and at ±10 % at 20 km/h. The “FaxCd” value had to be adapted from 0.6 to
0.7. A quite good result is shown in Fig. 3. Only at 20 km/h the required load couldn´t
be achieved within the limits.
As well the rotational masses of inertia of the dynamometer have to be checked and
compared to the required values.
measured load
allowable minimum load
allowable maximum load
load in N
velocity in km/h
Fig. 3. Measured load of the brake at the roller with FaxCd = 0.7
3 Consumption Measurement
3.1 NEDC (New European Drive Cycle) Und WLTC/WLTP (Worldwide
Harmonized Light Vehicle Test Cycle/Test Procedure)
The main differences between the NEDC and WLTC get obviously in Fig. 4 [5]. In the
WLTC significant higher velocities have to be driven for longer periods and the
acceleration values are corresponding to real driving conditions.
The main differences between the NEDC and

WLTC get obviously in Figure 4 [5]. In the
WLTC significant higher velocities have to be
driven for longer periods and the acceleration
values are corresponding to real driving condi-
tions.
Fig. 4. Comparison of NEDC and WLTC (WLTP) (Source: Pischinger, S. [5])
The cycles are shown in the Figs. 5 and 6 below.
3.2 Method of Measurement of the Fuel Consumption Driving the Test

Vehicle
The modifications aim for an operation of the vehicle either with the built-in gas tank or
with an external gas can. Changing between the gas tank and gas can is done by
three-way valves.
velocity in km/h
required velocity
driven velocity
time in s
Fig. 5. Test run according NEDC

velocity in km/h
driven velocity
required velocity
time in s
Fig. 6. Test run according WLTP/WLTC
Carrying out the NEDC- and WLTC-cycle related to the measurement of the fuel
consumption of a GM Vauxhall with a simple gasoline engine technology (indirect
injection, naturally aspired).
Three test runs of both cycles are performed. Significant influence on the con-
sumption measurement has the deviation of the driven velocity to the required velocity.
In the NEDC, the deviation is very low and this results in a low deviation of the
measured fuel consumption of only ±0.03 l/100 km. The run within the WLTC-/
WLTP-cycle is more difficult in comparison to the NEDC because of faster speed
changes. Here the deviation related to the fuel consumption is ± 0.085 l/100 km.
In Fig. 5, above, the straight lines with diamond-shaped points show the required
velocity in the NEDC. The velocities of the actual run (square points) have a small
deviation from the required velocities.
A consumption of 6.7 l/100 km gasoline is determined for the Vauxhall in the
NEDC.
In Fig. 6 the straight lines with square points show the velocity regulated in the
WLTP. The deviation of the driven velocities to the required velocities is obviously
larger than in the NEDC cycle. A consumption of 6.5 l/100 km gasoline is determined
for the Vauxhall in the WLTP. There aren´t large differences of the fuel consumption
related to the chosen cycle. The simple gasoline engine of the GM Vauxhall has neither
direct injection nor chargers or compressors which are working at their optimum at
specific speed ranges.
4 Determination of the Maximum Acceleration
The curves below are determined with a Mini Cooper on the chassis dynamometer
(direct injection, turbocharger). The engine power is measured on the dynamometer
and the excess traction force can be calculated as the difference of traction force from
the engine and traction force required for the rolling and air resistance. The maximum
acceleration in the 4th gear can be achieved at 70 km/h (Fig. 7).
traction force available

traction force in N
excess traction force
traction force required
velocity in km/h
Fig. 7. Traction diagram of a Mini Cooper, 4th gear engaged
5 Slippage at the Drive Wheels
The slippage s can be calculated according the following equation:

x rdyn v
s¼ ð3Þ
x rdyn
with v the velocity of the car and ω and rdyn the angular velocity and dynamic radius of
the wheel.
The slippage is closely linked to the traction force. The speed of the drive wheels is
measured by an optical system or can be alternatively calculated from engine speed and
ratio. Velocity v is derived from the roller speed by measuring the speed with an
inductive speed sensor. The results of test-runs with different loads (traction forces) are
shown in Fig. 8.
Having f.e. a wheel speed of 555 min−1 (about v = 70 km/h) and a traction force of
2,700 N (traction force at maximum acceleration in 4th gear), the slippage is about
0.57 %.
slipp age in %
wheel speed 555 min-1

wheel speed 898 min-1
traction force in N
Fig. 8. Slippage (Mini Cooper, 4th gear engaged)
6 Conclusion
Modern dynamometers allow a large number of different tests of vehicles. Several

possibilities of running load tests are presented in this paper.
Different test cycles as the NEDC and WLTC can be realized according the reg-
ulations and are well repeatable. Consumption tests of a small car with a simple
gasoline engine are presented as an example. Only small differences in fuel con-
sumption could be measured when comparing the two cycles
The full-load curve of the engine and the resistance curve - consisting of rolling
friction and air resistance - allow the determination of the maximum excess force.
A measurement with a vehicle having a gasoline engine with direct injection and
turbocharger is shown. Slippage can be calculated with the measured speeds of the
driven wheels and the roller of the dynamometer.
References
1. Butsch, M., Halim, A., Rothfuss, J.: Compararea consumului de energie al unei motociclete
cu propulsie electrică cu consumul uneia cu motor cu ardere internă de aceeaşi putere
(Comparison of energy consumption of a motorcycle with internal combustion engine and
with electric motor of same power). In: The Magazine: Ingineria Automobilului, Bucharest,
September 2010
2. Saur, F., Butsch, M.: Leistungsprüfstand (Dynamometer). HTWG Konstanz: unpublished
project, Konstanz (2011)
3. Paulweber, M., Lebert, K.: Mess- und Prüfstandstechnik (Technology of Measurement and
Testing). Springer Vieweg, Wiesbaden (2014)
4. Regulation no. 83 of the Economic Commission for Europe of the United Nations (UN/ECE)
5. Pischinger, S.: Getriebe für den zukünftigen Antriebsstrang (Transmissions for Future
Drivetrains). CTI-Getriebesymposium, Berlin (2014)
Validation Procedure for Worldwide
Harmonized Light Vehicles Test Cycle
via Hardware in the Loop - Real Time Testing
Bogdan Ovidiu Varga, Nicolae Burnete, and Calin Iclodean(&)

{bogdan.varga,nicolae.burnete,
calin.iclodean}@auto.utcluj.ro
Abstract. This paper studies the behavior of a vehicle’s power-train model

developed using IPG CarMaker software applications and connected to the
testbed with IC Engine by Hardware in the Loop (HiL) real time testing
methodology, for the homologation (due date 2017) driving cycle Worldwide
Harmonized Light Vehicles Test Procedures (WLTP).
The simulation task includes measuring the fuel consumption, power, and
torque and pollutant emissions over the usual classical vehicle model for a
WLTP driving cycle. Based on obtained results, a comparative analysis is
performed for virtual vehicle model in IPG CarMaker and the real IC Engine on
the testbed. Step-by-step simulation results are validated by integrating real-life
data from on-board measurement via CAN Bus. The system captures various
drivability related sensor and CAN bus signals. These input quantities were
achieved from Puma Open testbed application and sent to be implemented on to
the AVL InMotion equipment for further processing and analysis.
Keywords: HiL AVL InMotion IPG CarMaker RT simulation WLTP

driving cycle
1 Introduction
A globally harmonized standard for the determination of levels of pollutants and CO2
emissions, fuel or energy consumption, and electric range from light duty vehicles
(passenger and light commercial vehicles) is defined through The Worldwide harmo-
nized Light vehicles Test Procedures (WLTP).
With a final version under works, the WLTP is being developed by experts from the
European Union under guidelines of UNECE World Forum for Harmonization of
Vehicle Regulations [1].
The conditions regarding dynamometer tests and vehicle load, gear shifting, total
car weight, fuel consumption, ambient temperature and pressure were provided through
a strict guidance from the WLTP test procedure with WLTC test cycle.
The WLTC assigned to the highest power to mass ratio (PMR) vehicle category
(class 3) is composed by four speed phases (low, medium, high and extra-high) [1].
This PMR parameter is representative for the vast majority of European vehicles [1].

Validation Procedure for Worldwide Harmonized Light Vehicles 129
This paper verifies the accuracy and effectiveness of the HiL simulation and test
platform for WLTC driving cycle, by comparing the real-time simulation of the motor
driving system with the experimental results of the test bench, plus the analysis of the
HiL simulation results of the entire vehicle model.
Procedures of the development and testing environment described in this article, to
operate the vehicle under realistic conditions in the entire driving maneuver parametric
space for WLTC driving cycle by means of driving tests on a chassis dynamometer in
order to cover the consumption-relevant situations that occur during everyday
real-world use. Maneuver and event-based testing is the name of the method used for the
implementation of test cases on the AVL chassis dynamometer using AVL InMotion.
The base of this method is essentially the idea that driving a vehicle - the ultimate driver
of vehicle development - is a sequence of events and maneuvers [2]. Driver behavior,
route and environmental conditions in combination with the real model of the vehicle
(including the engine, powertrain and exhaust system) should be determined in order to
perform a more realistic simulation.
AVL InMotion provides a platform for such a simulation, using physical models for
each individual component of the vehicle. The use of physical models is a necessary
prior condition to have the possibility of extrapolating the operating conditions besides
the parameters required in the process of conducting a test [3].
2 Materials and Methods
2.1 Experimental Test Bed

The experimental test bed consists of an engine dynamometer coupled to an Internal
Combustion Engine (ICE) and auxiliary equipment (Fig. 1).
Fig. 1. Experimental test bed (Renault K9 K ICE and dynamometer)

130 B.O. Varga et al.
The ICE used for this study is a Renault K9K. Engine’s specifications and per-
formance characteristics are: total displacement 1461 cm3, maxim power 66 kW (at
4000 min−1), maxim torque 220 Nm (at 2000 min−1) and compression ratio of 15.5:1.
The DynoRoad 202 dynamometer is an asynchronous AC machine equipped with a
converter power module IGBT (Insulated Gate Bipolar Transistor) for direct connec-
tion to mains voltage. Power module uses a hybrid interface that facilitates control over
engine torque and speed.
2.2 Model in the Loop

The Model in the Loop (MiL) simulation captures the specified behavior of the model
that is to be implemented in C code.
The V model (Fig. 2) represents a development process that may be considered an
extension of the waterfall model, and demonstrates the relationships between each
phase of the development life cycle and its associated phase of testing [4].
The basis for this development is the well-established V model that is one of the
standard approaches in automotive software development [4].
Fig. 2. Structure of the V model for X-in-the-Loop (XiL) simulation procedures

2.3 Software in the Loop

Software in the Loop (SiL) is the inclusion of compiled production software code into a
simulation model. In SiL phase, the actual production software code is incorporated
into the mathematical simulation that contains the models of the physical system. This
is done to permit inclusion of software functionality for which no model exists, or to
enable faster simulation runs.
SiL means the code is generated and it replaces the controller blocks in the same
simulation model. The simulation results should be widely identical when compared to
the results of MiL simulation.
The IPG CarMaker platform is a virtual driving environment that offers a wide
range of applications from SiL operation to HiL tests. IPG CarMaker was designed to
support the development process from an early conceptual stage to hardware prototype
testing [5].
AVL InMotion is a solution for define the maneuver based on IPG CarMaker
product to which AVL adds SiL interfaces for different kinds of Units Under Test UUT
(Engine, Powertrain, Transmission etc.) [6, 7].
A virtual vehicle is a computer-modeled representation of a real vehicle with a
behavior that matches that of its real world counterpart. The model is parameterized
with data that relates directly to the vehicle to be studied (Fig. 3).
Linked to the models for the virtual vehicle AVL InMotion contains
three-dimensional road models and a fully parameterized driver model (IPGDRIVER)
that is able to perform closed loop maneuvers up to the stability limits. The simulation
of arbitrary traffic scenarios allows the user to conduct virtual driving tests overland and
in the city, that are exactly reproducible [8].
Fig. 3. IPG CarMaker simulation model

By this approach it is possible to use IPG CarMaker to test any vehicle with a
validated parameter set, and to easily switch between virtual vehicles by changing the
parameter data that are used in the vehicle model. The virtual vehicle contains all parts
of a real vehicle, including powertrain, tires, chassis, brakes, etc. It is also easy to
integrate real automotive controllers (ABS, ESP, and ACC) or software modeled
controllers into the virtual vehicle by using HiL or SiL [8].
The main component of hardware is the real time computer running the IPG
CarMaker. The executable is configured in way that not only allows communication
between the host computer (network card) and test bench (I/O modules), but a third part
is configured for communication with the AVL InMotion over a CAN bus to experi-
mental test bed.
The second piece of hardware is the host computer. It is set up exactly as in the
simple HiL configuration, except tools such as the AVL InMotion GUI or user defined
scripts are used to control the actions of the AVL InMotion.
The third piece of hardware is the test bench, which has been modified to allow a
direct connection to the real time computer, or to the AVL InMotion, or to a combi-
nation of both the AVL InMotion and the real time system. Adding or making changes
to the cabling does the modification of simulation environment.
The fourth piece of hardware is the AVL InMotion, which is configured to allow
the selected signals to be passed from the test bench to the real time system. The AVL
InMotion is told what to do through CAN communication to the real time computer,
which gets its instructions from user-defined commands or from the host computer
through the AVL InMotion dialog controlled by mouse clicks [9].
2.4 Model Implementation

The main characteristics of the WLTC class 3 (ver. 5) are given in the Table 1, and the
vehicle speed and acceleration are shown in Fig. 4 [10].
Table 1. Characteristics of the WLTC class 3 (ver. 5) [10]

Cycle phase Cycle Stop Distance Max Mean Min Max
duration time m velocity velocity acceleration acceleration
s s km/h km/h m/s2 m/s2
Low 584 156 3095 56.5 25.7 −1.47 1.47
Middle 433 48 4756 76.6 44.5 −1.49 1.57
High 455 31 7158 97.4 60.8 −1.49 1.58
Extra High 323 7 8254 131.3 94.0 −1.21 1.03
Total 1800 242 23262 – – – –
WLTC class 3 was implemented by maneuvers. Maneuver is the concept of IPG

CarMaker for the driving scenario (Fig. 5) and it is following maneuver definition,
which is split into several maneuver steps (acceleration, braking, stop etc.) [8].
Fig. 4. WLTC class 3 (ver. 5)
Fig. 5. Maneuvers for WLTC class 3
2.5 Hardware in the Loop

Hardware in the Loop (HiL) simulation is rapidly evolving from a control prototyping
tool to a system modeling, simulation, and synthesis paradigm combining many
advantages of both physical and virtual prototyping [11].
Through comparing the real-time simulation of the motor driving system with the
experimental results of the test bench, as well as analyzing the HiL simulation results of
the entire vehicle model, this paper verifies the accuracy and effectiveness of this HiL
simulation and test platform.
HiL preceded the PiL (Processor in the Loop) is a real-time simulation technology
which runs the simulation model using a real-time processor and simulates the motion
state of the controlled object in conjunction with some hardware. It connects the
under-test Electronic Control Unit (ECU) through an I/O port and conducts a real-time
test to the control strategies and control algorithms that are constructed [12].
The communication between AVL InMotion and the experimental test bed is done
using Controller Area Network (CAN) line (Peripheral-CAN). The Peripheral CAN
line realizes the communication with powertrain components. In addition to this CAN
line, several signals from sensors are fed into the system using analog and digital IO
(and vice versa with actuator signals to powertrain components) [13].
The ECU receives signals from the sensors and controls the ICE (Internal Com-
bustion Engine) of the experimental test bed. The dynamometer is an electrical drive
controlled by AVL P400 EMCON and is used to establish the simulated rotational
speed at the crankshaft of the ICE.
Puma Open is the operating system for command and control execution testing
tasks at the experimental testbed [13]. Through the CAN communication interface the
Puma Open system realizes the data exchange with the ECU and with the other
interaction points in the communication process of the control and command system.
The difference between MiL, SiL, PiL and HiL during the simulation process are
described in Table 2.
Table 2. The difference between MiL, SiL, PiL and HiL

Type Technique Controller Explanations
MiL Model in the Modelled All models are made in simulation tool
Loop ECU
SiL Software in Code Part of model in simulation tool, part in
the Loop Generated executable code (virtual ECU)
ECU
PiL Processor in Physical ECU Part of model runs in Real Time simulator, part
the Loop in physical hardware
HiL Hardware in Physical ECU Part of model runs in Real Time simulator, part
the Loop and ICE in physical hardware
HiL uses this proposed virtual model designing of a real time model, which will
additionally generate engine speed as output, depending upon throttle valve opening
percentage, gear position, vehicle speed, etc.
3 Conclusions
This paper underlines a number of advantages for MiL, SiL, PiL and HiL: reduce costs
to build a model able to change the structure of the model at any stage of the project
and adapt the characteristics of the model requirements imposed in a short time.
In the standard HiL configuration the hardware controller is interfaced to the system
sends and receives data exactly in the same style as if installed in a real vehicle.
The paper present that there is a reliable solution available for future vehicle’s
homologation requirements, putting together hardware and software solutions (pro-
duced by different companies). The methodology presented in this paper aims to give a
predictive solution, fast one, of the hottest theme in research at the moment. Using a
“mixture” of road and laboratory test will be possible to achieve immediate data
regarding the vehicle pollution level, in term of day-by-day use (according to usually
route pattern).
References
1. Marotta, A., et al.: Gaseous emissions from light-duty vehicles: moving from NEDC to the
new WLTP test procedure. Environ. Sci. Technol. 49(14), 8315–8322 (2015)
2. Pfister, F., Schyr, C., Le Rhun, F.: Total energy efficiency testing the chassis dynamometer
as a mechatronic development platform. ATZ Online, no. 11, vol. 111 (2009)
3. Mock, P., et al.: The WLTP: how a new test procedure for cars will affect fuel consumption
values in the EU. In: International Council on Clean Transportation (2014)
4. Balaji, S., Murugaiyan, M.: S.: Waterfall vs V-Model vs Agile: a comparative study on
SDLC. Int. J. Inf. Technol. Bus. Manage. 2(1), 26–30 (2012)
5. Ziegler, S., Höpler, R.: Extending the IPG CarMaker by FMI compliant units. In:
Proceedings 8th Modelica Conference, Dresden, Germany, 20–22 March 2011
6. Pfister, F., Schick, B.: The future has a sensor. Location awareness meets powertrain
controls. In: SAE-China and FISTA (eds.) Proceedings of the FISITA 2012 World
Automotive Congress, Volume 8: Vehicle Design and Testing (II), vol. 196. Springer,
Heidelberg (2013)
7. Leitner, A.: Critical System Engineering Acceleration, Milestone report, ver. 1, D307-011
(2014)
8. Varga, B.O., Mariașiu, F., Moldovanu, D., Iclodean, C.: Electric and Plug-In Hybrid
Vehicles Advanced Simulation Methodologies. Springer International Publishing Ed., Cham
(2015)
9. IPG CarMaker, User’s Guide Version 5.02, IPG Automotive, Karlsruhe, Germany (2015)
10. http://www.transportpolicy.net/images/f/fa/Wltp3.png
11. Fathy, H.K., Filipi, Z.S., Hagena, J., Stein, J.L.: Review of hardware-in-the-loop simulation
and its prospects in the automotive area. Proc. SPIE 6228, 62280E (2006)
12. Chuanxue, S., Feng, X., Silun, P.: Implementation of electric vehicle hardware-in-the-loop
test platform. Int. J. Multimedia Ubiquitous Eng. 10(11), 147–158 (2016)
13. Wasicek, A., et al.: Virtual CAN lines in an integrated MPSoC architecture. In: Proceedings
of the 17th IEEE Computer Society Symposium (2014)
Experimental Investigation of a Vehicle
Behavior Using Different Complex Data
Acquisition Systems
Ilie Dumitru(&), Racila Laurentiu, Tutunea Dragos, Matei Lucian,

Dima Alexandru, and Oprica Alexandru
Faculty of Mechanics, University of Craiova, Craiova, Romania

dumitru_ilie@yahoo.com, racila_laurentiu@yahoo.com,
dragostutunea@yahoo.com, mateiclucian@gmail.com,
alexandru1dima@gmail.com, alex.oprica91@gmail.com
Abstract. The manufacturer’s competitiveness and legislation restrictions

concerning pollution has imposed modern solutions for vehicles but also an
improvement of testing equipment. This paper presents the results of an
experimental analysis of a vehicle behavior during a dynamometer performances
test. During the test, for a better understanding of engine behavior, the engine
parameters that are read by the ECU are registered and analyzed. The evolution
of these parameters is followed constantly throughout the test.
Keywords: Dynamometer Performances ECU Behavior
1 Introduction
Dynamometers are widely used throughout the automobile industry for many research
and development purposes requiring performance testing of complete vehicles. Gen-
erally, these dynamometers have a supporting structure for a vehicle, driven rolls and a
power absorption unit connected to the rolls. Modern highway and traffic conditions
have resulted in increasing use of chassis dynamometers for all manner of vehicle
performance and durability testing. This type of testing falls into two general types; the
performance runs such as are made to check new vehicles coming off the assembly line
or for checking and adjusting vehicles following an overhaul, and durability runs
intended to simulate thousands of miles of road service.
The dynamometers used in prime-mover testing, of which there are several different
types, all resist, and thus measure, the torque produced by the unit under test. The
accuracy with which a dynamometer absorbs and measures torque and controls rota-
tional speed is fundamental to the power measurements and all other derived perfor-
mance figures made in the test cell. Perhaps the most difficult question facing the
engineer setting up a test facility is the choice of the most suitable type and size of
dynamometer. The matching of the power and torque-producing characteristics of the
engine, motor, or turbine with the power and torque absorption characteristics of the
dynamometer is vital for accuracy of data and control [1].

Experimental Investigation of a Vehicle Behavior 137
The idea of running a complete vehicle under power while it was at rest was first
conceived by railway locomotive engineers before being adopted by the road vehicle
industry.
Today the chassis dynamometer is used almost exclusively for road vehicles,
although there are special machines designed for fork-lift and articulated off-road
vehicles. The advantages to the designer and test engineer of having such facilities
available are obvious; essentially they allow the static observation and measurement of
the performance of the complete vehicle while it is operating within its full range of
power and, in most respects, in motion [1].
Systems development and distribution of data acquisition, briefly called data
acquisition systems (DAS) is closely linked to the development of art in general and
industry in particular. It will be appreciated that at the moment there are two main
categories of DAS: general purpose acquisition systems used in different purposes
(research, education, etc.) and acquisition systems used in the industry. If we refer to
general purpose DAS, their evolution has been decisively influenced by the develop-
ment of the personal computer, where they are generally connected. Such systems are
mainly intended to purchase a given process specific data for further processing,
processing taking place, as a rule; they are connected to the computer itself. At the
same time, the computer allows user interaction with SAD, via a specialized graphical
user interface [2].
2.1 MAHA Dynamometer Stand
General diagnosis for the propeller module of a vehicle follows the establishment of the
way for the transmission of engine power to the traction wheels. Due to the fact that the
diagnosis parameters have multiple connections with environmental parameters what
an engineer should search for when operates on a vehicle is wheel drag and fuel
consumption.
Wheel drag depends on the following parameters:
• air fuel ratio (λ)
• efficiency (ηi)
• engine mechanical efficiency (ηm)
• transmission efficiency (ηtr)
• braking power loss coefficient (ηbr)
The multiple connections of wheel drag with state parameters of powertrain module
proves that the information given by this parameter are not enough and as a result, the
vehicle testing only after this parameter is only general. Not obtaining the minimum
value of wheel power shows that technical condition of one or more assemblies is
inadequate requiring further detailed diagnostic elements.
The LPS 3000 stand is available in many versions for the purpose of cars perfor-
mance testing (Fig. 1). Depending on the chosen version, the wheel power varies from
260 kW to 520 kW with a maximum test speed of 250 km/h. For the loading of the
138 I. Dumitru et al.
dynamometer simulation it is used an eddy-current brake. The stand is built for both
Otto and Diesel engines [3].
Before the test begins, some initial data must be set. There are information about
License plate no., Inspector, the RPM thresholds and also certain engine parameters.
The user can pre-select from various load conditions which make possible the
determination of e.g. a specific engine diagnosis or vehicle climbing power. Depending
upon need, various load conditions can be simulated such as constant traction, a
constant speed, a constant RPM or a driving simulation [3]. These options offer various
possibilities for an engineer to explore and understand the engine’s working conditions.
Figure 1 shows a vehicle during testing with a print screen of main results.
Fig. 1. Evaluations and determinations on the MAHA LPS 3000 stand
2.2 OP-COM Interface

On-board diagnostics (OBD) is an automotive term that refers to a vehicle’s
self-diagnostic and reporting capability. OBD systems give to the repair technician or
vehicle owner access to the status of the various vehicle modules. The amount of data
available via OBD is varied, depending on the automobile level of technology. In the
early versions of OBD the fault was reported by the illumination of a malfunction
indicator light.
OP-COM is a diagnostic application package which consists in an interface con-
trolled by dedicated software (Fig. 2). It is used for many cars manufactures for
diagnosis. The package can be used even with new cars that employ CAN-bus-based
diagnostics.
CAN-bus (which stands for controller area network) is a type of protocol used in
the OBD-II standard of diagnostics [4]. Modern OBD makes use of 12 standardized
digital communications port that gives real-time information about the vehicle’s
system.
To evaluate and record real time parameters of the engine the next steps are
performed:
a. The OBD interface is connected to automobile’s OBD port. Red power LED on.
b. The USB cable from the interface is inserted into computer USB port.
c. The vehicle ignition key is turned to position II.
d. OP-COM software is launched, and the following instructions are followed:
• Diagnosis - Choose year of the car - Choose model - Choose system/module
(Engine in our case) - Choose engine type.
• The interface connects to the ECU (Engine Control Unit) via OBD Pin 7 and the
engine module panel is displayed.
• For recording the real time parameters, the Record button is selected, and after
the name of the file is typed the OP-COM starts to record real time data from
engine sensors via OBD port. The data are recorded in the.csv file which name
was typed ahead.
Fig. 2. OP-COM interface
3 Results and Discussion
The provided results from MAHA dynamometer are analyzed and then compiled with
the ones registered by the OBD – OP COM device to create a better understanding of
the engines evolution during various driving conditions [4].
Taking into consideration the manufacturer main vehicle parameters (Table 1), in
Table 2 it shows the final results for these parameters (power, torque etc.) for a specific
vehicle (in this case Opel Corsa).
The Tables 1 and 2 show testing possibilities on MAHA LPS 3000 stand, but not
always these equipment’s are available.
In this case another possibility is to use the OBD systems [5, 6].
Table 3 shows some selected parameters which can be measured with this kind of
equipment. All those parameter values are very useful for validating technical condition
of the car but also for possible mathematical modelling of main interdependence
parameters [7].
The figures below present measured values and graphic developments for various
parameters using the OBD equipment.
Table 1. Engine specification provided by manufacturer

Engine specifications provided by
manufacturer – example –
P-norm 48 kW M-norm 130 Nm
at 4400 rpm at 2000 rpm
Table 2. Final results

Final results using MAHA LPS 3000 chassis dynamometer
P-norm 49.3 kW Correction M-norm 124 Nm Correction
P-eng 49.3 kW acc. to at 2050 rpm acc. to
P-wheel 36.6 kW ISO 1585 97.6 km/h ISO 1585
P-drag 12.6 kW T-envir 27.3 °C
at 5000 rpm T-intake 25.7 °C
air
116 km/h H-air 59.7 %
p-air 1011.5 hPa
p-steam 21.6 hPa
Table 3. Values of parameters determined in various driving conditions, using mobile data
acquisition system
Calculated Throttle
Calculated Pedal
Relative Engine
Ignition Dwell
Vehicle Speed
Ignition Spark
Engine Speed
Engine Speed
Injector Time
Acceleration
Temperature
Position
Position
Coolant
Vehicle
Angle
Angle
Time
Load
sec % % RPM °C km/h RPM m/s2 ms ms °CA %

235.62 46 25 1.604 88 28 1.604 1.1 1.8 12.7 9.2 15
236.38 49 31 1.77 88 32 1.771 1.1 1.8 13.2 10.7 16
237.1 49 25 1.884 88 35 1.884 1.1 1.8 12.7 11.7 15
237.82 45 27 2.032 88 37 2.032 0.9 1.8 12.1 17.7 14
239.28 39 22 2.276 88 41 2.276 1.1 1.8 10 13.2 13
239.66 33 19 2.355 88 43 2.355 0.7 1.8 8.4 28.6 11
240 33 18 2.394 88 43 2.394 0.7 1.8 7.7 32.6 10
240.34 38 20 2.439 88 45 2.439 0.7 1.8 8.8 28.6 11
243.09 44 26 2.899 88 53 2.899 0.9 1.8 9.7 29.6 13
243.43 39 25 2.949 88 55 2.949 0.7 1.8 9.7 16.7 12
261.33 34 20 2.79 87 51 2.79 0.2 1.8 8 27.1 10
261.67 38 23 2.811 87 52 2.811 0.4 1.8 9.6 31.1 12
Figure 3 shows the possibility of corresponding evaluation between fuel injection

[0CA] and specific time [s].
Fig. 3. Fuel injection duration [CA˚] according to time [s]
Figure 4 presents the results of mass air flow estimation according to time unit.
Fig. 4. Mass air flow [kg/h] according to time [s]

In Fig. 5 and 6 there are presented external diagrams for the engine, using the
measurement issue from OBD equipment’s.
Fig. 5. Evolution of power at wheel [kW] according to engine speed [rot/min]
Fig. 6. Evolution of torque [Nm] according to engine speed [rot/min]

4 Conclusion
The importance of testing vehicles results from the fact that different types of tests as
part of the research, development, design, manufacturing, operation and repair of motor
vehicles contribute to their continuous improvement in all phases mentioned. Trials and
tests intended to verify whether major design parameters, technical and economic
indices, performance, operational safety and wear resistance of the vehicles correspond
with the technical documentation construction company.
The business units that build or repair vehicles, the testing attempts are necessary
for permanent control of manufacturing/repair, verification and improvement of pro-
totypes and new processes and materials. In enterprises operating tests are needed to
determine which type of vehicle is most suitable for certain working conditions, and to
determine the most advantageous turnover schemes, rate of fuel consumption, quan-
tities of lubricants and spare parts.
Evolution of modern automotive imposed a proper development of testing and
research of equipment’s.
In the paper the authors present two important laboratory systems (chassis
dynamometers stand LPS 3000 and complex data acquisition systems OP-COM
interface). So we have adapted and used for specific research methodologies for these
two systems.
Dynamic parameters, energy and environment parameters determined for a type of
vehicle permit to identify conclusions about the technical condition (very useful in
preventive maintenance), as well as making databases for relevant functionalities and
optimizations.
Tests were carried out which, based on the operating conditions and fidelity, give
results that can reflect functional status using both the dynamometer and acquisition
system.
References
1. Martyr, A.J., Plint, M.A.: Engine Testing, 4 edn. Elsevier (2012)
2. MC Measurements and Computing, “Data aquisition handbook”, third edition, USA (2004)
3. LPS 3000 - Dynamometer for Passenger Cars, HGVs and Motorcycles, Original Operating
Instructions, Maha Maschinenbau Haldenwang (2014)
4. Olsenn, H.: Vehicle Data Aquisition Using CAN. Denver, Colorado (2010)
5. Dumitru, I., Colici, F., Dima A., Mardarescu, V.: Studies regarding the evolution of pollutant
emissions related to certain transitory condition of an internal combustion engine using
mobile measuring systems. In: Current Solutions in Mechanical Engineering, Applied
Mechanics and Materials, vol. 823. Trans Tech Publication (2016)
6. Dumitru, I., Racila, L., Matei., L., Oprica, A., Valeanu, V.: Aspects of functional optimization
and performance evaluation using an experimental programmable engine control unit. In:
Automotive and Transportation Engineering, Applied Mechanics and Materials, vol. 822.
Trans Tech Publication (2016)
7. Dumitru, I.: Vehicles engines. In: Modern Methods of Optimizing the Energy Parameters of
Diesel Engines with Direct Injection (in romanian). Universitaria, Craiova (2008)
Validation of Kinematic Simulation
of Sprocket Contacts of Chain Links
by Experiments
Karim Shalaby(&), Florin Corciova, and Simona Lache
Mechanical Engineering, Transilvania University of Brasov, Brasov, Romania

{karim.shalaby,corciova.florin,slache}@unitbv.ro
Abstract. Considering a general chain driving system, the motion of chain link
along the sprockets still generates several problems when contact points are
considered. Therefore, the characterization of this motion is of utmost impor-
tance for the overall behaviour of the chain. This paper aims to perform a
comparative study on the kinematics of sprocket contacts of chain links, when
the analysis uses either experimental approach or numerical simulation. The
experiments are conducted on an analogous Huygens Pendulum. The results
should give a better understanding on how geometries and lacings can affect the
damping of the oscillations and magnitudes of contact forces between various
chain plates on the same sprocket. At the same time, it is expected a close
correlation between the experimental and numerical models, which will lead to a
validated model to be further used in numerical simulations.
Keywords: Analogous Huygens Pendulum Inverted Tooth Chain (I.T.C.)

Multibody dynamics Data acquisition
1 Introduction
Silent chains or Inverted Tooth Chains (I.T.C.) are considered an important tool for
decreasing noise in the chain drive system, reaching a lower noise level than that of
bush chain drive systems. Nowadays, designers are trying to increase the quality of I.T.
C.s by decreasing noises [1], frictions and contacts [2]. They use logarithmic profiles
for the teeth of I.T.C plates, illustrated in Fig. 5, and they arrange the plates with
different strategies called lacings as shown in Figs. 1 and 2. An important issue is to
know how each chain would react at its ‘free state’. The free state is considered when
letting a chain to oscillate on its adequate sprocket without trying to have an external
load other than that of gravity.
This paper illustrates which type of lacing and geometry all together will lose less
energy and will have less friction. For this purpose, an Analogous Huygens Pendulum
(A.H.P.) was created, Fig. 4. The A.H.P. is based on the theoretical Huygens pendu-
lum, Fig. 3, its effect being caused due to a fixed centroid around which the chain
oscillates against the sprocket. A mono-involute motion is created by the intersections
of the curves made by the motion of the chain [3].

Validation of Kinematic Simulation of Sprocket Contacts 145
Fig. 1. Lacing type (A) (Source: by the Fig. 2. Lacing type (B) (Source: by the courtesy
courtesy of Schaeffler Romania SRL) of Schaeffler Romania SRL)
Fig. 3. Huygens Pendulum [4] Fig. 4. Analogous Huygens Pendulum [5]
The method used is that of multibody dynamics which, as it is well known, has two
directions of analysis: rigid bodies and flexible bodies. In this case, rigid bodies of
multibody dynamics are used as there are many bodies that can affect the oscillations and
that should be taken into consideration. Both the simulation and experimental approach
have been considered to illustrate how different geometries and lacings can affect loss of
energy, magnitudes of contacts and the minute frictions between joints. This study is
expected to lead to a numerical model validated by the experimental results.
2 Experimental Approach
The research problem to be solved is to find the best solutions in a chain drive system
and the possibility to decrease the contact forces of the chain plates with the adequate
sprocket, as well as, the friction between the chain plates with their joints or pins. These
possibilities should lead to less energy consumption as the quality of the chain (from
geometrical, Fig. 5, and lacing points of view) by default would tend to loose less
energy. The idea of comparing two different chains of different geometries and different
lacings, as shown in Figs. 1 and 2, would give a better perspective to what type of
146 K. Shalaby et al.
Fig. 5. Different geometry contours on chain plates of types A and B respectively (Source: by
the courtesy of Schaeffler Romania SRL)
geometries and lacings would consume less energy. A smart camera is used to track the
markers created on the chains. This camera is capable of saving the pixel positions of
each marker, which would further process into coordinates and, in a fast amount of
time, would transmit the data to the PC. The LabView software is used in data
acquisition from experiments; it is considered an appropriate tool, as it gives a great
amount of flexibility in data acquisition and processing.
Each chain would be released from the exact angle and initial conditions so one
could study the behaviour of each chain individually. The experiments show differ-
ences between the two models. The chart presented in Fig. 6 illustrates the difference
between two chains of types A and B during oscillations. It is observed the slight
difference between the two types during contacts or, in other words, on the oscillation
of the chain plates towards the sprocket. The difference between the two chains gapes
wider during returning to their initial position. This clearly shows the loss of energy of
each type of chain after a contact between the I.T.C. plates and the teeth of the
sprocket. Figure 6 presents the damping of the oscillations and illustrates the dissi-
pation of energy during contacts of the chain plates and sprocket, as well as the links of
the chain plates and their joints.
32.000
Type B_+veº
Experimental approach Type B_-veº
22.000
Type A_+veº
Type A_-veº
12.000
Degree °
2.000
0 20 40 60 80 100 120 140 160 180 200

- 8.000
- 18.000
- 28.000 Number of Oscillations
Fig. 6. The difference of oscillations between the two types of chains (A) and (B)
1.000
Type B +ve° Type B at -ve°
0.800 0.83
Type A +ve° Type A at -ve°
0.600 0.63
0.400 0.43
0.200 0.23
0 50 100 150 200 0 50 100 150 200
Fig. 7. The oscillation of the chains away Fig. 8. The oscillation of the chains towards the
from the sprocket sprocket
The upper part of Fig. 6 shows the oscillation of the chain away from the sprocket,
while the lower part illustrates the oscillation of the chain towards the sprocket.
A normalization was used to better understand the change during oscillations at each
part of the graph, Figs. 7 and 8, and to make sure that they have the same starting point.
Figures 7 and 8 illustrate the behaviour and divergence of each chain during oscilla-
tions, in a more explicit way.
One can notice the difference between the two chains: the difference or the deflection
is 2.709 % during oscillation away from the sprocket, meaning that chain type B is
subjected to a greater loss of energy. Yet, during oscillation towards the sprocket one
can notice that there is a really slight difference between them, of 0.945 %.
3 Numerical Simulation
Considering the same initial conditions, in order to differentiate between the two dif-
ferent types of chains and to create a database that could easily compare them, the
MSC ADAMS software has been used. Two numerical models have been developed.
Within the first model, Fig. 9, each two coinciding plates are reduced to a single plate
of double thickness. This is done to reduce the number of joints needed and the number
of bodies, in general. In the second model, Fig. 10, the plates do not exactly coincide
with each other, so there is no need to reduce the number of joints. In both cases the
outer plates have been removed as they are not subjected to direct contacts with the
sprockets. The friction between the plates and their joints is neglected. Most energy
losses will actually occur during the contacts of the plates with the sprocket. In order to
develop the A.H.P. model, a variable integrator such as the Hiller-Anantharaman Stiff
Integrator (HASTIFF) is used. Its main benefit is that it is persistent during simulation
calculations and it typically requires fewer function evaluations to proceed with the
simulation. One of the main priorities of using HASTIFF integrator is that it is focused
on accuracy rather than speed. The HASTIFF results should have fewer numerical
discontinuities [6]. Yet, the models can be huge, involving hundreds of bodies. It is
recommended to increase the number of iterations and the margin of errors. This should
Fig. 9. Modified Lacing type (A) Fig. 10. Lacing type (B)
give more persistence in calculations and ease of processing. Creating a contact in

MSC ADAMS is based on the normal forces and the friction forces. The normal forces
depend on the damping and stiffness, the penetration depth allowed between the two
solids and the force exponent, which is function of the spring forces during impacts.
According to literature, the force exponent is indicated to be 2.2, for hard bodies such
as steel to steel solids. As for the friction forces, the Coulombian method has been used,
which takes into account the static and dynamic coefficients. For the current work, the
following values have been chosen for the friction coefficients: static coefficient = 0.3
and dynamic coefficient = 0.1.
In solving the contacts, the modified Hertzian contacts are used accompanied by the
friction laws of Coulomb even during stiction and slipping.
The analysis performed on the numerical model allows to compare the results
between the two types of lacings and slightly different geometries; at the same time, it
offers a better understanding of the difference between each setting (e.g. different
thickness, different type of joint uses for modelling). As illustrated in Fig. 11, the first
two oscillations can be neglected as the simulation then stabilizes and the error
decreases.
Fig. 11. Difference between the two types of chains, illustrated by numerical simulation
Fig. 12. The cycles of the two chains, as a result of numerical simulations
Fig. 13. The difference between the two models during oscillation away from the sprocket
The simulations were conducted for 30 s, in order to enable a differentiation

between the two chains. It can be observed that the chains are closely similar in
reaction. The shift in the cycles and length of each cycle differ slightly. The wavelength
difference according to time of each cycle is presented in Fig. 12. The difference
between the two models are better viewed in Fig. 13, where one can see that type B
chain oscillates at a higher frequency than type A chain.
The deflection between the two simulated models during their oscillation away
from the sprocket is of 0.276 %, while the similarity of the two models during their
oscillation towards the engagement or contact to the sprocket is of 100 %.
4 Correlation Between the Experimental Approach

and the Numerical Simulation
There is a slight difference between the experimental approach and the numerical
model, as shown in Fig. 14. A normalization has been used to assure the same starting
point for the simulated model and the experimental approach. By comparing the
experimental results with the ones resulted from numerical simulations, during the
oscillations of the chain away from the sprocket it is observed a loss of potential energy
in both cases. However, the differences may be structured in two directions:
(a) The time taken for a complete cycle to occur in a numerical simulation is much
shorter than that in the experimental approach. It is noticed that during simulations
17 cycles occur, while during the experimental approach 15 ½ cycles occur. The
time period considered in both cases, experiments and simulations, is of 30 s.
(b) There is a greater energy loss during the numerical simulations than in the
experimental approach.
(c) The divergence between the results obtained by numerical modelling and by
experiments is of 6 %. This may be due to a higher damping coefficient and a
Fig. 14. Correlation between the experimental approach and the numerical model
higher elasticity coefficient of the numerical simulations compared to that of the

experimental approach.
5 Conclusions
The results obtained during simulation show very close correlation to the experimental
approach. The overall damping curves, resulted from both experimental and numerical
approaches, being so similar corroborates the purpose of the paper and proves the
precision of the multibody dynamic analysis of a system in general. The paper also
shows that type A loses more energy during oscillating back to the initial position. The
loss of energy is due to the contacts of the I.T.C. plates with the sprocket. As the
geometries are quite similar it is safe to state that the importance of the lacing is
imminent in conserving or losing potential energy.
The close correlation between the experimental results and the numerical simula-
tions validates the numerical model; therefore, it is considered reliable for being used in
further analyses.
References
1. Masao, I., Hiroyuki, W.: Analysis and control of inverted tooth chain vibration. In: World
Automotive Congress, FISITA, Budapest, 30 May–June 2010, pp. 858–861 (2010)
2. Mulik, R.V., Joshi, M.M.: Dynamic analysis of timing chain system of a high speed three
cylinder diesel engine. Res. Invent Int. J. Eng. Sci. 4(5), 195–199 (2014)
3. Shalaby, K., Lache, S.: Kinematics of the analogous Huygens pendulum behaviour using
inverted tooth chain. Bull. Transilvania Univ. Brasov 8(57) no. 1, Series I, Engineering
Science (2015)
4. Emmerson, A.: Things are seldom what they seem Christiaan Huygens, the pendulum and the
cycloid. AWCC Bulletin No. 362, pp. 295–312, June 2006
5. Shalaby, K.H., Lache, S., Corciova, F.: Contact forces analysis of an analogous Huygens
pendulum using inverted tooth chain. Int. J. Mech. Mater. Manuf. 4, 195–199 (2016)
6. MSC Software: MSC ADAMS (2008). Release Guide (online). www.mscsoftware.com
Belt Losses Evaluation for a Push-Belt CVT
Valerian Croitorescu(&), Codrin Cuciurean, and Mircea Oprean
Scientific Research and Continuous Training Center for Sustainable Automotive

Technologies, University Politehnica of Bucharest, Bucharest, Romania
{valerian.croitorescu,codrin.cuciurean,
mircea.oprean}@upb.ro
Abstract. The transmissions are designed to transfer the power from the engine
or from the electric machine to the wheels and, as much as possible, without
power interruption and no shocks and no vibrations during shifting process. The
power flow from the engine or from the electric machine is not interrupted or the
interruption is insignificant to the driver and passengers, only for automated,
automatic and continuous variable transmissions. Being produced into large
number of units, the transmission should be used on vehicle during entire
vehicle life cycle. The powertrain efficiency depends mainly on the transmission
efficiency.
The continuous variable transmission (CVT) is advisable to the conventional,
hybrid and electric powertrain solutions. The continuous gear ratio range is one
of the major advantages the CVT because it is able to offer high comfort to all
passengers. Having continuous speed ratio, the CVT is bringing fuel economy to
the vehicle.
This paper highlights the possibility of developing new modular virtual
models in order to be computed the losses corresponding to the belt, during
different vehicle’s operating modes.
Keywords: Continuous variable transmission Push-belt Van doorne Metal

belt Energy losses Thermal losses
1 Introduction
The fastest automotive improvements evolution follows the current trends on electri-
fication and increasing the powertrain efficiency. Automotive systems are becoming
more complex and all the powertrain components represent energy efficiency reference
points for establishing the on-board vehicle energy management. Several strategies
have the objective to reach the best powertrain efficiency, using different powertrain
solutions and architectures. For all of them, the optimization comes with integrating
different devices from the early beginning development stages.
The research and development stages include the build of a virtual-mock-up as the
most cost-effective solution before manufacturing. The virtual mock-up development
consists of combining the mechanical systems behavior with the electronics and
building simulation models to accurate describe the investigated system behavior.

Belt Losses Evaluation for a Push-Belt CVT 153
This paper aims to develop the powertrain thermal virtual model in order to
describe the modeling development methodology and to investigate the heat exchange
focused on the belt from a push-belt continuous variable transmission.
2 Virtual Development Methodology
The mechatronic systems are complex and their development is based on modeling and
simulation before being physically manufactured. The push-belt continuous variable
transmission is a complex mechatronic system that needs special attention during all
the stages, from research stages to integration into the vehicles, and then during the
optimization stages.
2.1 The Transmission

The push belt continuous variable transmission (PBCVT) is known as being a
sophisticated mechatronic system that offers excellent comfort, reduced noise level and
helps reducing the CO2 emissions. The PBCVT can handle higher torque ratings than
other transmission type. It is versatile and especially suited for front-wheel-drive
vehicles, [7]. Compared with other continuous variable transmissions, the PBCVT has
very compact unit size and has fewer components, achieving better performances for
lower costs. Its control unit is integrated into the transmission, being placed inside the
transmission housing. The control unit can be linked to the vehicles’ energy man-
agement systems. Its mechanism and intelligent electronic control always provide an
optimal balance between the required speed and torque from the engine or from the
electric machine, thus allowing them to operate at the most efficient regime as possible
all the time.
The PBCVT performances include very good and robust dynamic acceleration
without reducing the tractive power at all. It provides a comfortable driving, elimi-
nating the gear shifting process and separate gears. It represents the convenient way to
reduce the emissions and the fuel (energy) consumption because it is able to use the
optimal engine or electric machine operation state. Because the engine or the electric
machine may operate at an ideal level of efficiency, it may ensure low energy con-
sumption and reduced CO2 emissions (from the engine).
The PBCVT consists of a driving shaft and a driven shaft, each of which having a
pulley splined onto it. Each pulley comprises two conical sheaves that face each other.
A metallic push-belt runs in the V-groove between the two conical sheaves of each
pulley. It operates using variable pitch radius. One sheave of each pulley can slide
along its shaft, continuously adjusting the pitch radius of the push-belt and the trans-
mission ratio. The moving sheaves are diagonally opposed, if the driving shaft has the
moving sheave on the left side, the driven shaft has the moving sheave on the right.
The push-belt is transmitting the drive power by means of shear force from the
engine or from the electric machine to the wheels. The push belts consist of two types
of components: rings and elements. The push-belt is made of hundreds of specially
pressed steel elements which are strung together between two rings made of layers of
154 V. Croitorescu et al.
high-alloy steel. This design makes the push-belt very flexible, and at the same time as
strong as a solid steel bar.
The components of the belt (Fig. 1) and the geometry of the belt (Fig. 2) are very
important to determine the mathematical relationships used to analyze their behavior in
different circumstances [2].
The fixed distance between the pulleys and the belt angle evolution for changing
the speed ratio is used to determine the push-belt state and in addition, the powertrain
regime. If the belt is perfectly tensed, the engine or the electric motor are in charge to
develop the demanded torque to the wheels.
Fig. 1. The push-belt components [1, 4]

Fig. 2. The push-belt geometry [1]
2.2 The Simulation Model

The transmission functional model is describing the kinematics of the PBCVT, taking
into account the speed of the engine or of the electric machine, the information received
from the main vehicle’s electronic control unit, which are filtered by the transmission
control unit, and the speed ratio current value. The PBCVT delivers information
regarding the torque at the vehicle’s wheels level. The information about he obtained
power is used for running the transmission thermal model, using also the external heat
flows.
The belt losses are mathematically modeled as friction losses caused by the slipping
between the belt segments and pulley, the segments and the band, and between the
bands. The mathematical models do not refer to the transmission efficiency under
realistic conditions such as an overdrive speed ratio. But, the normal behavior is taken
into account while defining the model [2].
The belt losses are influenced by friction that generates heat. During steady state
functioning, the belt loss can be quantified using the dependency between the oil
density, ρ, the specific heat at contact pressure, Cp, the belt lubricant inlet temperature,
Tin, the splash temperature from the primary pulley, Tout, and the belt lubricant vol-
umetric flowrate, Vlfbelt, as presented in Eq. (1): [5]
Pbelt ¼ q Cp Vlfbelt ðTout Tin Þ ð1Þ
The friction losses between the belt segment side and the pulley surface are based
on relative slipping and contact load, using the dependency of the friction coefficient,
μsp, of the normal load, Fsp, of the relative slipping speed for the Side-Pulley interface,
Vsp, as presented in Eq. (2):
Psp ¼ 2 lsp Fsp Vsp ð2Þ
The friction losses between the segment shoulder and the innermost band are
depending on the friction coefficient, μx-x, the normal load, N, when x-x refers to the
segment shoulder/innermost band interfaces in the arc part – sb-a or to the straight part
of the belt, sb-st, as presented in Eq. (3):
Psb ¼ lsb Nsba Vsba þ lsb Nsbst Vsbst ð3Þ
The friction losses between the bands, for “12” sheets of thin metal are determined
using Eq. (4):
X
12 X
12
Pbb ¼ lbb ½Nbba ðnÞ Vbba ðnÞ þ lbb ½Nbbst ðnÞ Vbbst ðnÞ ð4Þ
n¼2 n¼2
The total losses for the push-belt can be determined using Eq. (5):
Pbelt ¼ Psp þ Psb þ Pbb ð5Þ
The interaction between the push-belt layers and between the layers and the pulleys
influence the torque value that is transmitted by the driving pulley to the driven pulley.
Some assumptions were made: the push-belt layers are relatively thin and the distance
between them was neglected; the rotary speeds of the driving and driven pulleys
influence the push-belt deformation; but this deformation comes from the layers that act
individually. The transmitted torque was considered to be the same for each layer. Due
to physical considerations of the bands’ elastic extensions, the bands’ speeds are not
included [3] into the simulation model. A multi-layer push-belt installed on the driving
and driven pulleys is presented in Fig. 3.
Fig. 3. The push-belt multi-band operation [3]

The friction forces between the layers depend on the lubricant viscosity, η, the
oil-film thickness (estimated), tf, the layer width, B, and the relative sliding speed, Vrel,
[3] as presented in Eq. (6):
gB
f ðVrel ðuÞ; mÞ ¼ Vrel ðuÞ ¼ m Vrel ðuÞ ð6Þ
tf
The contact zones between the belt and the pulleys defines the tensions between the
layers in order to determine the tractive tension, T, which is part of the total tension, F,
that contributes to the torque transmission [3].
The simulation model (Fig. 4) was built using LMS.Imagine.Lab.AMESim by
selecting the appropriate models, sub-models and components in order to describe the
PBCVT behavior closer to reality. The transmission was taken from the existing
libraries and its parameters were modified accordingly. The added sub-systems are
using the mathematical equations to emphasize the power losses of the push-belt. The
belt losses can be computed if the input and output torques and velocities are available
through measurements. Each component is described using mathematical equations by
C-code fragments.
Fig. 4. The push-belt multi-band operation [3] (CVT – Continuously Variable Transmission,
ICE – Internal Combustion Engine)
The simulation model uses the parameters from an existing vehicle, engine and
transmission (the mass, the maximum power and torque, the ratio range, the distance
between the primary and the secondary pulley etc.). The transmission control unit uses
the speed and the torque in order to calculate the transmission ratio.
2.3 The Heat Exchange

The thermal effects between the push-belt components are investigated using real
parameters. Simple shapes are associated to the components in order to easily deter-
mine their volumes. The material density, the specific heat and the thermal conductivity
are defined for each of the push-belt component. A thermal mass is associated with
each component.
The heat exchange is made by conduction and convection. The conduction is met
between solid components, from the belt, to the pulleys, to the shafts, to the bearing and
to the frame, while the convection is met between the solid components’ contact with
the internal air and liquid.
The conduction is determined using the material thermal conductivity, λ, the
contact surface, A, the heat flow between the neighboring components, Q, and the
initial temperature, T, as presented in Eq. (7) and graphical described in Fig. 5:
dT
Q ¼ k A ð7Þ
dx
Fig. 5. The conduction between the push-belt layers
The convection is requiring the fluid motion or gas motion phenomena description
related to the operating push-belt inside the transmission frame. The convection
depends on the thermal conductivity between the solid component and the high speed
fluid, λ, the contact surface area, A, the temperature of the solid components surface
(estimated), Twall, and the fluid temperature measured near the already set distance to
the solid component, T1 , as presented in Eq. (8):
0
Q ¼ k A ðTwall T1 Þ ð8Þ
The conduction and convection heat exchange were investigated using different
assumptions for some dimensions that were not possible to be measured.
3 Simulation Results
The simulation was performed during 500 s, following a defined driving cycle,
starting-up the vehicle with 75 % tractive force. Four different oil temperature values
were considered before starting the simulation, 25, 50, 75 and 95 °C. Figure 6(a)–(d)
are showing the total power losses during the 500 s. The total losses increased almost
Fig. 6. The total losses evolution during a defined driving cycle of 500 s
six times between the oil temperatures from the beginning of 25 °C to the working
temperature of 95 °C.
The losses can be divided by each component. The simulation results concluded
that the belt losses are estimated to be between 15 to 25 % from the total losses
depending on the driving state.
4 Conclusions
Several possibilities for creating the virtual power losses were investigated. Taking into
account that the push-belt is made from different layers, the power losses were inter-
preted as thermal losses generated by the friction between each layer. The continuous
variable transmission push-belt losses were computed and its behavior was investi-
gated. The approach took into account describing the losses as thermal losses, by
defining all the push-belt components as thermal masses and by describing them as
simple shapes. The thermal transfer was considered to be made by conduction and by
convection. All components, including the solid components, the air from inside and
from outside the transmission were considered as thermal masses between which the
thermal transfer took place. The thermal model for each component used the material
properties and the motion between the neighboring components. The thermal transfer
between two neighboring components was described by using mathematical equations.
In order to reduce the power losses, several measures have to be taken into account:
the slip losses and the friction losses have to be reduced by using different materials and
lubricants, the actuated power have to be more efficient distributed by defining
improved control strategies, the balance between the demanded and the actuated power
have to be improved by defining improved control strategies, the actuated power have
to be improved.
The modeling and simulation approach by building virtual mock-ups for compo-
nents, sub-systems and systems used on vehicles are helping the entire research,
development and manufacturing industries allowing them to optimize the vehicles,
including the vehicles’ components and systems, long before it is physically built.
References
1. Akehurst, S., Vaughan, N.D., Parker, D.A., Simner, D.: Modelling of loss mechanisms in a
pushing metal V-belt continuously variable transmission. Part 3: Belt slip losses. Proc. Inst.
Mech. Eng. Part D: J. Automobile Eng. 218(11), 1269–1281 (2004)
2. Croitorescu, V., Anthonis, J., Oprean, M., Negrus, E., Andreescu, C.: Analysis of the CVT
efficiency by simulation. In: Proceedings of Multi-conference on Systems and Structures,,
Analele Universitatii “Eftimie Murgu” Reșița, Fascicula de Inginerie, Anul XVIII, nr. 1, 2011,
ISSN 1453–7394, Reșița (2011)
3. Kong, L., Parker, R.G.: Steady mechanics of layered, multi-band belt drives used in
continuously variable transmission (CVT). Mech. Mach. Theory 43, 171–185. Elsevier Ltd.
(2007, 2008). doi:10.1016/j.mechmachtheory.2007.02.003
4. Narita, K., Priest, M.: Metal-metal friction characteristics and the transmission efficiency of a
metal V-belt type continuous variable transmission. Proc. Inst. Mech. Eng. Part J, J. Eng.
Tribol. 221(1), 11–26 (2007)
5. Srivastava, N., Haque, I.: Transient dynamics of metal V-belt CVT: effects of band pack slip
and friction characteristic. Sci. Direct Mech. Mach. Theory 43, 459–479 (2008)
6. Xie, F., Wang, J., Wang, Y.: Modelling and co-simulation based on AMESim and Simulink
for light passenger car with dual state CVT. In: International Workshop on Automobile,
Power and Energy Engineering, 1877-7058. Elsevier Ltd. (2011). doi:10.1016/j.proeng.2011.
08.1096
7. http://products.bosch-mobility-solutions.com/specials/de/cvt/index.html?l=en#
Kinematic and Dynamic Study of a Mechanism
for a Vehicle Front and Rear Stabilizer Bars
Nicolae Dumitru(&), Cristian Copilusi, Ionut Geonea,

and Alexandru Margine
Applied Mechanics and Civil Constructions Department, Faculty of Mechanics,

University of Craiova, Craiova, Romania
nicolae_dtru@yahoo.com, igeonea@yahoo.com,
fam_margine@yahoo.com, cristache03@yahoo.co.uk
Abstract. In this paper, a method is described for theoretical and experimental

studies of a front and rear stabilizer bars used on vehicle steering and suspension
systems. Based on the fatigue experimental tests performed on a special test-bed,
a complex actuation mechanism from the test bed structure will be analyzed.
This mechanism through his kinematic and dynamic structure assures complete
conditions for simulating the dynamic behavior in a complex mode of the rear
stabilizer bar used on a vehicle steering and suspension structures.
Keywords: Kinematic Dynamic Actuation mechanism Stabilizer bar
1 Introduction
The essential key on a car steering and stability lays onto suspension and steering
systems. During time these systems were analyzed and continuously improved in order
to assure a proper roll characteristic of a car. Also it is well known that in steering
maneuvers, vertical loads on wheels at the outer track reach high values and those on
the inner track have small values, which is called lateral load transfer. Thus, when
moment equilibrium is broken at some specific conditions, a vehicle will lose roll
stability. In order to avoid this, a stabilizer bar was mounted transversely on spaced out
rubber bush swivel bearings to the chassis and having cranked arms attached to the
suspension system components. The main research scope for most research analyses
was to design this to proper dimensional parameters and also to find out the suitable
stiffness of this for a variety of vehicle models. Also another important aspect is
represented by the stabilizer bar material, which this has a crucial role on stiffness and
rigidity as it can be remarked in [1, 3, 4]. Most of the scientific research analyses were
concentrated on this element analysis from a theoretical viewpoint, by elaborating
mathematical models of the front stabilizer bar with the aid of multi-body systems
method [6]. But the research results were theoretical ones. The main target during time
was to analyze this in a dynamic mode on rough road conditions, by performing
experimental analyses with the aid of a special vehicle in [8, 9, 11]. Thus, on a vehicle
stabilizer bar where placed a series of transducers and sensors, which gives important
information about the stabilizer bar behavior. But these experimental tests are still

162 N. Dumitru et al.
irrelevant in case of trying to perform fatigue tests on this important component,

because in rough conditions this can be broke apart. Fatigue analysis can be accom-
plished by using dedicated software such as ANSYS and related work on this, were
performed in [12]. This is not so easy, due to introducing proper values of the specific
fatigue tests parameters, and in some cases the obtained results can be unsatisfactory.
The most relevant fatigue tests were represented by other researchers on developed
experimental test beds as in [10, 12].
Considering the related scientific works [5, 7], the aim of this paper is to elaborate a
mathematical model for analyzing the stabilizer bar fatigue mechanism in a dynamic
mode and compare the obtained theoretical results with the ones from an experimental
analysis on a stabilizer bar fatigue test bed. Thus, on the second section of this paper is
presented the fatigue test bed from a structural viewpoint. This is a universal one and it
can be parameterized and adapted on several stabilizer bar types. By considering only
the mechanism for the rear stabilizer bar of a vehicle, a dynamic analysis was per-
formed on third section of this research. The obtained theoretical results and also
numerical simulations were validated by performing experimental tests on this fatigue
test bed presented in the fourth section. Important conclusions which regard the
behavior in dynamic mode of the stabilizer bar were provided at the end of this
research.
The experimental setup consists on presenting and preparing the fatigue test bed were
all the experimental tests will have to be performed. The fatigue test bed presentation
will serve also as input data for elaborating the mathematical model of this, which will
be processed in a dynamic mode. His functionality resumes on fatigue experiments
represented by bending and torsion alternant symmetric cycles of the front and rear
stabilizer bars. Practically this fatigue test bed simulates the real and rough conditions
which these elements are supposed during a vehicle roll. This also gives the operator
possibility to adjust the motions amplitude and this is shown on Fig. 1.
Fig. 1. Fatigue test bed for stabilizer bars experimental tests

Kinematic and Dynamic Study of a Mechanism 163
The fatigue test bed technical data refers to: bending and torsion motions amplitude
which are applied to the stabilizer bars and it can be adjusted to a different values by
modifying the length of the actuation element; the fatigue test bed has a data acquisition
system for internal displacements and deformations measurement through a set of
tensometric sensors; counts the fatigue alternant symmetric cycles and also this can be
setup to a desired number when this will turn off automatically; gives the possibility to
turn off in case of an emergency, for example when a stabilizer bar was broke apart.
3 Rear Stabilizer Bar Test Bed Elasto-Dynamic Analysis
3.1 Mathematical Modeling

Based on research work from [2, 6] a similar mathematical model was built in order to
determine the longitudinal and transversal elastic displacements by overlapping the
solid-rigid motion with a solid-elastic one.
Fig. 2. Rear stabilizer bar test bed kinematic scheme and it’s virtual model (detailed view)
For the mechanism analysis from Fig. 2, the following input data were considered:
– The kinematic elements length are: O1A = 75 mm; connecting rod length AB =
550 mm; the rocker O2B = 200 mm.
– The kinematic elements mass: m1 = 11.62[kgs]; m2 = 3.05[kgs]; m3 = 9.63[kgs];
– The actuating element motion law: φ(t) = 18,587*t.
The mathematical models from [6] allows to perform an elasto-dynamic analysis on
a structural mechanism group R-R-R (revolute joints) type, identical with the one
(ABO2) from Fig. 2. For a numerical processing of these mathematical models with the
mentioned input data, a computational algorithm was elaborated in order to determine
the longitudinal elastic displacements u0 2k ðx; tÞ; u0 3k ðx; tÞ, depending on time, respec-
tively transversal elastic ones u00 2k ðx; tÞ; u00 3k ðx; tÞ during time, by considering the
kinematic joints reaction forces.
The transversal elastic displacement for the connecting rod (2), considered as
function of distance and time, were shown in Fig. 3, and for the rocker (3) these are
represented in Fig. 4.
Fig. 3. Rocker transversal displacements Fig. 4. Connecting rod longitudinal elastic dis-
[mm] during time [sec] placements [mm] during time [sec]
In case of elastic displacements on longitudinal direction, the maximum computed

value was of 0.0006 mm and the minimum value was −0.00006 mm obtained for the
connecting rod. On transversal direction for rocker element the maximum value was
0.10 mm and a minimum of −0.029 mm.
By having this important computed data, it can be remarked that the mathematical
model computed in a dynamic mode respect an alternative symmetric cycle, which this
was implemented at the beginning of the processed algorithm. In order to validate this
important results and the whole model, it is necessary to perform an experimental
analysis, by taking into account the essential conditions imposed to the mathematical
model which are: time, the parameterized model, an alternative symmetric cycle and
loads.
3.2 Numerical Simulations

In order to validate the mathematical model results, numerical simulations were per-
formed by creating a virtual model of the rear stabilizer bar fatigue test bed. The first
step for this, it was to create a virtual parameterized model with the aid of SolidWorks
software. After this, an interface was created for transferring this model on
MSC ADAMS environment, where the rear stabilizer bar was studied in a dynamic
mode as a deformable element, as it can be remarked in Fig. 2.
With the use of theoretical results, a kinematic parameters variation diagrams
depending on time were obtained. This was possible by overlapping the solid-rigid
motion with the one of solid-deformable in case of rear stabilizer bar. A part from the
Fig. 5. Rocker angular speed motion law during time
obtained results obtained through dynamic and durability analyses are presented in
Figs. 5, 6, 7 and 8.
From Fig. 5 it can be remarked that the rocker angular speed corresponds to an
alternant symmetric cycle for the proposed element. The angular speed amplitude has a
value of 20 rad/s and it has the same time interval as the one where the mathematical
model was processed.
Fig. 6. Rear stabilizer bar mass center transversal elastic deformations vs. time
In case of the transversal displacements from Fig. 6, for the rocker and implicit for
the rear stabilizer bar it can observe a maximum value of 0.11 mm and a minimum one
equal with −0.09 mm. These values are closer to the one processed in case of the
mathematical model, respectively Fig. 3.
For longitudinal elastic deformations in case of rear stabilizer bar, these are quite
small and also appropriate as values with the ones from the processed mathematical
model. As it can be remarked in Fig. 7, the highest recorded value was of 0.00035 mm
and a small one of −0.00025 mm.
An important result was characterized by the internal stress evaluation of the rear
stabilizer bar, due to the fact that this was considered as a solid deformable one. The
von Misses stress maximum value was around 300 N/mm2 and it can be seen in Fig. 8.
Fig. 7. Rear stabilizer bar mass center longitudinal elastic deformations vs. time
Fig. 8. Rear stabilizer bar Von Misses stress distribution during time
4 Experimental Analysis of a Rear Axle Stabilizer Bar
The experimental analysis aim is to obtain the internal displacements produced during
the torsion motion from the analyzed stabilizer bar. The torsion motion was given by
the test-bed and this is characterized by a symmetric alternant cycle. Thus, the obtained
experimental results can be compared with the one obtained on theoretical model. For
the experimental stress evaluation, was used a data acquisition system namely
MGCPlus, from HBM Company. This is shown on Fig. 9. A pair of tensometric
sensors LY – 10 types was used for the evaluation of the rear stabilizer bar internal
displacements, which were disposed at an angle of 45° between the rear stabilizer bar
axis and the generator cylinder curve. These are shown in Fig. 10 and were two sensors
placed in this way. The internal displacements produced inside the rear stabilizer bar
were recorded under the action of the tangential stress.
The experimental fatigue test-bed was started and there were recorded a total
number of 10 alternative symmetric cycles on 5 s. During this experimental test, the
acquisition data system records the internal displacements which are represented in the
diagram from Fig. 11. By analyzing these results it can be observed that the maximum
recorded values were 12.87 μm, and the minimum ones were −23.32 μm. Also it can
be remarked a small signal delay of the acquisition data system, due to the starting
point of the experimental test bed. This has a value of 1.5 s. The experimental analysis
effective time has a value of 4 s, when the test bed was turned off. But after turning this
off, the test bed mechanism continued to work due to the inertia forces which where
accumulated inside on the mechanical structure, for a period of 0.95 s.
Fig. 9. MGCPlus experimental data acquisition Fig. 10. Tensometric sensors placed on
system the rear stabilizer bar
Fig. 11. Internal displacements recorded during experimental analysis

5 Conclusions
The analyzed mechanism aim was to assure the necessary conditions for study the
fatigue experimental behavior of a rear stabilizer bar from a vehicle suspension system.
Also this fatigue test bed is conceived in a modular form, which is useful for experi-
mental analyses of different components from a vehicle structure. Thus, three aspects
were token in the sight of a complete research namely: dynamic response mathematical
modeling of the generated cycle mechanism with the longitudinal and transversal
variations identification during time which occur in case of the solid deformable
components motions; a fatigue test bed virtual prototyping with a complete assistance
on dynamic response for the analyzed element characterized by kinematic parameters
stress and deformations; fatigue experimental analysis of a rear stabilizer bar from a
vehicle suspension system with a controlled analysis in a dynamic mode for stress and
deformations, resulted from the symmetric alternant cycle.
The experimental analysis was performed by considering a fatigue test bed which
stresses a rear stabilizer bar through an alternant symmetric cycles. This stress is
represented to controlled bending and torsion motions. The obtained results certify that
the experimental tests will be performed with a small data acquisition error. In this way,
the recorded values during fatigue experimental tests of a rear stabilizer bar give the
possibility to increase or decrease the real conditions performed on experimental
analyses.
References
1. Wittek, A.M., Burdzik, R., Folega, P., Konieczny, Ł., Łazarz, B.: Influence of production
process and material factors on fatigue strength of tubular stabilizer bar. Arch. Metall. Mater.
60, 2485–2492 (2015)
2. Corless, R.M.: Essential in Maple. Springer, New York (1995)
3. Norton, R.L.: Design of Machinery, 4th edn. McGraw-Hill, New York (2008)
4. Juvinall, R.C., Marshek, K.M.: Marshek Fundamentals of Machine Component Design.
Wiley, New York (2003)
5. Dumitru, N., Margine, A., Catrina, Gh., et al.: Organe de maşini. Arbori şi lagăre. Proiectare
asistată de calculator, Editura Tehnica, Bucureşti. ISBN 978-973-31-2332-3 (2008)
6. Dumitru, N., Dumitru, S., Stoian, V., Geonea I.: Vehicle steering mechanism elastodynamic
analysis. In: Proceedings of International Conference of Mechanical Engineering ICOME
2015. Trans Tech Publications – Curent Solutions in Mechanical Engineering, pp. 241–246
(2016)
7. Dumitru, N., Dumitru, S., Copilusi, C., Geonea, I.: Modal dynamic analysis of the wicket
gate mechanism from a hydraulic turbine. In: Proceedings of the World Congress on
Engineering 2015, WCE 2015, Vol. II, pp. 1043–1048 (2015)
8. Marzbanrad, J., Yadollahi, A.: Fatigue life of an anti-roll bar of a passenger vehicle. World
Acad. Sci. Eng. Technol. Int. J. Mech. Aerosp. Ind. Mechatron. Manuf. Eng. 6(2), 407–413
(2012)
9. Nikravesh, P.E., Gim, G.: Ride and stability analysis of a sports car using multibody
dynamic simulation. Math. Comput. Model. J. 14, 953–958 (1990)
10. Shinde, P., Patnaik, M.: Parametric optimization to reduce stress concentration at corner
bends of solid and hollow stabilizer bar. Int. J. Res. Aeronaut. Mech. Eng. 1(4), 1–15 (2013)
11. Palmas, E.S., Santos, S.: Fatigue damage analysis in an automobile stabilizer bar.
J. Automobile Eng. 216(11), 865–871 (2002)
12. Liu, S.H., Li, F.: Fatigue life analysis of the stabilizer anti-roll bar using ANSYS. Adv.
Mater. Res. 383–390, 5894–5898 (2012)
Methods for Modeling an Elastic System
with Permanent Contour
Coupling Deformation
Hariton Poparad(&)

hariton.poparad@unitbv.ro
Abstract. This work proposes an analytical model that accounts for the
dynamic behavior of permanent coupling with elastic plate elements on contour.
Through CAE procedures, the influence of elastic elements and assembly node
rigidity on the coupling system can be established. Studies performed through
the use of CAE techniques allowed for practical application of the coupling
elastic system, the identification and monitoring of its performance, and vali-
dation of established analytical hypotheses. Additionally, MathCAD simulations
were used to build mathematical models of the coupling elastic systems. The
conclusions allow for an accurate prediction of the coupling system influence on
the mechanical transmission dynamic.
Keywords: Elastic system Displacement Coupling Misalignment
1 Introduction
The major functions of an elastic coupling system are to overtake spatial misalignments
in mechanical transmissions, absorb shock and torsional vibrations, and under the right
circumstances, simplify constructive solutions in the system. The coupling with elastic
elements on contour is classified under permanent elastic coupling systems with
metallic intermediary elements [1–3].
Elastic elements with curve plate geometry are displayed in a one-dimensional
plane, with their spatial arrangement depicted in Fig. 1. Studies performed through the
use of CAE techniques allow for practical applications of the coupling elastic system,
the identification and monitoring its performance, and validation of established ana-
lytical hypotheses.
The coupling properties are dictated by the rigidity parameters which further
determine the semi-coupling’s rotation angle φn at applied torsion force Mtn. The
analysis of the induced stress and the elastic deformation imply correlations between the
geometrical parameters (the radius of curvature R, the radius of connection r, central
angle 2α, height h, width b) shown in Fig. 1, and the number of elastic elements z [4].
Analysis conducted with MathCAD Software enabled the calculation of the dis-
placement of elastic nodes. The elastic system of the investigated coupling is estab-
lished through the assembly of at least two pairs of elastic plates in elastic nodes.

Methods for Modeling an Elastic System 171
Fig. 1. The coupling with elastic elements on contour A-A and C-C section view, B-B elastic
element section view, 1 – input shaft; 2, 3, 5,7 – elastic element; 4 – output shaft; 6 - spacer 8, 9,
10, 11 – attaching bolt; Mt – torque, R- radius of curvature, r - radius of connection, 2α - central
angle h – height, b – width, db- attaching bolt diameter, del-elastic plate hole diameter, D-elastic
plates mounting diameter, De-input shaft outer diameter
2 Modeling the Coupling Elastic Structure
The established model is characterized by the high values of the rigidity for the bending
constraint, as well as flexibility in tension constraint in the plane of the structure. The
working hypotheses formed for the elastic elements coupled in elastic nodes are:
• the rigidity in the plane of the structure, defined in Eq. (1), is secured through the
rigid beams in bending direction, and perfectly flexible both in tension as well as
compression directions;
0

K00 0
K00 ¼ 1 ð1Þ
0 K00
• elastic elements participate in deforming the structure.

According to the working hypotheses, modeling the elastic system of permanent
coupling is reduced to analyzing the system in two articulated beams (BEAM) and
establishing the two fundamental variants of a fixed elastic structure:
172 H. Poparad
2.1 Structure of Cylindrical Articulate Beams

(Rotation Is Permitted), Fig. 2
The displayed sequence allows the response of the static structure to be determined,
due to the applied force F on to node 1. According to the small displacements theory,
the forces applied to the nodes work on the undeformed rigid structure, leading to the
maintenance of the angle of articulation and the displacement of node 1 in accordance
to the circle marked with the center in C.
This hypothesis allows force F to be depicted by its two vectors components (E),
which are represented by the directions of the two beams (plate springs). Furthermore,
the theory expands to include the dominant contributors to the motion to be in tension
and compression stress.
The finite rotation in angle φ is designated by matrix R(u) from the Eq. (2), while
the node’s rigidity in the global reference system is represented by matrix KG, Eq. (4).
The displacement δ, of the elastic node is given by the product of the inverse matrix of
rigidity KG−1 of the applied force F and the dimensional factor d. The position of
node 1 is designated by the vector r.
RðuÞ ¼ eiwerternðeðuÞ; K eðuÞÞ ð2Þ

cosðuÞ
eðuÞ ¼ ð3Þ
sinðuÞ
X
KG ¼ Rðal Þ KLOC Rðal ÞT ð4Þ
l
d ¼ KG1 E \0 [ d ð5Þ
The application values for the structure of articulated beams in the Fig. 2 are:

0:785 1 0 0 1 1 0 0:9
a¼ ; KLOC ¼ ; K¼ ; KG ¼ ; d¼
0:785 0 0 1 0 0 1 0
Fig. 2. Structure of articulated beams

2.2 Structure of Fixed Beams, Fig. 3

This is an undetermined static structure that imposes the consideration of multiple
modes of node assembly. The matrix of nodal junctions I from Eq. (6), defines the
structure elements as composed of beams (the number of elements in the structure) and
lines (the first of which indicates the node from which the element originates, while the
second signifies the end point).
The structure’s elements are defined by several parameters. They include: the nodal
junctions and the incidence matrices, their planar position being designated by the polar
coordinates; Li the length of the elastic elements, and α the angle at the beam’s origin.

1
I \I [ ¼ ð6Þ
I þ1
SðIa;1;1Þ ¼ ð1Þa þ 1 ð7Þ
RTðIÞ ¼ Rðal Þ TdðIÞT ð8Þ
KL3I þ u;3I þ v ¼ K11ðIÞu;v ð9Þ
The application values for the structure of the fixed beam in the Fig. 3 are:
0 1 0 1
10 0 0 1 0 0
B C B C
K11ðIÞ ¼ @ 0 6 3 A; TdðIÞ ¼ @ 0 1 Ll A;
0 3 2 0 0 1
0 1 0 1
cosðuÞ sinðuÞ 0 1 0
B C 0 1 B C
RðuÞ ¼ @ sinðuÞ cosðuÞ 0 A; I¼ ; S ¼ @ 1 1 A
1 2
0 0 1 0 1
The rigidity matrix, constructed in the local base K11, is characterized by increased
compliance with increasing tension and compression, in comparison to the compliance
in bending. The preferred notations are: Td(I) the translational operator in a local base,
Fig. 3. Structure of fixed beams

174 H. Poparad
R(u) the rotational operator in a global reference system, S (Ia, l, l) the nodal incidence
matrix, and RT(I) the matrix of roto-translation.
Assembly of the structure is achieved through the intermediate operator SER (the
matrix of rotational incidence), as a product of the nodal incidence SI, l and rotation R
(αl). KL denotes the matrix of free elements, resulting in the following matrix of
assembly in a global reference system:
KGL ¼ SER KL SERT ð10Þ
The practical applications of the coupling with elastic plate elements on contour
have led to the feasibility of just five variants, represented by the matrix of free nodes
NL, and the number of equations NEQ, presented below in Figs. 4, 5, 6, 7, 8. Via the
operation in rigidity matrix KGF (where displacements are marked by 0 and the points
of beam intersection denoted by 1), the equilibrium equations are replaced with the
imposed displacements.
The exterior forces act on node 1, establishing its component on the x-axis. Loading
the structure results from the calculation of nodal displacement δl, as the product of the
inverse rigidity matrix KGF, the exterior force vector E, and the dimension d. The
calculated nodal displacements allow the deformed of the fixed beam structure to be
drawn using Eqs. (6) through (10) and practical examples as shown in Figs. 2 and 3. The
deformed structure was traced with cubical interpolators (HERMITE), while the main-
tenance of the deformations derived from the previous calculations with the vector ΚW,C.
By projecting the deformed structure in the local base of the beam, amplification
with the rigidity matrix KGF yields the forces in the beam direction from the node of
origin:
d
F ¼ KGL ð11Þ
d
The analysis of feasible solutions is presented below in Figs. 4, 5, 6, 7, 8:

1. c = 0 – represents the elastic structure with nodes 0, 2 fixed and 1 as translation
joint.
Fig. 4. CAE model for the elastic plate elements on contour, c = 0 constraints
2. c = 1 - represents the elastic structure with nodes 0, 2 fixed and 1 as rotational joint.
Fig. 5. CAE model for the elastic plate elements on contour, c = 1 constraints
3. c = 2, 3 - represents the elastic structure with nodes 0 as rotational joint, 1 as

translation joint and 2 fixed.
Fig. 6. CAE model for the elastic plate elements on contour, c = 2,3 constraints
176 H. Poparad
4. c = 4 - represents the elastic structure with nodes 0, 1 as rotational joint and 2 fixed.
Fig. 7. CAE model for elastic plate elements on contour, c = 4 constraints
5. c = 5- represents the elastic structure with nodes 0 fixed, 1 and 2 as rotational joints.
Fig. 8. CAE model for elastic plate elements on contour, c = 5 constraints

3 Conclusion
• To the best of the author’s knowledge, no work has been published related to
analytical modeling of the dynamic behavior of mechanical transmissions equipped
with permanent coupling with elastic elements on contour
• The analysis of feasible solutions in Figs. 4, 5, 6, 7, 8 allow the validation of the
analytical expressions deduced for the calculus of the working parameters of the
elastic coupling elements with curve plate geometry. The state of stress and
deformation in the curved plate springs is in dependence on the geometrical
parameters (the ratio of the two radius of connection - r/R, the central angle of the
curvature 2α).
• The results of the simulation of the deformed elastic structure of the coupling with
elastic plate elements on contour presented in this paper was optimized by choosing
the constructive practice module which led to the extension of the experimental
study of the hypothesis regarding the system of articulated beams in an elastic node.
• Practical applications of the coupling with elastic elements on contour presented in
this paper are limited by:
– the value of maximal torsion moment and its variations;
– the elastic characteristic of the coupling (rigidity level) limits the coupling
functional application range;
– the rotational angle of the semi-couplings;
– errors arising in assembly and execution which must be compensated without
exceeding the coupling components and assembly work capacity.
• This study mathematically showed that small displacement analysis alone is not
sufficient, thus further research is needed to approximate the image of the non-linear
response through the use of high displacements. This will be the basis of the
development of the dynamic response for the coupling equipped with plate elements
on contour.
References
1. Ugural, A.: Mechanical Design of Machine Components, 2nd edn. CRC Press, New York
(2015). ISBN 978-143887806
2. Beer, F., Russell, E.: Mechanics for Engineers – Dynamics, 5th edn. McGraw-Hill, New York
(2007). ISBN ISBN-10: 007246478X
3. Mott, L.R.: Machine Elements in Mechanical Design, 5th edn. Prentice Hall, Upper Saddle
River (2008). ISBN ISBN 10: 0135077931
4. Poparad, H.: Study of the motion of random axes using the coupling with curved plate
springs. In: The 11th IFToMM International Symposium of Science of Mechanisms and
Machines, Brasov 2013 (Romania), Ed. Springer, Heidelberg (2013). ISBN 978-3-319-
01844-7
Reinvestigation of the Electromagnetic Valve
Train (EMVT) Technology via Multidomain
Simulation
Christian Trebehs1(&) and Doru Groza2

1
Schaeffler Engineering GmbH, Werdohl, Germany
christian.trebehs@schaeffler-engineering.com
2
Schaeffler Romania S.R.L, Brasov, Romania
grozadru@schaeffler.com
Abstract. About 20 years ago electromagnetic valve trains have been topic of
many research projects in the automotive industry. Despite promising approa-
ches this technology had no chance to prevail on the market yet. In the meantime
reinforced market restrictions are introduced and improved technical boundary
conditions are available. The task of this project is to reevaluate the Electro-
magnetic Valve Train (EMVT) technology under these new circumstances.
A dynamic simulation environment in Dymola/Modelica is created to be able to
simulate the multidisciplinary system. This includes a magnetic, an electrical, a
mechanical and a controller subsystem. The biggest challenge regarding the
controller development is the high-dynamic valve movement. This is considered
by implementing a repetitive learning feed forward controller developed in
Matalb/Simulink/TargetLink via a FMI interface into the simulation. The sim-
ulation shows that a concept based on a conventional spring mass system with
additional lateral magnets is able to open the valves of an internal combustion
engine and simultaneously has a sufficient impact on velocity control.
Keywords: Electromagnetic valve train Multidomain simulation

Multiphysics simulation Dynamic system simulation
1 Introduction
The European Union legislation sets emission reduction targets for new cars. Target for
the year 2021 is a reduction of the carbon dioxide (CO2) emission to 95 g/km [1].
Besides the electrification of the powertrain, an efficiency improvement of conventional
internal combustion engine is a main approach to meet the target. One possibility to
optimize the combustion process is to increase the variability of the valve movement
and timing [2]. The behavior of conventional valve trains for internal combustion
engines strictly depends on the crankshaft angle. This implies that the valve movement
has a lift curve depending on the crankshaft angle. This raises two challenges: Con-
trolling the load and phasing the opening/closing times. The standard method to control
the load is the throttle flap. From a simplistic point of view the throttle flap is a
pneumatic resistance that reduces the gas flow in the cylinder. Considering the energy
consumption, this is an inefficient way to control the load. The phasing of gas exchange

Reinvestigation of the Electromagnetic Valve Train (EMVT) 179
valves for internal combustion engines has been a topic of interest for more than a
century [3]. An engine speed and load specific valve opening/closing timing is able to
reduce the losses during the gas exchange [2].
Variable valve trains are realized by different operating principles such as
hydraulic, pneumatic and electromagnetism, but only the electromagnetic valve train is
completely independent from the crankshaft [4]. Generally, an EMVT controls the
valve movement by an electromagnetic force, created by electromagnets. Major chal-
lenges of the development: Adequate valve seating velocity, fast travelling times and
proper size for automotive application. During the last 20 years different approaches of
an EMVT have been designed, though none has reached series production. But in the
meantime reinforced market restrictions are introduced and improved technical
boundary conditions are available.
For an EMVT development different engineering disciplines needs to be consid-
ered. The electric energy is stored in batteries and has to be turned into mechanical
energy. During the gas exchange the valve surrounding mass flow creates force acting
on it. Characterizing this force requires a CFD flow simulation. One key factor for
reaching an adequate landing speed is the controller technology. The fast valve
movement needs to be adjusted precisely.
This paper shows a possibility to reevaluate the EMVT technology using multi
domain simulations. For this purpose a simulation environment is created, which
includes magnetic, electrical, mechanical, gas dynamic and controller simulation. The
simulation process is demonstrated by an innovative concept. This concept consists of a
basic spring mass system with an additional magnet placed in the middle. Via the
additional magnet, further mentioned as middle magnet, it is possible to accelerate and
decelerate the valve during movement.
2 Basic Principles of a Spring Mass System Based EMVT
The introduced concept is basically a spring mass system with an additional lateral
middle magnet (Fig. 1). The spring mass system is a widespread concept [5]. The valve
is pre-loaded by two springs. The equilibrium position is centered between the closed
and opened valve position. Therefore the system is able to oscillate between the end
positions. During movement there is a conversion between kinetic and potential energy.
In each end position an electromagnet is able to hold the valve. Thereby the potential
energy is at maximum. During the movement mechanical losses occur. To reach the
other end position, it is necessary to compensate these losses by the catching magnets.
The losses are based on the mechanical friction, the remaining force of the holding
magnet and the gas forces. Consequent, the equation of motion is:
meff €s þ c_s þ ks ¼ FM ði; sÞ þ FG ð1Þ
The coordinate s describes the valve position. The effective mass meff includes all
moving masses of the oscillating system. For example the valve, the magnetic armature
and the reduced spring mass. FM is the resultant force of all magnets, FG is the force
180 C. Trebehs and D. Groza
due to the pressure inside the cylinder, k the spring stiffness, c damping constant and
i the current through the coil. On the left hand side of Fig. 1. the concept design is
shown. The different components are marked with numbers from 1 to 7.
Fig. 1 Concept design and force graph
On the right hand, the force course depending on the stroke is shown. The numbers
of the diagram are corresponding to the components on the left hand side. The left limit
of the diagram represents a closed valve, while on the right side the valve is completely
opened. The straight line 1 + 6 shows the resultant force of both springs. In both end
positions the force is maximum and at half stroke the force is 0 N. The curves 3 and 5
represent the force of each end magnet at constant current. The slope of the course rises
sharply to the end positions. The force of the middle magnet (curve 2) has its minimum
nearly at ¼ stroke and the maximum at ¾ stroke. To accelerate the valve it is necessary
to activate the magnet at beginning of movement and deactivate it before passing the
middle position. Curve 7 shows schematically the gas force at the exhaust valve, curve
4 represents the damping forces during movement.
In a first approximation, the opening and closing duration Δt depends on the
systems natural frequency. For the considered spring stiffness k and mass m the fol-
lowing applies:
p2 m
k¼ ð2Þ
Dt2
To hold the valve in end position, the magnetic force needs to be higher than the
maximum spring force.
3 Simulation method
To represent the overall system behavior of an EMVT the simulation environment

needs to cover:
• The mechanical system,
• The electrical system,
• The magnets,
• The gas forces and the mass flow around the valve and
• The controller.
The overall system simulation is implemented in Dymola/Modelica and comple-
mented by submodels created by FEMM, AVL Fire and TargetLink. In the following
the different subsystems are explained.
3.1 Simulation of the mechanical and electrical system

The electrical system for each magnet consists of a voltage source, two switching
outputs and the magnet itself.
Every switching output has two transistors and two freewheeling diodes. By closing
and opening the transistors a clocked current flow in both directions is realized. The
magnets consist of an ohmic resistor and an inductor. The voltage U over the magnet is
defined as:
d ði Lði; xÞÞ
U ¼ i Rþ ð3Þ
dt
The inductance L depends on the current i and the distance x between armature and
magnet. R describes the ohmic resistance of the coil. This dependence is given by a
magnet specific 2D lookup table created by the magnetic field simulation.
Fig. 2 2D Finite element model of the end magnet (left), middle magnet (center) and full 3D
model (right)
The mechanical part consists of the moving mass, the springs (include the damping
characteristic) and the hard stops. The spring force depends linear on the mass position
and the damper force linear on the mass velocity. The limitation is realized by a
spring/damper element, which is only considered in the end position. The parameters
are adjusted to the valve seat stiffness of a conventional internal combustion engine.
3.2 Magnetic field simulation

Electromagnets are based on the reluctance principle, where forces only act in one
direction [6]. The Lorenz force applied on a volume over charge distribution describes
both the electrical and magnetic part. The force is defined as:
ZZZ
F¼ ðqE þ qv lH Þ dV; ð4Þ
where q is electrical charge, E is the electrical field, ρ is charge density, ν is the velocity
vector of the charged particles, μ is the magnetic permeability, H is the magnetic field
intensity and dV is the volume element hence the triple integral.
These restrains prevent the possibility to define a general formula that predicts
magnetic force over the entire actuator. The formula can only be applied to elemental
units of volume that have thoroughly defined boundaries. Finite element analysis is
thus used to define regions with predictable behavior. The following explained 2D
model is built in the finite element software FEMM [7]. Fig. 2 illustrates the 2D
axisymmetric model of the top magnet (left), the 2D planar model of the middle magnet
(center) and a complete 3D benchmark model (right). The FEMM software simulates
static behavior. Multiple simulations at various armature positions have been carried
out. An equivalent dynamic simulation is built by merging the results. The result is
validated by comparison with a 3D dynamic simulation.
Magnetic simulation results are linked to the overall system simulation by two
lookup tables. The lookup tables describe the force and the inductance in relation to
current and armature distance.
3.3 Gas force analysis

During the gas exchange phase, the gas causes force acting on the valve. This force
influences the valve movement and due to this there is an impact on the EMVT design.
At the beginning of the opening process the pressure difference around the valve is at
its maximum and it decreases during the valve opening. For analyzing the gas forces
CFD simulation is used.
Thereby one challenge is the interaction between the force and the valve move-
ment, as the gas forces depend on the valve movement. The movement is defined by
the overall system simulation and depends on the gas forces defined by the CFD
simulation. Both simulations are performed sequentially. To ensure that the force and
valve movement of both simulations are identic, an iterative approach is chosen
(Fig. 3).
Fig. 3 Iterative process of the gas force analysis
The initial CFD simulation starts with a generic valve movement. The resulting
force is transferred to the dynamic simulation. The second CFD simulation uses the
valve movement out of the dynamic simulation. This procedure is repeated as long as
the difference between the calculated and the implemented valve movements are higher
than required.
3.4 Controller simulation

The controller has major influence to the overall system behavior as it adjusts the valve
movement. One requirement on the controller is to reduce the valve impact velocity
during seating.
Generally, an automotive controller is a time based system. Each calculation step is
repeated in a defined interval. This is in contrast to the overall system simulation
arithmetic. The Dymola solver determines the time steps by itself. Due to this a direct
implementation of the controller behavior in Dymola/Modelica is time-consuming.
Using the functional mockup interface (FMI) is a more efficient way. The functional
development of the controller is performed in TargetLink. TargetLink is a software to
build source code from Simulink models [8]. The resulting source code is compiled to a
functional mockup unit and is directly implemented to the overall system simulation.
This model can be integrated to the prototype control unit PROtroniC.
A system without taking friction, the magnets inductive behavior and the gas forces
into account is in the following referred as an ideal system. Considering an ideal system
starting from one end, the valve reaches precisely the other end position. There is no
temporal delay in the magnetic force build up. If the magnets are activated in the
moment, the armature arrives, the resultant impact velocity is 0 m/s. Furthermore, in an
ideal system the energy does not increase during catching the magnet.
In a real system the magnets compensate the losses. If the losses are precisely
compensated the impact velocity is near 0 m/s. The controller adjusts the voltage
applied to the magnets by triggering the switching outputs. A schematic voltage course
is given in Fig. 4.
If the armature is outside the magnetic field the magnets are deactivated. After
passing the activation position the voltage is set to maximum to build up the magnetic
Fig. 4 Schematic voltage course
flux. If the current reaches the defined threshold the voltage is reduced to uhol.
To release the armature the magnet is deactivated. The earlier the magnet is activated,
the higher the added energy is. Thus the controller task is to ensure that the magnets are
activated at the right time in order to add the energy which corresponds to the losses.
The valve velocity is too high to adjust the activation position during translation.
The controller analyses the valve movement for the current duty cycle and adjusts the
activation position for the next one (Fig. 5). If the impact velocity of the last duty cycle
higher than a defined threshold, the activation position is postponed. By this Repetitive
Learning Feed Forward Controller the impact velocity is reduced from duty cycle to
next duty cycle. Thereby it is necessary, that the valve always reaches the seat.
Otherwise the valve begins to oscillate. By increasing the velocity threshold, this risk is
minimized.
Fig. 5 Working principal of the Repetitive Learning Feed Forward Controller

4 Simulation results
In this chapter some simulation results are shown. The simulation scenario is consid-
ering an engine rotational speed of 6500 rpm. Fig. 6 shows the valve movement and
the magnets current course for one duty cycle.
Fig. 6 Valve movement and magnets current course
The duty cycle starts at 0° CA. This is defined as the top dead center next to the
ignition. At 130° CA the upper magnet is deactivated. Due to the spring force, the valve
starts to move from the closed to the opened position. At 164° CA the middle magnet is
activated and deactivated before the valve reaches the middle position. During this time
the valve is accelerated. At 254° CA the lower magnet is at maximum current and
catches the arriving valve. Energy losses occur during the valve movement. These
losses are compensated by the magnets. Thus the outputted and the inputted energy are
balanced (Fig. 7).
Fig. 7 Mechanical work acting on the valve
70 % of the energy losses are caused by the gas forces. The valve is opened against
the cylinder internal pressure. During the opening the valve is decelerated and the
kinetic energy decreases. Additionally 15 % of the energy losses are based each on the
friction and the magnets during releasing the valve (Lorentz force).
Most of the inputted energy (65 %) is handled by the middle magnet by acceler-
ating the valve during movement. In the end positions, the upper and lower magnets
attract the valve and increase the energy (20 % and 15 %).
5 Summary
The simulations results are not aligned with measured data, but they allow conclusion
of the overall system. The simulation environment considers all important technical sub
systems. So that it is possible to design the magnets and the controller before building a
prototype.
The spring mass system with additional middle magnet has advantages compared to
existing concept. The mechanical work balance shows, that this concept enables to
acting the valve even at the exhaust side. The energy losses by the gas forces are higher
than the energy inputted by the magnets in the end position. The middle magnet gives
an additional possibility to compensate the losses. Thereby the adjustment of the
middle magnet inputted energy is more precisely than the inputted energy of the end
magnets.
Furthermore the developed Repetitive Learning Feed Forward Controller handles
the valve movement. The actual valve movement adjusts the control signal for the next
valve opening. Thus it is possible to reduce the requirements for sensors and the control
unit for automotive application.
References
1. The European Parliament And The Council Of The European Union: Regulation (EC) No
443/2009 of the European Parliament and of the Council of 23 April 2009 setting emission
performance standards for new passenger cars as part of the Community’s integrated approach
to reduce CO 2 emissions from light-duty vehicles, EU (2009)
2. Mahle GmbH: Ventiltrieb: Systeme und Komponenten (Valve-train: Systems and Compo-
nents). Springer Vieweg, Wiesbaden (2012)
3. Brallsford, A., et al.: GB143769A: Improvements in and relating valve control mechanism
(1919)
4. Gutzmer, P.: Paradigms of Future Mobility. In: 10th Schaeffler Symposium, Springer Vieweg
(2014)
5. Scheidt, M.: Untersuchung des Betriebsverhaltens eines schnellen elektromechanischen
Schaltsystems (Operating Behavior Investigation of a Fast Electromechanic Switching
System). RWTH Aachen (1991)
6. Kallenbach, E.: Elektromagnete: Grundlagen, Berechnung, Entwurf und Anwendung
(Electromagnets: Basics, Calculation, Design and Application). Teubner Verlag, Stuttgart
(2003)
7. Finite Element Method Magnetics, http://www.femm.info
8. dSPACE GmbH, http://www.dspace.com
Study on Mixture Formation
at an Experimental Spark Ignition Engine
Dan Mihai Dogariu(&), Anghel Chiru, Stelian Țârulescu,

Marius Lazăr, and Vlad Ștefan Stancu

dan.dogariu@gmail.com,
{achiru,s.tarulescu,marius.lazar}@unitbv.ro,
vladss02@yahoo.com
Abstract. Important data on volumetric efficiency for a spark ignition engine

can be offered by CFD analyses on intake process. Thus, the combustion quality
can be estimated, which has to be verified by experimental research, on a test
stand. The engine subjected to testing is a research gasoline single cylinder
engine located at the Research Institute of Transilvania University of Brasov and
used for thermodynamic tests. In order to compare the results obtained during
simulation with the measurements on the running engine, the combustion
chamber was accurately reconstructed using high end touch probe 3D scanning
technology. The preliminary results obtained provide information on the flow
phenomena through the intake port. Knocking during testing stratified mixture
formation was under control, with the help of knocking sensors which are useful
for determining the critical spark timing.
Keywords: Engine testing Multiple injections Gasoline combustion

Mixture formation CFD simulations
1 Introduction
The flow phenomena during the intake process have a great influence on mixture
formation prior ignition, which is later the main contributor to engine performances and
emissions.
To study mixture formation using experimental techniques is difficult. Despite the
higher uncertainty that numerical simulation of mixture formation has compared to
experiments, it can offer more extensive information about complex processes in the
cylinder and in areas of interest at any place inside the computational domain [8].
In order to perform virtual analyses, especially for gas exchange and thermody-
namic simulations, and to compare the results obtained with the measured data on the
research engine, the virtual geometry should match accurately the one of the real engine
[1]. This kind of simulation wishes to determine details about the mixture formation,
ignition and combustion. A prediction on heat release rate and emission formation can
be made [8]. In order to realize such simulation, sub-processes of injection, fuel jet
break-up and evaporation of single liquid droplets and droplet dynamics are calculated
for the mass particles taking part in the simulation.

188 D.M. Dogariu et al.
In order to simulate fuel injection, the real process must be considered and sim-
plified during injection modeling. For example, two-phase fluid flow physics has to be
added to the simulation. This model type considers simulating mixing of two fluids
(fuel and air) having different properties.
For the combustion model, simplified the chemical components of the fuel must be
modeled. The more detailed the fuel composition is, more details regarding resulted
emissions can be obtained, at the cost of a higher computing time.
In a conventional spark ignition (SI) engine, the fuel is entering the cylinder
together with the air during the intake stroke and load regulation occurs quantitatively,
such that air and fuel always exists globally in the same stoichiometric ratio [5]. In
contrast to port fuel injection, in direct injection (DI) gasoline engines, the load reg-
ulation is realized both quantitatively and qualitatively [5]. In this way, the mixture
inside the integral cylinder may be lean, so a stratified charge is realized.
2 Prerequisites for Testing and Simulation
In order to obtain a high accurate 3D model of the real combustion chamber, the
geometry was obtained by 3D scanning using a touch probe scanner [1].
The main characteristics and parameters of the CMM (Coordinate Measuring
Machine) and touch probes used are presented in Table 1 [9–11].
Table 1. Main characteristics of the CMM and touch probes used [9–11]
Characteristics
Length measuring error 0.9 + L/350 [µm]
Repeatability range 0.8 [µm]
Scanning error 3.0 [µm]
Probing force Constant – programmable between 50 – 1000 [mN]
Smallest stylus tip diameter 0.3 [mm]
Travel speed 0–70 [mm/s], when motorized
Fig. 1. Measuring the engine components of the experimental singe cylinder on a bridge-type
CMM and visualization of the 3D Models
Study on Mixture Formation at an Experimental Spark Ignition Engine 189
Some parts were scanned preassembled for a better reproduction of assembly

dimensions, such as valve lift, spark plug geometry and position and injector nozzle, as
seen in Fig. 1.
With an exact 3D replica of the existing working engine, one could compare the
results obtained in the simulation with the results on the test bench, having a better
equivalence [2].
3 Experiments on Single-Cylinder Spark Ignition Engine
The initial conditions for testing and testing regime are presented in Table 2.
Table 2. Initial conditions for testing and testing regime

Initial Conditions
Atmospheric pressure at 600 m alt. 0.95 [bar]
Room temperature 25 [°C]
Test regime 1 Test regime 2
Dyno speed [rpm] 1500 3000
Dyno load [%] 5 5
Manifold pressure [bar] 0.6132 0.644
Injection type One main injection 2 injections and MPI
Fuel rail pressure [bar] 80 80
Duration of injection [µs] 1500 1st: 1500
2nd: 650
MPI: 2200
Injection advance [°Crank Angle] −300 1st: −300
2nd: −250
MPI: 300
Ignition [°Crank Angle] −18 −18
Data acquisition was performed with AVL hardware.

The equipment is electronically controlled by a management system of type
AVL-RPEMS, which uses CAN interface, called ETAS ETK 7.1. Through this system
the load is regulated, correspondingly to the two fundamental combustion principles
where the mixture is realized quantitatively and qualitatively. For the presented
gasoline application, this means a control of the throttle position.
Figure 2 presents the evolution of the pressure inside the cylinder at the first testing
regime.
There have been recorded a total number of 100 consecutive cycles, from which the
maximum pressure was 24.5 [bar] at cycle number 80, which is also shown in Fig. 2,
the minimum 16.51 [bar] and the average 20.13 [bar]. The pressure increase rate for the
presented cycle where the maximum pressure was achieved is 0.91 [bar/°Crk].
Figure 3 shows the cycle dispersion on several superimposed pressure cycles.
Fig. 2. Cylinder pressure, injection duration and timing and ignition timing for the test regime 1
Fig. 3. Cylinder pressure variation between the 100 measured cycles for the test regime 1
De-throttling at low speeds is known to have positive effects on fuel consumption.

The valve lifts are important in such a way that a vacuum that could cause load cycle
losses occurs in the cylinder during the intake stroke. An increase in the charge motion
in the cylinder can contribute to improving the mixture formation and thus making the
combustion process more efficient, especially at low loads [7] (Fig. 4).
Fig. 4. Mixture formation in direct-injection spark-ignition engines with wall-guided charge

stratification procedure [5]
4 Mixture Formation Prediction Through CFD Trials
The main advantage that DI offers is the qualitative load regulation at part load [5]. The
global air ratio obtained due to combustion requirements is lean and a so-called
stratification is realized, in such a way that ignitable mixture moves in the region of the
spark plug at the right ignition timing. However, the most important influence comes
from the combustion chamber design. Both the shape in the cylinder head and the
protrusions in the piston are strongly influencing the mixture formation inside the
cylinder during the piston travel [8].
The piston used in this application and the position of the injector under the intake
valves is creating a stratified mixture using a wall-guided procedure, in which the fuel
jet is led to the spark plug by the flow generated by the piston shape. Other strategies
are using high tumble and spray-guided methods [5].
The orientation of the spark plug has a major influence on combustion conditions
during flow guided mixture formation, reason why the engine components were
measured by scanning them assembled [6].
Combining port fuel injection and direct injection, a stratified combustion process
can be achieved, with an ignitable mixture at the spark plug and a lean mixture in the
outer combustion chamber.
One of the advantages of a stratified lean combustion is, for example, a higher
efficiency of the engine due to two factors: dethrottling and improvement of com-
bustion efficiency. The throttling losses can be reduced by increasing the air to fuel
ratios, at partial loads.
4.1 Prerequisites for the CFD Analysis

In order to realize a flow simulation through the studied engine, the 3D fluid domain
had to be considered. Moreover, the simulation model requires some values mea-
sureable on the test stand, which will be used as input data. The mesh for such a model
presented in this paper must contain typical key features for an analysis on thermo-
dynamic cycles [3].
Some of the boundary and load requirements are the valve lift curves and piston
displacement [4]. Figure 5 presents the used valve lifts, which were measured. These
graphs have to be imported in the simulation software which will use them to orient the
components at the right time step.
The prepared model features a model for fuel injection and automatic remeshing,
which is crucial during engine valves and piston displacement.
In Fig. 6 the mesh on the fluid domain can be observed.
In order to consider a simulation covering the complete four stroke engine cycle,
three meshes can be used. These meshes must consider the closed position of the
engine valves. Only one initial geometric model, having the engine valves in one
prescribed position may be exported and further used for this meshing. The close
position of the valves must be described when using the valve lifts curves. The resulted
meshes can be assembled into one single simulation model or used independently in
Fig. 5. Necessary valve lifts to be used in the analysis
Fig. 6. Fluid domain meshing
simulations, depending on the type of analysis. For example for a combustion simu-
lation, only the closed volume of the combustion chamber can be used.
4.2 Preliminary Results

A first analysis was realized only during the intake stroke.
Figures 7 and 8 show the velocity field within the domain at full intake valve
opening, where the piston is positioned at 450º crank angle. In order to observe the flow
and the influence that the piston head has, slices were made through the domain.
In Fig. 7, right side, there are slices perpendicular to cylinder axis in order to see a
possible swirl motion. This motion is limited, since the geometry of the piston head
assures high tumble motions, which can be better seen in the left side of the figure.
Figure 8 shows a frontal view on two of the five slices made in Fig. 7. One slice
passes though one intake valve section, and the next slice, shown in Fig. 8 at right,
shows the velocity field evolution in the middle of the cylinder, where the spark plug is
located. The tumble motion turns the flow towards the ignition point.
Fig. 7. Slices through the fluid domain showing the velocity field
Fig. 8. Detailed view on the flow evolution in one intake valve section and in the middle section
of the cylinder
5 Conclusions and Future Work
The accuracy of the scanned geometry can offer realistic results. When using models
for a CFD analysis, the level of detail is critical, but influences the computing time.
The observed flow phenomena observed in the resulted CFD simulation on the
intake stroke provides information on the tumble motion created inside the cylinder,
mainly due to the piston shape. Since the injector is located below the intake valves, the
injected fuel can be trained in the tumble motion together with the intake air.
Future work will continue on detailed mixture formation analysis using CFD. The
results obtained during CFD analyses must be compared to the results obtained at the
test stand, also using optical devices in order to gain insight into the mixing process.
Acknowledgements. This paper was supported by SC Schaeffler România SRL.
References
1. Dogariu, D.M., Andrei, C., Vieru, B.T., Plămădeală, R.A., Chiru, A.: A method for 3D
geometry scanning of a combustion chamber. In: Andreescu, C., Clenci A. (eds.)
Proceedings of the European Automotive Congress EAEC-ESFA 2015, pp. 577–584.
Springer, Heidelberg (2016)
2. Dogariu, D.M., Tănăsie, C., Chiru, A., Leahu, C.I., Stancu, V.Ș.: Model preparation for
structural FEA on main components of an internal combustion engine. In: Andreescu, C.,
Clenci A. (eds.) Proceedings of the European Automotive Congress EAEC-ESFA 2015,
pp. 455–462. Springer, Heidelberg (2016)
3. Juretić, F., Moser, W., Rainer, G.: Recent advances in AVL’s CFD mesh generation software
– ESE-tools. In: International Design Conference - DESIGN 2008, pp. 643–648 (2008)
4. AVL FIRE documentation (2014)
5. Merker, P.G., Schwarz, C., Stiesch, G., Otto, F.: Simulating Combustion. Springer,
Heidelberg (2006)
6. Law, K.C.: Combustion Physics. Cambridge University Press, New York (2006)
7. Solving the powertrain puzzle. Springer Vieweg – Schaeffler Technologies GmbH & Co.
KG (2014)
8. Baumgarten, C.: Mixture Formation in Internal Combustion Engines. Springer, Heidelberg
(2006)
9. CMM product overview, Industrial Metrology in Focus, Carl Zeiss brochure
10. PRISMO navigator sensor and accuracy overview, www.optecgroup.com
11. Touch probe specifications. www.zeiss.com/industrial-metrology/en_us/products/sensors
Aspects Concerning Modeling of Combustion
for Compression Ignition Engine
with Injection Management
Dan Mihai Dogariu(&), Anghel Chiru, Stelian Țârulescu,

and Marius Lazăr

dan.dogariu@gmail.com,
{achiru,s.tarulescu,marius.lazar}@unitbv.ro
Abstract. The current paper presents the results of the research obtained by the
authors in the study of the influence of the sequential injection, according to
different injection laws, on the main parameters of the indicated cycle, such as:
auto-ignition delay, combustion duration, maximum pressure and pressure
variation inside the combustion chamber. The objective of the study is to obtain
reduced fuel consumption in order to reduce the resulted emissions. The theo-
retical and experimental evaluations are realized on an experimental single
cylinder AVL 5402 engine.
The application for this proposed study is on hybrid vehicles, where the
experimental compression ignition engine is tested as a range extender. These
application requirements were considered when testing the experimental engine.
The paper also presents simulations for the tested regimes.
Keywords: Multiple injections Diesel combustion Mixture formation

Range extender
1 Introduction
Stringent emission legislation and smaller fuel consumption have defined the opti-
mization of internal combustion engines as the main task for the automotive industry in
the future, since they will continue to play a major role in vehicle propulsion. Solutions
today for internal combustion engine improvement include variable valve actuation
(VVA), cylinder deactivation (CDA), both static and rolling, thermal management and
other gas cycle related improvements by combining charging strategies with more than
one turbocharger.
A particular attention must be paid upon exhaust after treatment devices, which are
more complex for Diesel engines. However, modern engines using strategies for cut-
ting down emissions and fuel consumption with start-stop systems, which are included
in the hybrid category of vehicles, are in contradiction to classical exhaust gas after
treatment systems, where a constantly high exhaust gas temperature is required.
Methods for achieving the exigent emission legislations in the future suppose a
higher degree of electrification of the vital systems of the thermal engine, up to a

complete hybridization of the propulsion system. So, the challenge is to find the right
formula to efficiently combine the best of the thermal engines improvements and
electric drive.
Fig. 1. The powertrain configuration of a series hybrid [1]
In this way, the paper wishes to propose the possibility of using a Diesel engine in a
serial hybrid layout, with the focus on the advantages of this chosen configuration and
of the Diesel engine, which is sketched in Fig. 1 [1, 2].
Since the combustion engine has no mechanical connection to the wheels, new
potential design and layout options are conceivable [9].
Since the Diesel engine is used as a prime mover, only to generate electrical energy
with the help of a generator, and is not directly involved in the traction of the wheels, it
may operate at a constant high thermal efficiency regime, thus improving the fuel
consumption [3, 4].
The study will be carried on an experimental single cylinder AVL 5402 Diesel
engine from the test cell of the Department of Automotive and Transport Engineering
at Transilvania University of Brașov.
The possibilities of operation improvement of the tested engine for this application
are based on modeling different injection strategies while evaluating the resulted
exhaust gases.
2 Application Requirements
The stabilized working regimes for other range extender units are different depending
on the strategy chosen by the producer. The engines are usually small spark ignition
and mainly run under full load, having at least two main operating points. One point
may be at half of the power output at the crankshaft [9] and another at full power output
when the charge level of battery is to be retained even at high speeds [9].
The proposed study follows two working regimes of speed and load for the
experimental single cylinder compression ignition.
For the two working regimes, the injection parameters are modified such that the
engine fulfills the initial conditions of load and speed.
The injection law can be modeled in terms of injection timing and duration of
injection. The possibility to use multiple injections is also considered.
Aspects Concerning Modeling of Combustion 197
3 Experimental Compression Ignition Single Cylinder

Testing
The main parameters of the compression ignition (CI) tested engine are presented in
Table 1.
Table 1. Main parameters of the engine [7].

Engine parameters Value
Bore 85 mm
Stroke 90 mm
Swept volume 510.7 cm3
Offset from TDC 9.0 mm
Injector’s number of orifices 5
The equipment is electronically controlled by a management system of type

AVL-RPEMS. This system uses a control unit with CAN interface, called ETAS ETK
7.1. Through this system the load is regulated, correspondingly to the two fundamental
combustion principles where the mixture is realized quantitatively and qualitatively.
For the experimental compression ignition engine, this means the control of pressure
and injection duration correspondingly to a specific map. For gasoline, the system
controls the throttle position.
The test stand is monitoring the following parameters: ambient air temperature and
pressure, cooling water and oil temperature and pressure, fuel temperature and pres-
sure, hourly and specific fuel consumption, air to fuel ratio.
Table 2. Initial conditions for testing and testing regime.

Initial conditions
Atmospheric pressure at 600 m alt. 0.95 bar
Room temperature 25 ºC
Dyno speed [rpm] 1500 3000
Dyno load [%] 75 75
Injection type 2 injections 2 injections
Fuel rail pressure [bar] max. 1600 max. 1600
The data acquisition system comprises the measurement chain for the in-cylinder
pressure, crank angle marker and top dead center position, but also the current absorbed
by the injector, used to determine the duration of injection. The system also measure
the inlet and exhaust temperatures and pressures, which for the gasoline application can
differ from the ambient conditions and the blow-by flow.
For measuring the resulted emissions, a special particle detector was used, a micro
soot sensor, of photo acoustic measurement principle.
The system records a number of one hundred cycles, increasing the measurement
trust level, by offering a covariance factor.
All these information are recorded in parallel mode, which offers the possibility of
real time data acquisition.
From the testing cycles, there were chosen two testing regimes correspondingly to
the studied applications.
From the resulted tests, there has been selected one injection configuration for each
proposed testing regime that provided the best values for fuel consumption and soot
parameters.
The initial conditions for testing and testing regimes are presented in Table 2.
3.1 Evaluation of In-Cylinder Processes

Data acquisition was performed with AVL hardware.
Figure 2 presents the evolution of the pressure inside the cylinder for the two
testing regimes. At test 2, the duration of the main ignition is larger, due to the increase
in engine speed. Because the higher load and speed at test 2, some ripples may appear
close to the maximum pressure value. The ripples coincide with the combustion at
constant pressure, with respect to compression ignition engines combustion. They
usually appear when measuring raw data, without any filtering. They are visible also at
test 1, but because of the shorter combustion duration, they are not as obvious (Fig. 2).
Fig. 2. Cylinder pressure, injection duration and timing for the two testing regimes
Table 3. Resulted engine performances and emissions.

Test Results
1500 rpm/75 % load 3000 rpm/75 % load
Power output [kW] 3.7 6.05
Torque output [Nm] 23.5 19.4
Hourly consumption [kg/h] 1.092 1.849
Max. cylinder pressure [bar] 71 74
Lambda [–] 1.2 1.4
CO2 [%] 11.9 10.1
CO [ppm] 3248 1216
NO [ppm] 789 839
NO2 [ppm] 25 29
NOx [ppm] 814 868
When choosing the working regimes for the experimental engine, the requirements
for the electric generator application for range extender were considered. The injection
modeling with two fuel injections per cycle was performed such that the resulted
performances and emissions to be as good as possible. The summary of the results
obtained can be observed in Table 3.
The engine runs at the synchronous speeds in both tests [3]. The speeds of
1500 rpm, respectively of 3000 rpm are one of the best suited speeds for driving
electric generators. Using these speeds can also reduce complexity in the construction
of a range extender.
The next most important result is the torque output, which should remain as con-
stant as possible [3]. Figure 3 shows the torque and power evolution for 75 % loading
of the engine.
The regime of testing at 75 % load was considered instead of the 100 % load due to
concerns regarding safety and overloading of the engine components.
The hourly consumption increases for the second regime tested, since the speed
doubles and there is more fuel injected in the cylinder.
The obtained value for air to fuel ratio is very important, since the diesel after
treatment is complex due to the excess of air in the exhaust gases. Regeneration cycles
may require rich fuel to air mixtures [5]. However, the range extender application
considers the two working regimes at constant operation and thus, harsh accelerations,
where the fuel to air mixture goes rich, are not met.
Moreover, transient functioning from stop to the first regime and then to the second
and back should be carried out automatically using a well-defined engine map as well
as oxygen sensors on the exhaust stream, for the feedback system, and mass air flow
sensor after air filter, for the open loop system, to decide on fueling amount.
Fig. 3. Torque and power evolution at 75 % load
3.2 Simulation
In order to perform combustion simulations with different injection parameters, the
simulation model requires some values measureable on the test stand. Moreover, the
geometry of the combustion chamber is required and as much data about the used
injector as possible. The geometry of the piston is critical, since it incorporates the
combustion chamber.
Initial values of the engine kinematics, such as compression ratio, are recalculated
considering the resulted geometry. If there are small differences between these values,
the simulation may still run, because the model uses a compensation volume which is
added at the outer part of the geometry [8]. But for bigger differences, the geometrical
model should be revised. Differences may appear during components measurement and
reconstruction or omitting assembly variable components, such as compression ratio
shims. These shims are used to modify the geometrical compression ratio.
However, a measured pressure inside the cylinder after the compression stroke can
differ from the one calculated considering the geometrical dimension, mainly due to
valve positioning, but also blow by, in a certain percentage.
The two simulation models used in this study were created in dedicated combustion
simulation software, developed by the producer of the test stand, AVL Fire. The
models include the possibility of simulating the two injections per cycle which are
proposed in this study. The analyzed geometry is axisymmetric and meshed with spray
adapted grid, having the possibility of automatic remeshing for better element quality in
the complete domain during the piston displacement [6].
The conditions from testing regime 1 were considered in a combustion simulation
for observing the mixture formation process. Some snapshots taken from critical time
steps during the two injections can be seen in Fig. 4.
The velocity fields can outline the mixture shape and the mixture fraction shows the
fuel distribution within the cylinder.
When the main injection starts, there is still combustion due to the pilot injection.
This can be seen in the picture showing the flame speed, which presents the moment
where both flames exist at the same time.
Fig. 4. Injection simulation and mixture formation during two consecutive injections
4 Conclusions and Future Work
The two speed regimes chosen for this study, of 1500 and 3000 rpm correspond to
electrical generators requirements of synchronous type [3]. The electrical machines are
most efficient at these speeds. The load on the generator may be kept constant, while
the demand of power to the wheel is assured from the batteries. In some driving
regimes, the generator can supply the traction motors in a direct link and the battery to
be used as a buffer to compensate for the variable loading coming from the road. This is
usually the high speed driving regime, which is not the strength point of the range
extender application.
The proposed study used two injections per cycle for a smooth pressure increase
gradient and a better mixture formation. The application requirements do not demand
for the maximum attainable cylinder pressure, but a complete burning and high
combustion efficiency. The secondary injection increases moderately the pressure in the
cylinder and the burning is quiet. Here, the monitored lambda parameter can offer an
important information on how much the oxidation must be carried out prior after
treatment.
An obvious advantage of using the two injections per cycle is the quiet running of
the engine, also a requirement for range extenders.
The 75 % of the load seems to be suitable for this application as a high load applied
by the generator, offering a good torque output and consumption and emissions.
The simulation models offer the possibility of simulating two consecutive injections
per cycle and automatic re-meshing on the fluid domain inside a custom combustion
chamber.
Future work will continue on detailed mixture formation analysis through CFD.
The investigation pursues the influence of injection modeling on engine performances
and emission estimation. Also, the effect of EGR on the pressure increase gradient and
on the resulted lambda parameter is to be studied.
References
1. Dogariu, D.M., Chiru, A., Leahu, C.I., Lazar, M.: Experimental research on the hybridization
of a road vehicle. RoJAE 21(1), 15–20 (2015)
2. Mollenhauer, K., Tschöke, H.: Handbook of Diesel Engines. Springer, Berlin (2010)
3. Hodkinson, R., Fenton, J.: Lightweight Electric/Hybrid Vehicle Design.
Butterworth-Heinemann, UK (2001)
4. Kirsten, K., Golovatai Schmidt, E.: Consistent enhancement of variable valve actuation. In:
Schaeffler AG, Proceedings, Pitești (2013) (Unpublished)
5. Hill, L.: Emissions legislation review. In: Proceedings of the Conference Personalities of the
Automotive Industry, Brașov (2013) (Unpublished)
6. Baumgarten, C.: Mixture Formation in Internal Combustion Engines. Springer, Berlin (2006)
7. AVL Product Description, Single Cylinder Research Engines (2014). www.avl.com/
documents
8. AVL FIRE documentation
9. Mahle range extender engine and vehicle. www.mahle-powertrain.com/
Comparative Study on the Performances
of the One-Way Hydraulic Valves
Marius Lazăr(&), Anghel Chiru, and Dan Mihai Dogariu

{marius.lazar,achiru}@unitbv.ro,
dan.dogariu@gmail.com
Abstract. The purpose of use of the one-way valves is to achieve control

actuations or pressure equalizing in different types of circuits. The basic design
solutions for these valves include ball check valves, spring valves or reed valves.
In order to assure the highest response time of a valve, considering the oil
contamination and the degradation of some physical and chemical properties of
the hydraulic fluid, some requirements must be respected regarding changes in
flow sections, fillets, surface roughness, fluid velocity and response time.
Considering the above restrictions, the paper presents several technical solutions
used for configuring one-way valves, as well as a deep CFD study of some
commonly used one-way valves. The proposed simulations are based on com-
mon commercial designs, investigated by experimental procedures.
Keywords: Check valves CFD simulations Fluid structure interaction (FSI)
1 Introduction
One-way hydraulic valves can be found in many hydraulic control systems, such as
variable valve distribution systems for internal combustion engines or command sys-
tems for automotive automatic gearboxes. These valves can also be found in the
structure of hydraulic circuits for special destination vehicles. The purpose for using
one-way valves is for achieving different command actuation or pressure equalizing in
different types of fluid system components.
Stringent emission legislation and smaller fuel consumption have defined the
optimization of internal combustion engines as the main task for the automotive
industry in the future, since they will continue to play a major role in vehicle propulsion
[1, 6]. Modern control solutions of the internal combustion engine include variable
valve actuation (VVA) and other gas cycle related improvements by combining
charging strategies with more than one turbocharger [5].
Electromagnetic switching valves for connecting different hydraulic circuits are the
main functional components in assemblies which must assure different types of smart
actuation, such as VVA. These assemblies also require a lot of one-way valves through
their complex hydraulic circuits. One example of an actuator used in order to obtain
variable valve timing (VVT) and also lift can be seen in Fig. 1 [2].

204 M. Lazăr et al.
Fig. 1. Hydraulic actuator for variable camshaft timing (left side) and combination for variable
valve lift (right side) [2]
Methods for achieving the even more exigent emission legislations in the future
also consider a higher degree of electrification of the vital systems of the thermal
engine, up to a complete hybridization of the propulsion system [1]. The use of high
speed control systems in combination with already existing hydraulic circuits requires
the best performances from the hydraulics, meaning fluid properties, flow sections and
valves.
Consequently, this paper wishes to propose a study on some constructive solutions
for one-way valves, concluding with a classification of these valves depending on the
performances, with a special attention drawn on obtaining a smooth flow through the
valves with minimum resistance and turbulence, thus avoiding the possibility of liquid
aeration, and excessive pressure drop. Overheating can also occur, with negative effects
on the life time of the working fluid.
2 State of the Art and Application Requirements
The constructive solution for the one-way valves varies from ball check valves to
spring valves or reed valves.
In order to assure the highest response time of the valves, considering the presence
of contaminants and the degradation of the physical and chemical properties of the
hydraulic fluid, some requirements must be respected regarding changes in flow sec-
tions, fillets, surface roughness, fluid velocity and reaction time.
First of all, the response time of the valve must be very short and with little or no
overshoot at all. The settling time should be short also. The influencing factors are both
from the design of the flow section but also from the fluid properties.
Generally, in order to obtain very fast reaction time, the masses of the moving parts
should be as small as possible. However, in order to maintain the sealing function,
sometimes it is better to increase the force of the spring of the valve (spring preload).
Performances of the One-Way Hydraulic Valves 205
Figure 2 shows a cross-section through a small size oil pump using check valves to
assure flow in one direction only [4]. Its main advantage is that it can be used with a lot
of fluids and a large range of temperatures.
Fig. 2. Cross-sectional view through an electromagnetic pump which uses check valves [4]
The proposed study follows three types of one-way hydraulic valves.
3 CFD Simulation Models for the Proposed One-Way Valves
The main simulation model created for studying the proposed valves considers
incompressible laminar flow and a Newtonian fluid, having the properties of engine oil
[3].
Because of the displacement of the ball, valve and deformation of the reed valve,
adaptive mesh refinement was added to the simulation.
3.1 Simulation Methodology

For these two types of spring valves, a first simulation is carried on, considering the
final position of the moving component, the ball, respectively the poppet. This final
position corresponds to the full opening position of the valve. The flow is a steady one,
and the flow performances are evaluated during this step. In this situation, fully
developed flow gives information on the amount of fluid velocity and vorticity.
Fig. 3. The simulation setup for the ball check valve, at left and spring valve, at right
The next simulation considers the intermediate positions of the ball and, respec-
tively, poppet. These two elements are being pushed by the fluid towards the open
position and the spring force counteracts for this displacement. Due to fluid effects, the
ball and piston may have slight returns before they reach the maximum stroke. Figure 3
shows the initial conditions and boundary conditions for the first two cases of study.
For the reed valve, the fluid forces are acting on the metal sheet, displacing it so that
fluid flows pass through it. The reverse flow is stopped by its sealing capacity. Figure 4
shows the simulation setup for the reed valve.
Fig. 4. The simulation setup for the reed valve
3.2 Ball Check Valve and Spring Valve

Figure 5 presents comparatively the evolution of the flow velocities and pressures in
the mid-section of flow for the two simulated valves during maximum valve opening.
Performances of the One-Way Hydraulic Valves 207
Fig. 5. Fluid flow and pressures for the two simulated valves
Figure 6 shows the automatic mesh refinement with the ball in different positions as
it is moved by the fluid forces and balanced by the spring force.
Fig. 6. Adaptive mesh refinement at three different time steps during the simulation
3.3 Reed Valve

Figure 7 shows two different geometrical models of reed valve. The left model assures
a better sealing capability while the one from the right is more sensible during opening.
The high deflection of the first reed valve means higher energy consumption for valve
opening.
Fig. 7. Flow through a reed valve for two geometry versions
4 Conclusions
For a very fast reaction time the masses of the moving parts should be as small as
possible, keeping account on the sealing function and preload of the spring force.
The proposed study on two piston type one-way valves presents the difference
between the ball check valve and spring valve. The ball valve provides smaller local
flow velocities, thus a smoother flow. Moreover, the mass of the ball together with
spring is smaller as the masses of the spring valves, due to its simple construction,
which is another advantage of this type of valve.
Careful attention must be taken when designing reed valves, since too much sealing
capability means high obstruction of the fluid section and the deformation due to fluid
flow is higher. This also means higher energy consumption, head losses, turbulence
generation and noise. Also, the fatigue life of the valve must be considered at different
overloading cases. The low obstruction model presented at 3.1 works at small defor-
mations of the metal sheet and is more energy efficient.
Future work will include the influence of different fluid properties and different
materials for reed valves.
Acknowledgements. This paper was supported by SC Schaeffler România SRL.
References
1. Dogariu, D.M., Chiru, A., Leahu, C.I., Lazar, M.: Experimental research on the hybridization
of a road vehicle. RoJAE 21(1), 15–20 (2015)
2. www.schaeffler.de. Media gallery
3. Nakayama, Y., Boucher, R.F.: Introduction to Fluid Mechanics. Butterworth-Heinemann,
Oxford (1999)
4. www.eckerle.com. Product specifications – electromagnetic pumps
5. Kirsten, K., Golovatai Schmidt, E.: Consistent enhancement of variable valve actuation. In:
Proceedings, unpublished. Schaeffler AG, Pitești (2013)
6. Hill, L.: Emissions legislation review. In: Proceedings of the Conference Personalities of the
Automotive Industry, unpublished, Brașov (2013)
Estimation of Boiling Points of Brake Fluids
István Barabás(&), Adrian Todoruț, and Nicolae Cordoș

{istvan.barabas,Adrian.Todorut,
Nicolae.Cordos}@auto.utcluj.ro
Abstract. One of the most important properties of brake fluids is their boiling
point. A too low boiling point may permit the forming of vapor during frequent
and/or severe braking. Hereby the brake fluid becomes compressible and the
braking performance decreases.
There are standardized methods to measure the boiling point of brake fluids
(SAE J1703, ASTM D 1120), but they are inadequate for every-day practice. An
alternative method is measuring the water content of brake fluid and estimating
its boiling point.
In this work the electric conductibility, the permittivity and the density had
been correlated with the equilibrium reflux boiling points of DOT-3 and DOT-4
types of brake fluids which had controlled water content and had estimated the
water content of brake fluids from more than 70 motor vehicles. It was found
that the capacitive method provides more accurate results, since the conductive
method is strongly influenced by the conductivity of water.
Keywords: Brake fluid Water content Boiling point Conductivity

Permittivity Density
1 Introduction
The brake fluid has important roles in hydraulic braking systems: it transmits the force
developed by the driver to the brake pads, lubricates and seals the hydraulic system and
cools the high temperature parts. In order to transmit the pedal force to the brake pad
efficiently, the brake fluid has to be incompressible. For this reason, the boiling point of
brake fluid must be high enough not to boil when its temperature rises because of
severe and/or repeated braking [1].
Most of the brake fluids are glycol based, the most important constituent is the
hygroscopic ethylene glycol [4].
Currently three methods are used for estimation of brake fluid boiling point:
physical (direct measuring of the boiling point of a sample), electrical (specific con-
ductivity, permittivity) and optical (refractometry, spectrometry) [1].
The first method is standardized and measures the brake fluid boiling point directly
by boiling the brake fluid under controlled conditions. This method cannot be used in
everyday practice because of its complexity and the large amount of sample (it uses a
60 mL sample that can represent the whole content of brake fluid reservoir). Moreover,
the heated sample cannot be used anymore because of chemical reactions which come
off at elevated temperatures.
210 I. Barabás et al.
The electrical and optical methods are non-destructive and indicate the boiling
point indirectly. They use the dependence between the water content of brake fluid and
its other properties (electrical conductivity, permittivity, light absorption or refractive
index).
The main goals of this research were to compare the conductive and capacitive
methods in estimating of the boiling point of glycol based brake fluids and to evaluate
the possibility of using the density in order to estimate the water content and thus the
boiling point.
2 Experimental Section
2.1 Materials and Methods

In this research two types of brake fluids were used which were the same make: a DOT
3 (class 3) and a DOT 4 (class 6) and tap water. Using these liquids, 8 blends were
made with 2, 4, 6, 8, 16, 32, 50 and 75 % v/v water content. After dosing volumet-
rically, the blends were homogenized with a magnetic stirrer for 20 min. For removing
air bubbles, before the measurements the samples have been kept in sealed dark glass
vessels for three days.
Every measurement was repeated three times except for equilibrium reflux boiling
points which were read four times after the equilibrium was reached and were corrected
taking account of barometric pressure.
The capacitive and the inductive measurements were carried out at room
temperature.
2.2 Used Apparatus

For conductivity measurement a HI 9033 type conductometer was used with 0.01
µS/cm resolution and ±0.2 µS/cm accuracy in conductivity domain of 0-199.9 µS/cm.
Instead of permittivity the electrical capacitance of a cylindrical capacitor immersed
in the sample was measured. For this, a cylindrical capacitor was designed with a
measuring electric circuit showed in Figs. 1 and 2. The capacitive sensor consisted of
three parts: one measurement capacitor, Cm, and two compensator capacitors, Cc, used
to correct the electric field deformation at the ends of the middle capacitor. For
Fig. 1. The structure of the capacitor used in measurements

Estimation of Boiling Points of Brake Fluids 211
Fig. 2. Experimental setup for capacitance measuring
measuring the capacitance value an AD 7747 type 24-bit capacitance-to-digital con-

verter was used. The digital signals were processed by a microcontroller and the results
were shown on a display. This measurement system could measure capacitance up to
17 pF and had a repeatability of 0.2 pF.
The density of components and their blends were measured at atmospheric pressure
respecting the method described in EN ISO 12185, using an SVM 3000 type vis-
cometer, equipped with a vibrating U-tube densimeter and Peltier element to adjust the
temperature of the samples. The uncertainty of density was 0.0003 g/mL and of
temperature 0.03 K.
For direct measurement of equilibrium reflux boiling point (ERBP) the experi-
mental setup and the method described in 49 CFR 571.116, ASTM D1120 and ISO
4925 were used. The four read temperatures agreed with 2°C.
As it could be expected, the boiling point of samples decreased when their water
content increased (Fig. 3). This dependence can be described by a second order
polynomial function:
ERBPðvÞ ¼ a þ b w þ c w2 ; ð1Þ
where w is the water content of the samples in % m/m, a, b and c are constant whose
values were determined by polynomial regression based on experimental measure-
ments in case of pure brake fluid and the blends with 2, 4 and 6 % 75 % v/v water
content (Table 1).
It is important to evaluate the lowest boiling point of the brake fuel at which it can
be used securely. On the one hand it is regulated by specific standards but on other
hand it is resulted from the working conditions. Standards specify dry (without any
water) and wet (with 3.7 % (m/m) water) boiling points for each type of brake liquids
[6–8]. The wet boiling point can be considered a limit. Other sources consider that in
order to avoid vapor generation the boiling point of brake fluid has to be higher than
175–180°C [2]. There are opinions that the water content must be limited considering
Fig. 3. Variation of ERBP of brake fluid depending on its water content
Table 1. The constants of Eq. (1)

Type of brake fluid a b c Correlation coefficient Standard deviation
DOT-3 252.388 −35.099 2.750 0.999 1130
DOT-4 293.142 −37.215 2.561 1 0.243
the increasing corrosive behavior of brake fluid as water content increases. Thus, they
established that up to 1.5 % water the brake fluid is in good condition, between 1.5 and
3 % the quality of brake fluid is acceptable but has to be changed soon and over 3 %
the brake fluid must be changed urgently [3].
In the present work a boiling point of 175°C was considered as an alert level, when
the change of brake fluid is recommended, and the wet boiling point as the limit of
usage of brake fluid because of too low boiling point.
Results regarding the capacitance and conductance measurement are shown in
Fig. 4. It can be observed that both quantities increase almost linearly as the water
content of brake fluid increases. Values measured in case of DOT-3 are higher than
those measured in case of DOT-4.
In order to make these methods proper to evaluate the water content of brake fluids,
relations between the measured capacitance and water content, and conductivity and
water content were established (Fig. 5):
ERBPðxÞ ¼ A þ B ex=C ; ð2Þ
where x is the conductivity or the capacitance, A, B and C are constants determined by

nonlinear regressions (Table 2).
Fig. 4. Variations of electric capacitance and electric conductibility of brake fluids depending on
its water content at 20°C
Fig. 5. Variation of ERBP of brake fluids depending on their electric conductibility (a) and
measured electric capacitance (b)
Table 2. The constants of Eq. (2)

Methods Type of A B C Correlation Standard
brake coefficient deviation
fluid
Conductive DOT-3 125.330 1244.055 −4.360 0.999 1.843
DOT-4 118.150 569.576 −6.978 0.998 5.553
Capacitive DOT-3 118.888 37598.400 −1.323 0.999 2.213
DOT-4 130.528 14256.346 −1.520 0.999 1.007
At first the density measurements were carried out for the blends with 2, 4 and 6 %
water. After the evaluation of the results an increase in densities was observed when the
content of water increased. In order to identify the water content corresponding to the
maximum density, other five blends of DOT-3 type brake fluid (8, 16, 32, 50 and 75 %
water) were prepared and their densities were measured at four different temperatures.
The variations of densities of studied blends with their water content at four different
temperatures are shown in Fig. 6. As it can be observed, the densities of brake fluid
increase up to 25–30 % water content and then they decrease. This can be explained by
the negative excess volume of the blend [5], because of the water molecules are
incorporated into the glycol molecules due to ionic interactions between ions in a polar
water blend with glycols [4]. It means, if the initial value of density is known, the
current density of the brake fluid could be a good parameter to appreciate its water
content and thus its boiling point.
Fig. 6. Variation of brake fluid density depending on its water content
After the calibrations of these methods, the quality of DOT-3 and DOT-4 types of
brake fluids was checked in reservoirs of 74 passenger cars. A commercial, conductivity
type, low-cost brake fluid tester was used as well. This tester indicated three levels of
quality of brake fluid based on its conductivity (water content): good condition – green
light, alert, must be changed soon – yellow light and must be changed urgently – red
light. This research was fulfilled in eight car services. In the most cases, neither the brand
nor the age of brake fluids were known. The measurement procedure was as follows:
using the commercial tester the brake fluid had been evaluated and the type of brake fluid
and its quality level indicated by the tester were noted. After that the brake fluid was
evaluated by measuring its conductivity and the capacity of the capacitance immersed
into the brake fluid. Based on Fig. 5 the brake fluid quality was evaluated again.
The boiling point of 175°C was used as the alert level. The brake fluid was con-
sidered as being in acceptable condition if its boiling point was between this temperature
and its wet boiling point, but it had to be changed soon. The brake fluid was considered
in good condition if its boiling point was above 175°C. When the boiling point was
below its wet boiling point, the brake fluid needed to be changed urgently.
The results are presented in Fig. 7. It can be observed that in the case of 18 % of the
cars the change of the brake fluid was urgent and in just 14 % of the cases the quality of
brake fluid was good.
Fig. 7. Result of quality evaluation of brake fluids
The most changes were indicated by the commercial tester, followed by the con-
ductive and the capacitive methods.
The differences in the results can be explained by the large diversity of tested brake
fluids. Moreover, because of the fact that the composition of the brake fluids is not
standardized, their initial boiling point and water content were not the same.
4 Conclusions
From the point of view of the safety, the aptness of the brake fluid is determined by its
boiling point, its water content is a secondary quantity that can be used to estimate the
boiling point.
The results delivered by the conductive method, used in the majority of low-cost
commercial testers, are strongly dependent by the electric conductivity of water
accumulated in the brake fluid. Consequently, the results provided by this method
should be interpreted carefully.
The permittivity of water is eightfold higher than the permittivity of glycols and it is
not strongly influenced by the salt content of the water, the capacitive method can
deliver more accurate results than the conductive one.
The water content of brake fluid can be estimated by measuring its density if its
initial density is known. In order to clarify the dependence between the variation of the
brake fluid density and its boiling point further research is needed, including the most
representative brake fluid types and brands.
References
1. Chuantong, W., Katsunori, S.: A new method for on-line monitoring of brake fluid condition
using an enclosed reference probe. Meas. Sci. Technol. 18(11), 3625–3635 (2007).
IOPscience
2. Grant, M.: Brake Fluid for Classic British Cars. http://www.mossmotors.com
3. Kertay, N.: Féktelenül. Az Autótechnika fékfolyadéktesztje (Brakeless. The brake fluid test
carried out by the ‘Autótechnika’ journal). Autótechnika 8, 16–17 (2010)
4. Kao, M.J., Tien, D.C., Jwo, C.S., Tsung, T.T.: The study of hydrophilic characteristics of
ethylene glycol. In: 7th International Symposium on Measurement Technology and Intelligent
Instruments, pp. 442–445 (2005). Institute of Physics Publishing Journal of Physics:
Conference Series 13, Huddersfield
5. Zhang, J.B., Zhang, P.Y., Ma, K., Han, F., Chen, G.H., Wei, X.H.: Hydrogen bonding
interactions between ethylene glycol and water: density, excess molar volume, and spectral
study. Sci. China Ser B: Chem. 51(5), 420–426 (2008)
6. CFR 571.116 Standard No. 116 - Motor vehicle brake fluids
7. ISO 4925: Road vehicles - Specification of non-petroleum-based brake fluids for hydraulic
systems
8. SAE International J1703: Motor Vehicle Brake Fluid
Modelling and Optimization of Constructive
Form for GO-NOGO Exterior and Interior
Diameter Gauges
Iulian Alexandru Orzan(&) and Constantin Buzatu

{orzan.iulian,cobuzatu}@unitbv.ro
Abstract. Product quality verification is the operation integrated in the process

which feeds back in production if product standards, form, dimensions, color
etc. are according to the design. Because is non added value process people have
always tried to reduce the costs by finding optimized ways for verifying the
product, but sometimes an optimized device used in a wrong environment can
increase the costs by rejecting a number of good products judged bad or vice
versa.
Starting from standardized constructive form of the GO-NOGO gauges, the
paper aims to modulate with the finite element method the dependence between
thermal deformations of the exterior and interior diameter gauges active ele-
ments, form and dimensions in order to minimize as much as possible the
temperature influence over product quality control.
Keywords: GO-NOGO gauge Finite element Product quality control

Thermal deformations
1 Introduction
In every type of human activity it can be found a knowledge component whose main
purpose is to reflect almost perfectly the real technological processes through various
laws of progress, which then gives the possibility to improve in conformity with
desired and possible performances.
World products are always accompanied by measurement and control activities of
various properties and characteristics which determine the product quality.
Product quality verification by control is not directly adding value to the product,
this is done in order to approve the manufacturing process or the conformity of a
specific characteristic imposed by the customer of manufacturer of the product. On the
other way product quality verification results can have a direct influence over the
managers, technological, constructive, conceptual, executional and maintenance deci-
sions, verification being a reaction buckle (feedback) in every technological process for
the product quality assurance.
From economical point of view, quality verification is a non-productive time, which
increases the cost of manufacturing, for this reason this process must be limited to an
absolute necessarily in each process step of the product manufacturing.

218 I.A. Orzan and C. Buzatu
That’s way a major influence in selection of quality control devices should be the
environment where is operating, as most production facilities have now a production
floor and a verification floor, where the temperature can be different depending on each
product type.
2 Actual Principles for Selection of Methods and Instruments

for Control of Product Quality
Rational and efficient selection of methods and instruments for control of product
quality depends on a series of factors, conditions, metrological and technical-economic
characteristics, which can be found in the specialized literature and the most important
ones are listed below [1]:
1. The nature and character of the measured and controlled parameters [5];
2. Tolerance (dimensional, geometrical, micro-geometrical) of verified parameters [1];
3. Production volume (or the production batches: unique, serial, mass production) [2];
4. Error limit of the measurement/control instrument or method [1];
5. Safety of the measurement/control instrument [6];
6. Measurement/Control operation productivity [6];
7. Measurement/Control operation costs [3, 7];
8. Measurement method/instrument complexity and necessary operator qualifications
[4];
Taking in consideration all 5 M’s (man, machine, method, material and medium)
we can determine that control devices are not analyzed from environment point of view
when the need of a quality verification operation is imposed by the project. The most
common situation in production areas is when the product is verified for quality
assurance in different stages and places in the production floor and a final verification in
the quality/verification floor.
3 Finite Element Modelling for Interior and Exterior

Diameter Gauges
The modelling for the exterior and interior diameter gauges took in consideration a real
production situation, where the product is verified in different stages and different
locations during process. The modelled situation refers to a machining process, where
the product is controlled after the turning process, at the temperature of 37°C (un-
controlled environment conditions), then moves to a second operation, removing burr
for example, and the final control operated at a temperature of 20 ± 2°C (controlled
environment conditions).
For the simulation the gauge material used is RP3, which has the chemical com-
position listed in Table 1 and international equivalence in Table 2.
Modelling and Optimization of Constructive Form for GO-NOGO 219
Table 1. Brand and chemical compozition for high-speed steal [8]

Steal brand Chemical composition [%]
C W Cr V Mo Co
Rp1 0.90÷1.00 9.00÷10.00 3.8÷4.4 2.3÷2.7 ≤1.00 5.0÷6.0
Rp2 0.75÷0.83 17.5÷18.5 3.8÷4.5 1.4÷1.7 0.5÷0.08 4.5÷5.0
Rp3 0.7÷0.78 17.5÷18.5 3.8÷4.5 1÷1.2 ≤0.6 ≤0.6
Rp4 1.17÷1.27 6.0÷6.7 3.8÷4.5 2.7÷3.2 4.7÷5.2 –
Rp5 0.84÷0.94 6.0÷6.7 3.8÷4.5 1.7÷2.0 4.7÷5.2 ≤0.6
Rp9 0.95÷1.03 2.7÷3.0 3.8÷4.5 2.2÷2.5 2.5÷2.8 ≤0.6
Rp10 0.78÷0.86 1.5÷2.0 3.8÷4.2 1.0÷1.3 8.0÷9.2 ≤0.6
Rp11 0.97-1.07 1.5÷2.0 3.8÷4.2 1.8÷2.2 9.0÷9.2 ≤0.6
Table 2. High-speed steal symbols and international equivalance [8]

SR JIS GOST Werkstoff Bohler AISI/SAE UNE
Rp1 – – – – – –
Rp2 SKH3 – W.1.3255 – T4 F.5530
18-1-1-5
Rp3 SKH2 R18 W.1.3255 S200 T1 F.5520
18-0-1
Rp4 SKH52 – W.1.3344 S607 M3 F.5605
SKH53 6-5-2
Rp5 SKH51 (R6AM5) W.1.3343 S600 M2 F.5604
R6M5 6-5-2
Rp9 – – W.1.3333 – – –
Rp10 – – W.1.3346 S401 H41M1 –
Rp11 – – W.1.3348 S400 M7 F.5607
2-9-2
Rp10 Sp SKH59 – W.1.3247 S500 M42 F.5617
2-10-1-8
Catia V5 simulation tool has the next properties:

• Young Modulus = 2e11 N/m2;
• Poisson Ratio = 0.266;
• Density = 7860 kg/m3;
• Thermal expansion = 1.17e−5 K−1;
• Yield Strength = 2.5e8 N/m2;
For simulation purposes, the exterior and interior diameters were simulated for 7
specific dimensions intervals, from which the most relevant diameter was calculated
with relation (1), where Dmin is the minimal diameter from the group and Dmax is the
maximal one. The intervals and the calculation results are presented in Table 3.
Table 3. Diameter dimensions intervals

Diameter [mm] Min [mm] Max [mm]
7.745 6 10
23.237 18 30
44.721 40 50
109.544 100 120
129.614 120 140
149.666 140 160
169.705 160 180
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
D¼ Dmin Dmax ð1Þ
Modelling of the representative device is presented in Fig. 1 and for all the
remaining diameters the piece was modified with specific GO-NOGO gauge geomet-
rical parameters.
Fig. 1. GO-NOGO gauges simulated in Catia V5. a. interior diameter GO-NOGO Gauge b.
exterior diameter GO-NOGO gauge.
Simulation result, with Gauss R6 method, Clamp fixation at a temperature of

T1 = 310.15°K (37°C) considering the gauge at a temperature of T0 = 20°C and
translational displacement presented in Fig. 3, done on the OX axe as this is the method
of control, is presented in Fig. 2. Details about the GO and NOGO area of the shaft
gauge can be found in the Fig. 4.
The results obtained by modelling and simulating both gauges with the finite
element method are presented in Table 4 and plotted in graphic representation in
Fig. 5.
Fig. 2. Exterior and interior gauge finite element simulation

Fig. 3. Translational displacement on OX axe for the exterior diameter gauge
Finite element results showed a thermal expansion from 0.002 mm – 0.034 mm

which can sometimes, depending on the tolerance of the product, make the difference
between a good or bad judged piece. Most important is that the example used a turning
process as the manufacturing operation and it’s easy to understand that if the gauge is
used by hand and in higher temperature environment (most of machining environments
are hot due to the processes that generates heat) this kind of expansion will first give a
wrong input to the operator that has to setup the machine and the product will deviate
from the mean value (nominal tolerance) and second the production batch will have a
percentage scrapped when the products will be final inspected at T = 20°C.
Fig. 4. Details about the modelling of the GO and NOGO area of the shaft gauge
Table 4. Finite element simulation results for GO NOGO exterior and interior diameter gauge
Nrcrt D [mm] t [°C] Displacement Displacement Displacement Thermal
Magnitude Magnitude Left Magnitude Left expansion
Right (Bottom) Surface (Bottom) Surface [mm]
(Top) Surface “GO” [mm] “NOGO” [mm]
[mm]
1 7.745 37 0.001 0.001 0.001 0.002
2 23.237 37 0.002 0.002 0.002 0.005
3 44.721 37 0.004 0.004 0.004 0.009
4 109.544 37 0.011 0.011 0.011 0.022
5 129.614 37 0.013 0.013 0.013 0.026
6 149.666 37 0.015 0.015 0.015 0.030
7 169.705 37 0.017 0.017 0.017 0.034
Fig. 5. Thermal expansion for the GO-NOGO interior and exterior diameter gauge (with
trendline)
4 Constructive Form Optimization for Exterior Diameter

Gauge in Order to Reduce the Thermal Expansion
The finite element simulation above is the input that a need of optimization in the
method or device must be implemented in order to reduce the thermal impact over the
product.
For these reasons the constructive form of the shaft gauge was modified as:
• Optimization done by increasing the thickness, from 10 mm in the initial simulation
to 12 mm;
• Optimization done by removing material, trying to simulate a radiator;
• Optimization done by milling parts from the shaft gauge;
Fig. 6. Finite element simulation for the case of shaft gauge increased thickness
Fig. 7. Finite element simulation for the removing material case

Fig. 8. Finite element simulation for the milling parts from the shaft gauge case
For all these 3 cases the finite element simulation results are presented below in
Figs. 6, 7 and 8.
All 3 case presented below did not improve the thermal expansion to a level that
can influence, Δexpansion = −0.0001 mm. Resulting in the fact that the shaft gauge
functional design can’t be improved in order to reduce the thermal expansion and the
attention should move to a change in method, by adding an ABS or Bakelite coating to
reduce the heat exchange between human hand and the shaft gauge.
Interior diameter form optimization was not performed due to the results from the
exterior diameter gauge which show no high improvement in the thermal expansion.
5 Conclusions
Taking in consideration all presented in this paper, it can come off next conclusions and
recommendations:
(a) Depending on the diameter used the expansion due to higher temperature can have
values between 0.002–0.034 mm, which can sometimes make the difference
between producing a good part or a bad part, by giving wrong input to the setup
operator.
(b) Thermal expansion for control devices used in production floors with machines
that generate heat and without a method to protect the heat exchange between
operators hand and the device, will have a higher probability to increase wrong
machine setups and scrap levels.
(c) Where possible and where the cost of production allows it, control gauges can be
replaced with more complex measurement devices that can be protected against
the heat exchange between the environment and operators hands.
(d) High precision products are much more affected by temperature expansion then
the normal ones and by calculating a breakdown between products, high precision
vs lower precision, can raise a business case to improve devices, methods and
even the production environment.
(e) The cost for the equipment and method used for product quality verification must
be justified and balanced by the precision and productivity of them.
(f) The device and method complexity for product quality verification takes in
consideration: productivity, verification cost, but also the qualifications needed by
the quality inspector: a higher complexity could mean a simple method and an
increased productivity which doesn’t necessarily mean a higher qualification for
the quality inspector, instead a higher complexity and method means for the
quality inspector a higher qualification, which is only justified by a very high
productivity.
References
1. Buzatu, C., Lepadatescu, B.: Echipamente si tehnologii moderne de masurare si control a
calitatii produselor (Modern Equipment and Technologies for Product Quality Measurement
and Control). MATRIX ROM, Bucharest (2013)
2. Buzatu, C.: Elemente de proiectare tehnologica si management in fabricatia produselor din
constructia de masini (Elements of Technological Design and Management in the Product
Manufacturing in Machine Building). Matrix-Rom, Bucuresti (2012)
3. Olariu, C.: Costul si calculatia costurilor (Cost and Costs Calcultion). Editura Didactica si
Pedagocica, Bucuresti (1978)
4. Bralla, J.G.: Handbook of Product Design for Manufacturing. McGraw Hill, New York
(1986)
5. Mcmahon, C., Browne, J.: CAD/CAM Principle, Practice and Manufacturing Management.
Addison-Wesley, England (1998)
6. Walken, J.M.: Handbook of Manufacturing Engineering. Marcel Dekker Inc, New York
(1996)
7. Lane, C.: Modular Fixturing Handbook. Car Lane Mfg. Co., St. Louis (1991)
8. Micu, D.: Contribution and researchers regarding the application of thermo-mechanical
treatments on the tools steels and their implication for work safety and health on the operator,
Ph.D. thesis, Brasov, Romania (2010)
Systems Engineering Requires Digital Twins
of Machine Elements
Willi Theodor Gruender(&)
TEDATA GmbH, Bochum, Germany

grue@tedata.de
Abstract. Manufacturing processes are heavily influenced by virtualization and

increasing networking capabilities. The “Internet of Things” spreads across all
industries transforming isolated production cells to members of large commu-
nication, logistic and manufacturing systems. But not only production, also the
preceding engineering processes suffer from rigid procedural workflows, hier-
archically dominated decision-making structures and numerous media breaks. It
is therefore inevitable to include the product development within this new
concept of “Industry 4.0”, linking real and virtual products to the manufacturing
world. The aim is to generate structures and descriptions of products, which
define a new product completely in digital terms, so that it becomes transparent
and understandable in all its facets both in the initial stage of formation as well
as to its future. A digital image or model, also called virtual twin, therefore, must
be digitally accessible in all product life stages to all stakeholders and actors,
who are dealing with the product, so as to maintain service quality and func-
tionality even under changed conditions. The present paper considers the
associated changes in the future product development process.
Keywords: Product development MBSE Model based systems

engineering Virtualization Design process management
1 Introduction
In mechanical engineering recent decades have seen many new design tools and data
management systems. However, the innovations were targeting and supporting rather
the individual workplace, neglecting the aspects of interdisciplinary, of teamwork and
knowledge management. Thus, nowadays development work still reflects individually
generated contributions.
But at the same time, increasing complexity, lifecycle orientation and exploding
possibilities of information technology are leading to new methods and systematics in
product development. This new world of innovation is characterized by two main
trends: virtualization and process orientation. In the near future many authors expect
completely digitized products, represented by models and processes, which describe
the generics, the structure, the operational conditions, and last but not least the end of
life and the recycling phase.
In medicine such digital images are called virtual twins (Fig. 1). A digital image of
a human being has many facets (Fig. 2): genetic structures and data, historic data,

228 W.T. Gruender
Fig. 1. The virtual twin of a human being in 3D [1]
Fig. 2. The virtual twin in sequences and diagrams
actual measurements etc. Like images of human beings, virtual twins in mechanical
engineering provide numerous functional and behavioral views onto objects like
machine elements, assemblies, machines and complete production plants.
Why do we need virtual or digital twins? Because we are hungry to acquire more
and more knowledge about human beings and human life. Moreover, many of us want
to learn about the construction plan to eternal life, and other people want to know how
our body functions and how we keep our bodies and souls fit and healthy.
2 The Virtual Twin of a Machine Element
When we look at the virtual twin of a machine element or an assembly or a machine or

a complete production plant or the whole factory, we not just want to understand the
functionality and behavior of the objects, but we want to make them fully transparent
[2]. The flow of information and energy, which keeps up necessary operations and
influencing the behavior is one issue we must influence. On the other hand we aim at a
long operating life and we must know the limits of loads and many other limiting
conditions. Therefore we must monitor the actual state of the objects.
And last but not least we want to know how the object can be maintained and
renewed in case of wearing and breakdown. For that we need models to collect,
Systems Engineering Requires Digital Twins of Machine Elements 229
Fig. 3. Tools applied in engineering development in 19th, 20th and 21st century
formalize and revise data, information, experience and knowledge (Fig. 3). This has to
be done in all phases of the product lifecycle.
But let us shortly look back to the history of product development.
The base for all product development is education, supported by experience,
standardization and literature. For many centuries this background was the main source
for engineers. Eventually, drawing boards, calculators and formalized tables made the
job more comfortable. And then, only forty years ago the mostly manual work was
pushed to a new era by computer aided design. 3D modelling, FEM and multi body
simulation entered the scene and made highly efficient engineering tools and workflows
available in mechanical engineering.
At the same time, globalization, individualization, social issues and many national
and international regulations have led to a highly complex environment, engineers are
working in (see Fig. 4). And interdisciplinary and integrated cyber physical systems have
recently added another important engineering component to product development [3].
Fig. 4. Complexity of environment engineers are facing today

230 W.T. Gruender
Fig. 5. Product development process as a V-model
To describe the product development process in a multidisciplinary environment

German engineers have created a schematic picture, known as the V-model. Concept,
design and validation are happening in a concurrent process for all participating dis-
ciplines as shown in Fig. 5. In the light of circular economies, however, future pro-
cesses will cover a complete ring as indicated.
3 Model Based Systems Engineering
In the context of virtualization, the Systems Engineering approach, developed by

information technology researchers, provides very helpful assistance. Further refined to
Model Based Systems Engineering (MBSE), the method supplies a framework, which
spans all activities of product development. Starting with the requirements, all func-
tional relations and structures, logical links, algorithms and solution spaces are
assembled in a consistent systems model. SysML is the language to describe the
complete model and to provide interfaces to the partial structures and processes, which
analyses individual aspects and components. Beside integrated multi-phase product
planning, the intended procedure allows the integration of mechanics, electronics and
informatics, due to the fact that the information backbone for all disciplines follows the
same language.
In this way, all activities and team members are able to share previously obtained
results for their analysis and considerations, not only in the planning, but also in the
operational phase of the product lifecycle. The process-oriented component of SysML
on the other hand adapts elements of Business Process Management (BPM) allowing
the design engineer to generate continuous and data based workflows (Fig. 6). As a
result product development procedures are obtained, where digital data streams,
Fig. 6. SysML: the language to describe virtual product models [4]
starting as requirement profiles will form virtual product models, which match the
expected characteristics even before the first prototype has been produced.
Model Based Systems Engineering integrates isolated facets and descriptions of a
product, containing all functional and logical entities and relationships. The high
number of tools and diagrams, which have to be applied, demonstrates the degree of
complexity: Structure, Function, Geometry, Logic, Numeric, Analytics, Simulation of
Product, Manufacturing and Operations, Validation and Documentation. As a result,
we get a gap-free representation of all product characteristics starting with requirements
management up to the validation phase. The high amount of information, which is
produced in a completely virtual environment, requires an information architecture
which cares for an efficient balance between domain owned data and the overall data
back bone.
One of the most important issues in the next future is the mediation between
traditional engineering methods and future approaches. As the complete digital rep-
resentation of a product goes far beyond pure 3D modelling engineering staff requires
many new abilities. Thus, education and training will play a major role in transforming
the product development process to Model Based Systems Engineering. Figure 7
shows the infrastructure, the tools and the content, which will make the complete
environment.
232 W.T. Gruender
Fig. 7. Working environment of future product development in mechanical engineering
4 Conclusions
New technologies like Clouds, Big Data, Cyber-Physical Systems and Additive
Manufacturing will open up many unforeseeable opportunities and will cause disrup-
tion with many technologies. On the other hand, we face an integration of mechanical
and electrical components, driven by information technology, which requires a new
model based approach to create consistent and complete descriptions of new generation
products, systems and services.
Fig. 8. The thoughts of humans – Embedded into virtual systems!

Applying Model Based Systems Engineering methods to product development in

mechanical engineering, we will be able to provide digital representations of product
characteristics, which can be followed up and tested under various assumptions and
conditions at each stage of the development process. Creating such models, the design
engineer acts as a Knowledge Manager, defining the product in a complete digital
manner, including function structures, logic diagrams, 3D models, parts lists, simula-
tion and analysis workflows (Fig. 8).
At the same time, he transforms the core elements of his engineering knowledge
into a new generation of digital objects, which will act as components for digital twins
and as a base for future product and engineering systems modelling. However, suitable
tools and description languages have yet to be developed to step up product modelling
from 3D CAD to Model Based Systems Engineering.
References
1. N.N.: US soldiers could have their bones scanned and 3D printed in case of injury. http://
www.3ders.org/articles/20150216-us-soldiers-could-have-their-bones-scanned-and-3d-
printed-in-case-of-injury.html. Accessed 27 August 2015
2. Abramovici, M.: Produkte bekommen einen lebensbegleitenden virtuellen Zwilling, Eco-
nomic Engineering 4/2014, S 69 f (2014)
3. N.N.: VDI, cyber physical systems: Chancen und Nutzen aus Sicht der Automation: Thesen
und Handlungsfelder. Düsseldorf, März 2015
4. Wikipedia: Systems Modeling Language. https://en.wikipedia.org/wiki/Systems_Mode-ling_
Language. Accessed 23 June 2016
Collaborative Design and Use of Interactive
Simulations: Boost the Learning Environment
in Road Vehicle Dynamics Curriculum
Moein Mehrtash and Dan Centea(&)
Faculty of Engineering, School of Engineering Practice and Technology,

McMaster University, Hamilton, Ontario, Canada
{mehrtam,centeadn}@mcmaster.ca
Abstract. The purpose of this paper is to define a learning environment for

vehicle dynamics and to describe a set of computer simulation-based experi-
ments that have been developed to mimic real and standard road tests in vehicle
performance. The paper describes three experiments that have been developed
based on the learning outcomes of an automotive engineering course related to
vehicle dynamics. These simulation experiments include: (i) modelling the
standard coast-down to analyze the effect of head- and tail-wind and grade, in
determining the aerodynamic and tire coefficients of the vehicle; (ii) braking
performance evaluation based on the FMVSS 105 standard; and (iii) driver-
in-the-loop analysis that analyzes the performance of various drivers in
cornering manoeuvre at different conditions. These simulations are expected
provide students an authentic experience of performing specific driving
manoeuvres, modifying the vehicle parameters, and collecting the data with
minimal concerns about the cost, time, safety, and weather conditions.
Keywords: Interactive simulation Road vehicle dynamics Coast-down test

Brake proportioning Driver-in-the-loop
1 Introduction
Due to the challenges created by global competition in the automotive market, the
automotive industry currently uses modern tools for designing, prototyping, testing,
and manufacturing. Preparing graduates for such a fast evolving industry is a difficult
task if a traditional lecture-intensive approach is used. Alternative approaches include
student-centered experiential learning approaches. Experiential learning is a powerful
approach that allows learners to develop skills and knowledge by personal experience
and involvement. Experience plays a central role in the learning process. The experi-
ential learning theory and model developed in the early 1970s by David Kolb [1] gave
guidelines for designing and developing practical curriculum.
Recent advancements in computing and information technology have had a sig-
nificant influence in the development of a practical curriculum in vehicle dynamics
using an experiential learning environment. Computer simulations of vehicle dynamics
are an authentic experience of performing specific driving manoeuvres, modifying the

Collaborative Design and Use of Interactive Simulations 235
vehicles, and collecting simulation data. In this paper, a set of computer simulation-
based experiments have been developed to mimic real and standard road tests in vehicle
performance. The purpose of these simulations is to enable an interactive activity
without addressing computer programing and unnecessary detailed modelling. Various
commercial software packages have been used to develop simulation experiments that
(i) conceptualize the collaboration of designing the simulation and (ii) use the results
for constructive learning principles.
The course presented in this paper was developed using a series of expected
learning outcomes. Each learning outcome is addressed through lectures and through
hand-on simulations performed by each student. The approach of three hours of lecture
complemented by three hours of hands-on computer-based simulation labs per week
has been shown to provide a superior learning experience.
2 Simulation Studies of Coast-Down Testing
Coast-down is one of the most common tests for motor vehicles and comprises of
vehicle deceleration from a certain speed with the engine ungeared and concurrently
recording the speed and travelled distance until the vehicle stops. One of the main
purposes of this test is to assess the extent of the resistant aerodynamic and tire forces
acting on the vehicle at specific speed and road environments. A hands-on engineering
practice of coast-down testing can be employed for understanding of road vehicle
performance [1]. However, due to the concerns about the time, safety, and weather
conditions, an interactive simulation of the coast-down can be developed to ensure easy
adaptability to the classroom setting and to address the course expected learning out-
comes. This section includes a brief review of the learning outcome associated with this
topic, the background information, and the simulation procedure.
2.1 Learning Outcome: Coast-Down Simulation

The learning outcome of the coast-down simulation is to develop an understanding of
basic vehicle dynamics and losses associated with road loads. Students are expected to
become familiar with coast-down testing procedure and applications, and to gain
experience in data analysis and interpretation.
2.2 Background: Theory of Coasting Vehicle

To obtain the motion equation of the vehicle, some simplifying assumptions and a
simple dynamic model were implemented [2]. While moving, the vehicle interacts with
the environment. As a result, some forces acts against it (Fig. 1): the vehicle gravita-
tional force (the weight) W, tire rolling resistances (Rxf : front axle and Rxr: rear axle),
and aerodynamic drag Da. The resistant road loads, RRL, can be defined as [3].
RRL ¼ Rxf þ Rxr þ Da þ Wsinh ¼ Fr W þ 1=2qV 2 CD A þ Wsinh ð1Þ

236 M. Mehrtash and D. Centea
where fr is the coefficient of tire rolling resistance, ρ is air density, CD is drag coeffi-
cient, A is reference frontal area of vehicle, V is the velocity of vehicle, and θ is the
grade angle. The effect of grade resistance can be removed in the road load if the test
track has small grade. In practical application, coast-down test performs in two
directions on the same test track to remove the effect of grade. Thus, the effect of the
grade can be removed from the Eq. (1) ðh ffi hÞ. Moreover, a linear model of the
coefficient of tire rolling resistance can be used as fr = f0 + f1V. Thus, the deceleration
of the vehicle due to these resistant loads can be calculated as,
RRL
d¼ ¼ ðf0 þ f1 V Þg þ 1 ð2mÞqV 2 CD A ¼ f0 g þ f1 gV þ 1 ð2mÞqCD AV 2 ð2Þ
m
To find tire rolling resistance coefficients (f0 and f1) and aerodynamic drag (CD) the
deceleration of the vehicle with respect to velocity is required to be measured during
the coast-own testing. By fitting a second order polynomial then all resistant coeffi-
cients can be determined.
Fig. 1. Free-body diagram of the vehicle in coast-down test [4]
Fig. 2. Coast-down deceleration against velocity for case (1) using CarSim simulation
2.3 Simulation Methodology and Analysis

A commercial software CarSim [6] is employed to simulate the coast-down test.
CarSim provides a wide range of databases for vehicles, road environments, and
driving events. All these databases can easily be selected and modified by the students.
To investigate the effect of the unknown grade and wind direction, students perform
the coasting simulation from 120 km/h (specified as KPH in the Fig. 2) in five cases:
(1) without a grade; (2) with +0.1 % grade; (3) with a -0.1 % grade; (4) with a 10 km/h
head wind; and (5) with a 10 km/h tail wind.
Figure 2 demonstrates the simulation and fitted polynomial of the case (1). Table 1
demonstrates the full analysis of the five simulation cases. The averaging approach
shown in Table 1 is used to reduce the estimation error of the resistant force coeffi-
cients. Students can observe that the averaging approach is more effective in decreasing
the estimation error of the grade than wind effect.
Table 1. Simulation analysis results for five cases

Method/Parameters f0 f1 Cd Error f0 (%) Error f1 (%) Error Cd (%)
CarSim 0.0038 2.60E-5 0.30 0 0 0
Case (1) estimation 0.0037 2.54E-5 0.29 2.9 2.3 3.2
Avg. case (2) & (3) 0.0037 2.54E-5 0.29 2.9 2.3 3.2
Case (4) estimation 0.0039 −4.55E-5 0.24 2.1 275 20
Case (5) estimation 0.0055 14.9E-5 0.29 44.7 473 3.2
Avg. case (4) & (5) 0.0047 5.2E-5 0.27 23.4 99 11.7
3 Simulation Studies of Brake Proportioning System
Braking performance testing is one of the major performance tests in automotive

industries due to out of public concern for safety. The U.S. National Highway Traffic
Safety Administration publishes performance standard for new vehicles that would
increase safety on the highway. Among the many standard that have been imposed are
Federal Motor Vehicle Safety Standard (FMVSS) 105 [5] establishing braking per-
formance requirements for vehicle with hydraulic brake systems. The FMVSS 105
provides the standard procedures and requirement for the road test of braking system.
When these procedures are performed using simulation software they provide an
excellent learning environment. In this section, the learning outcomes, theory of brake
proportioning, and simulation methodology will be discussed.
3.1 Learning Outcomes: Brake Proportioning Simulation

and FMVSS 105
The learning outcome of the brake proportioning simulation is to develop an under-
standing of brake performance using proportioning. Students are expected to become
familiar with the brake performance test FMVSS 105 procedure and requirements.
3.2 Background: Theory of Brake Proportioning

The FMVSS 105 is quite detailed and complex; it provides the requirements for the
stopping distance performance into five effectiveness tests [4, 5]. In curriculum
development of the vehicle dynamic laboratory, two restricted situations, similar to
effectiveness in FMVSS 105, have been defined based on the stopping distance as
– effectiveness 1, stopping distance less than 56 m from 100 km/h on a dry road
– effectiveness 2, stopping distance less than 40 m from 50 km/h on a wet road.
The primary challenge in brake system design is the selection of proportion ratio to
bring both axle lockups simultaneously despite the variability in surface friction,
weight distribution, CG height, and brake conditions. A practical solution to meet
FMVSS 105 requirements is to incorporate a pressure-proportioning valve in the brake
hydraulic system. This valve provides equal pressure to front and rear axles brakes to a
certain level, and then reduces the rate of pressure increase on the rear [5].

The commercial software CarSim [6] has been used for this lab session to model
vehicle dynamics, brake proportioning rules, and various road surfaces. Students model
the hydraulic brake system first without the proportion valve, and then with propor-
tioning valve. The brake application pressure is increased gradually and the stopping
distance, vehicle speed, and wheel speed are measured. Figure 3 demonstrates the
simulation result of effectiveness 1 without the proportion valve. Figure 3(b) shows
that an equal pressure 4.2 MPa sent to all wheels leads to a stopping distance 65.1 m
that is does not satisfy the effectiveness 1 requirement. Increasing the application
pressure to 4.6 MPa in order to reduce the stopping distance causes the rear axle (wheel
R2 and L2) lockup. The lockup condition can be identified from the wheel speed,
shown in Fig. 4(a). The stopping distance for this simulation is 69.9 m.
Using the brake proportion valve, two different pressures can be sent to the front
and rear axles, as shown in Fig. 5(b). The wheel lockup does not occur and the
stopping distance of 49.5 m is achieved with the application pressure of the 7.2 MPa.
This condition meets the requirement for the FMVSS 105. Similar simulations can be
performed for the effectiveness 2, which addresses the wet surface condition.
Fig. 3. Simulation of effectiveness 1 for case 1, 4.2 MPa application pressure without the
proportion valve: (a) Wheel speed and vehicle speed, and (b) applied brake pressure for each
wheel
Fig. 4. Simulation of effectiveness 1 for case 1, 4.6 MPa application pressure without the
wheel
Fig. 5. Simulation of effectiveness 1 for case 1, 7.2 MPa application pressure with the
wheel
4 Simulation Studies of Driver-in-the-Loop
The interaction of the driver with the vehicle is an important input to the vehicle
dynamics. Thus, the driver behaviour is required to be considered as a part of vehicle
dynamics analysis. The use of a track test is not easily adaptable for classroom teaching
application. On the other hand various models suitable for modelling driver behaviour
are available in automotive simulation software packages like as CarSim, Adams,
MapleSim, etc. This section introduces a basic background for modelling of human
driver.
4.1 Learning Outcomes: Simulation Studies of Driver-in-the-Loop

The learning outcome of the driver-in-the-loop simulation is to develop an under-
standing of driver behaviour in the dynamics of a vehicle. Students are expected to
become familiar with using CarSim and Matlab/Simulink for collaborative modelling.
4.2 Background: Driver-in-the-Loop

There are various strategies for the modelling of the driver behaviour [7] such as path
following, vehicle slip control, two-loop control, etc. To implement driver strategies in
simulation, there are various approaches such as optimal control, logic control, PID
control, Fuzzy logic control, neural network control, and adaptive control. However,
the PID control is most appropriate for modelling of driver behaviour to usefully
influence the vehicle dynamics. In addition, students have deep knowledge of PID
control system and PID gain changings influence from their previous courses in control
theory.
Fig. 6. Modelling of the driver behaviour using Matlab/Simulink and CarSim


CarSim is a professional automotive software that provides robust communication with
the Matlab/Simulink software. One of the most convenient and effective methods of
modelling control system such as PID control system is to use Matlab/Simulink. The
driver behaviour can be modeled as a PID control system using Matlab/Simulink
software, while the vehicle dynamics is implemented on the CarSim software. These
two software packages communicate through the Matlab/Simulink interface or through
the CarSim interface. Figure 6 demonstrates the modelling vehicle and driver in
Matlab/Simulink interface. Students can easily change the PID gains via Matlab/
Simulink that represent the driver behaviour and observe the change in the dynamics of
vehicle.
5 Conclusion
The knowledge that students retain after completing a course and the skills gained
during the course are important elements that help graduates find employment. An
important approach that helps students enhance their knowledge and skills in engi-
neering education is the inclusion of an experiential learning methodology. This paper
presents the implementation of a vehicle dynamics curriculum that includes a signifi-
cant experiential learning component. The theory taught during the lectures is com-
plemented with weekly hands-on simulations that reinforce the theory. Although the
course includes ten simulations, this paper describes only a coast-down simulation, a
brake proportioning simulation, and a driver-in-the-loop simulation. The coast down
simulation allows students to vary the vehicle and environment parameters and obtain
results related to rolling resistance and aerodynamic drag. The brake-proportioning
simulation allows students to vary the pressure applied in the braking systems and to
use or not a brake-proportioning valve and observe the corresponding stopping dis-
tances. The driver-in-the-loop simulations allow students to model and modify various
parameters related to the driver and observe the dynamic response of the vehicle. All
these simulations allow students to retain the expected learning outcomes and to create
a better learning environment.
References
1. Kolb, D.A., Boyatzis, R.E., Mainemelis, C.: Experiential learning theory: previous research
and new directions. In: Perspectives on Thinking, Learning, and Cognitive Styles. Routledge,
Taylor and Francis Group (2011)
2. Cai, Z.T., Worm, J.J., Brennan, D.D.: Experimental Studies in Ground Vehicle Coastdown
Testing. American Society for Engineering Education (2012)
3. Preda, I. Covaciu, D., Ciolan, G.: Coast-down test–theoretical and experimental approach. In:
Proceedings of CONAT 2010, Brasov (2010)
4. Gillespie, T.D.: Fundamentals of Vehicle Dynamics, pp. 168–193. Society of Automotive
Engineering Inc. (1992)
5. Standard No. 105; Hydraulic Brake System. Code for Federal Regulations, Title 49, Part
571.105, pp. 199–215, 1 October 1990
6. Manual, CarSim User. Mechanical Simulation Corporation. Ann Arbor, MI 48013
7. Blundell, M., Damian, H.: The Multibody Systems Approach to Vehicle Dynamics. Elsevier,
Amsterdam (2004)
Experimental Researches
on the Magneto-Rheological Dampers
Response to the Control Parameters
Alexandru Dobre(&), Nicolae Vasiliu,

and Cristian Nicolae Andreescu
University Politehnica of Bucharest, Bucharest, Romania

alexandru.c.dobre@gmail.com
Abstract. The magneto-rheological (MR) dampers used in automotive sus-

pension improve the energy dissipation capacity for the vertical oscillations of
the car body and wheel of the vehicle by increasing the damping force devel-
oped by piston motion. The determining of the time constant of the
magneto-rheological fluid is essential in order to build a realistic dynamic model
of an actuator, needed for implementing a real time control system provided
with a comfort feedback. In this paper the authors compare the information
collected from the dedicated literature, which currently are relatively restricted,
to the results obtained from their own experimental determinations. The final
objective of the paper is to help the optimal design a control structure for the
assembly of the car magneto-rheological dampers.
Keywords: Magneto-rheological damper Experimental researches Control

Current
1 Introduction
The magneto-rheological fluids should have an acceptable low viscosity in the absence
of the magnetic forces, but at the same time, they should be able to develop a high
pressure drop when they are subjected at high values of the magnetic flux density.
Concerning the magneto-rheological dampers control, the most important aspect which
must be taken into account is the response time to the magnetic field change. The
response time depends on: the size and shape of the magnetic particles in suspension in
the fluid, the fluid viscosity, the conductivity of particles etc.
Two significant factors for controlling the damping force of a MR fluid shock
absorber through a damper controller are considered in the paper [1]: tracking ability of
the controlled damping force to the desired damping force, and energy needed for the
MR fluid shock absorber.
In the paper [2] it was studied the behavior of the magneto-rheological shock
absorber with a pneumatic suspension. The results showed that the tested suspension
obtains the same roll angle levels as in a comparable passive suspension while
improving the ride comfort by reducing acceleration by up to 30 %.

244 A. Dobre et al.
A bounded stochastic optimal semi-active control strategy for random excited hys-
teretic systems by using MR damper is proposed and illustrated by an example of a single
degree-of-freedom hysteretic system in paper [3]. The results showed that the control
strategy proposed by the authors has a high control effectiveness and efficiency [3].
The main characteristic of MR fluids is their ability to generate forces that can be
controlled with a magnetic field produced by an electromagnet or a permanent magnet.
Under the influence of an applied magnetic field, the MR particles join together,
forming a chain (Fig. 1a). This chain once formed, induces a reversible yield stress in
the fluid. The yield stress of the MR fluid is continuously and quickly adjustable,
because it responds to the intensity of the applied magnetic field. The magnetic field
applied leads to the formation of chains of particles in the direction of magnetic field
lines (Fig. 1b).
Fig. 1. The magnetorheological fluid microstructure: 1 and 2 – magnetic poles; 3 – ferromagnetic

particle; 4 – MR fluid; A – power source
Considering the fluid mechanics, the behavior of the MR fluid in the absence of a
magnetic field can be described by the Newtonian mechanics, while, in the presence of
the magnetic field, it exhibits the behavior of the Bingham model [4].
If the MR fluid is used as a controllable actuator in different application devices, the
fluid can behave into three different modes, as shown in Fig. 2. The MR fluids can be
also integrated into conventional structures (called smart structures). The complex
mode of intelligent structure can be controlled by adjusting the intensity of the mag-
netic field. The MR fluids must have high yield stress at maximum magnetic fields and
low viscosity in the absence of a magnetic field. Also, the MR effect should be stable
within a wide temperature.
Experimental Researches on the Magneto-Rheological Dampers 245
Fig. 2. The operating modes of the MR fluids
In general, the particles are in spherical shape. Its geometry is preferred for low
magnetic anisotropy, lubricity and durability [5]. The MR shock absorber is con-
strained in a specific volume and the optimization problem identifies the geometric
dimensions of the shock absorber which minimize an objective function. The objective
function refers to: the damping force, the dynamic range, and the inductive time
constant of shock absorber [6].
2 Test Bench Presentation
The static and dynamic performances of the MR shock absorbers are determined for
evaluating the efficiency of the vehicle suspension systems. In the Fig. 3 it is shown a
magneto-rheological damper. The stand is made up of a fast electrohydraulic ser-
vomechanism, provided with position, velocity, amperage and force transducers; the
servomechanism and damper tested are mounted in a rigid frame (Fig. 4).
Fig. 3. The tested magneto-rheological shock absorber (Chevrolet Corvette)

246 A. Dobre et al.
Fig. 4. Partial view of the magneto-rheological shock absorbers test bench
The stand includes state of the art electro-hydraulic components of high precision.
For the servomechanism control and for data acquisition is used an industrial type
electronic computer PXI, equipped with a data acquisition board produced by the
National Instruments Corporation. The excitation of the shock absorber is made with a
position signal, similarly to the real situation of operation. The input signal may be
sinusoidal, triangular or compound. The output is the damping force developed by the
shock absorber for different velocities.
3 Experimental Results
The dynamic behavior of the magneto-rheological damper shown in Fig. 4 was ana-
lyzed at different frequencies and different values of the electrical current intensity. The
input signals studied were: sine, triangular and rectangular. In this paper are presented
some results obtained from the damper test bench. In the Figs. 5, 6 and 7 are presented
the types of input signals depending on the time, as well as the displacement of the
magneto-rheological shock absorber. The period of the sine signal is 2 s and for
triangular and rectangular signals is around 3.33 s. Generally, the road profile is
described by a sinusoidal signal.
The results obtained from the experimental researches show a significantly
improved comfort when using a magneto-rheological damper compared to a classic
Fig. 5. The variation in time of the displacement and input signal for the sine input signal
Fig. 6. The variation in time of the displacement and input signal for the triangular input signal
Fig. 7. The variation in time of the displacement and input signal for the rectangular input signal
248 A. Dobre et al.
one, which is due to the fact that the MR shock absorber allows the changing of the
damping ratio depending on road conditions, thus maintaining the permanent contact
between the tire and the road as a consequence of the increased damping force.
In the Fig. 8 it is presented the velocity of the magneto-rheological piston
depending on the time, for different values of the current intensity. The period of the
sine signal is 2 s. The maximum velocity of the piston is 60 mm/s, practically constant
for different values of the current intensity.
Fig. 8. The variation in time of the piston velocity for different values of the current intensity for
sine input signal
Fig. 9. The damping characteristic F(v), for a magneto rheological and a classical damper
In Fig. 9 it is presented the damping characteristic for two dampers (classical and
MR), in coordinate force-velocity, for a sine signal input.
The maximum damping force for a classical damper is around 500 N while for
magneto-rheological damper it is 1 kN, for a current intensity of 1 A. It is observed that
the maximum damping force for the MR damper is two times higher than the force
developed by the classic shock absorber. The period of the sine signal for both dampers
is 2 s.
For a magnetizing system with four coils, the damping ratio could be increased by
up to three times for an excitation current of only 2 A. Such current could be reduced to
less than 1 A if the magnetizing system uses eight small cores. The fuzzy intelligent
controller reduces effectively the vertical vibration and improves the ride comfort and
handling stability of the vehicle [7, 8]. The magneto-rheological shock absorbers have a
semi-controllable damping force output which is dependent on the electric current
intensity applied to the shock absorber, as well as the relative velocity. The MR
dampers can be used also for motorcycles; high controllability and low power con-
sumption are two vital considerations for motorcycles [9, 10].
4 Conclusions
The maximum damping force developed by the MR damper increased considerably

compared to the classical damper for a velocity of the piston of 0.06 m/s. The hys-
teresis loop increased for the MR shock absorber. The MR damper is used specially for
sport and luxury cars, because the magneto-rheological fluid is expensive, but it offers
the possibility of adjusting the damping ratio to improve the comfort and the car
stability. The MR shock absorbers are used in semi-active suspensions. The input
signal type can cause a brutal functioning of the damper. The rectangular signal input
produces a hard functioning compared to the other two, like sine and triangular.
The design of the suspension controller using MR shock absorbers has to take into
account both typical road profiles and the dynamic behavior of the used shock
absorbers established by the experimental test bench containing all the components of
the suspension.
References
1. Wang, D.H., Liao, W.H.: Semi-active controllers for magnetorheological fluid dampers.
J. Intell. Mater. Syst. Struct. 16(11–12), 983–993 (2005). doi:10.1177/1045389X05055281
2. Morales, A., Nieto, A., Chicharro, J., Pintado, P.: A semi-active vehicle suspension based on
pneumatic springs and magnetorheological dampers. J. Vib. Control (2016). doi:10.1177/
1077546316653004
3. Huan, R.H., Li, X.P., Wu, Y.J., Zhu, W.Q.: Optimal bounded semi-active control of
hysteretic systems with MR damper. Adv. Struct. Eng. 13(6), 1199–1205 (2010). doi:10.
1260/1369-4332.13.6.1199
4. Choi, S.B., Han, Y.M.: Magnetorheological Fluid Technology-Applications in Vehicle
Systems. CRC Press, Boca Raton (2015). ISBN 978-1-4398-5673-4
250 A. Dobre et al.
5. Gołdasz, J., Sapiński, B.: Insight into Magnetorheological Shock Absorbers. Springer,
Heidelberg (2015). ISBN 978-3-319-13232-7
6. Nguyen, Q.H. Choi, S.B.: Optimal design of a vehicle magnetorheological damper
considering the damping force and dynamic range. Smart Mater. Struct. 18(1). IOP
Publishing Ltd. (2008). doi:10.1088/0964-1726/18/1/015013
7. Sassi, S., Cherif, K., Mezghani, L., Thomas, M., Kotrane, A.: An innovative magnetorhe-
ological damper for automotive suspension: from design to experimental characterization.
Smart Mater. Struct. 14(4). IOP Publishing Ltd. (2005) doi:10.1088/0964-1726/14/4/041
8. Rui, L., Chen, W.M., Yu, M., Liu, D.K.: Fuzzy intelligent control of automotive vibration
via magneto-rheological damper. Cybern. Intell. Syst. 1, 503–507 (2004). doi:10.1109/
ICCIS.2004.1460466. IEEE, Print ISBN: 0-7803-8643-4
9. Sandu, C., Southward, S., Richards, R.: Comparison of linear, nonlinear, hysteretic, and
probabilistic models for magneto-rheological fluid dampers. ASME J. Dyn. Syst. Meas.
Control 132(6), 164–172 (2010). Special Issue on “Physical System Modeling”
10. Ahmadian, M., Sandu, C.: An experimental evaluation of magneto-rheological front fork
suspension for motorcycle applications. Int. J. Veh. Syst. Model. Test. 3(4), 296–311 (2008).
doi:10.1504/IJVSMT.2008.025405
Innovative Solution for Motor Vehicles
Contributions in Experimental Research
Concerning Diesel Fuel Supply
and Lubrication in the Case of Comparative
Study Between Euro V and IV Common
Rail Engines
Doru Baldean1(&), Mihai Aurel Crișan2, and Adela Ioana Borzan1

1
{doru.baldean,adela.borzan}@auto.utcluj.ro
2
RAROM, Cluj-Napoca, Romania
mihai.crisan@rarom.ro
Abstract. The fuel efficiency and engine durability, besides the taxes added to
pollution, are nowadays some of the most important concerns in the mind of
vehicle owners. Lubrication and fuel consumption influences the pollution level.
Engine’s operating temperature, load, and speed, and fuel injected quantity,
affect the output performances. The present work outlines the research results
regarding the innovative methods for fuel consumption measurements in diesel
engines with electronic control module. The modern cars have more than one
electronic system for engine control, on-board data management and other
operating features. In order to efficiently use all the available data first they have
to be monitored. Optimizing the fuel consumption and engine thermal effi-
ciency, as well as the reduction of friction forces due to a better lubrication is a
desirable goal for the majority of motor vehicle’s owners and engineers.
Improving the fuel economy leads to a financial supplementary gain.
Keywords: Consumption Diesel Engine Fuel Lubrication
1 Introduction
The engine’s fuel economy and its operation durability, besides the taxes applied to
pollution level and to available fossil fuels, represent today few of the most significant
problems in the mind of all diesel owners. Lubrication process and fuel consumption
influences the pollution level. Engine’s operating regime, including coolant tempera-
ture, load, and crankshaft speed, fuel injected quantity and brake resistance of trans-
mission and vehicle’s body mass, leads to various energetic and dynamic
performances. The present paper shows the research protocol and some of the testing
results achieved in the real exploitation conditions designed at Automotive Engineering
and Transports Department in Technical University from Cluj-Napoca regarding the
innovative method for fuel consumption measurement in the case of diesel engine with
electronic control module and supplied through a common rail injection system.

254 D. Baldean et al.
The latest motor vehicles have more than one electronic system for engine working
control, on-board data management and other operating features and options. In order
to efficiently use all the available data in operation, first they have to be closely
monitored and rigorously recorded step by step, then analyzed.
Through an advanced electronic diesel control system with on-board management
module there are recorded and stored some of the most relevant actual values regarding
engine’s operating conditions and energy efficiency characteristics. Diesel fuel euro 5
with 5÷9 % volume participation of biofuel and 10W40 lubricant are used for optimal
engine operation. Experimental data post processing and calculations are closely
considered in the process of developing the proposed research. By taking into con-
sideration two available motor vehicles equipped with euro 4 and euro 5 common rail
diesel injection systems it was possible to determine the differences and similarities in
the field of fuel specific consumption, overall economy, thermal efficiency and envi-
ronmental impact for various operating conditions and loading regimes.
The study aimed results are foreseen in direction of substantial fuel economy,
environmental low impact and improved torque performance due to a better lubrication
and exhaust after treatment system operation. Optimizing the fuel consumption and
engine thermal efficiency, as well as the reduction of friction forces due to a better
lubrication and due to a better process management in operation, is surely a desirable
goal for the majority of motor vehicle’s owners and engineers. Improving the fuel
economy leads to a financial supplementary gain. The optimization of lubrication
process has an effect on engine’s operation expectancy.
State of the art equipment and high technologies are implemented today for testing
and monitoring new types of sensors and management systems on board and off board
in automotive field. Few of the latest technological achievements are synthetized in the
chapters of the present paper.
2 State of the Art Systems and Methods
2.1 Studies About Fuel Consumption and Lubrication

There is an important consideration of the latest research and development achieve-
ments in the field of information packages regarding fuel consumption and lubrication
process available for vehicle users on the front screen or on dash board in operation.
Diesel engine is one of the most efficient and reliable solutions for equipping the
automotive sector and the present day tractors or at least was considered in such
manner till the last year debate (that was before international Dieselgate scandal) [1].
The Volkswagen scandal has already raised questions in Europe about the accuracy
of emissions testing, putting other automakers on the defensive position [2].
The injection and combustion besides the supercharging have a major influence
upon the engine’s dynamic performances, fuel economy and pollution. Testing and
studying the injection and air-fuel supply systems on diesel engines (Fig. 1) may offer
engineering solutions and optimizing ideas in order to improve their environmental
impact and fuel economy, while the engine dynamic behavior will be little affected [3].
Contributions in Experimental Research Concerning Diesel Fuel 255
Fig. 1. Schematics of state of the art telecommunication system [3]. 1 – available vehicle; 2 –
OBD Bluetooth interface; 3 – short wave comunication device with Bluetooth; 4 – vehicle
dashboard; 5 – system user; 6 – mobile comunication device; 7 – access menu of the application;
8 – graphical interface of the comunication module.
Other research makes an experimental evaluation of some parameters related to fuel

consumption, engine speed, vehicle speed and environmental temperature in different
gear ratios and accelerator pedal positions, with sport utility vehicle Peugeot 4007
(Fig. 2) during the urban driving process. There isn’t an explicit study concerning the
consumption and performance in urban driving situation of all relevant SUVs. There
are many of them on the roads, but there should also be an interest concerning their fuel
economy and environmental impact. The engine and environment temperature, besides
Fig. 2. Equipment configuration for fuel consumption and pollution measurements [4].
1 – Peugeot 4007 SUV for experimental testing; 2 – dashboard; 3 – fuel consumption on
board display; 4 – engine assembly; 5 – injection system; 6 – Bosch electronic control module;
7 – KTS 650 (monitoring system).
the gear ratio and accelerator pedal position have a significant influence on the engine’s
energetic performance [4].
2.2 Mathematical Equations Related to the Fuel Consumption

Outlining all the important parameters related to the fuel consumption, vehicle speed,
engine power, and pollutant generation is significant in order to understand the
mathematics which stands behind the results.
Fuel consumption is determined with the mathematical relation [5]:
IV q Im
Ch ¼ 3:6 ¼ 3:6 ; ½kg=h ð1Þ
t t
in which: IV is the fuel injected volume, in [ml];
t – measuring time interval, in [s];
ρ – fuel density, [kg/h];
Im – the fuel mass, [mg].
Effective power output is calculated with the mathematical expression [6]:
pe n i Vs
Pe ¼ ; ½kW ð2Þ
30000 s
in which: pe is the brake mean effective pressure;
n – engine speed;
i – number of cylinders;
Vs – cylinder-unit volume;
τ – number of engine strokes.
Effective fuel specific consumption is calculated with the relation [5]:
Ch Im
ce ¼ ¼ 3:6 ; ½kg=kWh ð3Þ
Pe Pe t
in which: Ch is the hourly fuel consumption [kg/h];

Pe – effective power output [kW].
3 Materials and Method
3.1 Conventional Methods Applied in Fuel Consumption Investigation

Within all areas of transportation, solutions for economical and environmentally
friendly technology are being examined. Fuel consumption, combustion processes,
control and limitation of pollutants in the exhaust gas are technological problems, for
which guidelines like 98/69/EC and 99/96 determine the processes for the reduction of
fuel consumption and exhaust gas emissions. Fuel consumption of vehicles may be
expressed using the metric unit system in terms of consumed fuel per passenger
kilometer and passenger mile or per weight of transported cargo [7]:
– Fuel volume or fuel mass necessary per passenger kilometer in liters (passenger
km)−1 or (passenger kg)−1;
– Fuel volume or fuel mass per freight mass and distance, or freight volume and
distance in liters (kg km)−1 and liters (m3 km)−1 or kg (kg km)−1 and kg (m3 km)−1;
and
– Fuel volume or fuel mass per engine performance or engine thrust as Specific fuel
Consumption (SFC) in milliliters (kNh)−1 and milliliters (kWh)−1 or g (kNh)−1 and
g (kWh)−1.
Unlike the metric system (International System of Units (SI) or Système Interna-
tional d’Unités), the imperial measurement system gives details on attainable distance
per volume of fuel, i.e., mile per gallon or mpg consumed. In the metric system, the
amount of gases can be expressed in g km−1 in road transportation. The particles can be
characterized by the average diameter in mm or in inches, and the number of the
emitted particles without physical units [7].
In the present case investigating method was based on the digital equipment con-
nection to the vehicle’s OBD2 socket in order to read and store the actual values. One
of the most useful applications for realizing the experimental determinations was the
Torque acquisition system with a specific interface connected to a middle class pas-
senger car (type model Dacia Logan with diesel engine (K9k 792) euro 4 and (K9k
892) euro 5).
Applied method for measuring the energy efficiency and lubrication general
parameters is based on a program sketched in Fig. 3.
Two vehicle testing protocol

concept
Defining the vehicles configura- Realizing the field tests and

Configuration of tests track
tion processing experimental data
Fig. 3. Basic method for experimental tests
3.2 Materials and Test Procedures

In this paper is presented an innovative protocol for field operation testing of the engine
management system in real time working conditions by implementing state of the art
equipment and latest telecommunication features in order to read and store the two
engines eco-behavior by accessing the vehicle’s electronic control unit (VECU).
Field tests which are designed for this specific case require the implementation of
the following protocols:
• Operating at no load due to transmission engage;
• Operating tests in specific phases;
• Field tests for the two engine’s model;

• Specific engine start test;
• Engine operation mapping;
• Fuel supply methods for each Diesel Engine on test.
Because the experimental research is configured by the specified procedures some
testing results are determined and stored in the specialized equipment.
The connection through the electronic telecommunication interface has the purpose
of allowing an external evaluator (researcher, inspector etc.) to establish a scanning
protocol of parameters in which the electronic system of the motor vehicle, also the
engine works and manages the operating processes during some experimental tests, as
it is shown in (Figs. 4 and (5) [1].
For enquiring the field measurements are used ESI-tronic applications and KTS
Bosch 650 equipment. At their operating electronic base these programs have complex
mathematical models. Main apps are using the analog values measured by the sensors,
converted then to digital code. With the technology progress experimenting protocols
became very complex and the restrictions which were in the former systems are now
diminished.
Fig. 4. Setting testing communication protocol for K9K 792 ECU [1]. 1-Test track configu-
ration, 1a-Start/stop location, 1b-Track sector, 1c-Tested vehicle, 2-Injection pump,
2a-Temperature sensor, 2b-Fuel flow actuator, 2c-High pressure output, 2d-Injection unit,
3-Digital phonometer, 3a-Graphic display, 3b-Analogic display, 3c-Digital operation key,
3d-Digital display, 4-Digital label, 5-Analogic interface, 6-Electronic device, 7-OBD connection,
7a-OBD interface, 7b-OBD socket inside vehicle, 8-Portable gas analyzer, 8a-Digital display,
8b-Operation panel, 9-aquisition terminal, 10-Delphi injector, 10a-Nozzle, 10b-Electric connec-
tion, 10c-Injector return, 10d-high pressure connection.
In order to acquire reliable engine operational data the electronic diesel control
(EDC) module is activated which allow users to identify the management system.
After the engine identification procedure the user goes to actuators scanning and
actual values monitoring in order to read their proper operating state. Readings from
electronic control module corresponding to the actual values were accessed progres-
sively. Also, battery voltage, power supply for sensors and glow control were
monitored.
Fig. 5. Defining communication protocol for K9K 892 ECU. 1-experiment site map; 2-track;
3-start-point; 4-end-point; 5-vehicle tested; 5a-ON-BOARD display with total number of km;
5b-fuel quantity consumed during field tests; 5c-electronic prediction of distance in which the
fuel tank should be refilled; 5d-fuel quantity in the vehicle’s tank.
The tests were made during long term vehicle use. There were at least 4 oil changes
intervals at 10000 km for each car engine took into consideration. Lubrication process
monitoring system is highlighted to the user on vehicle dashboard by the specific
indicator lamps, shown in Fig. 6. The K9K 892 engine shows on the dashboard when is
the right moment for oil maintenance and filters, the specific lamp turns on.
In order to reset oil maintenance light there are specific procedures to be followed [8]:
1. Switch ignition ON;
2. Press and hold the throttle pedal;
3. Press the brake pedal 3 times (Perform all operations in less than 10 s. Oil change
warning light should turn off.)
Fig. 6. Defining lubrication general protocol for K9K 892 ECU. 1-indicator lamp for oil change;
2-diminished oil pressure and/or oil level, D-dashboard; T-tachometer.
4 Study of the Experimental Testing Results
In order to gain reliable fuel consumption data (Table 1) the engine was operated by
the technician in the crankshaft speed interval of 1500 ÷ 3600 rpm. The best perfor-
mances of fuel economy were achieved between 1550 ÷ 1800 rpm.
Table 1. Fuel consumption values in the case of two engines tested at partial loads (10 ÷ 60 %).
Nr. Temp. < 60°C Fuel Temp. > 60°C Fuel
Crt. consumption, consumption,
[l/100] [l/100]
1 Urban rush hour (E4), 6.4 Urban rush hour (E4), 5.3
[l/100 km] [l/100 km]
2 Urban rush hour (E5), 5.9 Urban rush hour (E5), 4.8
[l/100 km] [l/100 km]
3 Urban fluid (E4), 5.8 Urban fluid (E4), 4.9
[l/100 km] [l/100 km]
4 Urban fluid (E5), 5.5 Urban fluid (E5), 4.6
[l/100 km] [l/100 km]
5 Extra-Urban rush hour 5.9 Extra-Urban rush hour 5.2
(E4), [l/100 km] (E4), [l/100 km]
6 Extra-Urban rush hour 5.3 Extra-Urban rush hour 4.6
(E5), [l/100 km] (E5), [l/100 km]
7 Extra-Urban fluid 4.7 Extra-Urban fluid 4.2
(E4), [l/100 km] (E4), [l/100 km]
8 Extra-Urban fluid 4.4 Extra-Urban fluid 4
(E5), [l/100 km] (E5), [l/100 km]
9 Without trafic, (E4), 4.8 Without trafic, (E4), 3.9
[l/100 km] [l/100 km]
(Continued)
Table 1. (Continued)
Nr. Temp. < 60°C Fuel Temp. > 60°C Fuel
Crt. consumption, consumption,
[l/100] [l/100]
10 Without trafic, (E5), 3.9 Without trafic, (E5), 3.4
[l/100 km] [l/100 km]
11 Highway, 3.9 Highway, 3.5
(E4 < 100 km/h), (E4 < 100 km/h),
[l/100 km] [l/100 km]
(E5 < 100 km/h), (E5 < 100 km/h),
[l/100 km] [l/100 km]
(E4 > 100 km/h), (E4 > 100 km/h),
[l/100 km] > [l/100 km] >
(E5 > 100 km/h), (E5 > 100 km/h),
[l/100 km] > [l/100 km] >
Temp. < 70°C

Fuel Consumption, [l/100km]
8
Temp. > 70°C
6
0
1 2 3 4 5 6 Temp. > 70°C
7 8 9 10 Temp. < 70°C
11 12
Testing conditions 13 14
Fig. 7. Fuel consumption values in real road testing conditions
The consumption (Fig. 7) and emission (Fig. 8) measurements were represented on

graphics in order to easily study their comparative behavior.
Fuel consumption was monitored ON-BOARD in other conditions (Table 2).
It has been observed that the fuel consumption and the fuel injected quantity in all
cases of driving and operation conditions are close related to the engine’s temperature
and load. Also, the accelerator pedal position is essential for fuel injection manage-
ment. Fuel nominal pressure in the common rail system is at idle engine speeds
271 bar.
Table 2. Values for fuel consumption in urban driving conditions

Symbol Quantity Experimental
values
C1 Fuel consumption (FC) when defensive driving strategy is
applied
C2 FC in defensive driving, but engine is under 60°C coolant

temperature
C3 Rush hour defensive driving, engine is 60÷80°C coolant

temperature
C4 Air-conditioning on, engine is 60÷80°C coolant temperature
C5 Not defensive driving, engine is 60÷80°C coolant temperature

(headlights on,)
C6 Beginner style driving, engine is 60÷80°C coolant temperature

(headlights on,)
Fig. 8. Emissions variations in road testing conditions for the K9K 792 engine. MG-Motor Go
alone (no transmission); TR-Transmission Ratio 1; a-nitrous oxides variations; b-carbon oxide;
c-carbon dioxide; d-hydrocarbons variations.
5 Conclusion
As an observation derived from the considered problem, the air and fuel supply system
has a major influence on the engine operating economy, which may be downgraded by
injection characteristics and also by each auxiliary system improper performance which
may be controlled with the state of the art equipment that are almost everywhere
available today. ECU gives us the possibility to reset EGR adaptation values, rail
pressure sensor adjustment, fuel pump setting and to access other features in the
management system in order to clear the error memory and to increase the capacity of
the electronic management system in operation.
The experimental research paper presents an applied enquiry in the problem of fuel
consumption monitoring and engine management system on real diesel engines during
road traffic operation. Carbon foot print range, during a whole year of these vehicles
usage, is between 4500 and 7500 kg of CO2 emitted in atmosphere.
Future papers will cope with study of the alternative fumigation supply methods on
conventional injection systems and will provide emission values for K9K892 also.
Acknowledgement. The paper has been developed in the period of internal research project at
Technical University of Cluj-Napoca, with the reference code number UTCN 11/1.2/2015, in
which were financed some specific activities.
References
1. Baldean, D.: Contributions to the research and testing of modern diesel engines for
certification and homologation of performances and emissions. Inginerie agrară și transport
auto, 45, 146–150 (2015). Chişinău, R. Moldova
2. Ewing, J.: Volkswagen C.E.O. Martin Winterkorn Resigns Amid Emissions Scandal. New
York Times, International Business, SEPT, 23 (2015)
3. Baldean, D., Kocsis, L.B., Gaspar, F.: Research of fuel injection system parameters for the
engine of BMW 320d. Știință și Inginerie, XVI. AGIR, Romania, 30, 353–361 (2016)
4. Baldean, D., Burnete, N.V., Gaspar, F.: Contributions to the experimental evaluation of the
common rail fuel supply system and consumption from the Peugeot 4007 Sport utility vehicle.
Știință și Inginerie, XVI. AGIR, Romania, 30, 361–368 (2016)
5. Bataga, N.: Thermal Engines. Practical works (rom. Motoare termice. Lucrari practice). I.P.
C-N, pp. 57–60 (1988)
6. Bataga, N., Burnete, N., Rus, I., Cazila, A., Teberean, I., Sopa, S.: Internal combustion
engines (romanian) Motoare cu ardere interna. E.D.P, pp. 15–17 (1995)
7. Palocz-Andresen, M.: Decreasing Fuel Consumption and Exhaust Gas Emissions in
Transportation. Sensing, Control and Reduction of Emissions. Green Energy and Technology,
pp. 1–12 (2013). http://www.springer.com/us/book/9783642119750
8. ***, Oil reset service light indicator (2016). http://www.irls.eu/oil-reset-service-light-
indicator-dacia-duster
Experimental Research Regarding
the Possibility of Biofuel Fumigation Supply
Method on a Single Cylinder Compression
Ignited Engine
Doru Baldean1(&), Nicolae Vlad Burnete1, Dan Moldovanu1,

Ferenc Gaspar1, Adela Ioana Borzan1, and Mihai Aurel Crișan2
1
{doru.baldean,dan.moldovanu,
adela.borzan}@auto.utcluj.ro
2
RAROM, Cluj-Napoca, Romania
Abstract. Limits and specific disadvantages of using the conventional fossil

fuels, as well as their high price for the individual customer and the environ-
mental impact of their incomplete combustion are some of the main reasons for
taking new steps in research and development of new recipes, materials and
methods regarding the fuel supply for internal combustion engines. The present
paper encompass the research developed at Automotive Engineering and
Transports Department in Technical University from Cluj-Napoca regarding the
innovative possibilities for biofuel supply in compression ignited engines by a
fumigation method on a laboratory engine test bed. There were recorded and
studied the engine speed, torque variation at different operating regimes. The
obtained results from the conducted research are analyzed and discussed in close
detail. The bio-fuel fumigation method is compatible with the available diesel
engine and has a positive effect upon few behavior features in testing operation.
Keywords: Bio-fuel Diesel Economy Fumigation Tractor
1 Introduction
Today humanity faces the problem of limited access to conventional fuel or with higher
price at fuel selling stations. Each country makes the most of the available amount of
fuel. Individual user on the other hand has a restrictive access to the use of the fossil
fuels due to the increasing prices at selling points.
The present increases in atmospheric carbon dioxide concentration are in large part
the result of anthropogenic carbon dioxide emissions from fossil fuel combustion. Sce-
nario analysis is commonly used to generate projections of future carbon dioxide emis-
sions, the resulting atmospheric concentrations and climate impact. In most scenario
modeling published to date, carbon dioxide emission scenarios are based on demand-side
(socioeconomic and technology) factors. The fossil fuel resource is assumed ample
enough that supply-side factors do not drive future emission scenarios [1].

Single Cylinder Compression Ignited Engine 265
In the last decades some methods for the pollution reduction of ICE emissions have
been studied. The main oxygenated organic compounds are biodiesel, alcohols and
ethers. The biodiesel is produced from vegetable oils (fresh or used) and also from
animal fat. Today, the minimal quality conditions of biodiesel are regulated by the
European Standard for Biodiesel EN 14214. The use of biodiesel can be considered as
an alternative for CI engines, but some of its properties, as density and viscosity, are
higher than those of classic diesel fuel. These properties can be ameliorated by adding
bioethanol, which on one hand allows the biofuel’s level-increase in the whole blend,
and on the other hand brings the mentioned properties in standard diesel fuel prescribed
limits [2].
By a patented fumigation method of bioethanol in the intake manifold of a com-
pression ignited engine installed on a laboratory research test bed. Biofuel is fumigated
toward the engine intake collector pipe and improves the quality of the air-fuel mixture.
The present paper develops theoretical and experimental approach methods of the test
bed configured components from Lombardini compression ignited engine mounted on
a laboratory testing bench in correlation with fuel consumption and bio-fuel fumigation
parameters. The importance of applied research in laboratory conditions of the fumi-
gation method in tractor compression ignited engine is related with the state of the art
equipment and technology available today at Automotive and Transports Department
from Technical University of Cluj-Napoca for alternate fuel supply system delivered
directly inside the engine. Research expected results are aimed toward the fuel econ-
omy and dynamic performance. Improving the fuel combustion and energy efficiency
leads to evolution and progress in environmental carbon level.
State of the art technologies are used today for testing and implementing new types
of biofuels and supply systems on internal combustion engine. Some of the techno-
logical material available today is summarized in the following sections of the present
paper.
2 State of the Art Technology Available in the Research Field
2.1 Studies About Fuels and Biofuels Close Related to Research Field
There are some studies developed in order to increase the info data base concerning
conventional fuels and biofuels available today.
Alternative sources of energy for transportation are liquid fuels made from biomass,
such as biodiesel and bioethanol, considered as strong candidates for substituting fossil
fuels. Biodiesel can substitute diesel fuel, while bioethanol is used as a substitute for
petrol. An efficient method of solving some drawbacks of biodiesel is blending it with
substances which can compensate for their deficient properties. Using ethanol as an
extender for biodiesel is one method which has received close attention in research
lately [3].
Different mixtures of conventional and alternative fuels, as biodiesel–diesel–-
bioethanol (B–D–E) blends have become lately the subject of intense research which
targets mainly the evaluation of performance and pollution of compression ignition
engines (CIEs) fueled by such blends. Density is a very important fuel property because
it influences production, transportation, and distribution processes as well as all pro-

cesses that take place in the internal combustion engine. Elaborating models to describe
the density of biodiesel–diesel–bioethanol blends with a high content of biofuel
facilitates the production of blends that comply with the standard quality requirements
for density and the modeling of injection and combustion processes [4].
The versatility and popularity of conventional fuels (derivatives of plant and animal
fossils that are millions of years old, formed from the remains of the decayed plants and
animals of the carboniferous era) made them (i.e. coal, natural gas and oil/petroleum)
the major energy source that are being used in the world today. The fossil fuels are
non-renewable sources of energy as they are derived from pre-historic fossils and won’t
be available once they are fully used. Their disadvantages are: environmental hazards,
rising prices, acid rain phenomena, effects on human health, non-renewable, impact on
aquatic life by oil spill, coal mining results in ecological imbalance, need for huge
amount of reserves to carry out the process of generating power. Also these resources
are limited and they are depleting at a faster rate [5].
2.2 Relevant Studies Concerning Fumigation Method Applied in Diesel

Engines
Alcohols, being oxygenated fuel improve the combustion and reduce greenhouse gas
emissions, thus enhancing agrarian economies and encouraging national economy as a
whole. The objective of this paper is to investigate the thermal performance, exhaust
emissions and combustion behavior of small capacity compression ignition engine using
fumigated ethanol. Fumigated ethanol at different flow rates is supplied to the cylinder
during suction with the help of a simplified ethanol fuelling system. The rise in peak
pressure of cycle, heat release rate and ignition delay along with decrease in combustion
duration for different rates of ethanol fumigation were observed. Coefficient of Variance
of indicated mean effective pressure increases with ethanol fumigation [6].
Other studies concentrate on the impact of ethanol and n-butanol fumigation on the
combustion, performance, pollutant emissions, particle number concentration and size
distribution of an automotive diesel engine. Independently of engine load, both alcohols
exhibited higher premixed combustion peaks, faster combustion process, and higher
coefficient of variation of indicated mean effective pressure (imep) and reduced maximum
in-cylinder temperature, in comparison with ultra-low sulfur diesel (ULSD). Neither
n-butanol nor hydrous ethanol presented better brake thermal efficiency (bte) and brake
specific fuel consumption (bsfc) than ULSD. Engine performance with alcohol fumi-
gation is highly susceptible to operating mode. Therefore, it is necessary to optimize the
specific thermal conditions for implementing a fumigation strategy. Both alcohols
increased carbon monoxide (CO) and total hydrocarbons (THC) and reduced nitrogen
oxides (NOx) and particulate matter (PM), in comparison with ULSD fuel. However, the
magnitude of this reduction was highly influenced by engine operating mode [7].
Others examined the effects of ethanol fumigation on combustion, smoke and
nitrogen oxides (NOx) emission and performance of a turbocharged diesel engine [8].
Ethanol being an alternate fuel used in diesel engine either as a blend or fumigated
using a separate injector. Experiments were conducted on a diesel engine to examine
the effect of EGR temperature on NOx and other emissions. It was discovered that NOx
reduction up to 88 % was possible but at the cost of about 18 % loss in thermal
efficiency. This inconvenience of fuel loss caused by Exhaust Gas Recirculation can be
overcome by applying ethanol fumigation [9].
Other experimental results show that at low engine loads, the brake thermal effi-
ciency (bte) decreases with increase in fumigation ethanol; but at high engine loads, the
BTE is not significantly changed by fumigation ethanol. Fumigation ethanol can
effectively decrease in brake specific nitrogen oxides (BSNOx), particulate mass and
number emissions but significantly increase in brake specific hydrocarbon (BSHC),
brake specific carbon monoxide (BSCO) and other specific emissions [10].
A very close related study shows the effects of an experimental bioethanol fumi-
gation application using an experimental ultrasound device (Fig. 1) on performances of
a single cylinder compression ignited engine (CIE). Performance variations were
considered for three types of fuels in this investigation. Reductions in brake specific
fuel consumption and NOx pollutant emissions are correlated with the use of ultrasonic
fumigation of bioethanol fuel, comparative to use of biodiesel-bioethanol blend.
Considering the fuel consumption as Diesel engine’s main performance parameter, the
proposed bioethanol’s fumigation method, offered the possibility to use more efficient
renewable biofuels, with effects on environmental protection. Figure 1 represents the
schematic principle of ultrasonic fumigation process and technology [11].
Fig. 1. Schematics of ultrasonic fumigation technology [11]. 1-ultrasonic emitter; 2-fan, 3-air;
4-bioethanol; 5-intake pipe; 6-intake pipe air stream; 7-fumigated bioethanol; 8-fuel injector;
9-exhaust pipe.
In the middle of 2016 there was applied for a patent proposal concerning a method
for multi-fuel supply in the internal combustion engines with aerosols generated
through ultrasonic waves applied to biofuels based on alcohols and a study concerning
fumigation in the air-fuel supply system from CIE. Also the phases of development for
the internal research project at Technical University from Cluj-Napoca are treated with
complete responsibility in order to achieve the proposed results.
2.3 Mathematical Equations

In order to study all the engine’s parameters concerning the consumption, optimal
operation, output power, torque and brake specific consumption it is necessary to
highlight the mathematical models which are at the basis of each calculation after
experimental measurements.
Hourly fuel consumption is determined with the mathematical relation [12]:
Vq m
Ch ¼ 3:6 ¼ 3:6 ; ½kg/h ð1Þ
s s
in which: V is the consumed fuel volume in a given time [ml];
τ – the time in which the fuel volume V is injected, measured [s];
ρ – fuel density, in [kg/h];
m – the fuel mass in [g].
Brake specific fuel consumption is determined with the mathematical relation [12]:
Ch
ce ¼ ; ½kg/kWh ð2Þ
Pe
in which: Ch is the hourly fuel consumption [kg/h];

Pe – the effective power [kW].
2.4 State of the Art Research Technology Available

at Technical University
In Technical University of Cluj-Napoca, at the Automotive Engineering and Transports
Department, as well as in the most advanced testing and high-tech research entities, is
surely available a special configured Laboratory for Testing, Research and Certification
of the Internal Combustion Engines supplied with biofuel (Fig. 2).
Fig. 2. Advanced research test-cell in Technical University of Cluj-N [13]

General package of tests that can be managed in such a place are in the field of
testing internal combustion engines with classic diesel fuel as reference and also with
alternative biofuels, supplied with conventional methods and through alternative and
innovative procedures also, as well as engine’s reliability in long term tests [13]:
• Analyze of wear impact, operational performance and engine’s reliability;
• Analyze of operational irregularities with different fuels containing
bio-components;
• Determining the necessity of engine’s adjustment to operate with other fuels.
Specific engine tests that are developed on this advanced research test-cell require
the implementation of the following methods for study:
– Real time determination of exhaust emissions (engine pollutants);
– Real time measurements of particulate matter emissions;
– Real time determination of fuel consumption ();
– Study of the internal combustion phenomena;
– Evolution of engine’s parameters during testing procedures.
3 Materials and Method
3.1 Test Bench Components and Materials

In order to realize the testing battleground from available materials, components and
technological systems and in order to read and write all the engine’s significant
parameters the technicians and working staff designed a specified test bed with the
Lombardini internal combustion engine from engine’s laboratory, that operates on the
diesel fundamental cycle, with an air cooled cylinder block, supplied with direct fuel
injection and equipped with piezoelectric pressure sensor between the valves. Single
cylinder research engine is installed on a test bench electro-mechanically coupled with
a hydraulic dynamometer, as well as electronically and optically connected with data
read and write system. Also the engine exhaust line is connected with the specific
sensor of the exhaust emission analyzer. Figure 3 shows the specifically configured
engine test bed and the auxiliary components connected to the bench in order to
measure and record the operating parameters.
3.2 Method of Investigation

Analyzing the actual values and processing the engine operation diagrams during and
after the tests were realized with an auxiliary data acquisition system, which consists of
specific devices and optical units for digital versatile data storage. Methodology for
determining the operating and energetic actual values is designed considering a pro-
tocol sketch similar to the one presented in Fig. 4.
Fig. 3. Experimental research test bed at Technical University of Cluj-Napoca [14]. 1-single
cylinder research engine; 2-hydraulic passive dynamometer, 3-measuring and control systems;
4-recording and digital storage equipment; 5-fuel mass flow meter; 6-emissions analyzer;
7-workshop tools; 8-data post processing system; 9-engine accelerator handler.
Concept of ultrasonic fumigation process and technology

development
Theoretical Experimental Patent development for

documentation study measurements on test bed ultrasonic fumigation process
Fig. 4. Simplified program applied in the research methodology
3.3 Test Procedures

Laboratory tests which are applied on this specific test bench require the implemen-
tation of the following testing procedures:
• Steady operating state performance testing;
• Operating tests in sequence;
• Standard tests for specific engine’s model approval;
• Specific cold start testing;
• Pressure-Volume slope mapping;
• Fumigation supply procedure for Single Cylinder Research Diesel Engine.
When the applied study is developed by the mentioned procedures certain exper-
imental results are gained and may be stored with specific equipment.
4 Analyze of the Experimental Results

4.1 Engine Torque and Speed Measurement
In order to gain reliable torque measurements (Table 1) the engine’s crankshaft was
operated by a hydraulic brake and the indicated values were recorded. Engine speed
(Table 2) was determined with the sensor installed on the engine.
Table 1. Values for Engine torque in the case of two fuels tested at full load (100 %).
Symbol Quantity Experimental values, [N]
M100.100 %.M Engine torque determined in 1.657
the case of injecting in cylinder 5.021
conventional Euro 5 diesel fuel 10.127
15.087
20.099
25.461
B20.100 %.M Engine torque in the case of 1.662
20 % biofuel supplied in 5.026
alternative procedure 10.103
15.070
20.080
25.369
Table 2. Values for Engine Speed during testing.

Symbol Quantity Experimental values, [N]
M100.100 %.M Engine speed determined in 2977
the case of injecting in cylinder 2902
conventional Euro 5 diesel fuel 2856
2833
2716
2554
B20.100 %.M Engine speed in the case of 2974
20 % biofuel supplied in 2899
alternative procedure 2848
2821
2713
2570
4.2 Engine Parameters on Graphic Charts

The measurements were represented on graphics in order to easily study their trends
and engine behavior (Fig. 5). Engine load coefficient has high values when torque
riches the upper scale. At low torques the engine loads are similar, but at the maximum
1.2 1.7 B20.100%.Ch

B20.100%.χP M100.100%.Ch
Hourly fuel consumption,

1.0 M100.100%.χP 1.5
Engine Load, [%]
0.8
1.3
[kg/h]
0.6
1.1
0.4
0.2 0.9
0.0 0.7
0 10 20 0 10 20
Engine torque, [Nm] Engine torque, [Nm]
2.0 3.0
consumption, [kg/kWh]
2.5
1.5
Smoke, [1/m]
B20.100%.Ce
Specific fuel
2.0 B20.100%.k
1.0 1.5
1.0
0.5
0.5
0.0 0.0
0 10 20 0 10 20
Engine torque, [Nm] Engine torque, [Nm]
120.0
Nitrios oxides, [ppm]
100.0 B20.100%.NO
80.0 M100.100%.NO
60.0
40.0
20.0
0.0
0 5 10 15 20 25
Engine torque, [Nm]
Fig. 5. Engine’s actual values versus developed torque on hydraulic dynamometer
points the load differs between the biofuel and diesel fuel supplied to the combustion
chamber.
5 Conclusion
The present paper has realized a study concerning the importance necessity, opportu-
nity, and possibility for experimental testing concerning alternate fuel fumigation
procedure on a laboratory single cylinder diesel engine test bed. Some results which
were determined during experimental tests concerning the engine fuel economy were
also outlined and discussed. The most efficient operating regime is when the engine
torque is obviously put out in large amounts. In research developing process a par-
ticular phase was reserved to the environmental impact evaluation. Nitrous oxides are
at high level when supplying fossil fuel through conventional methods comparatively
with the situation of alternative fuels delivered through innovative procedures.
Future tests will concentrate on the importance of fumigation and technical appli-
cations and extensions of this patent proposal
Acknowledgements. The paper has been developed in the period of internal research project at
Technical University of Cluj-Napoca, with the reference code number UTCN 11/1.2/2015, in
which were financed some specific activities.
References
1. Vernon, C., Thompson, E., Cornell, S.: Carbon dioxide emission scenarios: limitations of the
fossil fuel resource. J. Procedia Environ. Sci. 6, 206–215 (2011)
2. Barabas, I., Todorut, A., Baldean, D.: Performance and emission characteristics of an CI
engine fueled with diesel–biodiesel–bioethanol blends. J. ScienceDirect Fuel. 89(12),
3827–3832 (2010)
3. Barabas, I.: Liquid densities and excess molar volumes of ethanol + biodiesel binary system
between the temperatures 273.15 K and 333.15 K. J. Mol. Liq. 204, 95–99 (2015)
4. Barabas, I.: Predicting the temperature dependent density of biodiesel–diesel–bioethanol
blends. Fuel 109, 563–574 (2013)
5. Rinkesh, K.: What are fossil fuels? Conserve energy future (2016). http://www.conserve-
energy-future.com/Disadvantages_FossilFuels.php
6. Jamuwaa, D.K., Sharmab, D., Sonib, S.L.: Experimental investigation of performance,
exhaust emission and combustion parameters of stationary compression ignition engine
using ethanol fumigation in dual fuel mode. Energy Convers. Manag. 115, 221–231 (2016).
http://www.sciencedirect.com/science/article/pii/S0196890416300851
7. Lópeza, A.F., Cadrazcoa, M., Agudeloa, A.F., Corredorb, L.A., Véleza, J.A., Agudelo, J.R.:
Impact of n-butanol and hydrous ethanol fumigation on the performance and pollutant
emissions of an automotive diesel engine. Fuel 153, 483–491 (2015). http://www.
sciencedirect.com/science/article/pii/S001623611500318X
8. Şahina, Z., Durguna, O., Kurt, M.: Experimental investigation of improving diesel
combustion and engine performance by ethanol fumigation-heat release and flammability
analysis. Energy Convers. Manag. 89, 175–187 (2015). http://www.sciencedirect.com/
science/article/pii/S0196890414008541
9. Hebbar, G.S., Bhat, A.K.: Control of NOx from a DI diesel engine with hot EGR and ethanol
fumigation: an experimental investigation. Int. J. Autom. Technol. 14(3), 333–341 (2013).
http://link.springer.com/article/10.1007/s12239-013-0037-8
10. Yao, C.D., Zhang, Z.H., Cheung, C.S., Xu, G.L.: Experimental study on the effect of
gaseous and particulate emission from an ethanol fumigated diesel engine. Sci. China
Technol. Sci. 53(12), 3294–3301 (2010). http://link.springer.com/article/10.1007/s11431-
010-3187-1
11. Mariasiu, F., Burnete, N.V., Moldovanu, D., Varga, B.O., Iclodean, C., Kocsis, L.: Effects of
bioethanol ultrasonic generated aerosols application on diesel engine performances. Therm.
Sci. 19(6), 1931–1941 (2015). http://www.doiserbia.nb.rs/img/doi/0354-9836/2015/0354-
98361500108M.pdf
12. Bataga, N.: Thermal Engines. Practical works (rom. Motoare termice. Lucrari practice). I.P.
C-N., pp. 57–60 (1988)
13. *** UTCN, Laboratory for Testing, Research and Certification of the Biofuel Internal
Combustion Engines (2016). http://www.testecocel.utcluj.ro/
14. Baldean, D.: Studies and researches concerning modeling the combustion process in
compression ignition engine when using biodiesel to improve engine performances. Ph.D.
thesis, UTPress, Cluj-Napoca, Romania (2011)
Study on a Stratified Charge Spark Ignition
Automotive Engine
Edward Rakosi(&), Sorinel Talif, Gheorghe Manolache,

and Adrian Sachelarie
“Gheorghe Asachi” Technical University of Iasi, Iasi, Romania

edwardrakosi@yahoo.com, tsorinel@yahoo.com,
gmanolache@yahoo.com, asachelarie@yahoo.com
Abstract. This paper is a synthesis of some theoretical and experimental

researches done during several years, in order to find the most efficient and
simple solutions for a spark ignition engines that function on inhomogeneous
poor mixtures. There are numerous stratification solutions for the mixture, both
in the unitary and divided combustion chambers. We have applied the divided
combustion chamber concept, using a flame jet in order to ignite the lean air-fuel
mixture, in the main combustion chamber. The secondary or auxiliary com-
bustion chamber, placed inside the cylinder head was fueled by direct fuel
injections, with a rich mixture.
Generally, gasoline direct injection imposes higher pressures than injection
into the inlet manifold. Meanwhile, the injection pump should have a long
service life while operating at high speeds. As a matter of fact, injection pump
service life was one of the factors that have restricted the use of direct injection.
Using mainly, the direct injection of gasoline, based on our own equipment we
studied this procedures to producing and igniting the poor mixtures on an
experimental engine. In the meantime, three types of spark plugs were tested.
The results experimental obtained on the test bed with this engine in homo-
geneous poor mixtures confirms the initial premises and proves the possibilities
of the stratification procedure that was used.
The article also contains the experimental results, as the reduction of the
specific fuel consumption for small partial loads, a domain in which the car
engine is often used and the reduction of some noxious emissions, especially CO
and HC.
Keywords: Internal combustion engines Direct fuel injections Lean air-fuel

mixture noxious emissions
1 Introduction
Taking into account fuel economy, the different constructive variants applied to the
spark ignition engine are limited due to the growth of the combustion zone behind the
ignition front. All the researches show that the use of homogenous lean mixtures
(characteristic to Diesel engines) could not improve the engine’s fuel consumption.
Under these circumstances, the use of stratified air-fuel mixtures seems to be a rea-
sonable issue, unless higher costs or constructive difficulties are implied.

Study on a Stratified Charge Spark Ignition Automotive Engine 275
Up to now, the flame jet engine, which infers a higher specific fuel consumption
and more elaborated constructive solutions then the single combustion chamber engine,
has been rather cautiously approached [1, 2, 6].
We consider that this type of engine still has unrevealed advantages. Thus the
divided combustion chamber allows the use of gasoline with a lower octane number
and even of the inferior fuels. In the meantime, researches have proved a favorable
behavior concerning the emission of primary pollutant products, especially NOx [3, 4,
8]. Finally, the solution we have adopted has removed the third valve with its driving
mechanism, using a more simple and reliable fueling system, containing parts that are
commonly used in internal engine construction.
2 The Constructive Solution
The authors have used a spark ignition engine, with four in line cylinders (bore = 76 mm,
stroke = 77 mm). The engine is water-cooled.
The divided combustion chamber has been achieved partly in the cylinder head and
partly in the piston head. The rest of the volume consists of the connecting channel
between the two parts of the combustion chamber. The cylinder head is equipped with
two vertical valves for each cylinder; the valves control only the main combustion
chamber [7].
The fueling system was a combined one. Thus, the auxiliary combustion chamber is
fueled with rich mixture through direct injection at approx. 8 MPa; the main com-
bustion chamber is fueled with a lean mixture. Thus we have avoided the use of more
elaborated solutions.
The electrically induced ignition takes place in the secondary combustion chamber.
After ignition, the pressure rises in the secondary chamber, and the gases are ejected at
high speed into the main combustion chamber. As most of these gases are burning, they
penetrate into the main chamber as a flame jet and produce a mass ignition of the
existing lean mixture. The result is a sure ignition and quick combustion, which are the
necessary conditions for a proper working process of the engine.
The main combustion chamber was built in two variants. The first one provides the
piston head with a profiled cavity, as shown in Fig. 1. Thus a centrifugal movement
during compression is ensured, creating a supplementary stratification of the lean
mixture. Facing the connecting channel between the two chambers, a round clearance
has been made in the main combustion chamber, in order to allow the discharge of the
jet coming from the secondary combustion chamber. This construction has led to a
compression ratio ε = 10.5.
The second type of main combustion chamber provides the piston with three
compartments, as shown in Fig. 2. A first turbulence threshold was achieved in the
passage area. Its role is to break up the jet and spread the active centers into the mass of
the lean mixture. The slope of the threshold ensures the reflection, while the semicir-
cular part imprints a divergent shape to the jet. The two symmetrical cavities are used to
create; a swirl movement, that entrains and diffuses the active centers from the flame jet
into the mass of the lean mixture.
276 E. Rakosi et al.
Fig. 1. Main combustion chamber, type I
Fig. 2. Main combustion chamber, type II
The achievement of the main combustion chamber into the piston head has imposed
the use of pistons with a thicker, head. The oil flow through the last piston ring channel
has been increased due to the presence of 8 holes with 0.5 mm diameter.
The second type of main combustion chamber has led to a compression ratio of 10
or 11.5, depending on the type of the secondary combustion chamber. In order to use
heavy fuels (as Diesel fuel and petroleum), a variant with ε = 7 has been achieved.
We have chosen three types of parts in order to create the auxiliary combustion
chamber, shown in Fig. 3. The design of these pieces has taken into account the
influence of thermal stress and specific mounting conditions; we have used temperature
resistant steel. In order to decrease the temperature of the secondary combustion
chamber, tighter yawns were used.
Fig. 3. Auxiliary combustion chambers (variants)
Considering the two types of main combustion chamber, and the three types of
auxiliary combustion chamber, we have finally obtained six constructive solutions for
the flame jet ignition engine. Figure 4 presents one of these constructive solutions.
Fig. 4. Variant Fl
The camshaft was also used to entrain the injection pump from its posterior end.
For this phase of the research we have used a classic ignition system. The ignition
distributor was placed on the injection pump using a piece that allows its rotation with
approx. 120o and thus enabling the adjustment of the ignition advance.
Recommendations made by other researchers as well as preliminary experiments
indicated that initiation of the flame in the rich mixture from the secondary combustion
chamber should take place in its central area [2, 5]. In order to achieve this purpose we
have used special spark plugs, equipped with long electrodes, as shown in Fig. 5. In
order to adjust more accurately the plug’s position into the combustion chamber, we
have used copper washers with different thickness. It is known that a large part of the
spark energy heats the electrodes or is lost through radiation and only 5–16 % is used
to heat the mixture to the ignition temperature level [9, 10]. This is the reason why we
have designed and achieved three types of electrodes, shown in Fig. 6. Type a com-
bines the classic shape with the above-mentioned requirements, while type b prevents
clogging due to a jet sprayed in the vicinity. The surface tension forces produced by the
curved shape of the electrode stops the formation of fuel droplets. The type c spark plug
uses two shaped ground electrodes; the discharge energy for this type of spark plug is
between 0.2 and 3 mJ, according to Bosch [10, 11].
Fig. 5. Long electrodes spark plug
Fig. 6. Spark plug variants

Figure 7 presents a view of the engine on the test bed.
Fig. 7. Engine mounted on the test bed
3 Experimental Tests
We have developed many tests in order to establish the most convenient type and the
dimensions of the two combustion chambers and connecting channel. Table 1 presents
the testing variants taken into account.
Table 1. Tested variants

Variant Main combustion Auxiliary Compression Fueling
chamber combustion chamber ratio
Fl one cavity Figure 3a εl = 9 combined fueling
F2 (Fig. 1) Figure 3b ε2 = 10.5 solution
F3 Figure 3c ε3 = 7
F4 three cavities Figure 3a ε4 = 10
F5 (Fig. 2) Figure 3b ε5 = 11.5
F6 Figure 3c ε6 = 8
We have started the tests with the variant marked Fl. Adjusting the position of the
spark plug inside the auxiliary combustion chamber, we have noticed that the steadiest
working process, both at idle and load, was obtained when the ignition point (electrode
tip) was placed at about 14–15 mm inside the chamber. When this position is exceeded
with ±3 mm the cyclic stability is compromised, which proves the existence of a zone
with rich mixture - probably a 6 mm diameter sphere.
The next step was to establish the most favorable injection timing. The criteria
taken into account were the starting ease, the stability of the engine’s working process
and the achievement of maximum power indices. The intake and the compression
stroke were explored; the most favorable moment for the start of the injection was
found to be at 75˚ B.T.D.C., during the compression stroke.
These adjustments and especially the injection advance proved to be also fitted for
the rest of the tested variants, so they were preserved for all the other experiments.
4 Experimental Results
The experimental results we have obtained are presented in Table 2. We have devel-
oped the analysis taking into account the following criteria:
• starting ease;
• stability of the working process;
• power indices;
• engine dosages and consumption;
• knocking tendency.
Table 2. Summarized experimental results

Variant Pn λ at ΔCh Δce Spark Remarks
kW idle % % plug type
Fl 37.8 1.92 18.1 12.4 Figure 6a Very stable working process
F2 40.5 2.10 21.3 13.6 Figure 6a Good working process, with high
knocking resistance and low
consumption
F3 – 1.56 10.8 – Figure 6c Unstable working process,
especially at high part load and
full load
F4 39.0 1.98 18.7 13.2 Figure 6b Good behavior but without any
remarkable performances
F5 42.6 2.50 29.2 15.0 Figure 6b Best performances for this solution
F6 22.7 1.71 11.2 – Figure 6c Low performances, tendency
towards unstable
The tests have emphasized that idle and low load working regimes could be ensured
only by injecting fuel in the secondary combustion chamber, especially at low engine
speed.
We noticed the engine could be operated for all the loads with the throttle com-
pletely open. Under these conditions, load adjustment was achieved by modifying, the
dosage of the lean mixture. This allowed us to avoid equipping the injection pump with
a cyclic dose correction device.
In order to present more details about the most significant results, we have selected
variant F5, which seemed to be the best one, satisfying most of the initial criteria.
Figure 8 presents the fuel dosage at idle for both standard and fuel injection engine;
λ stands for the excess air coefficient. We noticed important differences between the
two fueling solutions. Thus, for the flame jet engine, the engine had a steady working
process even for λ = 3.3, proving the efficiency of the stratifying process. The mean
dosages we obtained were as follows:
• λm, F5 = 2.5 for the tested engine;
• λm, c = 0.66 for standard engine.
Fig. 8. Air excess at idle
The lower temperature of the exhaust gases presented in Fig. 9, confirms the fact
that the engine is working with lean mixtures.
Fig. 9. Exhaust gases temperature at idle

The engine may work with leaner mixtures, as reflected in the engine’s fuel con-
sumption characteristic at idle (Fig. 10). Figure 11 shows the fuel consumption
decrease ΔCh of the jet flame engine compared to the classic engine.
Fig. 10. Idle characteristic
Fig. 11. Fuel consumption decrease at idle
Because the engine works without throttling the inlet manifold, the cylinder filling
process has improved. This aspect is presented in Fig. 12, where ηv marks the filling
coefficient.
Fig. 12. Cylinder filling coefficient at idle
Figure 13 presents the engine’s BSFC at different part loads. The most important
decreases in the specific fuel consumption have occurred at 25 %, 40 % and 70 % load.
The minimum specific fuel consumption was 240 g/kW.h.
Fig. 13. Specific fuel consumption
The influence of the engine speed over the decrease of the BSFC, at 25 % and
40 % load, is shown in Fig. 14.
Figure 15 presents the CO content of the exhaust gases at 25 % and 70 % load,
while Fig. 16 shows the HC emissions; the mean air excess coefficients are also shown.
From this point of view, the jet flame engine has a different behavior compared to the
Fig. 14. Variation of specific fuel consumption
Fig. 15. CO emissions, F5 variant
classic spark ignition engine, for which λ should decrease at low loads and at idle. For
the jet flame engine, leaner mixtures led to higher HC emissions.
Tests made at very low loads with the F3 variant of the same engine have revealed
the possibility of using less volatile fuels without knocking, even for a slightly higher
ignition advance.
Fig. 16. HC emissions, F5 variant
5 Conclusions
1. The stratified charge solution provided the engine with lean mixtures, that could
lead to a specific fuel consumption lower than 245 g/kWh for the best variant of the
flame jet ignition engine.
2. Running at idle and low loads is ensured, at low speed only with fuel injection in
the secondary combustion, chamber. All the other regimes require dual fueling: a
small amount of fuel (independent of load and speed) is injected in the secondary
combustion chamber, while the main combustion chamber is fueled with a very lean
mixture. Under these circumstances, the injection pump does not need a flow
correction device.
3. All the working regimes are achieved with the throttle completely open. Load
adjustment is attained only, by modifying the lean mixture’s dosage, similar to a
Diesel engine.
4. A good cyclic stability for the engine running with lean mixtures is accomplished
when the spark plug electrodes have a length of 14–15 mm; this proves the exis-
tence of a zone with rich mixture (probably a 6 mm diameter sphere).
5. Fueling the engine with stratified mixtures leads to a good cyclic stability even for
overall dosage coefficients λ = 3.3. This is a characteristic of the flame jet ignition
engine and confirms its framing among stratified charge engines.
6. At low loads the mean BSFC reduction is comprised between 14.6 % and 18 % and
a fuel consumption lower than 300 g/kWh is achieved over a relatively wide speed
range.
7. At low part load, a significant reduction of the CO emissions has been noticed, for
the entire engine speed range. The CO emissions are lower than the ones recorded
for the direct injection spark ignition engine. The HC emissions are higher than the
ones for the direct injection engine, due to the leaner mixtures used in order to fuel
the flame jet ignition engine.
8. Although we have used a rather high compression ratio (ε = 11.5) knocking has not
been reported; this conduct is also characteristic for the flame jet ignition engine.
For another constructive solution, with the compression ratio ε = 7, the starting and
running tests we made have shown that less volatile fuels (as petroleum) could be
used in order to fuel the engine.
References
1. Brandstetter, W.R.: Experimental results from Volkswagen’s prechamber stratified charge
engines. In: Institution of Mechanical Engineers Conference Publications, pp. 71–80 (1976)
2. Gaiginschi, R., Rakosi, E.: Au sujet d’un moteur a jet de flamme. Bulletin of the Technical
University Iasi, torn XLII (XLVI), fasc. 1–2, sect. V, pp. 57–61 (1996)
3. Grunwald, B.: Theory, Design and Construction of Automotives Engines. Didactical and
Pedagogical Publishing Hhouse, Bucharest (1980)
4. Hensinger, D.M., Maxon, I.A., Horn, K., Oppenheim, A.K.: Jet plume Injection and
combustion. SAE Paper 920414 (1992)
5. Meurer, S.J.: Present experience with stratified charge engines working with initial
separation of mixture components. In: Institution of Mechanical Engineers Conference
Publications, pp. 81–88 (1976)
6. Oblander, K.: Stratified charge engines with open and subdivided combustion chamber. In:
Institution of Mechanical Engineers Conference Publications, pp. 89–94 (1976)
7. Overington, M.T.: A new stratified charge engine based on the Ricardo Cornet design. In:
Institution of Mechanical Engineers Conference Publications, pp. 103–108 (1976)
8. Pischinger, F.: Experimental and theoretical investigation on a stratified charge engine with
prechamber injection. In: Institution of Mechanical Engineers Conference Publications,
pp. 1–6 (1976)
9. Stan, C.: Directeinspritzsysteme fur Ott- and Dieselmotoren. Springer, Berlin (1999)
10. Aleiferis, P.G., Taylor, A.M.K.P., Ishii, K., Urata, Y.: The relative effects of fuel concentration,
residual-gas fraction, gas motion, spark energy and heat losses to the electrodes on
flame-kernel development in a lean-burn spark ignition engine. Proc. Inst. Mech. Eng. Part D:
J. Automobile Eng. 221(7), 845–857 (2007)
11. Aleiferis, P.G., Taylor, A.M.K.P., Ishii, K., Urata, Y.: BOSCH: Electronische Zundsysteme
EZ, Technische Kurzinformation (1987)
Researches on Cooling Air Flow Control
Devices Using on Cars with Internal
Combustion Engines
Spiridon-Cristian-Dan Dăscălescu, Marius Receanu,

Laurenţiu Dimitriu(&), and Liliana Vornicu-Albu
“Gheorghe Asachi” Technical University, Iasi, Romania

laurentiu_dimitriu@yahoo.com
Abstract. The cooling airflow control devices changes optimally the cooling
air flow passing through the cooling radiator using a blades system that can be
oriented conveniently. The use of these devices has the effect of reducing fuel
consumption, shortening the engine time starting and consequently reducing
wear. The experiments were done on a moving vehicle equipped with a petrol
engine, whose radiator crossing section was opened and closed successively. It
were measured simultaneously coolant temperature at the inlet and at the outlet
of the cooling radiator and air temperature entering the air filter. The time was
also measured, and fuel consumption was determined by the volumetric method.
For experimental data recording was used an acquisition board. It was found that
a convenient varying of the air flow passing through the cooling radiator, results
in fuel saving, reducing the time cold start engine and therefore a reduced wear
on cold start.
Keywords: Air flow cooling Fuel consumption
1 Introduction
A cooling air flow control device [1] works on a blades system which extreme positions
[2] are: closed at the beggining of cold starting (under 5°C) and are opened when the
engine thermostat is open. The paper presents a series of tests performed on a Dacia
1310 car in motion which run consecutively with open and closed cross section of the
cooling radiator, moving on the same section of road, with the same load, with the
same speed and the same ambient temperature.
The engine heat balance, in case of movement with open cross section of the
cooling radiator is:
Chn Q1 ¼ Qp þ qr ce ðt1 t2 Þ þ Qge þ Qrez þ Qin ð1Þ
where:
Chn – the engine fuel consumption per hour;
Q1 – calorific power of fuel
QP – the amount of heat converted into mechanical energy;
qr – the coolant flow;
288 S.-C.-D. Dăscălescu et al.
ce – the heat capacity of the coolant;

t1 – the coolant temperature at radiator inlet;
t2 – the coolant temperature at radiator outlet;
Qge – the lost heat with exhaust gases;
Qrez – the residual losses;
Qin – the heat losses by incomplete combustion of fuel;
In case of movement with the closed radiator cross sections, may be considered that
Qp, Qge Qrez and Qin have the same values, in the case of travel with the same speed of
the car, the same aerodynamics resistance of the car and the same road.
In this case, the energetic balance of the car becomes:
0

Ch0 Qi ¼ Qp þ qr ce t1 t2 þ Qge þ Qrez þ Qin ð2Þ
where:
Ch0 – the fuel consumption with closed cross section of the radiator;
0
t2 – the coolant temperature at radiator outlet with closed cross section of the
radiator;
0
In case of movement of the car with the open cross sections of the radiator, t1 \t2 ,
because cooling air flow rate is higher than the case of movement with the closed cross
sections.
Comparing the relationships (1) and (2), results that Ch0 < Chn.
To verify this theoretical conclusion and to appreciate comparatively heating speed
of the engine in both cases, with open and closed cross sections of the radiator, have
been performed the same tests with the moving car.
0
The temperatures t1, t2 and t2 , the consumptions of fuel Ch0 and Chn, the intake air
temperature in the air filter with open cross section (t02) and closed cross section (t01) of
the radiator have been measured.
2 Experimental Method
The tests were accomplished [3] with a Dacia 1310 car, in motion, successively with
closed cross section and open cross section of the radiator, on the same road with a
speed of 11.11 m/s and 16.16 m/s.
Ambient temperature of −3°C was measured with a mercury thermometer. The
temperature of inlet in the air filter was measured by a digital thermometer.
The coolant temperatures at the inlet (t1) and at the outlet (t2) of the cooling radiator
have been measured using two thermocouples. The temperature of the lubricant in the
oil sump was also measured using a thermocouple.
For the acquisition of signals from thermocouples, a National Instruments 6062E
acquisition board and a laptop have been utilised. The signals were amplified and
filtered, and the data was stored as files.
In Fig. 1 are shown the details about the mounting on the car of the system with
thermocouples for measuring temperatures.
Researches on Cooling Air Flow Control Devices Using on Cars 289
Fig. 1. Details mounting of measuring system on vehicle.
Experimental data acquisition board were loaded in the laptop using a virtual tool,
as shown in Fig. 2. The signals were processed through filtration using virtual tool
shown in Fig. 3.
Fig. 2. Acq&Graph Voltage.Int Clk.vi.Block Diagram/LabVIEW 8.2
Fig. 3. Signal processing experimental program-Block Diagram

3 Experimental Data
A first series of tests were performed with a Dacia 1310 car, which ran on the same
road, with the closed cross section of radiator, at an ambient temperature of -3°C.
The car was ran at a speed of 11.11 m/s, loaded with a mass of 350 kg.
The experimental data were processed using programming languages
MATLAB/Simulink and LabVIEW.
Figure 4 shows the variation in temperature of the coolant at the inlet in the
radiator, measured by thermocouple, depending on time.
In Fig. 5 represents the evolution of simultaneous coolant temperature exiting the
radiator and at engine entrance, measured with thermocouple.
T1 (deg)
t (min)
Fig. 4. The variation of the coolant temperature at the entrance of the radiator (T1 degrees)
measured with a thermocouple, a function of time t (min)
T2 (deg)
t (min)
Fig. 5. The variation of the coolant temperature at the outlet of the radiator (T2 degrees)
measured with a thermocouple, a function of time t (min)
The variation of coolant temperature, may be approximated with polinomial

functions [4].
For an overall assessment, lubricant temperature variation depending on time in
sump is shown in Fig. 6.
T3 (deg)
t (min)
Fig. 6. Temperature variation of engine lubricating oil (T3-degrees), depending on time t (min)
Figure 7 shows the variation in air temperature entering the air filter, with the open
cross section of the radiator, when the tested car is moving in third gear, with the speed
of 11.11 m/s and at an ambient temperature of −3°C, and Fig. 8 shows the variation
with the closed cross section of the radiator.
Fig. 7. Temperature variation inside the air filter with open cross section of radiator
In the Table 1 are shown time values corresponding to a cold start, at an ambient
temperature of −3°C, until the reaching the values of 40°C and 80°C for coolant
temperature, with the car in motion with the cross section of the radiator closed and
respectively open, with the car speed of 11.11 m/s and the same mass load.
Fig. 8. Temperature variation inside the air filter with closed cross section of radiator
Table 1. Coolant temperatures values during travel.

Elapsed time after Coolant Notice Ambient Speed
cold start [s] temperature temperature [m/s]
[oC] [°C]
150 40 Closed cross section -3° C 11.11 m/s
400 80 of the radiator
264 40 Open cross section
558 80 of the radiator
Table 2 shows the fuel consumption values for the car in motion with the cross
section of the radiator closed and respectively open and the car speed of 11.11 m/s and
16.66 m/s.
Table 2. Fuel consumption values during travel.

Fuel Time Fuel Fuel Fuel Fuel E Notice Moving
consump-tion [s] consump-tion onsumption consumption consumption [%] speed
[l] Chl [l/h] Chkg [kg/h] l
C100 kg
C100 [m/s]
[l/100 km] [kg/100 km]
1 770 4.67 3.586 11.67 8.945 14.82 Closed cross 11.11
section of the
radiator
0.85 770 3.974 3.051 9.935 7.627 Open cross 11.11
section
0.84 681 4.44 3.409 7.4 5.682 19.06 Closed cross 16.66
section of the
radiator
0.68 681 3.549 2.759 5.99 4.599 Open cross 16.66
section
It is also mentioned the relative percentage fuel saving, when the car in motion with
the closed cross section of the radiator, compared with the case of open cross section of
the radiator, calculated with the relationship (3):
Chn Ch0
E¼ 100½% ð3Þ
Chn
4 Analysis of Experimental Data
By analyzing graphs of Figs. 4 and 5, it follows that the difference between the tem-
perature of the coolant entering the radiator (t1) and temperature of coolant exiting the
0
radiator t2 with the closed cross section of the radiator is about 6°C. This difference is
small after 6 min because the car is moving with open cross section of the radiator.
Evolution of the lubricant temperature in the sump of the engine (see Fig. 6) shows a
lower temperature level compared to the coolant temperature in the same moments
during cold start.
The data in Table 1 indicate that the time required for heating the engine in cold
start, with the closed cross section of the radiator, untill to 80°C of coolant, is 400 s. By
comparison, the same test with the car moving with the open cross section of the
radiator, the time required for heating the coolant is 558 s, so with a difference of
158 s. Consequently, the engine is warming faster by 39.5 %. This results in reducing
pollution and wear when the car is moving with the closed cross section of the radiator.
If the car is moving with closed cross section of the radiator, air temperature entering
the air filter is 12°C (see Fig. 8) after 700 s from the cold start. In the case moving with
open cross section of the radiator the air temperature is 2°C (see Fig. 7). As a result, for
engine working with the closed cross section of the radiator, combustion reactions
inside of engine cylinders will intensify and will release a larger amount of heat, which
causes a more rapid heating of the engine during cold start.
By analyzing fuel economy values (E%) as is shown in Table 2, the resulting fuel
saving is of 14.82 % when the car is moving at the speed of 11.11 m/s, at an ambient
temperature of −3°C and with closed cross section of the radiator.
If the car moves on the same road with a speed of 16.66 m/s at an ambient
temperature of −3°C, fuel economy is 19.06 %, with the closed cross section of the
radiator, compared with the situation with open cross section. This represents a con-
siderable value.
5 Conclusions
1. The closing of cross section of the cooling radiator during engine cold start with
airflow cooling control devices it has the effect of improving the intake air tem-
perature for the engine cylinders and an improvement in the combustion process,
which causes a greater release of energy during combustion. This results in a more
rapid heating of the engine during cold start and therefore reducing wear and
pollution in the same conditions of movement of the car.
2. In the case of closing of the cross section of the cooling radiator using airflow
cooling control devices, in the same conditions of moving speed, mass loading and
on the same road, the fuel saving is substantialy. The fuel saving is about 19 % for a
car moving at a speed of 16.66 m/s at an ambient temperature of −3°C. It is
expected to improve fuel saving when the ambient temperature is lower.
References
1. Dăscălescu, S.C.D., Dimitriu, L.: Advantages of Cooling Airflow Control Devices Using by
Internal Combustion Engines. 2001, SAE Paper 2001-01-1703 (2001)
2. Dăscălescu, S.C.D., Dimitriu, L.: Sistem de control electronic pentru reglarea debitului de aer
de răcire la motoare termice” (Electronic controlled system for thermal engine cooling air flow
regulation), Brevet RO 113753, (1998)
3. Receanu, M.: Contribuţii privind optimizarea regimului termic al motoarelor cu ardere interna
prin utilizarea dispozitivelor de reglare a debitului de aer de răcire. Doctorate disertation,
“Gheorghe Asachi” Technical University of Iaşi, Romania, (2010)
4. Dăscălescu, S.C.D, Dimitriu, L., Sachelarie, A., Vornicu, L., Damian, I., Receanu, M.:
Method for thermal losses by cooling system radiator measurement at automotive vehicles. In:
CONAT 2004, International Automotive Congress, Brasov, Romania (2004)
Modeling the Energy Evaluation
for an Electric Machine
Valerian Croitorescu(&) and Alexandru Manta
Scientific Research and Continuous Training Center for Sustainable Automotive

Technologies, University Politehnica of Bucharest, Bucharest, Romania
{valerian.croitorescu,alexandru.manta}@upb.ro
Abstract. The regulations concerning CO2 emissions and the need for fuel
economy and low harmful emissions press forward the vehicle manufacturers in
their electric vehicles development tendencies. Currently, most of the vehicle
manufacturers are producing and selling hybrid and electric vehicles, being
acceptable solutions against pollution inside crowded urban areas.
The electric machine for the electric vehicles is not only the energy consumer,
but also the energy producer during regenerative braking. Its design and type
vary, being chosen for achieving the established performances. The electric
vehicle powertrain efficiency depends on the electric machine performances.
Therefore, by investigating the functional behavior for an electric machine as
part of an electric powertrain new research issues appear.
This paper highlights the needs for evaluating the energy efficiency for
electric machines long before being produced, by using virtual mock-up-ing
approach. The evaluation consists in determining the thermal losses using
simulation.
Keywords: Electric machine Energy losses Thermal losses Heat transfer

Electric vehicle
1 Introduction
The current automotive industry evolution is marked by both electric and autonomous
vehicles. The powertrain development is based on modeling and simulation for all the
systems, having its own limitations. The continuous improvement is based on the
accurate results and the computational performances of the simulation tools. The pow-
ertrains are developed virtually during more than one level of design, starting from the
control strategies, to physical investigation and mathematical algorithms definition. The
modeling results consist in the virtual mock-up, which is used to simulate different
processes during different driving cycles. The virtual mock-up is a virtual device of a full
sized system being able to provide at least part of its functionality and to test its behavior.
The virtual mock-up includes all the systems and sub-systems that are part of the
powertrain for allowing a realistic investigation very early in the development process.
The virtual mock-up is a sum of information which describes the system behavior, being
easy to be changed and updated when necessary, [4].

296 V. Croitorescu and A. Manta
The advanced methods used for developing and investigating the virtual mock-up
consist in integration during all the development phases of the mathematical equations
and the interpolation functions in order to analyze and to optimize the components
using different states of the parameters.
The powertrain energy efficiency is a target to be achieved by using virtual
mock-uping. The electric machine is the system under evaluation, therefore its func-
tional behavior have to be translated into mathematical models and to be simated taking
into account its construction and parameters.
The objectives of this paper include the basic functional and thermal modeling for
the electric machine, as part of an electric powertrain. The electric machine type defines
the configuration, while the electric machine power defines the geometry and the size.
Based on these, both the functional and the thermal model are developed in order to
investigate the energy efficiency by computing the energy losses as thermal losses.
2 Modeling and Simulation
The simulation approach require to understand the engineering processes regarding the
mechanical and mechatronic systems behavior. As the current systems are more
complex, the command and the control systems are moving towards to high complexity
models.
The modeling and simulation software are also increasing their performances as
well as the required computational resources. For developing the electric machine
mock-up system a dedicated simulation tool was used. The LMS.Imagine.Lab.AME-
Sim allows developing and running different multi-disciplinary models of systems or
sub-systems to analyze their complex behavior and to offer the possibility to create
control systems from the early design stage, [4]. The used simulation tool consists in an
intuitive graphical interface that presents individual icons for each component or
system, easy to be recognized by its shape. Behind each icon, a fragment of C-code is
describing the behavior of the component, using mathematical equations. The simu-
lation model is developed by linking existing models, sub-models and components in a
proper manner in order to easily describe the appropriate wanted system. Each com-
ponent parameters are able to be updated according to the type, size, integration
capacity and others. The simulation results consist in graphical evolutions of different
parameters, according to the selected variables.
2.1 The Electric Machine Functional Model

The electric vehicle equipped with a 3-phase permanent magnet synchronous motor
(PMSM) is simulated using LMS.Imagine.Lab.AMESim. The software helps to ana-
lyze and predict the performance of the system. The virtual model is defined by
connecting different existing components from the software libraries, for which the
parameters are particularly set.
Modeling the Energy Evaluation for an Electric Machine 297
The electric machine has simple design and control, being able to act both as motor
and generator. It can generate torque from zero speed because of its permanent magnets
and high torque density, having small frame size, [5]. The motor pole pairs’ parameters
define the ratio between mechanical and electrical quantities, [3]. The motor pole pairs
represent the number of north and south segments the rotor contains. The magnetic
material saturation is neglected in the simulation model.
The battery package is modeled as variable voltage source and an internal resis-
tance, its behavior being evaluated by its state of charge and temperature. The battery
package is designed as different banks in series and parallel, while each bank is being
made by different cells.
The three-phase inverter is used to convert the direct current from the battery
package to alternating current to empower the three-phase PMSM.
The electric machine control unit consists in a control block, additional included for
using the torque command to define the corresponding stator phase current commands.
The control is based on an algorithm that require the parameters of the electric machine
for a proper operation. The desired control-loop uses the machine electrical parameters.
The control is characteristic to achieve the necessary data files. The control is also
based on the angular speed and the developed power.
The modeled virtual electric machine (Fig. 1) is using a rotary load connected to
the shaft in order to simulate the motion resistances. In addition, the electric machine
offers the information about its temperature using a thermal mass between the inverter
and the motor.
Fig. 1. The electric machine functional model and the torque control
The parameters used in simulation are presented in Table 1. The values are taken
from existing components that have proper operation and fulfill the performances
demands for the electric vehicle model.
Table 1. The electric vehicle parameters used for simulation.

Component Parameter Value
Electric Machine Type PMSM
Maximum power <80 kW
Maximum torque 140 Nm
Battery Output voltage 440 V
Cells in series 114
Nominal capacity 85,4 Ah
Transmission Type Single speed
Gear ratio 5
2.2 The Electric Machine Thermal Model

The electric machine is described in simple parts, each of it having its own shape and
volume. For each part a thermal mass [6] was associated. The thermal mass were
defined on the made off of the materials and the shaped volumes, [2]. Simple geo-
metrical shapes were used to define each of the selected parts.
The electric machine construction is used for defining the heat transfer between the
electric machine parts. The heat transfer is defined by conduction and convection
(Fig. 2). The conduction is taking place between all solid parts of the electric machine,
while convection is present between all the electric machine parts and the internal air
and between the stator and the external air.
Fig. 2. The electric machine model and the torque control
The major heat is generated in the windings and is exchanged by the other parts of
the electric machine. Considering that the electric machine cooling is using liquid, the
heat transfer is influenced by the liquid maintained temperature.
The geometrical parameters are used in describing the thermal effects and as input
parameters of the model components.
The thermal model is individually developed (Fig. 3), in a modular manner, being
able to be introduced as separate component of the virtual electric vehicle.
Fig. 3. The electric machine thermal model
The heat flow is defined by measuring the real time current intensity through the
three phases of the motor and based on that the windings produced heat is simulated.
The heat flow starts from the windings and the signal is divided between the stator and
the rotating parts of the machine. Assuming that 80 % of the heat generated from the
windings reaches the permanent magnets and the difference is transferred through
conduction in both parts of the stator, which is water and air cooled, the stator thermal
behavior is modeled taking into account its geometrical shape. Conductive heat transfer
generates different ways to determine the resistance [7], due to the shape of the
investigated part (for example the hollow cylinder presented in Fig. 4).
Each solid part has its own thermal properties with respect to the manufacturing
materials. The materials densities, the specific heat and the thermal conductivity have
the temperature dependencies that cannot be neglected. Taking into account that each
electric machine part has its own shape, the heat transfer between each of them is
modeled. Thermal properties such as density, specific heat and thermal conductivity are
defined for each solid component.
The thermal flow that is induced in the permanent magnets travels through con-
duction in the rotor and the bearing of the machine. Due to the fact that the outer ring of
Fig. 4. Radial heat transfer for a hollow cylinder [2]
the ball bearings has contact to the stator, a part of the heat will be transferred by
conduction back to the stator.
The internal air temperature distribution is possible between all the electric machine
parts that get in contact with the internal or the external air. The air properties are
changing when the temperature is modifying. The turbulent air flow are making the
temperature more unstable, introducing high difficulties in modeling the thermal
behavior, [2]. Therefore, the density of the internal air, the constant pressure specific
heat, the perfect gas constant, the absolute viscosity and the thermal conductivity are
very important to develop the thermal model corresponding to the internal air.
Similar assumptions as for the conductive heat transfer regarding the determination
method for the resistances were taken into account for the convective heat transfer.
Radial conductive exchange is used to represent the radial conduction through the
walls of the stator, [1]. The real parameters used are the mass of the material, the
internal diameter, the length and the wall thickness. For the other electric machine
parts, the thermal masses compute the temperature dynamics of their solid mass with
respect to the incoming heat flux.
2.3 The Electric Machine Simulation Model

The both functional and thermal models are linked in order to have a complete electric
machine simulation model (Fig. 5). The thermal model is using the three phases cur-
rent. The resulted model consists in two different modular models, independently
modeled and then adapted to run on the virtual electric vehicle.
The approach used in this paper to evaluate the electric machine efficiency by inves-
tigating the temperature exchange between electric machine parts have been validated.
The simulation considers that all the electric machine parts have the same temperature
of 20°C at the beginning of the simulation and the air relative humidity of 40 %.
In the first phase (Figs. 6.a), the simulation runs during 250 s, while the torque
value was kept at 50 Nm. Immediately after, during the second phase, the torque
Fig. 5. The electric machine simulation model
(a) (b)
Fig. 6. The electric machine rotary velocity for reaching the 50 Nm in 250 s
(a) (b)
Fig. 7. The rotor and the stator temperature evolution for reaching (a) the 50 Nm in 250 s and
(b) the 140 Nm in 250 s
command is cut off and the electric machines parts show the temperature decrease
(Fig. 7.a). The temperature for the rotor and the stator had a fast increase followed by a
decrease after cutting of the torque command. The temperature for the stator external
part that is in contact with the external air is higher by around 3°C than the temperature
for the stator internal part (Fig. 7.a).
The electric machine was also simulated to develop the maximum torque of 140 Nm
and to continuously run during 250 s at this value for the same general conditions
(Fig. 6.b). The maximum achieved torque was almost three time higher while the
temperature for the stator and the rotor was almost ten times higher. The difference
between the temperature of the stator internal part and external part reached very high
values. The temperature for the stator external part that is in contact with the external air
is higher by around 100°C than the temperature for the stator internal part (Fig. 7.b).
In both situations, after the torque command was cut-off, a significant temperature
drop can be seen on the interior of the stator as it dissipates the heat to the rotor and the
cooling liquid. The exterior of the stator continues to increase in temperature even after
the current was interrupted.
4 Conclusions
Detailed functional and thermal models were developed for being used for controlling
the electric machine and evaluating its efficiency. The detailed electro-mechanical
model was created by means of integrated approach that uses functional design for the
electric machines. In order to investigate the electric machine efficiency, its behavior
was investigated taking into account the geometry, the materials used for manufac-
turing the parts and the operation following the dedicated control. For each part of the
electric machine simple shapes were assigned and the solid parts were assumed as
thermal lumped masses. Using the lumped thermal model, the important role of the heat
transfer was highlighted. The geometrical shapes for each part has been investigated
with respect to any influences that may appear during operation. The thermal properties
of the electric machine parts were investigated in order to determine how the materials
properties influence the heat transfer and the energy distribution. The electric machine
configuration was defined by the thermal model complexity. The thermal exchange
between neighboring components was investigated. The air inside the electric machine
was investigated also. The thermal properties for the materials used to manufacture the
electric machine components were mandatory for developing the thermal models for
each component. The thermal behavior was described in simulation models using
mathematical equations.
The resulted models have modular design being appropriate for integration on
different electric vehicles’ architectures. The thermal model can be used to evaluate the
energy efficiency during different driving cycles, for the studied type and configuration
or the electric machine. The proposed solution can be easily replicated for other types
of electric machine.
The modeling and simulation approach by building virtual mock-ups for compo-
nents, sub-systems and systems used on vehicles are helping the entire research,
development and manufacturing industries allowing them to optimize the vehicles,
including the vehicles’ components and systems, long before it is physically built.
References
1. Anthonis, J., Santos, F., Croitorescu, V., Van der Auweraer, H.: Multiscale thermal and NVH
models for EV/HEV Integration of an SR-based Drivetrain. VPC, Pune, India (2011)
2. Croitorescu, V., Anthonis, J., Vasiliu, N.: Thermal modeling of an electric motor used on road
vehicle powertrain. In: Proceedings of the 10th Industrial Simulation Conference, Brno,
Czech Republic, EUROSIS-ETI Publication, pp. 223–231 (2012). ISBN: 978-90-77381-71-7,
EAN: 9789077381717, INSPEC, SCOPUS Code 104315
3. Croitorescu, V., Croitorescu, I., Danciu, G.: Functional modeling of an electric machine used
on road vehicles. In: Proceedings of the 8th International Symposium on Advanced Topics in
Electrical Engineering (ATEE), Bucharest, Romania (2013). ISBN: 978-1-4673-5980-1,
978-1-4673-5979-5, Accession Number: WOS:000332928500172, doi:10.1109/ ATEE.2013.
6563518
4. Croitorescu, V.: Vehicle Powertrain Simulation: a Useful and save costing methodology. In:
Proceedings of the European Simulation and Modelling Conference (ESM 2013), Lancaster,
United Kingdom, pp. 5–9 (2013). ISBN: 978-90-77381-79-3, ISI Web of Science, Accession
Number: WOS:000344910500001, INSPEC, SCOPUS Code 104385
5. Emadi, A.: Handbook of Automotive Power Electronics and Motor Drives. Illinois Institute of
Tehnology, Taylor and Francis (2005)
6. Faiz, J., Iranpour, R.: Thermal model for a switched reluctance motor of TEFC design during
steady state and transient operation. Electr. Mach. Power Syst. 26(1), 77–91 (1998)
7. Rouhani, H., Faiz, J., Lucas, C.: Lumped thermal model for switched reluctance motor
applied to mechanical design optimization. Math. Comput. Model. 45, 625–638 (2007)
Balancing of a Single Stage Reciprocating
Compressor with Elastic Elements
Doru Groza(&), Ioan Călin Roșca, and Gheorghe Alexandru Radu
Department of Mechanical Engineering, Faculty of Mechanical Engineering,

Transilvania University of Brașov, Brașov, Romania
dorugroza88@gmail.com, {icrosca,radugal}@unitbv.ro
Abstract. The nature of most reciprocating machines generates unbalanced

loads. Those are one of the main criteria in comparing performance of com-
mercial and domestic applications. Such applications include internal combus-
tion engines, hydraulic pumps, gas compressors, etc. The load variation depends
on parameters such as: power status (acting or driven), frequency, mechanism
geometry, component mass, etc. These parameters must be strictly controlled in
order to investigate their effect. Therefore an application that operates at con-
stant speed and load is preferred for further investigation. In the present paper a
balancing method is presented that uses elastic elements. The method is applied
on the slider-crank mechanism from a piston compressor. The method is
investigated via a Multibody simulation and the results are compared with the
original unmodified design.
The paper aims to analyze NVH behavior and wear reduction possibilities.
A realistic approach in achieving such a design is also considered.
Keywords: Dynamic balance NVH Spring balancing
1 Introduction
The dynamic loads that occur during the operation of the slider-crank mechanism imply
consequences on two different levels. From the user approach, an unbalanced machine
is noisy and generates unpleasing vibrations. The noise is in fact a criterion that is
legally regulated regardless of coming from household appliances or transportation
vehicles [1]. From the designer’s perspective load variations lead to increased
mechanical stress and wear. The wear has also two downsides: component fatigue and
parasitic energy loss. Therefore a balanced mechanism leads to improved comfort,
extended lifetime and reduced friction of the machine that uses it.
Friction occurs in four areas that concern the slider-crank mechanism: the journal
bearing; the crank bearing; the pin bearing; the piston guidance. The implied losses are
the main friction energy consumer in the machines that use the slider-crank mechanism.
This is valid even in complex structures such as the internal combustion engine from
which the piston friction being the main loss [2].
Further in this paper, a spring balancing method will be applied on a multibody
model of a piston compressor [3]. The force reaction in each joint and between the

306 D. Groza et al.
piston and the cylinder will be measured without and with various elastic elements. The
result will be finally discussed.
2 Multibody Model
The multibody simulation was carried out in MSC Adams. The software provides a
proper environment for the dynamic behavior along with additional excitations [4].
2.1 Dynamics
Components. Each component was designed according to a commercial product. Each
geometry, mass and center of mass was measured on the original components. The
simulated components do not replicate their genuine appearance but they maintain the
relevant physical parameters and the surfaces that generate friction.
A particularity of the piston compressor is the acting motor. The chosen model has
its crankshaft fixed on the rotor from a 220 V A/C motor. Therefore the crankshaft and
the rotor have been designed as a single component.
The kinematic input is the rotational speed of the rotor (Fig. 1).
Fig. 1. Multybody setup (left) long with its real geometric equivalent (right)
Links. All moving components have been linked according to the mechanism and the
complete mechanism has been linked to a body that has the same mass as the original
compressor body. The interface between the mechanism and the compressor body is
made via the sliding surface between the piston and the cylinder and two bearing: one
journal bearing and one bearing on the other side of the rotor.
Balancing of a Single Stage Reciprocating Compressor 307
As previously stated, the piston friction is the main contributor for frictional losses.
Therefore the piston-cylinder interface will be analyzed in terms of wear reduction
effects, while the other links will be monitored only for their absolute load value. The
rotational joints will also be reduced to ideal joints in order to have a precise mea-
surement on the piston friction force.
Additional forces. There are two other forces that act on the piston.
Gas force. A diagram was calculated considering the nominal parameters of the
compressor. Since the precision of the diagram is not an object in this paper, the
mathematical statements have been considered in ideal conditions. The gas pressure
was development was considered as in the next sequence: atmospheric pressure at
bottom dead center; in polytropic compression up to the nominal output pressure;
maximum pressure up to top dead center; polytropic expansion up to atmospheric
pressure; intake at atmospheric pressure. Figure 2 shows the pressure vs. piston dis-
placement that was input in the simulation.
Fig. 2. Pressure diagram for the piston compressor
Elastic force. The elastic element is coupled between the piston and the compressor
body. The spring is subjected to compression with the maximum force reached when
the piston is at bottom dead center. The spring aims to store the kinetic energy of the
reciprocating mass and take over part of the gas force, ultimately to decrease the load
acting on the bearings.
A commercial spring that fits the available space was considered.
3 Results
The multibody simulation has been carried out in MSC Adams. All variable data has
been defined relative to the crankshaft position. Measurements concerning bearing
loads have been made on vertical and horizontal directions. The normal force between
the piston and the cylinder head was also considered.
308 D. Groza et al.
Fig. 3. Variation of loads acting on the journal bearing (top), crank bearing (middle) and pin
bearing (bottom)
There is a visible improvement in the piston normal force reduction. The reduced
force implies reduced friction. Another aspect is the curve smoothness. Each pike on
the interrupted line curve indicates contact disruption. Each contact is accompanied by
noise. A smoother curve means that the mechanism is less noisy.
Balancing of a Single Stage Reciprocating Compressor 309
Fig. 4. Variation in normal force on the cylinder wall between the original design and the spring
balanced design.
4 Conclusions
Based on the above simulations one can conclude the following:

(a) The proposed balancing method has potential in reducing loads on all four links of
the slider-crank mechanism;
(b) The vertical loads acting on the bearings are reduced. However the peak value of
the load is the least influenced. The proposed construction has its counteracting
spring force at minimum value when the excitation has its maximum value
(Fig. 3);
(c) The lateral loads acting on the bearings are greatly reduced. From those the pin
bearing being relieved of lateral loads exception being in the proximity of top
dead center (Fig. 3);
(d) The contact force acting on the piston is greatly diminished. The greatest differ-
ence occurs roughly at midpoint between TDC and BDC (Fig. 4). The friction
losses are accordingly reduced;
(e) A spring balanced piston has potential in generating less noise than a conventional
piston.
The multibody model presented in this paper is an incomplete version of a proper
multibody model. A first step in further developing the model is designing the bearings
considering their clearance and hydrodynamic conditions. Furthermore a single spring
design has been considered for the simulation. A deeper investigation is required in
order to determine what spring rate is optimum for this application. Other applications
than a piston compressor should be investigated.
310 D. Groza et al.
References
1. Hoshi, M., Baba, Y.: A study of piston friction force in an internal combustion engine. ASLE
Trans. 30(4), 444–451 (2008)
2. Groza, D.: Echilibrarea motoarelor cu ardere interna ce folosesc mecanismul biela-manivela.
Transilvania University of Brasov (2015)
3. Ma, Z.D., Perkins, N.C.: An efficient multibody dynammics model for internal combustion
engines. Multibody Syst. Dyn. 10, 363–391 (2003)
4. * * * DIRECTIVE 2002/49/EC OF THE EUROPEAN PARLIAENT AND OF
THE COUNCIL of 25 June 2002 relating to the assessment and management of environmental
noise
The Calculation Algorithm
for the Determination of the Temperature
at the End of the Intake Stroke
for GDI Engines
Tudor Mitran(&), Vasile Blaga, Sorin Moca, George Dragomir,

and Dinu Fodor
University of Oradea, Oradea, Romania

tudor_mitran@yahoo.com, vblaga.ar@gmail.com,
sorinmoca1@gmail.com, {gdragomir,dfodor}@uoradea.ro
Abstract. The paper presents an algorithm for the calculation of the parameters
of the working fluid at the end of the intake stroke for a GDI engine. This is only
a part of a program developed by the authors, based on an own designed cycle
model. The intake process determines the further evolution of the working fluid
during the entire engine cycle. The initial parameters depend on the ones at the
end of the previous cycle. This algorithm calculates the temperature at the end of
the intake stroke and the adiabatic exponent of the intake process.
Keywords: Adiabatic exponent Polytropic exponent Internal energy

Residual exhaust gas coefficient Volumetric efficiency
1 Introduction
A calculation program was developed, in order to calculate and to represent the

tridimensional variation: of the adiabatic and the polytropic exponents, of the tem-
peratures and of the pressures in the characteristic points of the engine’s cycle, the
volumetric efficiency, the excess air factor, the thermal input of the unit volume of the
cylinder, the pressure increase ratio in the isochoric burning, the volumetric increase
ratio in the afterburning, the technical-economical parameters of the engine and the
injection’s duration. The program includes two sub-programs: one to calculate the SI
engine’s parameters (depending on the rotational speed and on the air-fuel ratio at
temperatures between –35 and +45°C and at environmental pressure) the other one to
calculate the SI engine’s parameters (depending on the rotational speed and on the
environmental temperature at an air-fuel ratio λ = 1 and at environmental pressure).
The algorithm presented in the paper is a part of the program, and it calculates the
parameters at the end of the intake stroke (the pressure and the temperature): depending
on the rotational speed and on the air-fuel ratio at temperatures between –35 and +45°C
and at environmental pressure and depending on the rotational speed and on the
environmental temperature at an excess air factor λ = 1 and at environmental pressure.
Also the adiabatic exponent for the intake process is calculated.

312 T. Mitran et al.
2 The Operational Scheme of the Proposed Propulsion

System
The energetic balance of the intake is realized based on the scheme presented in
Fig. 1 [1].
Fig. 1. The calculation scheme of the fresh charge intake temperature
where: p, V, T and m are: the pressure [MPa], the volume [dm3], the temperature
[K], the mass of the working fluid; the index (r1) refers to the initial state α (the piston
at TDC at the beginning of the intake stroke) and the index (a) refers to the final state β
(the piston in BDC at the end of the intake stroke);
The equation of energetic balance between the initial state α and the final one β of
the system cylinder-intake manifold is:
dQ ¼ dU þ dL ð1Þ
where: δQ – is the elementary heat transfer; dU – the infinitesimal variation of the

internal energy of the considered system; δL - is the mechanical work exchanged with
the environment;
The total heat exchanged during the intake stroke is:
Qpa ¼ DU þ Lab ð2Þ
where: ΔU – is the total variation of the internal energy of the considered system during
the intake stroke; Lα−β is the total mechanical work exchanged with the environment
during the intake stroke;
The Calculation Algorithm for the Determination of the Temperature 313
Denoting with:
Qpa Lab
na ¼ ¼ 1þ ð3Þ
DU DU
the coefficient that takes in consideration the influence of the heat exchange with the
walls Qpa on the variation of the system’s internal energy ΔU, one can obtain the
coefficient:
Lab
ua ¼ ð4Þ
DU
The mean adiabatic exponent of the intake process:
ka ¼ 1; 438 0; 525 104 ðTo þ Ta Þ:
where: T0 [K] is the environmental temperature; Ta [K] is the temperature at the end of
the intake stroke;
Further, the parameters from the Eq. (2) will be determined [1, 3];
The mechanical work exchanged by the system with the environment for 1 kg. of
intake fuel is:
!
Lab 1 1 To 1 1 To
lab ¼ ¼ ð1 þ cr Þ R Ta 1 1=k þ ð5Þ
mad e w e Ta e w Tr
where: mad [kg] is the intake fuel mass; γr is the residual exhaust gas coefficient; ε is the
compression ratio; ψ is the global pressure losses coefficient; Tr [K] is the exhaust gas
pressure; ke is the exponent of the exhaust process;
The initial internal energy of the system for 1 kg of intake fuel is [3]:
!
To 1 1 To 1 1 cvg
ua ¼ ð1 þ cr Þ cva Ta þ 1=k ð6Þ
Ta e w Tr e w e cva
where: cva is the mean specific heat of the intake air or fuel mixture; cvg is the mean
specific heat of the burned gases;
The final internal energy of the system for 1 kg of intake fuel is [3]:
ub ¼ ð1 þ cr Þ cva Ta ð7Þ
By introducing (5), (6) and (7) in (3), it results:
R 1 1e w1=k
1
e
TToa þ 1e w1 TTor
/a ¼ To 1 1 To 1 1 cvg ð8Þ
cva T e w T þ e 1=ke c 1
a r w va
knowing that:
R R
cva ¼ ; cvg ¼ :
ka 1 ke 1
where R – is the constant of the thermal agent, considered to be the same for air and for
burned gases;
the equation for φa becomes:
1 1e w1=k
1
e
TToa þ 1e w1 TTor
ua ¼ ka 1 To ð9Þ
Ta 1e w1 TTor þ 1e w1=k
1
e
kkaek
1
e
Denoting: na ¼ 1 þ kau1; the adiabatic coefficient of the intake process, finally it

a
results an equation with two unknowns Ta and Tr:

To 1 1 To 1 ka ke 1 1
¼ 1 1þ 1=k ð10Þ
Ta e w Tr na k e 1 e w e
The volumetric efficiency is:
e To e 1 1 To
gv ¼ ð1 wo Þ ð1 w o Þ ð11Þ
e 1 Ta e 1 e w Tr
The first link relationship between Ta and Tr is:
To
Ta ¼ gv e
ð12Þ
1wo e1 þ 1e w1 TTor
And the second one is [2, 4]:

k1
Tr 1 1 pd e
¼ ð13Þ
Ta ur w pr
By solving the system consisting of the Eqs. (12) and (13), it results:
( 1=ke )
e 1 wo ur ðe=dÞnd dnu
Ta ¼ To 1 ð14Þ
e1 gv e w a enc
The computing goes through the entire engine cycle for several times, performing
the operations until the optimum values for each equation introduced in the program are
obtained, in the condition of an imposed degree of accuracy. The initial data were
introduced for a variation of the rotational speed from 100 to 100 rot/min on an interval
from n = 500 RPM to n = 5500 RPM and for a variation of the environmental
temperature from 5 to 5°C on the interval from t0 = –35°C to t0 = 45°C. The following
parameters were calculated: the pressure in the intake pipe (pga), the intake pressure
(pa), the adiabatic coefficients of the engine’s cycle processes, the polytropic exponents
of engine’s cycle processes, the volumetric efficiency (ηV), the excess air factor (λ),
the coefficient (ξ) that represents the ratio between the quantity of air necessary for the
burning of the fuel and the intake quantity of the thermal agent, the thermal input of the
swept volume unit (qcb), the pressure increase ratio in the isochoric burning; the vol-
umetric increase ratio in the afterburning and the temperatures in the characteristic
points of the engine’s cycle [2].
In Fig. 2 is presented the variation of the adiabatic exponent of the intake process
(na) depending on the rotational speed (n) and on the environmental temperature (t0) for
an excess air factor λ = 1.
Fig. 2. The variation of the adiabatic coefficient of the intake process depending on the
rotational speed and on the environmental temperature
In Fig. 3 is presented the variation of the adiabatic exponent (na) in the intake
stroke depending on the rotational speed, for an environmental temperature t0 = 20°C
and an excess air coefficient λ = 1. One can observe that the adiabatic coefficient seems
to drop as the rotational speed rises. In fact, the variation of the adiabatic exponent for
the intake process is only of 0,03 %. The drop can be explained by the fact that,
although, the heat taken over by the fresh charge from the cylinder walls is smaller
because of the time in which the heat exchange takes place is shorter, the operational
temperature regime rises as the rotational speed increases.
In Fig. 4 is presented the variation of the adiabatic exponent depending on the
environmental temperature for the maximum power rotational speed (n = 5250 RPM)
and an excess air coefficient λ = 1.
In Fig. 5 is presented the variation of the intake temperature depending on the
rotational speed and on the environmental temperature for an excess air factor λ = 1.
For the rotational speed the values were taken in the interval from 500 to 5250 RPM,
and for the environmental temperature the interval is from -40 to 50°C.
rotational speed
environmental temperature
Fig. 5. The variation of the intake temperature depending on the rotational speed and on the
environmental temperature
Fig. 6. The variation of the intake temperature depending on the rotational speed
Figure 6 presents the variation of the intake temperature depending on the rota-
tional speed for constant environmental temperature and excess air factor.
As can be observed in Fig. 6, the intake temperature rises as the rotational speed
increases. This is a consequence of the fact that the operational temperature regime
rises as the rotational speed grows.
Figure 7 presents the variation of the intake temperature depending on the envi-
ronmental temperature for constant rotational speed and air excess factor.
As normal, the intake temperature rises as the environmental one increase.
Fig. 7. The variation of the intake temperature depending on the environmental temperature
3 Conclusions
All these values obtained by completing the engine’s cycle show a rather insignificant
decrease of the adiabatic exponents depending on the engine’s rotational speed, some-
thing subtle on the graphics, so they can be considered to be constant. The temperatures
in the characteristic points of the engine’s cycle increase by a few units as the rotational
speed raises, and the pressure in the intake manifold as well as the intake pressure have a
decrease with one-two tenths as the rotational speed rises [2].
In this program all the mathematical relationships were written under the form of
matrix depending on the rotational speed and on the excess-air factor, so that all of
them can be graphical represented depending on the rotational speed, on the excess-air
factor and on the environmental temperature [2].
References
1. Bataga, N.: Thermal Engines (Motoare Termice). Didactical and Pedagogical Publishing,
Bucharest (1979)
2. Blaga, V.: Contributions to the modeling of the gasoline injection at SI engines (Contributii la
model area injectiei de benzina la motoarele cu aprindere prin scanteie). Ph.D. thesis,
Technical University of Cluj Napoca, The Faculty of Mechanics, Cluj Napoca, Romania
(2000)
3. Delanette, M.: The engines with injection (Les moteurs a injection), Thechnical Edition for
Automotive and Industry (Edition Tehniques pour L’automobil et L’industrie) (1989)
4. Grunwald, B.: The Theory, the Calculation and the Construction of the Engines for Road
Vehicles (Teoria, calculul si constructia motoarelor pentru auto vehicule rutiere). Didactical
and Pedagogical Publishing, Bucharest (1980)
5. Radcenco, V.: Optimization Criteria for Thermal Processes (Criterii de optimizare a
proceselor termice). Technical Publishing, Bucharest (1977)
Diagnosing the Operation of a Locomotive
Diesel Engine Based on the Analysis of Used
Oil in the Period Between Two Technical
Revisions
Cătălin V. Zaharia1(&), Rodica Niculescu1, Victor Iorga1,

Cătălin Ducu1, Adrian Clenci1, and Bogdan Aron2
1
University of Pitești, Pitești, Romania
{catalin.zaharia,rodica.niculescu,victor.iorga,
catalin.ducu,adrian.clenci}@upit.ro
2
CFR IRLU, Bucharest, Romania
aron.bogdan@irlu.ro
Abstract. Technical inspection of a motor vehicle is an activity that should not

be neglected and that is deemed to be binding both within the warranty period,
but especially post - warranty. An engine malfunction may be due to unsuitable
use of fuels and lubricants, resulting in a decrease in lubricating performances
and protection of the oil. On the other hand, the wear degree and wear nature of
engine lubricant can provide clues related to possible engine faults, for example:
leaks for the combustion chamber, incomplete burn of fuel, improper thickness
of the oil film, etc. The paper aims to determine viscosity, density, water content,
fuel dilution and solid compounds of used oil. Based on these data will perform
diagnostic analysis of locomotive diesel engine operation.
Keywords: Diesel engine Diagnostic Wear Lubricant Water content

Fuel dilution
1 Introduction
Finding the problem that arises when complex systems are not functioning properly is
relatively easy to do, if there can be found a certain level of training. This knowledge
involves the following: understanding the system where the problem occurs, knowing
the ability to apply a diagnostic flowchart, determine their appearance: what the user
(system) notes, finding the fault or error that caused the symptom.
It is necessary to recognize that something is not functioning properly, by applying
the idea of knowing the system and then by combining knowledge with diagnostic
skills, better said, to be able to find out what’s not normal.
Diagnostic methods and maintenance, in the field of rail transport, are more
important than in the automotive field, since the purchase of new conveyance cannot be
done, for economic reasons, very often. For this reason, knowing the state of degra-
dation of lubricants can lead to determination of particular faults of the system. Thus,
the process of diagnostic is in close dependency with lubricant analysis. In this paper,

320 C.V. Zaharia et al.
the authors are describing a methodology of diagnosing the operation of a diesel

hydraulic locomotive engine (DHLe), methodology based on analyzing the used oil in
the period of two technical revisions.
Lubricants are very important in the life of oil machinery, for our case in the life of
locomotive diesel engine. Thus, it is important to follow and analyze oil degradation
state, such as contaminants (water content, soot, ash, glycol, fuel dilution, wear metals)
or change of physico-chemical properties, [1]. Some contaminants are suspected of
being active in generating conditions of faults developing and require a different
response than simple action of oil changing. For example, damage of seals leads to fuel
dilution, or glycol contamination cannot be remedied by changing the oil or switching
to high quality oil. So, the value of oil analysis in the early detection of problems,
nowadays, it has become mandatory. For that, an immediate approach can be repre-
sented by the in–service oil analysis. This action can provide trace information about
machine wear condition, lubricant contamination as well as lubricant condition. With
such parameters it can be possible to emit maintenance decisions, based on the oil
analysis results, in order to avoid oil mix up, have control on contamination and failure
analysis.
A comprehensive oil analysis suite may include tests such as Ferrography, or
SEM/EDX (Scanning Electron Microscopy - SEM with Energy Dispersive X-Ray
Analysis – EDX, [4]) which are both time consuming and expensive. These tests
provide detailed and definitive information about machinery wear, such as what the
wear particles are made of, where they come from, and how severe they are. Such
information provides reliability professionals with information on a past or imminent
failure.
There are several ways to perform in-service oil analysis. The most common ones
are outsourcing to an off-site laboratory, using an on-site laboratory, or performing
route-based oil analysis using portable tools. Modern predictive maintenance practices
require real-time data and immediate feedback so reliability engineers can make
maintenance decisions on the spot. Advances in oil analysis instrumentation tech-
nologies also make it possible for reliability engineers to perform on-site oil analysis
without years of training or deep knowledge of oil chemistry and testing. Outsourcing
oil analysis to an off-site oil laboratory is the oldest and most ordinary approach in the
industry. A common process flow involves collecting oil samples from source and
sending them to a specialized laboratory. Such a study, considered off-site analysis
scheme, on oil samples, it was carried out in the laboratory Fuels and Lubricants of
University of Pitesti and will be presented in the following chapters.
Thus, in this paper the authors aim to identify the faults which can appear on the
operation of a diesel engine locomotive and their causes, based on analyzing the
contaminants of the oil used for lubrication.
2 Monitoring on Lubricating Oil – Case Study on DHLe
In service oil analysis, for DHLe, involves checking to determine oil change period of
time and observation of possible operational problems, [2, 3]. To prevent negative
influences that the factors listed above could produce on diesel engine oil is necessary
Diagnosing the Operation of a Locomotive Diesel Engine 321
to check the oil daily or every two day, in terms of diluting and water content. Oil
sample is taken during the operation of the diesel engine, leaving at the beginning to
flow approx. 2–3 L. The oil sample will receive only 1 L of oil.
During the operation of diesel engine, even the best oil is subject to the phe-
nomenon of aging and soiling, for which the entire quantity of oil must be replaced
from time with new oil. In the case study, for DHLe the oil change is done at one and a
half year or 150000 km. In the case of oil dilution and emulsification the cause for the
occurrence of these phenomena will be searched and the faults will be remedied.
3 Chemico - Physical Characterization of Oil
3.1 Viscosity
Using oil analysis laboratories, lubricant manufacturers rely on viscosity measurements
as a key parameter in determining the condition of used oil, or specification of new oil.
For example, changes in viscosity of used oils can be an indication of lubricant
degradation (increase/decrease in viscosity), oil dilution (decrease in viscosity).
Equipment description. The U-VIsc is a fully automatic system that uses glass capillary
viscometers to measure the kinematic viscosity of Newtonian fluids. It can be applied
for mineral and synthetic oils (fresh and used) and for any fluid within the viscosity
range of the instrument. Figure 1 shows the equipment and its principal components.
Fig. 1. U - VIsc – viscometer
Method description. Standard Test Method for Kinematic Viscosity of Transparent and
Opaque Liquids D 445 is under the jurisdiction of ASTM Committee D-2 on Petroleum
Products and Lubricants.
The time is measured for a fixed volume of liquid to flow under gravity through the
capillary of a calibrated viscometer under a reproducible driving head and at a closely
controlled and known temperature. The kinematic viscosity is the product of the
measured flow time and the calibration constant of the viscometer, [5].
The time it takes for the sample to flow through a section of the tube is proportional
to viscosity as follows:
m¼Ct ð1Þ
where: ν = kinematic viscosity, [mm2/s], C = viscometer constant, [mm2/s2], t = flow

time, [s].
In order to correct for this kinetic energy effect, the result should be corrected.
Following Hagenbach and Couette theory, [6], the result should be corrected as
follows:
m ¼ C t E=t2 ð2Þ
The kinetic energy factor E [mm2•s], though is not a constant, can be derived from
the tube’s dimensions, such as capillary length and diameter.
3.2 Flash Point and Dilution

Flash point data can indicate the possible presence of highly volatile and flammable
material in a relatively non volatile or non flammable material, such as the contami-
nation of lubricating oils by small amounts of diesel fuel or gasoline.
Equipment description. In order to determine the oil flash point and dilution degree
Eraflash apparatus produced by Eralytics, Austria was used. Eraflash measurement is
directly under the latest and safest standards ASTM D6450 & D7094. It uses the
Continuous Closed Cup method for finding the Flash Point (CCCFP). In Fig. 2 is
presented the Eralytics analyser. Figure 3 shows the measuring chamber.
Fig. 2. Eralytics analyzer Fig. 3. Measuring chamber
Method description. The standard method which was used (ASTM D7094) is an
extremely versatile flash point method that can be used for the accurate flash point
measurement and also, to determine the influence to the flash point of highly flammable
fractions in samples like gasoline in diesel or diesel in oil or gas in oil. When the start
temperature is reached the sample cup is lifted. After the sample cup is lifted the
measurement is initiated. The electric arc ignites in the pre-programmed interval steps.
For this case, the interval is 1 s. When the pressure after ignition, in the measurement
chamber, is 20 kPa or higher, the flash point was detected.
In this paper, we determine oil dilution within diesel fuel. Thus, it was necessary to
measure reference fuel dilutions and draw the dilution curve.
3.3 Density
Density is an important property of petroleum products being part of product specifi-
cations. Materials are usually bought and sold on that basis or if on volume basis then
converted to mass basis via density measurements.
Equipment description. In Fig. 4 is presented the Density Meters DS7800 made by A.
KRÜSS OPTRONIC – Germany.
Fig. 4. Density meter DS7800 with drying unit and peristaltic pump
Method description. The Density Meters DS7800 works according the frequency
oscillator principle and covers the determination of the density or relative density of
crude oils with high vapor pressures (to prevent vapor loss during transfer of the
sample to the density analyzer). A small volume of liquid sample is introduced into an
oscillating sample tube and the change in oscillating frequency caused by the change in
the mass of the tube is used in conjunction with calibration data to determine the
density of the sample, [7].
3.4 Water Content

Contamination by water leads to emulsifying oil and reduction of lubricating capacity.
It also increases the viscosity of the oil, and oil additives may be separated leading to a
drastic reduction of lubricating capacity of oil.
Equipment description. In Fig. 5 is presented the Aquamax KFCoulometric made by
GRScientific – United Kingdom.
Method description. Combining the Coulometric technique with Karl Fischer titration,
Aquamax KF titrators determine the water content of the sample by measuring the
amount of electrolysis current necessary to produce the required iodine. This is an
absolute technique which does not require calibration of the reagents. The standard
calculation mode is:
Result ðppm or %Þ ¼ microgram count=ðV SGÞ ð3Þ
Where: V = volume of sample [ml], SG = specific gravity [g/cm3]
Fig. 5. Aquamax KFCoulometric. (1) Printer; (2) Display; (3) Keypad; (4) Generator BNC
Connector; (5) Detector BNC Connector; (6) Vessel Holder; (7) Vessel Retaining Thumb Screw
3.5 Determination of Oil Degradation - Oil Stain Analysis

Oil analysis on filter paper is known as the “oil stain method”. This analysis enables
assessment of capacity detergent - dispersant, the degree of contamination of the oil
with particles of carbon and chemical degradation by oxidation engine oil increasing.
Method description. The method is based on viewing the stain that a drop of oil leaves
on a paper filter. Usually, a stain of used oil consists of three zones (Fig. 6). The size
dispersion, [8]:
MD ¼ ðd=DÞ 100 ð4Þ
Where: d is the diameter of zone 1; D is diameter of zone 2.
The central zone 1 - indicates by his blackening the

degree of contamination with carbon substances
d
D Diffusion zone 2 - indicates the dispersion power of oil
Zone 3 - composed of oil separated from solid particles,

indicates the degree of oil oxidation and may be colored
from yellow to brow
Fig. 6. Oil stain

3.6 Metal Analysis X-Ray Fluorescence Spectrometry

The determination of debris in used lubricating oils is a key diagnostic method prac-
ticed in machine condition monitoring programs. The presence or increase in con-
centration of specific wear metals can be indicative of the early stages of wear.
Equipment description. EDXRF spectrometer, SPECTRO MIDEX M model, config-
uration: 30 W-Molybdenum X-ray tube generating a spot of 2 mm in diameter and
SDD detector with a resolution of 160 eV on Mn-Kα line.
Method description. Terms of the acquisition were:
1. direct excitation into atmosphere of He, slot 1 mm, operating voltage of
U = 44.70 kV, current intensity I = 0.30 mA, measuring time on acquisition 180 s
and the number of energy channels used is 2048.
2. direct excitation into atmosphere of He, slot 2 mm, operating voltage of
U = 18.87 kV, current intensity I = 0.30 mA, measuring time on acquisition 300 s
and the number of energy channels used is 1024.
The oil used in the case study was M25W40. The first step for dilution determination is
to realize the dilution curve (Fig. 7). The dilution value is shown in Table 1.
Fig. 7. Dilution curve Fig. 8. Filter paper oil stain
The obtained filter paper oil stain is shown in Fig. 8 (see also Fig. 6). We can see
that the chemical degradation by oxidation is not too high, the dispersion power of oil
is very good (MD tends to 100), diameters of central zone 1 and diffusion zone 2 are
almost equal, but the degree of oil contamination with particles of carbon is very high
(blackening stain). We can also notice that the circles of the stain are not smooth; this
means that the water content is considerable, in concordance with the water content
value obtained from Karl Fischer titration (Table 1). The other results obtained are
shown in Table 1.
Table 1. Test results on M25W40 oil

Property Flash Kinematic Dilution Density Water
point viscosity@400C [%] [g/cm3] content
[0C] [cSt] [%]
Used oil 184.1 95.44 0.90 1.485 0.2629
Fresh oil 220.7 113.6 0 0.890 0
Admissible 180.0 ±25 % 5 – 0.200
limit, [9]
About the density (Table 1), it is clear that high growth is a normal process, taking
into account solid contamination degree.
The kinematic viscosity and flash point decreased (Table 1), as expected, because
of the dilution with diesel fuel.
The qualitative analysis EDXRF (Energy dispersive X-ray fluorescence) of
understudy oil samples is shown in the below spectra. From the analysis of fluorescence
spectra we can notice the presence of Cu in the used oil sample, as well as a great
amount of Fe (Figs. 9, 10).
Fig. 9. Acquistion mode 1 – specific spectra Fig. 10. Acquistion mode 2 – specific spectra
5 Conclusions
Comparing our results to the limits of the values described in Table 1, it is possible to
emit the following conclusions:
– regarding FLASH POINT the measured value (184,1°C) it is very close to the limit
(180°C); it can be proposed that the oil be changed in order to avoid explosion risk,
– in terms of FUEL DILUTION the obtained value (0,9 %) is not to high; according
to Table 1, the maximum value is 6 %; by analyzing the components of injections
system, it seems that the injection pump has some sealing problems, which can
cause the appearance of fuel dilution,
– the DENSITY raised from 0.89 to 1.485 g/cm3 and here we can say that metal wear
and contaminants contributed to this phenomena; for example, in engines, iron can
come from cylinders liners, rings, crankshaft, camshaft, oil pump gear. Copper can
come from main and rod bearing, turbo bearing. The anti-wear additive can be the
source for zinc, but in our case the additives from the used oil are not yet trans-
formed in contaminants,
– regarding the WATER CONTENT it can be noticed that the obtained value is
higher than the limit, the used oil must be changed and there can be proposed a
investigation process of cooling system in matter of sealing,
– the blackening of OIL STAIN may be the result of a improper combustion process
and it’s calibration is recommended to be done.
References
1. Henning, P., Walsh, D., Yurko, R., Barraclough, T., Shi, A., Yuegang, Z.: Oil analysis
handbook for predictive equipment maintenance. Spectro Scientific (2014)
2. Cheșa, A.: Locomotiva diesel hidraulică de 1250 CP - întreținere și exploatare, Editura ASAB,
București (2001)
3. Juoperi, K.: Lubricating oil quality – safe power plant operation depends on it. Energ. News,
Issue 14, 12–15 (2012)
4. Jensen, C.C.: Clean Oil Guide. The importance of oil maintenance, Tryc Team, Svendborg,
Denmark, ver. 007 (2015)
5. An American National Standard, British Standard 2000 Part 71:2000, Designation D: D 445 –
97 1990, Standard test method for kinematic viscosity of transparent and opaque liquids - The
calculation of dynamic viscosity (2000)
6. Hochschule Anhalt, J.W., Magdeburg, H.K., Rheinfelden, J.H.: Theory and Praxis of
Capillary Viscometry (Dieter Wagner Schott – Gerate Gmbh)
7. Kishore, R.: A: Guide to ASTM Test Methods for the Analysis of Petroleum Products and
Lubricants. Second edn. ASTM Stock No. MNL 44-2nd (2007)
8. Gheorghisor, M.: Carburanti si lubrifianti pentru autovehicule, Editura TIPARG (2012)
9. Ministerul transporturilor, Ordinul nr. 363/2008 privind aprobarea Normei tehnice feroviare -
Vehicule feroviare (2008)
Heavy and Special Vehicles
Simultaneous Influences of Tyre Pressure
and Steering Geometry upon the Theoretical
Speed Ratio in the Running Gear System
of Four-Wheel Drive Tractors
Mircea Nastasoiu(&) and Nicolae Ispas

{m.nastasoiu,inicu}@unitbv.ro
Abstract. The paper develops a mathematical model in order to evaluate the

theoretical speed ratio in the running gear system of four-wheel drive tractors,
taking account of both the tyre pressure and the turning radius. In using this model,
we may emphasize situations when the recirculating power (parasitic losses
power) occurs in tractor’s transmission due to inequality of wheels’ theoretical
speeds on the two axles. The paper presents applications of the mathematical
model developed for the tractor MAT 81, manufactured in Craiova, Romania.
Keywords: Four-wheel drive tractor Steering geometry Theoretical speed

ratio
1 Mathematical Modelling of the Theoretical Speed Ratio

in the Running Gear System of Four-Wheel Drive Tractors
in a Straight Line Motion
In order to study the theoretical speed ratio of four-wheel-drive tractors with rigid
interaxial coupling the specialist literature uses the theoretical speed ratio, defined as
ratio of peripheral speeds of the wheels on the two axles:
vt1 ¼ vt2 or xm1 r1 ¼ xm2 r2 ;
where: xm1 ; xm2 are the angular speeds of front and rear wheels respectively,
expressed in 1/s; r1, r2 are the corresponding dynamic radii, expressed in m.
However, we must mention that by using the theoretical speed ratio as such, we
cannot take into account the simultaneous influence of several factors upon the theo-
retical speed ratio. In other specialist papers [1] this ratio is defined as follows:
1 d2
K¼ ; ð1Þ
1 d1
where: δ1 - slip of front wheel; δ2 - slip of rear wheel.

The K ratio is called the theoretical speed ratio [2, 3]. The operation of
four-wheel-drive tractors indicates a maximum efficiency provided K = 1.
332 M. Nastasoiu and N. Ispas
For a particular value of the K ratio, there is a linear dependence between the slips
of the two axles:
d2 K 1
d1 ¼ þ : ð2Þ
K K
Many papers [1–3] demonstrate that in such cases, when K ≠ 1, the phenomenon of
interior power flow occurs in tractor’s transmission, known as parasitic power.
In case the tractor is in straight line motion, the theoretical speed ratio becomes [1, 2].
1 d2 i2 r1
kr ¼ ¼ : ; ð3Þ
1 d1 i1 r2
where: i1 is the gear ratio of the front axle transmission; i2 - is the gear ratio of the rear
axle transmission; r1 and r2 – corresponding wheels radii.
Taking account of the radial deformation of tyres, which has the greatest influence
upon the change in the wheels radius, we obtain the following relation for Kr ratio:
i2 r01 Z1 =2kr1
kr ¼ ; ð4Þ
i1 r02 Z2 =2kr2
where: r01 and r02 are the free radii of the wheels with no load to act upon (not even the
wheels’ own weight), expressed in m; Z1 and Z2 are the normal load on the front and
rear axle, expressed in N; λr1 and λr2 – radial elastic constant of tyres, expressed in
N/m.
The elastic constant λ may be determined with Hedekel formula [1], written under
the form:
pffiffiffiffiffiffiffiffiffi
kr ¼ 2ppa r0 rb ; ð5Þ
where: pa - tyre pressure, expressed in Pa; rb - tyre rolling radius in cross section,
expressed in m.
Using relation (2), we can determine the front axle slip δ1:
d2 Kr 1
d1 ¼ þ : ð6Þ
Kr Kr
2 Mathematical Modelling of Theoretical Speed Ratio

in the Running Gear System of Four-Wheel Drive Tractors
During Cornering
In most four-wheel drive tractors of low and average power, the cornering is completed
by changing the heading of the front wheels through the steering system. Many
deviations from the kinematic concordance occur in the running gear system during
cornering. This results from the cornering kinematics (Fig. 1).
Simultaneous Influences of Tyre Pressure and Steering Geometry 333
Fig. 1. Steering geometry.
The speed of each axle centre may be determined depending on the angular speed
ωv during cornering and the corresponding radii R1 and R2 to the centre of rotation.
v1 ¼ xv R1 ; v2 ¼ xv R2 :
On the other hand, the same speeds may be determined with the relations:
v1 ¼ xm1 r1 ð1 d1 Þ; v2 ¼ xm2 r2 ð1 d2 Þ:
Consequently, the ratio between speeds v1 and v2 will be:
v1 R1 xm1 r1 1 d1
¼ ¼ :
v2 R2 xm2 r2 1 d2
Using the definition relation (1), we obtain the following expression for the theo-
retical speed ratio in the running gear system during cornering:
1 d2 xm1 r1 R2
K¼ ¼ : ð7Þ
1 d1 xm2 r2 R1
The ratio xm1 r1 =xm2 r2 ¼ kr represents the theoretical speed ratio for a tractor in
straight line motion and, as it results from Fig. 1, R2 =R1 ¼ cos a ¼ k. Therefore, we
obtain the following mathematical relation for the K ratio:
ai þ ae
K ¼ kr cos a ¼ kr cos ¼ kr k; ð8Þ
2
where: αi - steer angle of inside front wheel; αe - steer angle of outside front wheel
(Fig. 1).
In tractors with all steering wheels as well as in tractors with articulated chassis,
R1 ¼ R2 . Consequently, K ¼ kr ; that is to say the theoretical speed ratio remains the
same during cornering as when driving the tractor in straight line.
3 Effect of Theoretical Speed Ratio on the Slips of the Front

and Rear Tyres of a Four-Wheel-Drive Tractor with Rigid
Interaxial Coupling
The study on the interaction between the tractor’s running gear system and the ground
shows that when the wheels slip is positive they become driven wheels. If, for example,
the slip of the front wheels is negative (d1 \0), the sign of the tangential force and of
the torque on these wheels will change, which means that the front axle brakes the
tractor. In this case instead of pushing the tractor body, the front axle is pushed by it.
Due to the negative moment of torsion of front wheel, the power on the front wheels
flows in a closed circuit: front wheels – transmission – rear wheels – tractor body –
front wheels. This power does not work in case of a supplementary traction force and,
therefore, it is not a useful but a detrimental power, which is the reason why it is often
called parasitic power (or recirculating power).
The flow of power in closed circuit is caused by the inequality of the theoretical
speeds of front and rear wheels respectively, included in a closed system: the above
mentioned flow plus the ground.
If the rear wheels slip is negative (d2 \0), the phenomena mentioned above are
alike.
Consequently, the negative sign of the slip may be used as criterion for the
occurrence of the phenomenon of parasitic power flow.
For a particular value of the K ratio the change of the slip sign may be analytically
explained using relations (2) and (7). For instance, in the particular cases below:
K1
ða) if d2 ¼ 0; d1 ¼ ; ð9Þ
K
ðb) if d1 ¼ 0; d2 ¼ 1 K: ð10Þ
4 Applications of the Mathematical Model Developed
The following 4 figures graphically illustrate applications of the mathematical model

developed. The applications are for tractor MAT 81, with a nominal engine power of
58 kW (manufactured at MAT Craiova, Romania), with front tyres 11.2 R 24
(r01 = 546 mm) and rear tyres16.9 R 30 (r02 = 740 mm), with a total weight of
3627 kg, distributed as follows: 1540 kg on the front axle, 2087 kg on the rear axle. In
this tractor the ratio i2/i1 = 17.077/12,217 = 1.398.
The theoretical speed ratio kr was determined for different levels of tyres pressure
(deviations of pressure of ±50 kPa as compared to the values recommended by the
manufacturer). The theoretical speed ratio k was determined for various turning radii R2
(R2 = 5…50 m).
The tyres pressure (as recommended by manufacturer) when the tractor is stopped
is p01 = p02 = 140 kPa. The theoretical speed ratio kr (in straight line motion) was
determined for the following cases:
1. p1 = p01 − 50 kPa; p2 = p02 + 50 kPa; in this case we obtained kr = 0.982;
2. p1 = p01 + 50 kPa; p2 = p02 − 50 kPa; in this case we obtained kr = 1.083;
3. p1 = p01; p2 = p02 − 50 kPa; in this case we obtained kr = 1.066;
4. p1 = p01; p2 = p02 + 50 kPa; in this case we obtained kr = 1.016;
5. theoretical model: kr = 1.000, in this case k = K.
Therefore, the following matrix was used for the coefficient kr in calculations:
kr ¼ ½0:982 1:000 1:016 1:066 1:083:
As seen in Fig. 2 the values of the theoretical speed ratio kr, taking into account the
tyre pressure are significant, but the variation of the theoretical speed ratio considering
the turning radius is obviously constant. Instead, the variation of the theoretical speed
ratio k depending on the turning radius is considerable. At minimum turning radii, the
value of this ratio is just below 0.9.
Fig. 2. Variation of k and kr ratio with turning radius R2
Figure 3 presents the variation of the coefficient K (K ¼ kr k). It represents the

simultaneous influence of tyre pressure and cornering upon the theoretical speed ratio
depending on the turning radius R2.
Fig. 3. Variation of ratio K = krk with turning radius R2, for 5 values of the kr ratio
The slip values δ1 and δ2 were also determined depending on the turning radius R2
(Figs. 4 and 5). The situations when the slip is negative, see Sect. 3, illustrate the
phenomenon of parasitic losses power (recirculating power).
Fig. 4. Variation of front wheels slip with turning radius R2, for 5 values of the kr ratio
Fig. 5. Variation of rear wheels slip with turning radius R2, for 5 values of the kr ratio
5 Conclusions
The main outcome of the paper is the development of a mathematical model that allows
us to analytically study the simultaneous influence of both tyre pressure and cornering
upon the theoretical speed ratio of four-wheel-drive tractors with rigid interaxial
coupling.
The mechanics of these tractors presents important particularities. An essential
particularity is that in these tractors the inequality of wheels’ theoretical speeds on front
and rear axles may considerably influence the mechanics of the tractor, that is to say a
driven axle may become a braked axle. In this case the phenomenon of interior power
flow occurs in the tractor’s transmission, known as parasitic losses power. The
mathematical model developed in this paper allows us to pinpoint the conditions that
favour this phenomenon.
The negative consequences of the theoretical speed ration in the running gear
system explains the manufacturer’s recommendations to rigorously consider the values
prescribed for tyre pressure and not to operate the tractor with the coupled front axle on
roads with good adhesion. Likewise, the front axle should not be coupled in case of
tight turning radius. On roads with bad adhesion the theoretical speed ratio is coun-
terbalanced by different slips of wheels on the two axles.
The authors of this paper have remarked that in case of the tractor subjected to
analysis, MAT 81, the theoretical speed ratio is significant even with the tyre pressure
recommended by the manufacturer (kr = 1.031). If, for example pressures p01 = 120
kPa and p02 = 160 kPa were to be used, the theoretical speed ratio would approach to
the value 1 (kr = 1.014).
References
1. Nastasoiu, S., Popescu, S., Andreescu, C., Fratila, G., Cristea, D.: Tractoare. Editura
Didactică şi Pedagogică, Bucureşti (1983)
2. Nastasoiu, M., Ispas, N.: Influence of tyre pressure and weight distribution on axles on the
theoretical speed ratio in the running gear system of four-wheel drive tractors. In: Andreescu, C.,
Clenci, A. (eds.) Proceedings of the European Automotive Congress EAEC-ESFA 2015,
pp. 695–701. Springer International Publishing, Cham (2016)
3. Wong, J.Y.: Theory of Ground Vehicles. John Wiley & Sons Inc., New York (1999)
Analysis of the Comfort of the Seats Used
in Light Military Vehicles
Marian Mitroi1(&) and Anghel Chiru2

1
Ovidius University, Constanta, Romania
ejju_marian@yahoo.com
2
Transilvania University of Brasov,
Brasov, Romania
achiru@unitbv.ro
Abstract. The main postural stress that affects the human body during long
journeys, off roads, are generated by the vibration induced to the undercarriage
and vehicle structure via the seat as well as by posture in the chair, which affect
the spine. The amortization of shock and vibration occurring during the travel,
associated with a variable speed movement, and also braking or the maneuvers
performed, is a very important element to ensure a high degree of comfort. The
comfort of persons from the vehicle exposed to vibrations depends on the entire
suspension system of the vehicle, but especially by the solution chosen for seats
design.
Keywords: Comfort Vibration Body pressure
1 Introduction
The comfort feeling could be appreciated as an assembly of mental and physical

reactions due to all the factors involved in this state which act on it in different ways.
Driving a military vehicle during specific mission is a complex action characterized
by the quick succession of the external factors and the reaction rate of the drivers. The
effect of the vibrations on the vehicle crew must not be neglected [1].
The rugged terrain generates random vibration of different amplitudes, in general
with very high level. These vibrations may increase seriously the effect on the human
body, in the condition of the driving under emergency. The route of driving (off road)
being very varied and complicated like physical structure, the stresses that occur are
multidirectional and create resultant forces much higher than the stationary, on level
terrain. It can be said that the characteristic factors of displacements accentuate stresses
physical, psychological and biological, on the human factor inside the car and gener-
ates a strong discomfort.
The research objective is represented by identifying vibration values perceived by
the driver of a special vehicle, when driving in a very rugged terrain, under emergency
situations, values that affect the level of comfort.

340 M. Mitroi and A. Chiru
The objectives set by the authors is to identify risk values of vibration to the human
body, and also to find a new solution to reduce dangerous values on which the driver is
subjected on the seat, and to increase the level of comfort.
2 The Stresses Transmitted to Seat
The comfort offered by cars, in all its aspects, it was approached by many specialists
over the years. Nowadays, the specialists it attaches a great attention because of the
increasing demand of the consumer society.
The existing seats in each vehicle, no matter the form, structure or fixing system
take over from body platform, a number of vibrations. They are vertically transmitted
to the human body, reducing its level of comfort during driving.
The seats existing in military vehicles, not present in general same comfort like
existing on civilian vehicles. Light military vehicles are made on civilian structure and
after were adapted for the fulfillment of goals military. During research, has been taken
into account analysis of the comfortable level generated by the driver’s seat in such a
vehicle, by observation the frequency values and the amplitudes of vibration, and the
changing posture in the chair. The displacement was performed on a accessible land
under time limit, of anti-terrorism intervention, with a VW T4 car 130 horsepower.
Next image (Fig. 1) presents the vehicle under investigation.
Fig. 1. VW T4 light military vehicle
During the travel, over the seat system have been transmitted a very wide range of
vibrations, which have been identified by vibration sensors and accelerometers installed
by the authors, in a custom configuration, on the vehicle. An accelerometer was
mounted on the surface of the bottom bracket in close proximity to seat, a vibration
sensor on the metal support of the chair, and a vibration sensor on the seat of the chair.
The vehicle seats are mounted in normal configuration (Fig. 2), on the metal
backing, and having sponge structure.
Analysis of the Comfort of the Seats 341
Fig. 2. The seats of VW T4 used, and the vibration sensor (type 4447)
The signal received on the surfaces of bottom bracket generates values shown in the
following chart (Fig. 3).
Fig. 3. The signal measured in the time domain on vertical direction, on the bottom bracket
The signal received on the surfaces of seat of the chair with the help of sensor 4447
generates values shown in the following chart (Fig. 4).
In both graphs it is observed that the values recorded, in general, are in the area of
perception of the state of discomfort for the driver: more than 1,6 m/s2 RMS (ISO
2631:1).
Fig. 4. The vertically signals measured in time, on the driver’s seat
3 Data Analysis and the Influence Over Comfort
Looking at the graphs shown in the figures above, for the road segment analyzed it can
be stated that:
– Generally, the accelerations have values between 1,6–2,5 m/s2 RMS; however there
are many higher values in the structure of bottom bracket of the car, which are
reduced at the level of the seat.
– According to ISO 2631-1, as from the value of 1,6 m/s 2 RMS, uncomfortable state
is increasing, which means that during the experiment the driver has felt different
conditions of discomfort. In the same way (ISO 2631-1, Annex C and BS 6841), in
the following table (Table 1) can be observed weighted values of the accelerations
and their influence on the comfort.
Table 1. Guidelines ISO 2631 and BS 6841 for vibration and human perception
Values (m/s2) Perception
<0.315 Not uncomfortable
0.315–0.63 A little uncomfortable
0.5–1.0 Fairly uncomfortable
0.8–1.6 Uncomfortable
1.25–2.0 Very uncomfortable
>2.5 Extremely uncomfortable
The graph on the recorded values at the level of the seat (Fig. 5) demonstrates that
these are reduced to the passage by seat structure but are refound to the limit values for
discomfort, in these conditions of road. They meet and extremely high values in some
portions of the road, which are the equivalent of shocks products when driving, being
correspondences of some unevenness, extremely pronounced.
The routes with terrain rough generating a lot of variation of vibration, and have the
effect of negative impact on staff comfort in the vehicle. Vertical vibrations are
transmitted through the bottom bracket of the car and the seats, to the human body,
following the axis of the spine [2]. The line of transmission of vibrations and shocks
towards spine in seating position is shown in Fig. 5.
Fig. 5. The line expressions of shock in a seated position
These values have negative effects on the spine, by compressions which occur in the
vertebrae L1 - L5. An increased pressure in car seat is generated by the bending of the
spine, through muscle contraction and ligaments. A high pressure in lumbar spine creates
an increased state of discomfort, especially in military applications. It should not exceed
the value of 0.8 N/cm2 [3]. Also, the expression of body pressure in the seat and the
human perception of discomfort were analyzed: Andreoni-2002 [4], Kyung-2008 [5],
according to the following image (Fig. 6).
Fig. 6. The level of pressure in seat: left, regular person; right, heavy person
However, in case of the military body, the pressure exerted at the level of seat is
much higher because the soldiers have a more robust physical construction, but also as
a result on the weight of the military equipment used in missions (Fig. 7, Table 2).
Fig. 7. The changing military position in the car seat
Table 2. The additional weight which affects the pressure exerted by the military body on the
seat surface
Materials Weight
Protection helmet 3,7 kg
Protective helmet visor 1,2 kg
Bulletproof vest (level IV) 15,0 kg
Guns 1,7 kg
Specific equipments 0,8 kg
TOTAL 22,4 kg
The increased pressure of body in the seat generates at the same time an increase in
blood pressure to a high level, so an amplification of the discomfort (Table 2).
3.1 The Multidirectional Analysis of the Vibration

The human body receives the vibrations transmitted by the vehicle structure according
to an orthogonal coordinate system, whose accelerations ax, az, ay are specified in ISO
2631. The rough trails create major difficulties of vehicle regardless of their con-
structive characteristics. At the same time it generates a high degree of discomfort for
everyone due to the low frequency transmitted: 0.5–5 Hz, frequency to which the
Fig. 8. Multidirectional vibration - sequence of the route taken
human body is very sensitive. The following image (Fig. 8) shown a sequence with
multidirectional vibrations recorded in the experiment.
It the above diagram, there can be noticed the sections in which accelerations is
manifested predominantly in the directions of propagation x (horizontal) and y (lateral).
There are situations when these values are growing instantly very much because of
pronounced unevenness of the pitch.
However, the main problem is generated by situations when these values are
combined with vertical accelerations (az), amplifying their effect of them.
The discomfort associated with such areas is very pronounced, and in the case of
constant travel or prolonged are appearing negative effects on the human body.
The vibrations generated by vehicles on the movement time they are harder to
define, expressing accelerations via RMS and RMQ. This is due to the fact that these
are mean values and do not grow at the same time with increasing duration at steady
state signals, they decrease with increasing duration measurement, in the case of
non-stationary signals. Because of this fact, in the case of rough trails the most
appropriate method for determining values is VDV (vibration dose value).
This method exploits the r.m.q. without the divide the exposure times, so it can be
used to measure the vibrations received, of any kind.
Relations between the three methods of analysis are shown below:
Z t¼T 12
1
R:M:S: ¼ a2w ðtÞdt ð1Þ
T t¼0
Z t¼T 14
1
R:M:Q: ¼ a4w ðtÞdt ð2Þ
T t¼0
Z t¼T 14
VDV ¼ a4w ðtÞdt ð3Þ
t¼0
where
aw(t) is the frequency weighted acceleration;
T is the period during which the person is exposed to vibration.
4 Conclusions
In order to obtain conclusive results concerning vibrations that have direct influence on
the state of comfort, proper analysis is needed of all vibrations transmitted from the
road surface up to the level of the chair seat.
Ensuring an adequate level of comfort of staff in the vehicle during travel on rough
terrain depends on the entire depreciation system, but especially by the solution chosen
for the achievement of the seat.
The comfort associated to the seats comprises two states:
– Static - related to postures in the seat, inclination angle of the backrest.
– Dynamic - related to vibrations transmitted through the seat. Here, intervenes
depreciation structural and internal materials amortization.
These characteristics must be analyzed in relation to the vehicle and passengers
because the assessment is purely subjective.
The effects oscillation on the human body depends on the value of accelerations,
duration of exposure and the composition of the respective oscillation frequencies.
For the analysis of the frequency response of the human body to vibration during
travel in rough terrain, through the chair, it is indicated for use speed of the vehicle as a
reference parameter. The comfort of human factor while driving depends on the
propagation direction of oscillation, the vertical direction (az) being the most important.
4.1 Research Directions for Improving Comfort

Taking into account those presented and by referring to the frequency values with
negative influence on the human body we can establish some lines of research, to
increase the level of comfort of the seats for these types of vehicles.
Thus we can indicate the following measures:
– Use of a large number of sensors to providing conclusive data as to oscillatory
phenomena of the seat.
– Use of new structural components to the seat and the backrest to prevent vibration
range 4–8 Hz.
– Development of composite materials for the chair seat, to reduce blood pressure in
the thighs, on the long traveling.
– Structural adjustment of the seat backrest, in the lumbar spine zone to reduce
pressure on the L1-L5, shock and vibrations of the road.
– Realization of a seat suspension system, so as to obtain a natural frequency as low
(<1.5 Hz) but also, a more effective damping of vibrations.
Fig. 9. Scheme of influential parameters of the military comfort, in the vehicles
By analyzing the data resulted after the experiment and the setting of general
conclusions, the authors identified factors that influence the comfort of military per-
sonnel, but they also, made a diagram of them (Fig. 9).
According to those presented, the authors think that the most realistic method to
assess the level of comfort of the driver and passengers during journeys in very rough
areas is correlating various vibration analysis techniques (BS6841, VDI 2057, ISO
2631, VDV). It is also very important, field data correlating of the vehicle, and
appreciation of military personnel.
References
1. Aramă, C., Mitroi. M., Aramă, L., Aramă, M.: Considerations regarding the improvement of
the military drivers seat comfort for the military special intervention vehicles. In: International
Conference Scientific Research and Education in the Air Force, AFASES 2016, Brasov
(2016)
2. Mitroi, M., Aramă, C.: The implications of random vibrations generated by rough trails and
hardly accessible on military body. In: International Conference Scientific Research and
Education in the Air Force, AFASES 2016, Brasov (2016)
3. Zenk, R., Franz, M., Bubb, H.: Spine load in the Context of Automotive Seating. Society of
Automotive Engineers. Inc., Warrendale, PA, USA, Technical Paper No. 2007-01-2485
(2007)
4. Andreoni, G., et al.: Method for the analysis of posture and interface pressure of car drivers.
Appl. Ergon. 33(6), 511–522 (2002)
5. Kyung, G., Nussbaum, M.A., et al.: Driver sitting comfort and discomfort (p.II): relationships
with and prediction from interface pressure. Int. J. Ind. Ergon. 38(5), 526–538 (2008)
Aspects Regarding the Kinematic Optimization
of a Tracked Military Vehicle’s Transmission
Octavian Alexa(&), Radu Vilău, Ovidiu-Constantin Ilie,

Daniela Voicu, and Marin Marinescu
Military Technical Academy, Bucharest, Romania

alexa.octavian@gmail.com
Abstract. This paper concerns the development of a set of simulating models

able to assess gear sets. The modeled gear set is an intermediate gear drive, a
component of the drive line of a military armored tracked vehicle. It is
responsible of transferring the power from the vehicle’s engine to the hydraulic
torque converter of the hydro-mechanical transmission of the vehicle. The
intermediate gear drive has been modeled with the aid of a specialized simu-
lation module of Matlab® programming environment. The components of the
intermediate gear drive make energetic network, able to reproduce both the
physical structure and the real working mode within the basic system. The paper
concludes by comparing the dynamic performances of the tracked vehicle for
both the options: with the basic layout of the intermediate gear drive and the
suggested one within the developed simulating operations.
Keywords: Simulation Torque Tracked vehicle
1 Introduction
Many papers study the dynamic behavior of the gearing, as a specific case of the
vehicle’s dynamics [1–3]. In this paper we developed a simulation model to assess the
working modes of a driveline’s gear set. The modeled intermediate gear drive belongs
to a complex transmission of a battle tank and it is responsible to transfer the power
flow from the engine to the torque converter and to drive some auxillary devices. The
models were developed using the specialized Simscape module of MatLab®. They
reveal, by their network, the elements that effectively take part in transferring the power
flow from one point to another. The input is the engine’s torque and the model allows
computing the torque, the angular speed and the power at any point of the network,
both in steady and transient regimes.
After performing the analysis it was made a layout simplification, getting a mass
decrease of about 10 %. Implementing the new layout could lead to an improvement of
the vehicle’s performances. Obviously, the fuel consumption could decrease. To reveal
the modifications, Fig. 1 depicts the kinematic structures of the both layouts (old and
new).

Aspects Regarding the Kinematic Optimization 349
Fig. 1 Kinematic layout of the transmission (left - old layout, right - new layout)
2 Intermediate Gear Drive Analysis by Simulating Its Work

Within the General Simulation Model
The general simulation model (Fig. 2) is designed to simulate the work of the vehicle in
driving conditions. The taking-off process was studied.
Fig. 2 General virtual simulation model of the vehicle, realized in Matlab Simscape
350 O. Alexa et al.
The models of the basic intermediate gear drive and of the new simplified one
(Figs. 3 and 4) were developed in the Simscape module of MatLab®. The gears’
behavior has been described by the equations providing the mechanical elementary
gear work [2, 5]:
MB ¼ MF gFB iFB
Min þ IF eF þ MF ¼ 0 iFB ¼ xxFB ¼ eeFB ð1Þ
MB þ IB eB þ Mout ¼ 0
where: MB, MF are the torques on the driving element of the gear (or bevel gear)
affected by the inertial torque in the accelerating process of the driving and the driven
element, respectively IB and IF; ηF-B is the efficiency of the gear; iF-B is the gear ratio;
Min, Mout are the input torque on the driving element and the output torque on the
driven element respectively; ωF and εF are the angular speed and acceleration of the
Fig. 3 Virtual simulation model of the intermediate-gear drive’s present layout

Fig. 4 Simplified model of the intermediate-gear drive’s proposed layout
gear’s driving element; ωB and εB are the angular speed and acceleration of the gear’s
driven element.
The signals from the energetic network (values of the vehicle’s speed, acceleration
and displacement during simulation) were taken and saved in order to plot the com-
parison graphics.
3 Comparing the Data Obtained by Simulation
By modifying the intermediate gear drive (hence, reducing of the spinning masses
inertia) the proposed solution improves the dynamic performances of the vehicle. New
data were obtained by simulation. To make a good comparison between the old and the
352 O. Alexa et al.
Fig. 5 Comparison between the performances of the vehicle for the two layouts of the
intermediate gear drive (up - the speed; down - the take-off distance) [4]
new layouts with respect to the vehicle’s performances, in the Fig. 5 there were
overlapped respectively the speed and take-off (starting) distance for the two solutions.
One could notice that the new layout slightly improves the simulated dynamic
performances of the vehicle. Hence, the new solution improves the mobility. As could
be seen, the speed increases when the inertia of the spinning masses decreases. Thus,
the new layout allows the vehicle to reach 30 km/h instead of 28 km/h in 10 s. The
speed of 40 km/h is reached in 18.41 s instead of 19.77 s. Also, the 500 m distance is
reached sooner with the new layout.
4 Conclusions
The virtual simulation model of the new intermediate gear drive was replaced in the
battle tank’s general model and the dynamic characteristics were plotted for both
options. By simplifying the intermediate gear drive, new, improved performances were
obtained. Thus, the accelerating ability increased with 6 % for the first 10 s. Time
needed to reach 40 km/h is 7 % shorter and the 500 m distance is reached 2 % sooner.
The increase in dynamics is explained by a decrease of the power needed to accelerate
the inertial masses. Implementing the new layout it could lead to an increase in
vehicle’s fighting abilities. But what may be rather important is the saving in fuel
consumption. According to performed ground tests, the battle tank burned around 21 L
of Diesel fuel to achieve a speed of about 40 km/h when starts from rest [4]. Even yet it
wasn’t possible to realize any simulation or tests with respect to the fuel consumption,
the authors estimate that the battle tank with the proposed intermediate gear drive will
gain about 5…10 % savings. At the first glance, 1…2 L per each take-off wouldn’t
seem to be that much, yet since the transient regimes are typical for this vehicle, the
overall fuel consumption savings could matter.
Acknowledgments. This paper has been financially supported within the project entitled
“Horizon 2020 - Doctoral and Postdoctoral Studies: Promoting the National Interest through
Excellence, Competitiveness and Responsibility in the Field of Romanian Fundamental and
Applied Scientific Research”, contract id POSDRU/159/1.5/S/140106. This project is
co-financed by European Social Fund through Sectoral Operational Programme for Human
Resources Development 2007-2013. Investing in people!
References
1. Alexa, O., Marinescu, M., Trută, M., Vilău, R., Vînturiş, V.: Simulating the longitudinal
dynamics of a tracked vehicle. Adv. Mater. Res. 1036, 499–504 (2014)
2. Feng, S., He, J., Zhang, L.: Modeling vehicle dynamics based on modelica. Int. J. Multimed.
Ubiquit. Eng. 8(3), 307–317 (2013)
3. Hedrih, K., Nikolić-Stanojević, V.: A model of gear transmission: fractional order system
dynamics. Math. Probl. Eng. 2010, 1–23 (2010). Article ID 972873
4. Alexa, O.: Contribuţii la studiul influenţei coeficientului maselor de inerţie asupra dinamicii
longitudinale a autovehiculelor militare pe şenile (Contributions to the study of the influence
of the mass inertia coefficient on the straight motion dynamics of the tracked military
vehicles). Doctoral thesis, Military Technical Academy (2015)
5. Gorianu, M.: Mecanica Autovehiculelor cu Roți și Șenile (Mechanics off the wheeled and
tracked vehicle’s). Military Technical Academy Publishing House (1992)
Automobile and Environment
Research on Gasoline Homogenous Charge
Compression Ignition (HCCI) Engine
Corneliu Cofaru(&)
Transilvania University Brasov, Brașov, Romania

ccornel@unitbv.ro
Abstract. Homogenous charge compression ignition engine (HCCI) represents

a new idea of combustion technology that combines the main advantages of the
gasoline engines (the homogeneous charge) with the main advantages of the
diesel engines (higher compression ratios which determinates higher efficien-
cies). An important advantage is the low temperature combustion that offers
higher thermal efficiency and lower emissions of Nitrogen Oxides (NOx) and
Particulate Matter (PM) in comparison to spark ignition and diesel engines. For
such engines, a major problem is to control the start of the combustion because
the air and fuel need to be mixed homogeneously before the starts of combustion
and, the air fuel mixture auto ignite due to the temperature increasing at the end
of compression stroke. This paper presents the work to develop a HCCI engine
starting from a spark ignition engine platform, for this the engine was modified
as follow: the carburetor was replaced by a fuel injection system, the com-
pression ratio was increased, and variable valve timing was introduced. The
research work was focused on to determine the parameters which control the
combustion timing o. HCCI engine to obtain the best energetic and ecologic
parameters.
Keywords: HCCI engine Air-fuel ratio Combustion control EGR
1 Introduction
World urban areas face a number of major environmental challenges, their scale and
intensity can vary, thus, in this case, a common set of issues can be identified. They are
related to poor air quality, traffic volumes and congestion, high levels of ambient noise,
lack of recreational areas, high level of greenhouse gas emissions or other factors of
this kind. These environmental challenges constitutes a major problem and have sig-
nificant impacts on human health, environment and economic performance. Likewise,
it is generally agreed that transport is one of the major contributors to pollution of the
atmosphere.
The emission increasing was observed for both categories: passenger transport and
freight transport. The values of this phenomenon are mainly due to an ever growing
transport demands, characterised by large increases in passenger kilometres and tonne
kilometres. As for the passenger road transport, a relative decrease in the use of public
transport is also noteworthy. Even the high level of efficiency on the passenger cars has
not been sufficient to counteract this trend. As for the freight road transport, an

358 C. Cofaru
increased share of road freight transport as opposed to other transport modes supple-
ments the increased transport demand for goods. Therefore, the modal shift can lead to
a wrong direction.
There are two ways to achieve the tasks regarding diminishining fuel consumption,
pollutants and CO2 emissions by:
• improving the internal combustion engine processes using the latest technologies;
• using new fueling systems and alternative fuels in ICE.
The conventional fuels based on petroleum will continue to be in the front of line of
mobility in the coming years, due to the main properties consisting in the extremely
high energy density, ensuring large distances covered by using a relatively small
volume of fuel instead, the simplest scenario remains to develop more efficient com-
bustion engine technology.
In the trend of efficient combustion engine technology, the Homogeneous Charge
Compression Ignition (HCCI) engines can be enrolled. The HCCI engine combines the
advantages of the spark ignition engines which operate with homogenous air-fuel
mixture with the advantages of the compression ignition engines which operate with
lean air-fuel mixture. The result is an engine with low raw emissions and high fuel
economy. The HCCI combustion may be achieved in two modes: starting from a
gasoline engine platform or starting from a diesel engine platform. When the HCCI
combustion is obtained starting from a gasoline engine some measures have to be taken
to increase the temperature of the air-fuel mixture. This can be achieved by increasing
the compression ratio and trapping a part of the burnt gases in the cylinder (internal
EGR). Due to the homogeneous charge and compression ignition, the combustion is
almost simultaneous in all the charge. The burning speed is lower than in the con-
ventional engines but, as the whole mixture is burning at the same time, the heat release
is shorter. Then, the HCCI engines can work only at partial loads. When the HCCI
combustion is obtained by modifying a conventional diesel engine, some measures
have to be taken to avoid the premature auto ignition of the air-fuel mixture due to the
high compression ratio. This can be achieved by using cooled EGR gases. The com-
pression ratio is higher, similar to the compression ratio of the diesel engines, with
benefits on the fuel consumption. The load is controlled by modifying the quality of the
mixtures. The air-fuel mixture is homogenized before the combustion, so there are no
zones with rich mixtures, leading to lower soot emissions. Due to the high air excess
ratio the temperatures from inside the cylinder are lower, and they determine lower
nitrous oxides emissions. Different strategies were used to obtain the homogeneous
combustion. The biggest difficulty in HCCI engines development is the control of
combustion because are no direct means for controlling ignition timing. In this case the
control of combustion (ignition timing) can be achieved indirectly through adjustments
in cylinder mixture preparation. In case of HCCI engines the most important param-
eters that control the ignition timing can be summarized: the intake fresh charge
temperature; the quantity of burnt gases existent in fresh charge (residual gas and EGR
or internal EGR); cylinder wall temperature; the fuel-air equivalence ratio; fuel com-
position; variable compression ratio [8–13].
Research on Gasoline Homogenous Charge Compression 359
Combustion process in Homogenous Charge Compression Ignition (HCCI) engine

is radical different in comparison with combustion process in spark ignition and
compression ignition engines. Due to the combination of the diluted and premixed air
fuel mixtures with multiple ignition sites in the combustion chamber diminish the level
of combustion temperatures zones and limit the production of NOx and particulates.
HCCI combustion can reduces simultaneously NOx and particulate emissions for the
engine starting from a diesel platform and, a reduction of fuel consumption and NOx
emissions for the engine starting from a petrol engine (SI). HCCI gasoline engine, due
to usage of lean or more often diluted air fuel mixtures with burnt gases, permits the
unthrottled operation obtaining a higher engine efficiency and better fuel consumption
than conventional SI engine.
2 Gasoline HCCI
2.1 Principle of Engine Operation

In the gasoline HCCI the mixtures air fuel and burnt gases are prepared in similar way
with the conventional SI engine during the intake stroke. The premixed air fuel and
burnt gases mixtures are then compressed and towards of the end of compression
process the combustion is initiated by auto ignition as in conventional diesel engines.
The level of cylinder charge temperatures at the beginning of the compression stroke
need to be higher than in conventional SI engine to reach the auto ignition conditions.
This can be realized by heating the incoming air or by keeping part of hot burnt gases
results from previous cycle in cylinder. These two strategies are applied to increase the
charge temperature during compression process to determine the rise of chemical
reactions speeds that lead to the start of combustion process of homogeneously air fuel
and burnt gases mixtures. In this case, the main heat release starts when the charge
temperature reaches a value of 1050–1100 K.
In an ideal gasoline HCCI engine, the auto ignition and combustion will develop
simultaneously throughout the combustion chamber resulting in a rapid heat release.
Comparing the characteristics of combustion heat release in ideal conventional SI
and those of HCCI engine some differences result.
In a conventional SI engine the flame front divides the charge of combustion
chamber into burnt and unburnt regions and the heat release is confined to the reaction
region and, in this case, the heat released is a sum of the heat released by a certain mass,
dmi, in the reaction zone and it can be expressed as:
ZZ
Q¼ q:dmi ð1Þ
0
Where: q is heating value per unit mass of air fuel mixture and, Z represents the
number of reaction zones.
360 C. Cofaru
In an ideal HCCI engine, combustion process develop simultaneously chemical

reactions in entire content of combustion chamber and all the mixture participates in the
heat released process at any instant of the combustion. In this case, the cumulative heat
release is a sum of heat released from each combustion reaction, dqi, of the complete
mixture in combustion chamber and can be expressed as:
Z K
Q¼ m:dqi ð2Þ
1
Where K is the total number of heat release reactions, and qi is heat released from ith
heat release reaction involving unit per mass air-fuel mixture [13].
In practice, due to the homogeneities in mixture composition and temperatures, in
real engine, the heat release process will not be uniform throughout the air fuel mixture.
In consequence, faster heat release takes place in the less diluted mixtures or high
temperature regions.
2.2 Control of HCCI Combustion

The gasoline HCCI combustion is achieved by controlling the levels of the cylinder
charge temperatures, pressures, and mixtures composition through the following
parameters:
– the EGR rate,
– the air fuel ratio,
– the compression ratio,
– the intake temperature,
– the inlet manifold pressure,
– the injection timing,
– the fuel properties (volatility of the fuel, auto-ignition properties of the fuel),
– the air-fuel homogenization,
– the heat transfer to the engine (coolant temperature).
In order to obtain the HCCI combustion, the cylinder charge needs to have a high
temperature and sufficient dilution at the start of compression process. In cylinder, air
fuel mixture temperature must be at high level to initiate and sustain the chemical
reactions to determine auto-ignition process. The air-fuel mixture dilution is necessary
to control the rates of heat release reactions.
The air fuel mixture can be diluted using two technologies:
1. using air, in this case the level of mixture temperature is reached by air heating by
additional system in the intake system;
2. using burnt gases, this method meets both requirements, rise the mixture temper-
ature and realize the mixture dilution.
The specific heat of the burnt gases is higher than for air due to presence of
substantial fractions of CO2 and H2O.
The use of burnt gases as mixture diluents to control the combustion process
presents the following effects:
• the charge heating effect, the hot burnt gases mixes with the inlet cooler air fuel
mixture increasing the cylinder charge temperature;
• the dilution effect, the presence of burnt gases in engine cylinder reduces the oxygen
concentration in the cylinder charge mixture;
• the heat capacity effect, the total heat capacity of cylinder content will be higher
with burnt gases due to the presence of carbon dioxide and water vapor which
presents higher specific heat capacity values;
• the chemical effect, the combustion products present in the burnt gases can par-
ticipate in the chemical reaction leading to auto ignition and to developing the
combustion process.
The analytical and experimental studies showed that in gasoline HCCI engine the
ignition is dominated by the charge heating effects instead the combustion duration is
dominated by the dilution and heat capacity effects. The heat release is influenced by a
combination of charge heating effect and by the dilution and heat capacity effects.
Taking in account these influences on ignition and combustion processes in gasoline
HCCI, it is important to manage the using of burnt gases, because it is advantageous to
use burnt gases at as high temperature as possible to determine auto ignition of air fuel
mixtures especially at low loads.
In case of gasoline HCCI engine, the management of burnt gases for HCCI com-
bustion can follow two strategies:
(a) Residual gas trapping, known as negative valve overlap strategy (NVO). The
great amount of residual gas is trapped in cylinder when the exhaust valve is
closed early, during the exhaust stroke. In order to keep the trapped burnt gases in
cylinder and to avoid the back flow of these gases into inlet manifold, the intake
valve is opened later, after TDC at the beginning of the inlet stroke.
(b) This strategy involves the recirculation of burnt gases after they left the engine
cylinder, and the valve timing permit the positive valve overlap. The burnt gases
recirculation can be realized using so called internal gas recirculation (IEGR)
method. The method consist in sucking the exhaust gases existing in the exhaust
system by reopening for a short period of time the exhaust valve during the intake
stroke or by closing of the exhaust valve later during the inlet stroke.
Comparing these two strategies, the result is that burnt gases used by second
method (b) have a lower temperature due to the heat losses involved during the exhaust
process than in case of the burnt gases trapped by method (a) negative valve overlap.
The higher charge temperatures obtained by using the negative valve overlap
strategy can be advantageous to extend the combustion at low loads, but at high speeds
and loads the ignition can be too advanced and the result being a very fast pressure rise
[11–13].
362 C. Cofaru
The technical data of the test engines, platform engine and HCCI engine are presented
in Table 1. The HCCI test engine was realized starting from a SI engine equipped with
carburetor as platform, modifying the compression ratio, providing a variable gas
distribution by using a cam phaser system and by replacing the carburetor by direct
injection system.
The inlet air has the possibility to be conditioned in conformity with research
requirements to obtain the control of combustion timing. During the tests one could
register data regarding: ambient, intake and exhaust pressures, injection and cylinder
pressure, oil pressure; ambient, inlet, exhaust, cylinder, oil, water temperatures; torque;
the valves timing and, pollutant emissions (HC, CO, NOx). For tests the standardized
equipments were used [1–8].
Table 1. Engine’s specification

Engine type SI HCCI [-]
Fuel Gasoline Gasoline [-]
Displacement 652 652 [cm3]
Compression Ratio 9.7:1 11.5:1 [-]
Bore 100 100 [mm]
Stroke 83 83 [mm]
Con-rod length 149 149 [mm]
Valve position DOHC DOHC -variable [-]
No. valves 4 4 [-]
Fueling system Carburetor Direct injection [-]
Injection pressure – 10 [MPa]
4 Research Results
In order to use the trapped residual gas as to obtain HCCI combustion, the set up of the
valves timings was due to realize a negative valve overlap. During the engine test to
determine the influence of quantity of burnt gases on combustion process the engine
speed, start of injection, fuel injected mass and the intake valve timing were kept
constant.
As variable parameter to modify the mass of residual gas was chosen the timing of
closing of the exhaust valve which ranges 639 – 649[CA], (Fig. 1).
In Figs. 2 and 3, the influence of exhaust valve closing timings on the cylinder
temperatures and heat released rates are presented. The results have shown that, by
closing the exhaust valve earlirer, the quantities of residual gases trapped are greater
and the combustion process is initiated rapidly and a vigurous heat is released deter-
mining, beside the heating of fresh mixture made by an incresed mass of residual gas,
the rising of level of the cylinder’s temperature occurs.
Fig. 1. Cylinder pressure/vs/Exhaust valve closing timing
Fig. 2. The influence of exhaust valve closing timing on the cylinder’s temperature
The measured NOx emissions follow the values of cylinder’s temperature detem-
ined by the amount of residual gases trapped (Fig. 4) and, the values of NOx decrease
by rising of dilution of air fuel mixture with residual gases trapped in cylinder due to
closing earlier of the exhaust valve.
The intake valve timings presents a great influence on air fuel mixture and com-
bustion processes in HCCI engine. The opening of the intake valve is influenced by the
closing of the exhaust valve. If the exhaust valve closes earlier, the burnt gases existing
in cylinder are compressed and, at the opening of the intake valve due to the pressure
364 C. Cofaru
Fig. 3. Influence of the exhaust valve closing timing on heat release rate
Fig. 4. Influence of the exhaust valve closing timing on pollutant emissions

difference between the cylinder and intake manifold, causes a back flow, determining a
diminishing of volumetric efficiency. A diminshing of volumetric efficiency also
appears at the retarding of the opening of the intake valve. During the experimental
research, three values of intake valve closing timing were chosen, 204[CA], 210[CA]
and 215[CA], (Fig. 5). The influence of the intake valve closing on heat release rate
and polutants emission are presented in Figs. 6 and 7.
Fig. 5. Cylinder pressure vs. Exhaust valve closing timing
Fig. 6. Influence of the intake valve closing timing on heat release rate.
366 C. Cofaru
Fig. 7. Influence of the intake valve closing timing on pollutant emission
When the opening and closing of the intake valves take place earlier, the increasing
of fresh charge due to the improved volumetric efficiency determines the rising of the
heat release rate and the cylinder’s temperatures has higher values. The pollutant
emissions are influenced by the values of cylinder’s temperatures, the increased tem-
perature values determine lower values for unburnt hydrocarbons (HC) and carbon
monoxide (CO) and higher value for NOx.
An important factor to control the HCCI processes is represented by the air fuel
ratio. During the experimental research, the air fuel ratio was modified by increasing
the injection duration (ID) from 2400 μs to 2700 μs keeping constant the start of
injection, engine speed and residual gases trapped controlled by the valves timings. By
modifying the injection duration, the engine load is changing.
The influence of modifying the quantities of fuel injected on the cylinder’s tem-
perature is presented in Fig. 8, by increasing the injection duration the cylinder’s
temperatures, the heat release rate increases, (Fig. 9). The polutants emissions are
prominently influenced by the cylinder’s temperatures (Fig. 10).
The start of injection (SOI) is an importat parameter used to obtain homogenous
mixture in HCCI engines, for this reason the point was placed during the exhaust stroke
before the opening of intke valve. The fuel is than injected into compressed burnt gases
where evaporated and mixes with them, after the opening of the intake valve the burnt
gases and evaporated fuel mixes with air, and this mixture is homogenized during the
intake and compression strokes. Increasing the injection advance from 325 [CA] to 350
[CA] the heat release rate increases (Fig. 11).
Fig. 8. The influence of injection duration (ID) on the cylinder’s temperature
Fig. 9. The influence of injection duration (ID) on heat release rate
The air fuel ratio influences both oxygen concentration and fuel concentration.
However, since HCCI engines operate lean, the air fuel ratio has a stronger influence on
fuel concentration than on oxygen concentration. The dominating effect of decreasing the
air fuel ratio, thus, is an increase in fuel concentration, which will result in a higher
reaction rate (Fig. 9). Thus, increasing the air fuel ratio serves to advance ignition timing.
The experimental tests have shown that the increasing of engine’s speed determines
the rising of heat release rate.
368 C. Cofaru
Fig. 10. The influence of injection duration (ID) on pollutant emission
Fig. 11. The influence of starting of injection (SOI) on heat release rate
5 Conclusions
During the experimental research the negative overlap strategy to control the engine
processes by residual gas trapped in cylinder was adopted. The amount of trapped burnt
gases has a huge influence on the combustion process.The amount of residual gas
trapped in cylinder was changed by modifying the exhaust valve closing timing earlier
during the exhaust stroke. The test was done for EVC values: 639[CA], 644[CA], 646
[CA] and 649[CA]. The quantity of the residual gas trapped increase which influence
the heat release rate by combination of charge heating effect and by the dilution and
heat capacity effects. The trapped gases are replacing a part of the fresh air which can
be aspirated into the cylinder, so the performances of the engine can be reduced by
using a higher EGR rate. When the EGR rate is too low the conditions needed for the
auto ignition of the air fuel mixture are missing, so there will be no combustion.
Closing the exhaust valve earlier during the exhaust strokes, the burnt gases are
than compressed and this process determines the increasing of the pumping work.
The intake valve timings was chosen in order to keep the trapped burnt gases in
cylinder and to avoid the back flow of these gases into inlet manifold, the intake valve
is opened later, after TDC at the beginning of the inlet stroke. The closing of the intake
valve was adopted after TDC to maintain the volumetric efficiency in an acceptable
range. For this reason, the intake and exhaust valves timing needs to be tuned and
depends on the engine’s load and speed. The test’s results demonstrate that the
strategies adopted regarding the gas exchange processes are viable.
The engine load was modified by changing the injection duration from 2400 [μs] to
2700 [μs]. The increasing of quantities of injected fuels per engine’s cycles determines
the rising of the heat release rate.
The obtaining of homogenous mixture in HCCI needs to select carefully the
advance of injection (start of injection). Within the experimental tests, the point
(SOI) was placed during the exhaust stroke after the exhaust valve had been closed, but
before the intake valve had been opened. The fuel is injected into hot compressed burnt
gases where the evaporates and mixes with them. After the opening of the intake valve,
the burnt gases and the evaporated fuel mix with air, and this mixture is homogenized
during the intake and compression strokes. Increasing the injection advance from 325
[CA] to 350 [CA] the heat release rate increases.
In order to have a good combustion process, the quantities of injected fuels per
engine’s cycles need to be correlated with the advances of injection depending on the
engine’s speed.
References
1. Cosgarea, R., Cofaru, C., Aleonte, M., Scutaru, M.L., Jelenschi, L., Sandu, G: An approach for
modeling the valve train system to control the homogenous combustion in a compression
ignition engine. In: Proceedings of 12th WSEAS International Conference on Neural
Networks, Fuzzy Systems Evolutionary Computing & Automation, pp. 131–136. Transilvania
University, Brasov (2011). ISBN978-960-474-274-9
370 C. Cofaru
2. Cosgarea, R., Aleonte, M., Cofaru, C., Dahnz, C., Velji, A., Spicher, U.: The influence of the
internal exhaust gas recirculation on the combustion characteristics in a gasoline
homogeneous charge compression ignition engine. In: Proceedings of 13th EAEC European
Automotive Congress, Valencia, Spain, 13–16 June 2011
3. Cosgarea, R., Aleonte, M., Cofaru, C.: The influence of the valve lift strategies on the
combustion characteristics of a homogeneous charge compression ignition engine model.
Int. J. Mech. 5(3), 191–201 (2011)
4. Cosgarea, R., Aleonte, M., Cofaru, C.: The influence of the internal exhaust gas recirculation
(EGR) on the pumping losses. Bull. Transilvania Univ. Brasov 4(53), 1–7 (2011). No. 1 –
2011 Series I, ISSN 2065-2119 (Print), ISSN 2065-2127 (CD-ROM)
5. Cosgarea, R., Cofaru, C., Aleonte, M., Sandu, G.: An approach for modelling the
homogeneous charge compression ignition (HCCI) combustion using the vibe functions. In:
Proceedings of the 10th International Automotive Congress “Automotive and Environment”,
CAR 2011, Universitatea din Pitesti (2011). ISSN 1453-1100
6. Cosgarea, R., Cofaru, C., Aleonte, M.: The influence of the EGR rate on a HCCI engine
model calculated with the single zone HCCI method. In: International Congress on
Automotive and Transport Engineering, CONAT 2010, pp. 9–14, Brasov, Romania (2010).
ISSN 2069 – 0401
7. Cosgarea, R.: Cercetari privind utilizarea de sisteme avansate de ardere prin folosirea
amestecurilor omogene/partial omogene la motoarele cu aprindere prin comprimare.Teza de
doctorat, Universitatea Transilvania din Brasov (2011)
8. Sandu, G., Cofaru, C., Cosgarea, R., Aleonte, M.L.J.: The impact of the valve variable
timing upon the pumping work in the internal combustion engine. In: Proceedings of the
10th International Automotive Congress “Automotive and Environment” CAR 2011
Universitatea din Pitesti (2011). ISSN 1453-1100
9. Subbotin, M.V., Knierim, K.L., Park, S., Kojic, A., Ahmed, J.: Modeling and control of a
two stroke HCCI engine. In: 2008 American Control Conference Westin Seattle Hotel,
Seattle, Washington, USA, 11–13 June 2008. ISBN978-1-4244-2079-7
10. Bidarvatan, M., Thakkar, V., Shahbakhti, M.: Grey-box modeling and control of HCCI
engine emissions. In: 2014 American Control Conference (ACC), Portland, Oregon, USA,
4–6 June 2014. ISBN 978-1-4799-3271-9
11. Killingsworth, N.J., Aceves, S.M., Flowers, D.L., Espinosa-Loza, F., Krstic, M.: HCCI
engine combustion-timing control: optimizing gains and fuel consumption via extremum
seeking. IEEE Trans. Control Syst. Technol. 17(6), 1350–1361 (2009). ISSN 1063-6536
12. Aziz, N.A., Rahman, N.A., Innayatullah, O.: Control of homogeneous charge compression
ignition (HCCI) engine through stochastic variables. World Appl. Sci. J. 27(3), 359–366
(2013). ISSN 1818-4952
13. Zhao, H.: HCCI and CAI Engines for the Automotive Industry. Woodhead Publishing
Limited, Cambridge (2007)
Study of the Diesel Engine Cycle Variability
at LPG Fuelling
Liviu Nemoianu, Alexandru Cernat(&), Constantin Pana,


liviu.nemoianu@yahoo.com, cernatalex@yahoo.com,
constantinpana@yahoo.com,
niculae_negurescu@yahoo.com, cristi_cmt@yahoo.com
Abstract. Efficiency improves and pollutant emissions decrease of the auto-

motive diesel engine can be assured by Liquid Petroleum Gas fuelling. Paper
target is the study of cycle variability at LPG fuelling, which shows new,
important, considerable implication and effects of LPG use at diesel engine,
completing the general experimental investigation of engine running at fuelling
with alternative fuel for a clean environmental. The K9K EURO 4 diesel engine
was mounted on the test bed, equipped with instrumentation and fuelled with
LPG by diesel gas method. The investigated engine operating regimes were
55 % and 70 % load at speed of 2000 rpm, at fuelling with diesel fuel and LPG
at different substitute ratios. The increase of LPG cycle dose leads to the increase
of cycle maximum pressure and maximum pressure cycle dispersion, thus the
dispersion of successive running cycle’s increases. This fact may become a
limitation criterion for the LPG substitute ratio value.
Keywords: Cycle variability LPG dose Maximum pressure Lean mixture

Homogeneity
1 Introduction
Nowadays, in the global content regarding the impact on the environmental of the gases
emissions resulted from the fossil fuels combustion, an interest aspect discussed also on
the 21st Session of the Conference of the Parties (COP21) from the 2015 Paris Climate
Conference in November 2015, and the gradual diminution of the worldwide oil
reserves contribute to the necessity of searching of alternative energy from durable and
renewable resources. The replacement of the classic fuels of oil origin by alternative
fuels was initiated many decades ago and represents one of the most actual preoccu-
pations linked by the possibility of pollutant emissions decrease and efficiency increase
of the diesel engine and necessity of fossil fuel saving [1]. These issues impulse the
research’s for developing and using new fuel sources which will allow the substitution,
even a partial one, of the classic fuels. Liquid Petroleum Gas (LPG) is a worldwide
alternative fuel used on a large scale for internal combustion engines [2–6]. The
physical-chemical properties of LPG show difficulties for CI engine use (a lower cetane
number, high autoignition temperature) that implies special solutions for mixture

372 L. Nemoianu et al.
forming and combustion control [7–11]. The air-LPG mixture in a higher state of
homogeneity, due to propane higher vapor pressure will lead to a high velocity com-
bustion process, after the flame nucleus is forming inside the cylinder next to the diesel
fuel jet boundary, [1–3, 6]. The initial very lean mixtures of air-LPG achieved inside
the cylinder before combustion starts will influence the combustion cycle variability
and engine mechanical running. Ashok [10] the performance, combustion and emission
characteristics of the LPG diesel dual fuel engine have been reviewed from the various
experimental studies and indicate that the part load characteristic can be improved by
optimizing the engine operating parameters and design factors such as engine speed,
load, pilot fuel quantity, injection timing, intake manifold condition and intake gaseous
fuel compositions. Increasing the compression ratio while using LPG in dual fuel mode
leads to excessive combustion noise and cyclic variation at high loads in addition to
loss in indicated mean effective pressure, [11]. Thomas Renald [12] shows that at the
fuelling with 60 % of diesel and 40 % of LPG the thermal efficiency of the diesel
engine increases with about 5 % when and the mechanical efficiency was increased
about 5 % [12]. Santoso [13] analyses 160 consecutive cycles in terms of
cycle-by-cycle variability for diesel engine fuelled with hydrogen gas at the running
regime of 2000 rev/min and 5 Nm, [13]. Cyclic variability tends to increase with the
increase of hydrogen enrichment and is reduced at higher engine load operation.
Santoso shows the higher cylinder pressure variation, in terms of maximum pressure,
for 36 consecutive cycles in diesel and dual fuel operation at engine load of 5 Nm,
comparative to 20 Nm regime [13] and presents IMEP variation when hydrogen was
introduced at two the flow rates and explains there is a significant variability between
these two fuelling and running cases, the mean values being almost the same, [14].
Also, the researcher affirms that the cyclic variability tends to increase with an increase
in hydrogen addition at low load, [13]. Poonia and Mathur [14] use stationary, single
cylinder water cooled diesel engine to study the cycle variability at LPG fuelling,
analyzing 100 pressure diagrams for the regime of 3.7 kW at 1500 rev/min diesel
engine. Cyclic variations in IMEP are less than 6 % with diesel comparative to dual
fuel operation and decreases with increase in power output. The author affirms that the
amount of heat released in the second stage of combustion is mainly responsible for the
cyclic variations. The cyclic variations in peak cycle pressure with straight diesel were
found to be less than 2 %. In the dual fuel mode, an increase in load increases the COV
till 80 % load. It is common opinion among the investigators that the higher the
burning rate and lower the combustion duration lesser are the cyclic variations. It is
seen that in the dual fuel mode, while COV in peak pressure increases with load, the
COV in IMEP decreases with load generally. The COV in peak pressure is a result of
the variations in the heat release rate in the first and second phase while the COV in
IMEP is a result of the whole combustion process. Regarding the effect of load on
cyclic variations, Poonia [14] explains that the COV of peak pressure and Indicated
Mean Effective Pressure (IMEP) with change in load depicted. It is seen that COV of
IMEP and peak pressure has no relationship. The COV of IMEP, a tool of the
drive-ability test, is found to be lowest at 100 % load and increases with decrease in
output. The COV of IMEP is less than 10 %, which indicate good drive-ability. Poonia
[14] affirms that the COV of peak pressure is higher than diesel mode at all temper-
atures and outputs but the values within 4 % at up to 60 % loads [14]. COV of peak
Study of the Diesel Engine Cycle Variability at LPG Fuelling 373
pressure is mainly affected by the irregularities in the combustion of the gas air mixture
and not by the combustion of the pilot diesel. Only second and third stages of com-
bustion are responsible for cyclic variations and not the first stage. Up to 60 % load, the
cyclic variation in peak cycle pressure and IMEP are lower with optimum EGR as
compared to with no EGR. Higher COV is observed at loads higher than 60 % with
EGR, [14] the reason may be slower flame propagation due to the presence of EGR. Up
to 60 % load, the increase in cyclic variations with 5 mm plungers is more due to larger
pilots at these loads. Unmatched injection deteriorates the atomization, vaporization
and mixture formation of pilot fuel. At 80 and 100 % loads, the size of pilot is small
and hence the effect of cyclic variations in these loads due to lower rates in injection is
less. With well-matched injection system, cyclic variations are due to the second phase
of combustion but with the unmatched injection system, both the phases are responsible
for variation, [14]. Paper target is the study of cycle variability at LPG fuelling, study
that shows new, important, considerable implication and effects of LPG use at diesel
engine, completing the general experimental investigation of engine running at fuelling
with alternative fuel for a clean environmental. Were analyzed individual pressure
diagrams, in terms of maximum pressure, maximum pressure angle and IMEP for a
1,5 L direct injection supercharged engine, type K9K from Dacia Logan. The engine is
dual fuelled with diesel fuel and LPG by diesel gas method, the LPG cycle dose being
injected inside the inlet manifold and electronic controlled by a secondary ECU con-
nected back to back with the main engine ECU (Engine Control Unit).
2 Cycle Variability Study
The cycle variability can be characterized by coefficients of in-cylinder pressure

variation (COV). The intensity of the cycle variety phenomena is defined by the
coefficient of cycle variability, as relation (1) shows [5, 6, 8].
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u 0 P 12
un
n
a
uP ai i¼1n i
t @ A
n1
i¼1
ðCOVÞai ¼ P
n 100 % ð1Þ
ai
i¼1
n
where n is the number of cycles, a is the parameter of which variability is study and is
defined by indicated mean effective pressure IMEP, maximum pressure pmax, maxi-
mum pressure rise rate (dp/dα)max and the angle where maximum pressure occurs,
αpmax in the cycle number “i”. The coefficient of cycle variability is defined as a relative
average deviation of maximum pressure values [5, 6, 8]. For n consecutive cycles, if is
considered a normal distribution of the deviation probabilities, the squared average
deviation is calculated and the cycle variability coefficient is defined. Generally, the
way of cycle variability evaluation for regimes with spark timing closer to the value of
spark timing for MBT (Maximum Brake Torque) the COV of maximum pressure is
suitable. The COV of maximum pressure angle, when the maximum pressure occurs,
is used for characterization of the combustion cycle variability during the initial phase
of combustion. The variation of the IMEP, appreciated by (COV)IMEP, is the most
suitable instrument to define the engine respond to the combustion process variability
[3, 10]. From this point of view the limit value of (COV)IMEP define practically the
limit of mixture leaning. This cycle coefficient can also indicate the variability of flame
development during the initial phase of combustion. The quantity of LPG injected into
the inlet manifold is established by an energetic substitute ratio xc which take into
consideration the lower heating values of both fuels:
ChLPG :HiLPG
Xc ¼ :100 ð2Þ
ChDF :HiDF : + ChLPG :HiLPG
where: ChDF, ChLPG diesel fuel and LPG fuel consumptions, HiDF, HiLPG lower heating
values of diesel fuel and LPG.
3 Results
For different substitute ratios, at the regime of 2000 rpm and loads of 55 %, 70 %, 150
consecutive pressure diagrams were measured by AVL Indimodul 621 data acquisition
system and forward analyzed in terms of maximum pressure, maximum pressure angle,
indicated mean effective pressure and angles of 5 %, 10 %, 50 % and 90 % of mass
fraction burned (MFB) is achieved. Following the aspects that lead to the establishing
of a maximum substitution ratio, can be also taken into consideration the values of
cycle variability coefficients for in-cylinder maximum pressure and IMEP, (COV)pmax
and (COV)IMEP, respectively. The cycle variability can be very well appreciated by the
values of these two coefficients, and they are related with the engine adaptability for
traction. Generally, the normal automotive manoeuvrability is assured if (COV)pmax,
respectively (COV)IMEP, don’t exceeded 10 %. [3, 8].
The coefficient (COV)pmax don’t exceed values of 1.8 % at 2000 rpm speed and
55 % partial load, Fig. 1, being registered a slightly increase of the coefficient value
with the LPG dose. At 2000 rpm speed and 70 % load regime, the higher dose of LPG
leads to a much significant continuously increase of (COV)pmax, especially in the range
of xc = 9.6…12. The value registered for xc = 19, load 55 %, is 4.2 times bigger
comparative to the value of standard engine. Even the maximum value doesn’t exceed
1.76 %, with no reason for xc limitation, the further increase tendency of (COV)pmax
with xc, especially for xc > 10 %, can be take into consideration, Fig. 1.
(COV)pmax for xc = 12 %, load 70 %, is 2.47 times bigger comparative to the value
of standard engine. The maximum value doesn’t exceed 1.72 %, with no reason for xc
limitation, the further increase tendency of (COV)pmax with xc, especially for xc > 9 %,
can be take into consideration, Fig. 1. Also, the angle value for maximum pressure,
αpmax, is registered closer to the TDC, with is generally related to a rapid, brutal
combustion. The variability of maximum pressure angle slightly, Fig. 2, increases
comparative to reference regime for the maximum substitute ratio. At 55 % load, for
xc = 1.2…19 the (COV)αpmax increases up till 23 %…74 % comparative to diesel
fuelling. The increased value of this coefficient for xc up till 19 % can show the
Fig. 1. Coefficient (COV)pmax versus xc at 2000 rpm, 55 % and 70 % load
increasing of the cycle variability intensity during the initial combustion phase. For
70 % load, (COV)αpmax continuously increases from 10.9 % (xc = 0) up till 23 %
….35 % for xc = 6.5…12, showing a continuously increasing of the cycle variability
during initial phase of combustion. The increasing tendency of (COV)αpmax for xc
higher values can be related with the increased cycle variability during the initial phase
of combustion process at this two loads regime.
The limitation of (COV)IMEP maximum value leads to the reduction of the vari-
ability of flame development during the initial phase of combustion. For this speed
regime, IMEP cycle variability coefficient (COV)IMEP is 2 times higher for all LPG
cycle doses comparative to diesel fuelling, 55 % load, Fig. 3. At 70 % load, coefficient
(COV)IMEP is 1.5 times higher for maximum LPG cycle dose comparative to diesel
fuelling, the cycle variability of the flame developing during the premixed phase of
combustion not being significantly affected by the increase of the substitute ratio. For
55 % and 70 % load and 2000 rpm regime, the increase of LPG cycle dose leads to the
increase of cycle maximum pressure and maximum pressure cycle dispersion, thus the
dispersion of successive running cycles increases, explained by their coefficients,
Figs. 1, 2 and 3. For xc = 9.6…12 % the value of coefficient (COV)IMEP remains the
same, and is 1,7 times bigger comparative to reference value for xc = 5, 6 %, Fig. 3.
Fig. 2. Coefficient (COV)αpmax versus xc at 2000 rpm, 55 % and 70 % load

Fig. 3. Coefficient (COV)IMEP versus xc at 2000 rpm, 55 % and 70 % load
Fig. 4. The COV of 5, 10, 50 and 90 % of MFB versus xc at 2000 rpm, 55 % and 70 % load
Even the values of cycle variability coefficients continuously increase for dual fuelling,
the limitation of the LPG dose at a specific xc value, in order to maintain the normal
engine running at dual fuelling, is not necessary due to the fact that COV of maximum
pressure and IMEP don’t exceed the admitted value of 10 %.
If the admitted limit is [(COV)αpmax]max = 10 %, then appears a limitation issue for
xc at 1.2 % for 55 % load regime and at 6 % for 70 % load regime.
If the substitute ratio is limited to xc = 19 % the maximum pressure and maximum
pressure rise rate are limited at pmax = 74 bar and (dp/dα)max = 3.26 bar/CAD,
respectively, values measured in different individual cycles, for 55 % load and at
pmax = 84.5 bar and (dp/dα)max = 3.58 bar/CAD, at 70 % load. This phenomenon is
related with a combustion process closer to TDC (Top Dead Centre) and leads to the
increasing of the maximum pressure rate during the combustion process. The cycle
variability is reflected also on moment per cycle were 5, 10, 50 and 90 % of mass
fraction burned is achieved. The start of combustion increases its variability, COVα5 %,
with 6.15 times at 55 % load and with 1.7 times for 70 % load, comparative to the
reference regime, Fig. 4. The end of combustion has a higher variability for dual
fuelling, the COVα90 % increases 24.5 times for 55 % and by 8 times for 70 % load
comparative to xc = 0, Fig. 4.
4 Conclusions
The experimental study shows the main characteristics of the engine running for diesel
and diesel-LPG fuelling. The main conclusions can be formulated:
1. The diesel and LPG engine fuelling leads, comparative to classic diesel fuelling, to
the increase of in-cylinder maximum pressure and of its cycle variability coefficient
(COV)pmax increasing. At both loads regimes the dual fuelling increases the cycle
variability in the same measure.
2. The increased values of maximum pressure cycle variability coefficient registered at
the rise of LPG dose, over xc = 19 % at 55 % load and xc = 12 % at 70 %, are
related with the variation tendency of (COV)αpmax that show the increase of the
cycle dispersion between the values of angles were maximum pressure occurs,
especially at high load regime. This aspect may reflect the increase of the com-
bustion cycle variability during the initial phase of combustion at the increase of xc
over the maximum values.
3. The diesel and LPG engine fuelling leads, comparative to classic diesel fuelling, to
an increase tendency for (COV)IMEP increasing. The cycle variability increases with
the rise of xc, the increasing being more significant at load of 55 %, were xc also
take maximum higher values.
4. The diesel fuel substitute ratio by LPG, xc, is limited by the maximum values or by
the maximum variation interval of same running characteristic parameters of the
engine, as is follows: if the maximum admissible level of the cycle variability is
(COV)IMEP = 7.7 %, for IMEP, then the substitute ratio will be limited to xc =
19 %; if the maximum pressure rise rate is limited to the value (dp/dα)max =
3.26 bar/CAD, then the LPG substitute ratio will be limited to xc = 19 %, value that
will lead to a maximum acceptable in-cylinder pressure level of 74 bar, at 55 %
load; if the maximum admissible level of the cycle variability is (COV)IMEP =
7.7 %, then the substitute ratio will be limited to xc = 12 %; if the maximum
pressure rise rate is limited to the value (dp/dα)max = 3.58 bar/CAD, then the LPG
substitute ratio will be limited to xc = 12 %, value that will lead to a maximum
acceptable in-cylinder pressure level of 84.5 bar, at 70 % load.
5. Dual fuelling leads to the increase of the cycle variability of the beginning and the
ending of the combustion, evaluated by angles of α5 %, α90 %, explained by the
raised values of the COVα5 %, COVα90 %, for all loads regimes, comparative to
classic fuelling.
Acknowledgements. The work has been funded by the Sectoral Operational Programme
Human Resources Development 2007–2013 of the Ministry of European Funds through the
Financial Agreement POSDRU/159/1.5/S/134398. The authors would like to address special
thanks to the AVL GMBH Graz Austria for providing the necessary equipments.
References
1. Popa, M.G., Negurescu, N., Pana, C.: Diesel Engines. Processes. MatrixRom, Bucharest
(2003)
2. Pana, C., Negurescu, N., Popa, M. G., Cernat, A.: Experimental aspects of the use of LPG at
diesel engine. In: Scientific Bulletin of UPB, no. 1, Bucharest (2010)
3. Polk, A.C., Carpenter, C.D., et. al.: Diesel-ignited propane dual fuel low temperature
combustion in a heavy-duty diesel engine. J. Eng. Gas Turbines Power 136(9) (2014).
doi:10.1115/1.4027189
4. Negurescu, N., Panǎ C., Popa, M.G., Racovitzǎ, A, Boboc, G., Cernat, A.: Performance of a
diesel engine fueled by LPG. In: 5th International Colloquium FUELS 2005 Proceeding,
12–13 January 2005, ISBN 3-924813-59-0, pp. 143–148. Technische Akademie Esslingen,
Germania (2005)
5. Pană, C., Negurescu, N., Popa, M., Boboc G., Cernat, A.: Reduction of NOx, smoke and
BSFC in a diesel engine fuelled with LPG. In: MECCA, Journal of Midle European
Construction and Design of Cars, vol. III, no. 4, ISSN 1214-0821, pp. 37–43, Praha, Czech
Republic (2005)
6. Ergenç, A.T., Koca, D.O.: PLC controlled single cylinder diesel-LPG engine. Fuel 130,
273–278 (2014)
7. Negurescu, N., Pana, C., Popa, M.G., Cernat, A., Boboc, G., Soare, D.: GPL – combustibil
alternativ pentru motorul diesel. In: ZILELE ACADEMICE TIMISENE, Editia a IX-a,
ISBN 973-625-225-6. Editura Politehnica, Timisoara (2005)
8. Heywood, J.B.: Internal Combustion Engines Fundamentals. McGraw-Hill Book Company,
New York (1988)
9. Goto, S., Lee, D., Harayama, H., Honjyo, F., Ueno, H., Honma, H., Wakan, Y., Mori M.:
Development of LPG SI and CI engines for heavy duty vehicles. F2000A171, Seoul 2000
FISITA World Automotive Congress, Seoul, Korea (2000)
10. Ashok, B., Ashok, S.D., Kumar, C.R: LPG diesel dual fuel engine – a critical review.
Alexandria Eng. J. 54, 105–126 (2015). Elsevier
11. Leemakers, C.A.J., Van der Berge, B., Luijten, C.C.M., Goey, L.P.H., Jaasma, S.: Direct
injection of the diesel – butane blends in a heavy duty engine. SAE Int. J. Fuels Lubricants
4(2), 179–187 (2011)
12. Thomas, R.C.: Ja., Somasundaram, P.: experimental investigation on attenuation of emission
with optimized LPG jet induction in a dual fuel diesel engine and prediction by ANN model.
Energy Procedia 14, 1427–1438 (2012)
13. Santoso, W.B., Rosli, A., Bakar, R.A., Ariyono, S., Cholis, N.: Study of cyclic variability in
diesel-hydrogen dual fuel engine combustion. Int. J. Mech. Mechatron. Eng. IJMME-IJENS
12(04), 52–56 (2012)
14. Poonia, M.P., Mathur, Y.B.: Effect of EGR injection rate and engine load on cyclic
variations of an LPG diesel dual fuel engine. J. Environ. Res. Dev. 7(2), 746–755 (2012)
Direct Injection of Diesel and Ethanol
in a Diesel Engine - A Numerical Analysis
Nicolae Vlad Burnete(&), Călin Iclodean, Dan Moldovanu,

and Nicolae Filip
Faculty of Mechanics, Department of Automotive and Transport,

nicolae.vlad.burnete@auto.utcluj.ro
Abstract. The objective of this study is to obtain an insight inside the com-
bustion chamber of a Diesel engine running with a blend containing 90 % v/v
Diesel and 10 % v/v Ethanol (D90E10) compared to pure Diesel (D100)
operation. Since the experimental conditions are kept the same for both cases the
simulation boundary conditions are almost identical. Following the calibration
of the model and the analysis of in-cylinder images it was noted that the start of
injection for the D90E10 case is slightly advanced while the total duration
remains the same. Combined with an improved combustion this led to higher
in-cylinder pressure and temperature values which increased the NO emissions.
However, the soot emissions reduced.
Keywords: Combustion Diesel Simulation Pollution Blend
1 Introduction
Current trend and legislation pursue the idea that more fossil fuel must be replaced with
renewable fuels [1]. Although there have been numerous studies regarding Diesel–
Ethanol blends, the results presented are contradictory [2, 3]. Also, there are very few
numerical studies that provide insight in the combustion process differences that occur
when using Diesel–Ethanol blends compared to pure Diesel operation. Mixing the two
fuels without any additive but at a high enough temperature (to maintain blend stability
[4]) can help achieve good engine performances. Even though the NOx emissions
increased, an optimized injection strategy and the employment of exhaust gas recir-
culation can help reduce fossil fuel consumption, lower soot emissions and improve the
combustion process.
2 Methodology
The experimental results were carried out in the “Laboratory of research, testing and
homologation of internal combustion engines that run on biofuel” of the Department of
Automotive and Transport within the Technical University of Cluj-Napoca on an
engine with the main technical specifications presented in Table 1. Based on the
gathered data a simulation model was set up using the AVL ESE Diesel simulation

380 N.V. Burnete et al.
environment which allows for an engine sector simulation (Fig. 1) in case of sym-
metrical combustion chambers. The main simulation parameters are presented in
Table 2. For a better comparison of the phenomenon, the test parameters were kept
constant for both fuels (Table 3).
Table 1. Engine data and test parameters

Parameter Value M.U.
Bore x Stroke 90 × 85 mm x mm
Compression ratio 17.1 –
Rated speed 4200 min−1
Maximum power 6 kW
Injection system/Injector Bosch CRDI/Central injector
(8 × 0.12 × 162°)
Table 2. Simulation parameters [5]

Phenomenon Model
Turbulence k-ζ-f
Wall heat transfer Standard
Combustion ECFM-3Z combustion model
Two-stage autoignition
Spray Multi-component evaporation
WAVE break-up
Table 3. Test cases

Case Engine Pilot Pilot Main Main Rail
speed timing quantity timing quantity pressure
[rpm] [°CA] [mm3] [°CA] [mm3] [bar]
1 1500 20.625 2.34 5.625 22.85 700
2 2000 22.125 2.34 8.625 25.78 800
3 2500 23.25 2.34 10.500 23.42 800
Fig. 1. Simulation mesh (45° engine sector)

Direct Injection of Diesel and Ethanol in a Diesel Engine 381
3 Validation and Results
In order to validate the simulation model the following parameters were considered:
in-cylinder pressure (p) and temperature (T), rate of heat release (dQ/dt) and the NOx
emission level (the red dotted line in Figs. 2 and 3 represents the measured NOx value
in the exhaust gases and not its variation). The experimental thermodynamic values
were determined based on the pressure trace using the following equations [6]:
pV
T= ½K ð1Þ
mR

dQ 1 dV dp J
¼ cp þV ð2Þ
dt c1 dt dt RAC
where: m is the in-cylinder mass, V the cylinder volume, R the gas constant and γ the
polytropic coefficient. The validation study (Figs. 2, 3) provided valuable information
about the effects of a diesel-ethanol blend on the Diesel engine operation.
Fig. 2. Model validation for the 1500 min−1 case with D100 injection
It was noted that the combustion adjustment parameters are the same for both fuels,
which proves that these parameters depend only on the engine operating conditions.
Another important observation was that the start of injection for D90E10 was slightly
advanced compared to the D100 case even though, the user defined injection param-
eters were kept constant in the test bed control software. This was observed in the
images captured inside the engine combustion chamber with the Visioscope equipment
from AVL. The idea is backed up by the recorded pressure traces. Although Ethanol
addition leads to an increased ignition delay [7] the pressure traces show a slight
advance in the pressure rise due to the combustion of the main injection, an effect
which was attributed to a smaller start of injection delay as well as an improved
combustion.
The analysis of the accumulated heat release values showed that the total heat
released in the D90E10 case was with about 2,5 % higher than in the D100 case even
though, the addition of ethanol led to a 3.7 % decrease in the lower heating value of the
resulting blend. It is believed that this effect is a result of a combination between the
improved combustion and a higher fuel quantity injected (due to the changes in the
injection parameters).
Fig. 3. Model validation for the 1500 min−1 case with D90E10 injection
The comparative simulation results for the 2000 and 2500 min−1 cases are pre-
sented in Figs. 4 and 5. The engine operation with the D90E10 blend yielded higher
pressure and temperature values with important engine performance benefits and lower
soot emissions. However this comes at the expense of increased NOx emissions. It can
be observed that combustion duration is not significantly influenced by the addition of
ethanol.
Fig. 4. Comparison of D100 and D90E10 simulation results at 2000 min−1

Figure 6 shows the fuel distribution as well as the temperature field inside the
combustion chamber at TDC, 10 and 20 °CA after TDC respectively. It can be
observed that there is a better fuel distribution for the D90E10 case. This favors an
improved air utilization and thus a more complete combustion and increased engine
efficiency. Also, the maximum temperature field covers a much larger area in the blend
operation case which would explain the higher NOx emissions.
4 Conclusions
The purpose of this study was to use computer simulation in order to estimate the
in-cylinder conditions and to understand the causes that led to the obtained experi-
mental results regarding the influence of a blend consisting of 90 % v/v Diesel and
10 % v/v Ethanol on the performances and emission pollutants of a Diesel engine
compared to pure Diesel operation. By maintaining the fuel temperature at 35 °C, no
additive was required and hence no additional influences had to be considered for the
simulation. The general conclusions that can be drawn from the obtained results are:
– The addition of ethanol led to a smaller start of injection delay which caused an
increase in the peak cylinder pressure and temperature.
– The better fuel distribution and earlier start of injection favor a better air utilization
and thus a more complete combustion.
– By increasing the temperature and reducing the carbon content of the fuel the soot
emission values decrease.
– The NOx values increase in the D90E10 case as a result of the higher temperature
values.
Acknowledgement. This work was partially supported by the strategic grant POSDRU/
159/1.5/S/137070 (2014) of the Ministry of National Education, Romania, co-financed by the
European Social Fund – Investing in People, within the Sectorial Operational Programme Human
Resources Development 2007-2013.
This work would have not been possible without the hardware and software of the AVL LIST
GmbH Company, available as part of the partnership between AVL LIST GmbH Company and
the Technical University of Cluj-Napoca.
References
1. European Parliament C of the EU. Directive 2009/28/EC of the European Parliament and of
the Council of 23 April 2009 (2009). http://eur-lex.europa.eu/legal-content/EN/ALL/?uri=
CELEX:32009L0028. Accessed 24 Jul 2015
2. Burnete, N.V., Filip, N., Barabás, I.: Diesel-ethanol blends and their use in diesel engines.
J. Automot. Eng. 89, 88–106 (2015)
3. Agarwal, A.K.: Biofuels (alcohols and biodiesel) applications as fuels for internal combustion
engines. Prog. Energy Combust. Sci. 33, 233–271 (2007). doi:10.1016/j.pecs.2006.08.003
4. Torres-Jimenez, E., Jerman, M.S., Gregorc, A., Lisec, I., Dorado, M.P., Kegl, B.: Physical
and chemical properties of ethanol-diesel fuel blends. Fuel 90, 795–802 (2011). doi:10.1016/j.
fuel.2010.09.045
5. AVL List GmbH. AVL Fire Manual. Graz, Austria (2014)
6. Heywood, J.B.: Internal Combustion Engine Fundamentals. Mcgraw-Hill, New York (1988)
7. Rakopoulos, C.D., Antonopoulos, K.A., Rakopoulos, D.C.: Experimental heat release
analysis and emissions of a HSDI diesel engine fueled with ethanol-diesel fuel blends. Energy
32, 1791–1808 (2007). doi:10.1016/j.energy.2007.03.005
The Effects of a Diesel-Ethanol Blend Without
Additives on Diesel Engine Operation
Nicolae Vlad Burnete(&), Călin Iclodean, George Popescu,

and Nicolae Filip
Faculty of Mechanics, Department of Automotive and Transport,

nicolae.vlad.burnete@auto.utcluj.ro
Abstract. The objective of this study is to establish the influence that a blend of
90 % v/v Diesel and 10 % v/v Ethanol without additives has on the operation of
a Diesel engine. In order to address the blend stability issue, the temperature of
the fuel was maintained at 35 °C. For the comparison with pure Diesel operation
the same test conditions were applied. The acquired data showed an increase in
engine performances (power, torque) due to an increased indicated mean
effective pressure. The higher values of the pressure come from an improved
combustion, a conclusion which is supported by the pollutants measured in the
exhaust gas: lower CO and O2, higher CO2 and higher NOx (as a result of the
increased combustion temperature). The engine efficiency increases slightly.
Keywords: Combustion Ethanol Diesel Pollution Blend
1 Introduction
Emission regulations are getting ever tighter and therefore, the car manufacturers task
of finding solutions to this issue is increasingly harder. In-cylinder and aftertreatment
measures are not always enough to comply with the pollutant emission limits and
researchers have to turn their attention to fuels as well. Also, there is a global desire to
reduce fossil fuel dependency which lead to the mandatory use of biofuels. If the
current limit of biofuel content in Diesel is 5 %, by 2020 this limit will increase to
20 % [1]. For this study, the use of ethanol as biofuel component in Diesel-ethanol
blend was considered as a possible solution. There are numerous studies regarding the
use of Diesel-Ethanol blends in Diesel engines [2–4]. However, these studies rely on
additives to maintain the blend stability. The diversity of additives used is believed to
be one of the causes that leads to contradictory results in this research area [2, 5].
Studies that focused on the parameters influencing the blend stability showed that the
blend stability has a strong temperature dependency [6, 7]. This means, that the
maximum ethanol percentage that can be mixed with Diesel fuel without the occur-
rence of phase separation is determined by temperature. In order to overcome these
problems this study focused on maintaining blend stability without the use of additives.

2 Methodology
The experiments were performed in a controlled environment using a single cylinder

research engine from the “Laboratory of research, testing and homologation of internal
combustion engines that run on biofuel” of the Department of Automotive and
Transport within the Technical University of Cluj-Napoca. The main technical speci-
fications of the engine and the test parameters are presented in Table 1.
Table 1. Engine data and test parameters

Parameter Value M.U.
Bore × Stroke 90 × 85 mm × mm
Compression ratio 17.1 –
Rated speed 4200 min−1
Maximum Power 6 kW
Injection system/Injector Bosch CRDI/Central
injector
(8 × 0.12 × 162°)
Test parameters
Ambient pressure 960 mbar
Ambient temperature 24 °C
Relative humidity 25 %
D100 density 837 kg/m3
D90E10 density 831 kg/m3
The equipment used for data acquisition and to maintain a controlled environment
during the experiments comprises of:
• ambient pressure and temperature sensors;
• dynamometer – used to start and turn the engine with speeds up to 12000 min−1;
• fuel conditioning unit:
– AVL Fuel Mass Flow Meter FlexFuel (735) – destined to measure the fuel mass
flow, consumption and density;
– AVL Temperature control FlexFuel (753) – capable of maintaining a fuel
temperature between 10 and 80 °C with a maximum deviation of ±1 °C;
• coolant conditioning unit – maintains the coolant temperature to the desired value in
the interval 40–120 °C with an error of ±5 °C;
• indicated pressure sensor (type AVL GU22C);
• PUMA and INDICOM control systems
– PUMA is the system used to control all test chamber parameters and for the
acquisition of engine performance data;
– INDICOM is a software connected to the engine control unit (ETAS ETK 7.1)
that allows the user to control injection parameters like timing, quantity and rail
pressure.
The Effects of a Diesel-Ethanol Blend Without Additives 389
The major issue when using Diesel-Ethanol blends is phase separation. There are
various methods proposed in the literature to address this problem [2, 8]. The majority
of authors that studied the influence of ethanol addition on the performances of Diesel
engines opted for a stabilizing agent (biodiesel, tetrahydrofuran, AG Reitz etc.). For the
present study, the solution used to maintain the stability of a 90 % v/v Diesel and 10 %
v/v Ethanol blend was to raise the temperature of the blend thus eliminating the
influence of additives on the combustion process (Table 2).
Table 2. Test cases

Case Engine Pilot Pilot Main Main Rail
number speed timing quantity timing quantity pressure
[rpm] [°CA] [mm3] [°CA] [mm3] [bar]
1 1500 20,625 2,34 5,625 22,85 700
2 21,750
3 22,875
4 20,625 6,750
5 21,750
6 22,875
7 20,625 7,875
8 21,750
9 22,875
10 2000 22,125 2,34 8,625 25,78 800
11 23,250
12 24,375
13 22,125 9,750
14 23,250
15 24,375
16 22,125 10,875
17 23,250
18 24,375
19 2500 23,250 2,34 10,5 23,42 800
20 24,375
21 25,500
22 23,250 11,625
23 24,375
24 25,500
25 23,250 12,750
26 24,375
27 25,500
Based on the obtained performance values the engine efficiency (g) was calculated
with the formula [9]:
P
g¼ ð1Þ
Vfuel qfuel LHV
where P is the engine power, qfuel the fuel density, Vfuel the injected fuel volume and
LHV the Lower Heating Value.
3 Results
Test revealed that by adding ethanol, which is a fuel with high oxygen content, the
combustion of the Diesel is improved. As a result, more fuel is burnt, the heat release is
prolonged (Fig. 1) and there is a considerable increase in the indicated mean effective
pressure (IMEP) for all the test cases (Fig. 2). Also, for the studied cases, there was no
noticeable influence of the injection timing on the differences between the D100 and
D90E10 operation. By varying pilot and main injection timing the same variation trend
could be observed for both fuels.
Fig. 1. Heat release rate for case 1
The idea of the improved combustion is supported not only by the increase in
pressure but also by a higher brake thermal efficiency (BTE) as can be seen in Fig. 3.
Pure Diesel combustion is a slow process (part constant volume, part constant pressure)
that leads to high heat losses to the cylinder walls. As a result there is an efficiency loss
proportional to the amount of heat lost. Ethanol combustion is a single stage process
that develops with superior velocities [10] and therefore can be considered as a constant
volume process. By replacing Diesel fuel with ethanol a part of the heat is released
faster than it can be transferred to the cylinder walls, leading to an increase in effi-
ciency. The BTE was determined from the brake power values measured by the
dynamometer and the theoretical quantity of heat contained in the injected fuel.
Fig. 2. Indicated mean effective pressure as a function of injection parameters
Fig. 3. Brake thermal efficiency
The evolution of the oxygen (O2), carbon monoxide (CO) and carbon dioxide
(CO2) content in the exhaust gases during the experiments is shown in Fig. 4. Due to
the fact that the measured values for CO are very close to each other and under the limit
of the measuring device resolution, there isn’t a clearly distinguishable difference in the
graphs. However, based on the O2 and CO2- content of the exhaust gases it can be
concluded that by replacing 10 % v/v Diesel with ethanol, the CO emissions decrease.
As stated before this is a result of an improved combustion which leads to a higher
conversion factor of CO into CO2 (up to 10 % increase for D90E10) with consumption
of O2 (up to 4 % lower values for D90E10).
In spite of the higher oxygen concentration in the combustion chamber due to
ethanol, there is a decrease in the oxygen content of the exhaust gases. This trend is
valid for all test cases.
The downside of the improved combustion is that there is also an increase in the
mean temperature inside the combustion chamber (Fig. 5). This trend is supported by
the higher exhaust gas temperatures which can be observed in Fig. 6. As a result of the
Fig. 4. O2, CO and CO2 emissions
Fig. 5. Comparison of in-cylinder temperature for case 1
Fig. 6. NO, O2 and exhaust gas temperature

Fig. 7. Power and torque
higher temperature the formation process of thermal NO is favored and a significant

increase in NOx (5–18 %) can be observed. For retarded pilot injections the difference
between the results obtained with the two fuels decreases while for high injection
advances, the difference increases significantly. This shows that even in small pro-
portions, the higher heat of vaporization of ethanol [9] has a noticeable influence on the
combustion temperature.
Due to the benefit of the improved combustion, the engine performances increased
for all test cases with values between 3.6 and 28 % as can be seen in Fig. 7.
4 Conclusions
The purpose of this paper was to determine the influence of a mixture consisting of
90 % v/v Diesel and 10 % v/v Ethanol without additives on the performances and
pollutant emissions of a Diesel engine compared to D100 operation. The main novelty
of the paper is that the blend stability was assured by maintaining the temperature of the
blend at 35 °C. The general conclusions that can be drawn from the obtained results
are:
– The addition of ethanol led to an increase in engine power and torque.
– There is an improvement of the fuel combustion which is supported by the com-
position of the exhaust gases namely, a reduction of the O2 and CO content and an
increase in CO2.
– Due to the improved combustion the maximum and indicated mean effective
pressure increase, which translate into higher temperatures for longer times (higher
exhaust gas temperatures) that ultimately lead to an increase in NOx emissions.
Acknowledgement. This work was partially supported by the strategic grant POSDRU/
159/1.5/S/137070 (2014) of the Ministry of National Education, Romania, co-financed by the
European Social Fund – Investing in People, within the Sectorial Operational Programme Human
Resources Development 2007–2013.
This work would have not been possible without the hardware and software of the AVL LIST
GmbH Company, available as part of the partnership between AVL LIST GmbH Company and
the Technical University of Cluj-Napoca.
References
1. European Parliament C of the EU: Directive 2009/28/EC of the European Parliament and of
the Council of 23 April 2009 (2009). http://eur-lex.europa.eu/legal-content/EN/ALL/?uri=
CELEX:32009L0028. Accessed 24 July 2015
2. Burnete, N.V., Filip, N., Barabás, I.: Diesel-ethanol blends and their use in diesel engines.
J. Automot. Eng. 89, 88–106 (2015)
3. Wang, X., Cheung, C.S., Di, Y., Huang, Z.: Diesel engine gaseous and particle emissions
fueled with diesel–oxygenate blends. Fuel 94, 317–323 (2012)
4. Moon, S., Tsujimura, T., Oguma, M., Chen, Z., Huang, Z., Saitou, T.: Mixture condition,
combustion and sooting characteristics of ethanol-diesel blends in diffusion flames under
various injection and ambient conditions. Fuel 113, 128–139 (2013). doi:10.1016/j.fuel.
2013.05.060
5. Agarwal, A.K.: Biofuels (alcohols and biodiesel) applications as fuels for internal
combustion engines. Prog. Energy Combust. Sci. 33, 233–271 (2007). doi:10.1016/j.pecs.
2006.08.003
6. Torres-Jimenez, E., Jerman, M.S., Gregorc, A., Lisec, I., Dorado, M.P., Kegl, B.: Physical
and chemical properties of ethanol-diesel fuel blends. Fuel 90, 795–802 (2011). doi:10.1016/
j.fuel.2010.09.045
7. Lapuerta, M., Armas, O., García-Contreras, R.: Stability of diesel-bioethanol blends for use
in diesel engines. Fuel 86, 1351–1357 (2007). doi:10.1016/j.fuel.2006.11.042
8. Hansen, A.C., Zhang, Q., Lyne, P.W.L.: Ethanol-diesel fuel blends - a review. Bioresour.
Technol. 96, 277–285 (2005). doi:10.1016/j.biortech.2004.04.007
9. Rakopoulos, C.D., Antonopoulos, K.A., Rakopoulos, D.C.: Experimental heat release
analysis and emissions of a HSDI diesel engine fueled with ethanol-diesel fuel blends.
Energy 32, 1791–1808 (2007). doi:10.1016/j.energy.2007.03.005
10. Saxena, P., Williams, F.A.: Numerical and experimental studies of ethanol flames. Proc.
Combust. Inst. 31, 1149–1156 (2007)
The Influence of Biodiesel on Engine
Combustion and CO2 Emissions
Bogdan Cornel Benea(&)
Mechanical Engineering Faculty, Automotive and Transportation Department,

b.benea@unitbv.ro
Abstract. This paper discusses the influence of the biodiesel obtained from
grape seeds oil, palm oil and sunflower oil on the energetic performances and
combustion process of a Diesel engine. In this study pure diesel fuel and two
concentrations (6 % and 10 %) of blends with biofuels were used for experi-
mentally tests on a Renault K9K diesel engine. During experimental tests were
observed: engine power, cylinder pressure, heat release and the amount of CO2
emissions. The results obtained with biodiesel are compared with the results
obtained with pure diesel fuel. Better oxidation of the biofuels induces a better
combustion in cylinder and more CO2 emissions. The CO2 emissions are either
influence by the lower carbon content of biofuels. The results of this study
sustain that using 6 % and 10 % of blended biofuels with diesel fuel decrease
the pollutant emissions of the diesel engine. Deviations between experimental
and the simulation results confirm the validity of the mathematical model
adopted for the simulation.
Keywords: Biofuels Pollution Diesel engine Heat release rate
1 Introduction
In terms of pollution, fuel combustion represents a major source of emissions of

greenhouse gases, especially CO2. There are numerous alternative options for replacing
oil with natural gas, coal, biomass, hydropower, wind power, solar energy, marine
energy and nuclear energy. All this can be converted into electricity and hydrogen.
Carbon dioxide is the main greenhouse gas effect emitted due to human activities:
burning fossil fuels (coal, gas, oil) for energy and transport, industrial processes and
land use change.
Demand for fossil fuels is considerable. Assessment is that reserves will be
exhausted in the next 100 years (the current global reserves of fossil fuels is about 1383
billion barrels of oil and 187.1 trillion m3 of natural gas [1]). These reserves are
unevenly distributed around the globe, over 50 % of the oil hovering in the Middle
East.
As oil prices have an upward trend, and fossil resources are limited, it requires the
production of alternative fuels and exploitation of alternative resources that have cer-
tain advantages: reducing emissions of greenhouse gases and other pollutants, securing
supply through diversification of energy producers, creating new jobs.

396 B.C. Benea
Due to the increasing price of fossil fuels, using biomass for energy in industry,
construction and transport experienced a superior use in the recent years. Biomass and
bioenergy represents a priority in energy policy by the advantages it presents: avail-
ability of raw materials, reduction of CO2 emissions.
This paper present the influence of blends of biodiesel from sunflower oil, palm oil
and grape seeds oil to engine power, the amount of heat release and CO2 emissions.
2 Simulation Model
For the modelling of the combustion process in the cylinder was used AVL Boost
program. For the simulation of combustion in the cylinder it is assumed that the fuel
injected into the cylinder ignites instantly, combustion products mix with the rest load
in the cylinder, leading to variation in air/fuel ratio between a high values at the start
burning until the final value at the end combustion.
The heat transfer to the walls of the cylinder, cylinder head, piston, cylinder liner is
calculated from [2]:
Qwi ¼ Ai aw ðTc Twi Þ ð1Þ
Qwi – wall heat flow (cylinder head, piston, liner)

Ai – surface area (cylinder head, piston, liner)
αi – heat transfer coefficient
Tc – gas temperature in the cylinder
Twi – wall temperature (cylinder head, piston, liner)
For the calculation of heat transfer coefficient at partial load operation it was used
the heat transfer model Woschni 1990 [3] given by:
( " #)0;8
0;2 0;53 VTDC 2 0;2
aw ¼ 130 D pc Tc
0;8
c1 cm 1 þ 2 IMEP ð2Þ
V
D – cylinder bore
cm – mean piston speed
pc – cylinder pressure
Tc – cylinder temperature
VTDC – TDC volume in the cylinder
V – actual volume
IMEP – indicated mean effective pressure
For the gas exchange process it used next equation:
aw ¼ 130 D0:2 p0:8 0:53
c Tc ðC3 cm Þ0:8 ð3Þ
cu
C3 ¼ 6:18 þ 0:417 ð4Þ
cm
The Influence of Biodiesel on Engine Combustion and CO2 Emissions 397
αw – heat transfer coefficient

cu – circumferential velocity
During the process of gas exchange is essential to consider the transfer of heat from
the inlet and outlet channels of the cylinder head [4]. To calculate heat transfer from the
valve and valve seat it is using the calculation model Zapf, modified according to the
relation:

a p
Aw mc
_ p
T d ¼ ð Tu T w Þ e þ Tw ð5Þ
Heat transfer coefficient depends on the direction of the gas flow (to and from the
cylinder).
The heat transfer coefficient in case of evacuation of gas is calculated with the
relation [5, 6]:

hv
ap ¼ C4 þ C5 Tu C6 Tu2 Tu0:44 m_ o:5 dvi1:5 1 0:797 ð6Þ
dvi
The heat transfer coefficient for gas intake is calculated by the formula:

hv
ap ¼ C7 þ C8 Tu C9 Tu2 Tu0:33 m_ 0:68 dvi1:68 1 0:765 ð7Þ
dvi
αp – heat transfer coefficient in the port, Td – downstream temperature, Tu –

upstream temperature, Tw – wall temperature, Aw – port surface area, m_ – mass flow
rate, cp – specific heat at constant pressure, hv – valve lift, dvi – inner valve seat
diameter, C4, C5, C6, C7, C8 – given constant.
Figure 1 presents the model used for engine simulating.
Fig. 1 Virtual model of engine

398 B.C. Benea
3 Experimental Equipment and Test Description
Tests were performed in the Department of Automotive and Transport of Transilvania

University of Brasov. They ware made using mineral diesel fuel and blends of biodiesel
produced from grape seeds oil, palm oil and sunflower oil. The mixtures were made in
two different volumes of biodiesel blended with diesel fuel. The volume ratios of
biodiesel fuel were 6 % for B6 and 10 % for B10. Fuels properties are presented in
Table 1.
Table 1 Tested fuels properties

Fuel D B6GS B10GS B6P B10P B6SF B10SF
Density at 20 °C (kg/m3) 840.2 841.6 843.4 842.1 844.1 841.9 843.8
Viscosity at 20 °C (mm2/s) 5.34 5.1 4.93 5.32 5.25 5.27 5.1
Cetane number 51.1 57.7 62.5 58.3 62.7 54.5 57.6
Flash point (oC) 67 71.2 73.4 76.6 73.4 67.2 67.8
Caloric value (MJ/kg) 43.16 42.51 42.07 42.58 42.19 42.58 42.19
D – Diesel fuel, B6GS – 6 % biodiesel from gape seeds oil, B10GS – 10 % biodiesel
form grape seeds oil, B6P – 6 % biodiesel from palm oil, B10P – 10 % biodiesel from
palm oil, B6SF – 6 % biodiesel from sunflower oil, b10SF – 10 % biodiesel from
sunflower oil
The choice of mixture of 10 % was made in accordance with Directive 2009/28/EC

which stipulates that by 2020 the level of biofuel in fossil fuels should be 10 %.
The experimental tests were made on a 4 cylinders Renault diesel engine. This
engine is equipped with a common-rail injection system.
The engine specifications are presented in Table 2.
Table 2 Engine properties

Engine type Renault K9K
Number of cylinders 4
Bore (mm) 76
Stroke (mm) 80.5
Total displacement (cm3) 1451
Compression ratio 15.3
Fueling Common-rail
The engine is mounted on a Horiba Titan 250 engine test bed. The engine test bed
is equipped with an electric Dynas3 LI250 dynamometer, which is designed for
operated within a range of 0–8000 rotations per minute. It can measure engine power
up to 250 kW with an accuracy of ±2 %. Pressure in combustion chamber was
measured with a Kistler 6005 piezoelectric pressure transducer. This pressure trans-
ducer can measure up to 1000 bar and it has an accuracy of ±0.8 % [7].
For the measurement of CO2 emissions was used the HGA 400 Pierburg chemi-
luminescence analyser. This analyser can measure CO2 emission between 0 and 20 %
vol, with an accuracy of 0.01 %.
The engine tests were made for full throttle position for two different speeds:
2700 rpm (maximum torque speed) and 3700 rpm (maximum power speed).
4 Results
Figure 2 shows the experimentally and numerically results of engine’s power. The
highest experimentally power was obtained for B6 from sunflower oil. For all test with
blend of biodiesel the obtained power is higher than mineral diesel. Due to the higher
density of biodiesel blends it is injected in cylinder a greater mass of fuel that com-
pensates the lower calorific power.
76
D_EXP
74 D_S
B6SF_EXP
72
B6SF_S
70 B10SF_EXP
Power [kW]
B10SF_S
68 B6GS_EXP
B6GS_S
66 B10GS_EXP
B10GS_S
64
B6P_EXP
62 B6P_S
B10P_EXP
60 B10P_S
2700 3700
Engine speed [rpm]
Fig. 2 Numerical and experimental engine power
In Fig. 3 is represented the maximum cylinder pressure.

Figures 4 and 5 shows the heat release for 6 % and 10 % blends. It can be observed
that the blends have a shorter ignition delay. In first phase of burning heat release rate is
higher for blends than for mineral fuel because of higher volatility and better oxidation.
400 B.C. Benea
115
D_EXP
110 D_S
Pressure [bar]
105 B6SF_EXP
B6SF_S
100
B10SF_EX
95 P
B10SF_S
90
B6GS_EXP
85 B6GS_S
2700 3700
Engine speed [rpm]
Fig. 3 Experimental and numerical maximum cylinder pressure
Fig. 4 Heat release for 6 % blends
CO2 emissions were higher for all blends to mineral fuel. Shorter ignition delay of
blends leads to a more complete combustion. The presence of the supplemental oxygen
in biodiesel leads to a supplementary oxidation of CO obtained after fuel burning,
contributing to the growth of CO2 emissions (Fig. 6).
Fig. 5 Heat release for 10 % blends
Fig. 6 CO2 emissions
5 Conclusions
The effect of biodiesel from sunflower oil, grape seed oil and palm oil were investi-
gated in this paper. By experimental tests and numerical simulations were analyzed the
relationship between fuel properties and combustion process and engine emissions.
• Higher density of blends causes a higher mass of fuel injected in the cylinder. The
lower vapor content cause a shorter delay of auto ignition.
402 B.C. Benea
• The heat release for blends is higher than for mineral fuel.
• The higher oxygen content of blends increases the CO2 emissions. However, due to
the oxygen consumed during photosynthesis process, the global emission of CO2
decreases.
• The best performances of the engine were obtained with B6 from grape seeds oil.
This biofuels has the highest price from all biofuels tested.
• Deviations between the values obtained from simulations and real tests are small.
The mathematic model may be useful for theoretical research for other types of
fuels.
References
1. http://www.bp.com/content/dam/bp/pdf/statistical-review/statistical_review_of_world_
energy_2013.pdf
2. AVL Boost 2010, Theory, AVL List GMBH Graz
3. Woschni, G.: Einfluß von Rußablagerungen auf den Wärmeübergang zwischen Arbeitsgas
und Wand im Dieselmotor. In: Proceedings to Der Arbeitsprozeß des Verbrennungsmotors,
Graz (1991)
4. Kegl, B.: Influence of biodiesel on engine combustion and emission characteristics. Appl.
Energy 88, 1803–1812 (2011)
5. Chmela, F., Orthaber, G.: Rate of Heat Release Prediction for Direct Injection Diesel Engines
Based on Purely Mixing Controlled Combustion, SAE Paper 1999 01 0186 (1999)
6. Merker, G.P.: Combustion Engines Development: Mixture Formation, Combustion, Emis-
sions and Simulation. Springer, New York (2012). 9783642029516
7. http://www.kistler.com/ca/en/product/pressure/6005
Modeling the Performances of a Vehicle
Provided with a Spark Ignition Engine Using
a Double Energy Source, Gasoline
and Liquefied Petroleum Gas
Valentin Nişulescu1(&), Florian Ivan2, Gheorghe Fratila1,

and Gheorghe Banca1
1
University Politehnica Bucharest, Bucharest, Romania
nisulescu.valentin@yahoo.com
2
University of Piteşti, Pitești, Romania
Abstract. The paper contains in the first part a comparative study of the
consequences involved by the use of two power sources for the spark ignition
engine respectively the gasoline and liquefied petroleum gas, highlighting the
criteria of dynamicity, economy and pollution. In the second part, the paper
presents the results of the modeling of the fuel consumption, respectively the
criteria for economy, results reported and conducted using the European running
cycle NMVEG for both power sources, gasoline and liquefied petroleum gas. In
the third part is presented the practical case, based on a project study which
involved the identification of all operating points of the engine after the Euro-
pean running cycle NMVEG and respectively the estimation of the polluting
emissions resulted from the combustion process taking place both with gasoline
and with liquefied petroleum gas for a vehicle provided with a spark ignition
engine, using as the processing and simulation software AMESim “Advanced
Modeling Environment for Simulation of engineering Systems”.
Keywords: Engine Gasoline Liquefied petroleum gas Pollution

Modeling
1 Introduction
Considering the larger and continuously expanding number of vehicles manufactured at a

global level, provided from the factory with a double power source respectively gasoline
and liquefied petroleum gas, and also the numerous technical re-conversions with liq-
uefied petroleum gas installations mounted in the post-sale period, the liquefied petro-
leum gas presents itself as an alternative power source for the engines of the cars [2].
The current studies show a significant increase of the interest of the car manu-
facturers and also of the users to the liquefied petroleum gas, at the current level the
production and distribution networks for the LPG being very developed in the com-
munity, extra-community and global areas. At the same time it imposes on the market
low procurement prices for the final beneficiary due to the low production costs, being
a derivative product resulted from the refinement processes for crude oil or directly

404 V. Nişulescu et al.
from the extraction, after the separation from the natural gas or crude oil with which it
is mixed in the oil deposits.
Also studies conducted on the internal combustion engines using as an alternative
power source the liquefied petroleum gas showed the following technical aspects
related to it:
• The liquefied petroleum gas has a higher octane number than gasoline, between 102
and 110, which allows for an increased compression ratio, thus being permitted the
increase of the thermal efficiency of the engine [4].
• The life span of the LPG engine is increased compared to the gasoline one because
it does not wash the mirror of the cylinder during the injection, does not allow the
formation of deposits on the components from the combustion chamber, and does
not compromise the electrodes of the sparkplugs as ignition elements.
• The combustion of the liquefied petroleum gas also increases the life span of the
components for the treatment of the exhaust gases.
• Respectively oxygen probes, catalyst or nitrogen oxides filter.
• From the perspective of the pollution the liquefied petroleum gas presents important
reductions compared to the gasoline for all polluting emissions, HC, CO, NOx and
CO2, resulted from the combustion process.
Along with the advantages presented above the LPG also has some disadvantages
compared to the base fuel, respectively:
• Increased consumption, as previously shown, between 10–20 %
• Reduced power and torque compared to the gasoline up to a maximum of 10 %.
• Increased hazard in the case of improper sealing of the storage and injection
installations because it is superior as weight compared to the air weight, the max-
imum gas concentrations being found at the ground level [4].
1.1 Description of the Model

As specified above, as the processing and simulation software was employed the
program AMESim. The program is run for a vehicle of middle class, equipped with a
spark-ignition engine and manual gearbox with ratios parameterized according to the
model of the chosen vehicle. The modeling is based on internal schematics for torque,
polluting emissions and fuel consumption identified based on an actuation field for an
engine of 1.6L MPI 75 kW.
Were also defined for model the following components:
– A component that simulates the driver component employed to ensure the control of
the vehicle while running the NMVEG cycle. (Fig. 1)
– A component that simulates the exhaust gas treatment system specific for a spark
ignition engine respectively three-way catalyst.
– A component that simulates the engine control unit.
The components listed above represent the main input values for the model, but at
the same time it also uses other subcomponents, which provide information regarding
Modeling the Performances of a Vehicle Provided with a Spark Ignition Engine 405
Fig. 1. AMESim model for a vehicle provided with a spark ignition engine
the moment of inertia, engine torque, coupling stages, temperatures, masses of pol-
luting emissions, which are the basis of the internal calculation algorithms. The model
employed is intended to identify, based on the internal mathematical algorithms, the
differences of performance, economic and ecological based on the consecrated for-
mulas, differences shown in agreement with the running conditions imposed by the
European cycle NMVEG for the two power sources [3].
The two types of fuel are characterized by different values of density and calorific
power.
2 Experimental
In the first part the modeling supposed the identification of all operating points of the
engine according to the European running cycle NMVEG and respectively the estima-
tion of the dynamic performance, fuel consumption and the mass of pollutant emissions
resulted from the combustion process taking place both with gasoline and LPG.
The torque of the engine is obtained from the matrix torque – engine speed “torque.
data” relation (1)
M ¼ Mtab þ Mpc Mpr ð1Þ
where:
– M is the engine torque [Nm].
– Mtab is the torque read from the database “torque.data” [Nm].
– Mpc is the friction torque at high engine temperature from “FMEP.data” [Nm].
– Mpr is the friction torque at low water temperature or oil temperature read from the
Table 1D “FMEP.data” [Nm].
During the acceleration phases to the value of output torque is applied a correction
to be able to consider the actual environmental conditions, the calculation method
being described in relation (2).
qa
Me1 ¼ Md ð2Þ
1; 205
where:
– Me1 is the output torque at port 1 [Nm].
– Md is the dynamic engine torque [Nm].
– qa is the air density [kg/m3].
With regards to the average effective pressure corresponding to output maximum
torque for the p me brake, it results from the following relation (3):
4 p Me1 102
pme ¼ ð3Þ
Vt
where:
– pme average effective pressure for braking [bar].
– Me1 the torque for the engine shaft [Nm].
– Vt the engine displacement [ l ].
Eventually the effective power is calculated using the formula (4)
Pe ¼ Me1 x 103 ð4Þ
where:
– Pe the effective power [kW].
– Me1 the actuating torque for the shaft [Nm].
– ω the angular velocity of the crankshaft [rad/s].
The determination of power losses by friction, it is assumed to be identical for both
power sources employed, this type of losses being calculated using the formula (5).
Pp ¼ Mpr x ð5Þ
where:
– Pp power losses by friction[kW].
– Mpr the friction torque at low water temperature or oil temperature read from the
Table 1D “FMEP.data” [Nm].
– ω the angular velocity of the crankshaft [rad/s].
With regard to the fuel consumption and the mass of CO2 emissions is required the
establishing of the flow of exhaust gases, which is made by determining the fuel
consumption and respectively the equivalence ratio using the relation (6).

A
m_ ev ¼ m_ c þ1 ð6Þ
F
where:
– m_ ev the flow of the exhaust gases [g/s].
– m_ c the fuel consumption in identifiable in the table “conshot_i.data” [kg/s].
– FA the stoichiometric ratio air/fuel.
– the equivalence ratio air/fuel.
The calculation of the flow of exhaust gases is made only when the fuel injection is
executed. During the operation of the combustion engine with engine brake, the
injection is stopped for the duration of the whole process which means the flow of
exhaust gases is proportional with the revolution of the engine and the exhaust gases
flow specific for the idle operation [3].
For the spark ignition engine the exhaust gases flow during the idle operation is
calculated from the consumed fuel quantity for the idle operation and an equivalence
ratio established at 1.
During the engine brake when the value of the injected fuel is 0, the value of the
polluting emissions is also 0 and the temperature of the exhaust gases is established to
be identical to the temperature of the cooling fluid. [1] The fuel consumption is cal-
culated as follows relation (7)
tw tmin
at ¼ ð7Þ
tmax tmin
bt ¼ 1 at ð8Þ
m_ c ¼ ðat b kc Þ m_ ctab ð9Þ
where:
– m_ c fuel consumption [g ].
– Tw temperature of the cooling fluid [°C].
– Tmax the maximum threshold of the temperature of the cooling fluid, parameter
corresponding to the temperature of the cooling fluid at the operation at combustion
status [°C].
– Tmin the minimum threshold of the temperature of the cooling fluid, parameter
corresponding to the temperature of the cooling fluid at the operation at cold status
[°C].
– kc the correction coefficient corresponding to the fuel consumption for the cold
operation with the condition (Tw ≤ Tmin),
– m_ ctab the fuel quantity consumed from the table “conshot_i.data” [g ].
At last, the CO2 emissions are estimated as directly proportional with the fuel
consumption (conversion factor parameter) as specified in the Eq. (8)
m_ CO2 ¼ m_ c aCO2 ð10Þ
where:
– m_ CO2 the mass of CO2 emissions [g].
– m_ c fuel consumption [g].
– aCO2 transfer coefficient CO2
The calculation of the polluting emissions of HC, CO and NOx resulted from the
combustion process considers the air flow admitted in the engine cylinders and the
conversion factors for the efficiency of the catalytic converter. Each conversion effi-
ciency has a value between 0 and 1 and can be defined by a formula depending on the
temperature of the catalyst wall twC [°C], the equivalence report and the stationary
time for the gases within the catalytic converter tTWC [s] (9).
All the values for the catalytic efficiency are checked by the program and mini-
mized to 0 and maximized to 1. The global stationary time inside the catalytic converter
is calculated as follows:
mgaz
tTWC ¼ ð11Þ
jdm2 j
where:
– tTWC the stationary time for the gases within the catalytic converter [s].
– mgaz the mass of gas from the catalytic converter [kg].
– dm2 the quantity of air entered in the engine cylinders [kg/s].
With regard to the maximum standing time inside the catalytic converter, it is
limited to 105 s. The amount of gas from the catalytic converter is calculated with
formula (10):
pV
mgaz ¼ ð12Þ
T R
where:
– p the pressure inside the exhaust ducts [Pa].
– V represents the open volume [m3].
– T the temperature of the exhaust gases in the port [3].
– R the specific gas constant 287.58 [J/kg K].
The heat due to the conversion of pollutants and transferred to the monolith of the
catalytic converter (Fig. 2) and dhreac is calculated based on the input and output of the
catalyst and from various exothermal internal chemical reactions (relation 11)
ðCOi COe Þ H reac CO

dhreac ¼ þ ðHCi HCe Þ
28
ðNOxi NOxe Þ H reac NOx
inc HC þ ð13Þ
30
where:
– COi mass flow of the CO in the port 4 [g/s].
– COe mass flow of the CO in the port 3 [g/s].
– HCi mass flow of the HC in the port 4 [g/s].
– HCe mass flow of the HC in the port 3 [g/s].
– NOxi mass flow of the NOx in the port 4 [g/s].
– NOxe mass flow of the NOx in the port 3 [g/s].
– H_reac_CO 2; 83 105 [J/mol],
– inc HC: the heat resulted from the HC conversion [kJ/kg].
Fig. 2. Input and output ports of the Catalyst
With regard to the results obtained from the modeling, several graphs were prepared,
highlighting the differences between the fuel consumption, engine torque and polluting
emissions of CO, HC, NOx for the two power sources of the studied engine. At the
same time were presented the torque losses generated by the internal frictions from the
engine mechanism.
Fig. 3. Diagram of the engine torque - friction losses

According to the diagram from Fig. 3, the maximum at the operation with the
power source gasoline is approximately 5 % larger than the maximum torque for the
operation with the power source LPG.
Table 1. Fuel consumption for the NMVEG cycle

Fuel consumption
Cycle Measuring NMVEG NMVEG LPG Difference
units gasoline gasoline-LPG
Distance traveled [km] 11058 11058 –
Fuel consumption [g] 558.3609 635.0938 12,08 %
In Table 1 are presented the compared results of the fuel consumption for the
NMVEG cycle, for the two sources of energy, observing a decrease of the fuel con-
sumption in the LPG operating mode of 12.08 % from the gasoline operating mode.
Table 2. CO emissions for the NMVEG cycle

CO polluting emissions
Cycle Measuring NMVEG NMVEG Difference gasoline
units gasoline LPG - LPG
Distance traveled [km] 11.058 11.058 –
Engine emissions [mg] 40261.95 33472.87 16,86 %
Exhaust emissions [mg] 2876.553 2265.83 21,23 %
Catalyst efficiency [%] 92,85 93,23 0,38 %
Table 3. HC emissions for the NMVEG cycle

HC polluting emissions
Engine emissions [mg] 8773.438 6860.296 21,8 %
Exhaust emissions [mg] 739.5619 557.0177 24,68 %
Catalyst efficiency [%] 91.57 91.88 0,31 %
In addition in the Tables 2, 3 and 4 are presented the compared results of the
polluting emissions CO, HC and NOx, results obtained on the NMVEG cycle for the
two power sources. Are presented the values of the polluting emissions measured
directly at the engine exit before the input to the catalyst and the results after the
treatment measured at the level of the exhaust. At the same time can be observed the
percentage differences for each polluting emission between the two power sources both
for the catalyst input and at the exhaust level.
Table 4. NOx emissions for the NMVEG cycle

NOx polluting emissions
Engine emissions [mg] 17485.57 5636.286 67.76 %
Exhaust emissions [mg] 1186.885 393.4694 66.84 %
Catalyst efficiency [%] 93.21 93.01 0,2 %
Based on the two estimations of the emissions before the input of the catalyst and at
the level of the exhaust was established the global efficiency of the catalyst.
In addition were prepared graphs for the fuel consumption, which show the dif-
ferences for the two power sources (Fig. 4).
Fig. 4. Comparative consumption gasoline - LPG
In Figs. 5, 6 and 7 were presented the emissions at the exit from the engine
respectively the output from the exhaust cumulated for the two power sources from
which can be calculated the efficiency of the catalyst for each individual polluting
emission.
Fig. 5. CO cumulated emissions gasoline - LPG
Fig. 6. HC cumulated emissions gasoline - LPG

Fig. 7. NOx cumulated emissions gasoline - LPG
4 Conclusions
Given the results of the modeling for the dynamic performance, economy and polluting
emissions for the NMVEG cycle of the studied engine for both power sources gasoline
and LPG were established the following conclusions:
1. From the perspective of the dynamic performance results, compared to the power
source gasoline, the engine moment using the LPG source is reduced by approxi-
mately 5 %, but the stable operation area of the curve for LPG observes the curve
shape for the gasoline.
2. With regard to the fuel consumption resulted that during the LPG operation the
quantity consumed is larger than for the gasoline operation with approximately
12.08 %.
3. For the CO polluting emissions resulted that during the LPG operation the emis-
sions are reduced compared to the gasoline operation with approximately 21.23 %.
4. For the HC polluting emissions resulted that during the LPG operation the emis-
sions are reduced compared to the gasoline operation with approximately 24.68 %.
5. For the NOx polluting emissions resulted that during the LPG operation the
emissions are reduced compared to the gasoline operation with approximately
66.84 %.
6. At the same time, with regard to the efficiency of the catalyst for the reduction of the
polluting emissions can be observed that between the two power sources there are
no significant differences, the differentiable values being of at most 0.4 %.
7. Given the prior conclusions can be easily stated that the liquefied petroleum gas
represents a very good and current alternative power source for the spark ignition
engines from all perspectives, considering the indicators for dynamicity, economy
and pollution.
8. The liquefied petroleum gas presents an increasing interest for the car manufac-
turers, the number of vehicles provided from the factory with LPG systems is
continuously expanding, having a very large coverage among the manufacturers.
References
1. Ivan, F., Lita, D., Busoi, A.: Methods and means of depolution of the engines for vehicles
(2014). ISBN 978-606-25-0038-2
2. Mamidi, T., Suryawnshi, J.G.: Investigations on S.I. engine using liquefied petroleum gas
(LPG) as an alternative fuel. IJERA 2, 362–367 (2012). ISSN: 2248-9622
3. Help AMESim Rev12
4. Patil, P.N., Deshmukh, D.S., Patil, J.G.: A review on liquid petroleum gas as a fuel in spark
ignition engine application. In: International Conference on Global Trends in Engineering,
Technology and Management (ICGTETM-2016), pp 124–128. ISSN: 2231-5381
Modeling the Performances of a Vehicle
Provided with a Hybrid Electric Diesel
Propulsion System (HEVD)
Gheorghe Bancă1(&), Florian Ivan2, Gheorghe Frățilă1,

and Valentin Nișulescu1
1
banca_2004@yahoo.com, ghe_fratila@yahoo.com,
nisulescu.valentin@yahoo.com
2
florianivan2002@yahoo.com
Abstract. The paper deals in the first section with a comparative study of the
implications of the use of a PDHEVP (Parallel Diesel Hybrid Electric Vehicle
Powertrain) on a vehicle from the perspectives of dynamicity, economy and
pollution.
In the second part the paper presents the results of the modeling of the fuel
consumption using the cycle NMVEG (New Motor Vehicle Emissions Group)
for a vehicle equipped with a PDHEVP.
The practical case is based on a project study involving the identification of
the engine operating points on the cycle NMVEG and the estimation of the
polluting emissions for a vehicle provided with a PDHEVP using the deploy-
ment of simulations using the simulation platform AMESim (Advanced
Modeling Environment for Simulation of Engineering Systems).
Keywords: Hybrid Modeling Economy Pollution NMVEG
1 Introduction
In the context of the awareness regarding the harmful effects of the internal combustion
engines on the environment, the severity of the anti-pollution norms and the reduction
of the fuel consumption come as a natural consequence to limit the irreversible dam-
ages that the automotive industry development can bring to the environment we live in.
The reduction of the CO2 emissions is currently the largest challenge at a global
level, all industrial sectors being directly involved and having clear objectives for the
reduction of CO2 emissions, especially the automotive industry, the CO2 emissions
being directly linked to the fuel consumption [1].
For this, the electric hybridization of the vehicles powered by an internal com-
bustion engine has proven to be one of the viable solutions to reduce the fuel con-
sumption and the polluting emissions of the vehicles, benefiting at the same time from a
preserving or even an improvement of the dynamic performances and their autonomy.

416 G. Bancă et al.
Even if numerous power management strategies were developed, like the ones for
dynamic programming (DP) and minimization strategies based on the Pontryagin
principle (PMP), the reduction of the fuel consumption of a HEV does not guarantee
the reduction of the polluting emissions, the optimal control that minimizes the fuel
consumption could lead to some operating situations for the engine which present high
emissions of pollutants [2].
From another perspective, the one of the deactivation (removal from operation) of
the combustion engine in order to reduce the CO2 emissions, this leads to disadvan-
tages like the removal from the thermal operation of the internal combustion engine and
the loss of priming of the post-treatment systems for the exhaust gases (loss of the
temperature required for the system to effectively treat the gases) like in the case of the
catalyst.
In this paper is presented the optimization of the fuel consumption, evolution of the
polluting emissions and of the dynamic performances, in order to determine a mixed
criterion for obtaining a good compromise for a realistic model of a vehicle provided
with a hybrid electric diesel propulsion system, having two sections:
– the first section conducts a comparative study of the implications of the use of a
PDHEVP on a vehicle from the perspectives of dynamicity, economy and pollution;
– in the second section the paper presents the results of the modeling of the fuel
consumption using the cycle NMVEG for a PDHEVP.
2 The Presentation of the Model
The model is specific to an SUV (Sport Utility Vehicle) middle class equipped with a
diesel parallel hybrid electric propulsion system (Fig. 1). The model was executed
using the multi-domain simulation platform, AMESim, platform specialized in the
modeling, simulation and analysis of complex systems.
Fig. 1. Structure of a parallel hybrid electric vehicle

Modeling the Performances of a Vehicle 417
The AMESim model proposed in Fig. 2 is composed from 8 large components, as

follows:
– parameterized vehicle based on a medium class SUV;
– compression combustion engine - Diesel (component based on schematics for
torque, polluting emissions and fuel consumption identified based on an actuation
field for an engine of 1.5 dCI 81 kW Euro 6);
– mechanical transmission with ratios parameterized according to the chosen vehicle;
– a driver component which is used to control the vehicle on the cycle;
– an electric machine (motor/power generator) of approx. 36 kW;
– a battery of accumulators with a voltage of 310 V;
– a computer VCU-PH (Vehicle Control Unit – Parallel Hybrid) which manages the
power and the operating mode for the combustion engine and the electric engine;
– an exhaust line that includes the post-treatment systems for the exhaust gases
specific to a Diesel vehicle (oxidation catalyst, NOx-trap, particle filter).
Along with these components the model also presents elements like: type of cycle
used for testing, clutch, elements simulating the inertia and frictions and various sensors
to recover information regarding the status, torque, temperatures and polluting emissions.
Fig. 2. Model AMESim proposed for a vehicle provided with PDHEVP
Description of the model: because in the construction of the model were employed
several components that were presented in another paper “Aspects Regarding the
Evolution of the Depollution Norms and Test Cycles in Order to Determine Polluting
Emissions” from the international conference EAEC – EFSA 2015 within this paper we
will only detail the new elements or the ones not presented in the aforementioned paper
which are the post-treatment line for polluting emissions and the high voltage network
and its elements: the battery, computer for managing the power flow and the electric
machine (motor/generator).
The post-treatment line of the presented model is composed from: Diesel oxidation
catalytic converter - DOC, this being the only current solution which ensures the
compliance of the pollution norms for the HC and CO emissions for the compression
combustion engines.
Fig. 3. Ports (inputs-outputs DOC)
DOC Conversion Efficiency

Conversion Efficiency [%]
Conversion efficiency HC %
Conversion efficiency CO %
Monolith temperature [°C]
Fig. 4. Conversion DOC depending on the temperature
The HC and CO emissions at the engine output are used as input for port 4 of the
component and are transferred to port 3 (Fig. 3) after applying the conversion yield of
the catalytic converter depending on the temperature of the monolith, (Fig. 4). Its
conversion yield for each pollutant is between 0 and 100 % and can be defined by the
expression presented by the relation:

m_ pi
gx ¼ 1 100 ð1Þ
m_ pe
where m_ pi and m_ pe are the specific flows of the each pollutant, for the input, respec-
tively, the output, the chemical reactions taking place in a real DOC being given by the
formulas 2 and 3. The main variable that influences the efficiency of this conversion is
the monolith temperature (Tmon).
1
CO þ O2 ! CO2 ð2Þ
2
HC þ O2 ! CO2 þ H2 O ð3Þ
The post-treatment system for the nitrogen oxides, diesel, NOx-Trap, [3] ensures the
storage and reduction of the NOx (NO and NO2). The flow and ratio air/fuel are
transferred from the port 4 to port 3 without changes. The flow of NOx at output of port
3 depends on the input flow for port 4, and the storage, release and reduction ratio is
based on a state variable (concentration of NOx [ppm] defined as an internal variable in
the component model). The NOx storage ratio depends on the NOx mass stored and the
oxygen contained by the exhaust gases passing the NOx-trap, and the calculation of the
oxygen content is based on the equivalence ratio air/fuel. The NOx release ratio
depends only on the quantity of NOx stored in the NOx-trap, and the NOx reduction
ratio depends on the content of NOx and the equivalence ratio air/fuel. All these ratios
(storage ratio, release ratio and reduction ratio for NOx) are based on the Arrhenius law
[4], relation 4:

Ea
k ¼ A exp ð4Þ
R TNOxTrap
where:
– A: pre-exponential factor (given by the user);
– Ea: activation energy (given by the user);

– R: ideal gas constant 8; 3114 molK
J
;
– TNOx-Trap: temperature NOx-Trap in Kelvin.
The post-treatment system for PM (mechanical particles, DPF – Diesel Particulate

Filter), is represented by a submodel 0D particle filter with square channels, which
considers the particle loading speed and the regeneration, the internal and external heat
exchange. The HC mass and flows of NOx entering the port 5 are sent to port 3 without
changes, but the CO flow and the equivalence report air/fuel at port 3 are considered for
the impact on the oxidation of the particles.
The volume temperatures of the DPF and the wall are state variables. They are
determined by using the energy balance equations for gas and solid phases. The con-
vective heat exchanges inside the monolith and with the environmental air are considered,
ports 1 and 2 being dedicated to the external flowing conditions (port 1 being the air speed
around the DPF casing and port 2 being the air temperature around the pipe).
For this model the particle mass is a state variable, depending on the ratio between
the loading speed and the particle oxidation. The storage of the particles can be uniform
or at the end of the filter, depending on the choice of the user, the induced
counter-pressure being calculated in both cases, and the soot oxidation report can be
linear or according to the law of Arrhenius (relation 4).
The porosity and permeability of the filter kf is given by the relation 5 recom-
mended in [3] and vary between kempty and kfull , depending on the covering state of the
filter channels walls with particles:
1
kf ¼ ð5Þ
wp w
w0 kfull
1
þ 1 wp0 kempty
1
where:
– wp : thickness of the particle deposit;
– w0 : thickness of initial particle deposit.
Permeability of the particle deposit - the impact of the temperature on the per-
meability of the particle deposit kp is considered as being of the order of the power
1.125 according to formula 6 [3]:
1;125
Tf
kp Tf ¼ kp ð25 CÞ ð6Þ
298
where:
– kp : permeability of the particle deposit;
– Tf : temperature of the particle filter.
Evolution of the particle deposit - the loading speed or the time derivative of the
particle mass depends on the ratio of the particle flow reaching DPF, being given by the
relation 7:
d
mpinc ¼ eFaP Qp ð7Þ
dt
with:
– mpinc : particle mass deposited in DPF;
– eFaP : filtering efficiency DPF;
– Qp : flow of particles in the exhaust gases.
The oxidation of the particles represents the derivative of the particle mass varied
with time, depending on the oxidation ratio of the particles roxi , given by the relation 8:
d
mpoxi ¼ roxi ð8Þ
dt
where roxi can be a linear function depending on the temperature (relation 9, paper [3])
or given by the law of Arrhenius (relation 10, paper [3]):
Tf Tp
roxi ¼ k0 mp ð9Þ
Tampl
or

Ea
roxi ¼ k0 mncarb exp ð10Þ
p
R Tf
with:
– k0 : constant factor;
– Tf : filter temperature;
– Tp : temperature threshold;
– Tampl : temperature amplitude (difference between maxim and minim temperature);
– mp : particle mass deposited in DPF;
– ncarb: carbon reaction order - the order of the reaction shows the influence of a
reagent (or carbon to the case) in the kinetics of the reaction [5];
– Ea : activation energy;
– R: constant of the ideal gas.
As a consequence of introducing the DPF along the exhaust line we have additional
restrictions with regard to the pressure drop for this element (filter), the evolution of the
particle mass strongly influencing the counter-pressure in the filter after relation 11:
!
wf 1 wp 4 F L 2 l V_ Dc þ wf 2
DP ¼ log 1 2 þ Dc 2 w p
kf Dc kp Tf Dc 3 2V
ð11Þ
with:
– wf : thickness of the filter channel [m];
– kf : permeability, porosity of the filter [m2];
– Dc : width or the equivalent diameter of the channel [m];
– kp : permeability of the particle deposit [m2];
– Tf : filter temperature [K];
– wp : thickness of the particle deposit [m];
– F: factor equal with 28.454 - represents the friction loss in channels;
– L: DPF length [m];
– _ volumetric flow [m3/s];
V:
– V: DPF volume [m3];
– l: dynamic viscosity [kg/(ms)].
The high voltage network and its elements - the battery, computer for managing the
power flow and the electric machine (motor/generator).
The battery - it is modeled as being formed from elements connected in series and
in parallel, each element being composed from cells. The voltage for the open circuit
cells V0cell and the internal resistance Rcell of the cells are read from the data files.
The voltage used with the load Q [As] (or voltage removed from the battery) is
calculated using the relation 12:
dq
¼ I ð12Þ
dt
where:
– I is the current ½A.
The State Of Charge of the battery SOC ½% is a derivative state variable that is
calculated with relation 13:
dSOC dq 100
¼ ð13Þ
dt dt Cnom
where:
– Cnom is the rated capacity ½As . This derivative is limited so that the loading state of
the battery will be in the interval ½0 100 %.
Computer for the management of the power flow - the command unit receives
information from the vehicle driver like acceleration position, coupled ratio of the
transmission, break, torque of the electric motor, torque of the combustion engine,
loading state respectively voltage of the battery and speed of the vehicle. The computer
analyzes them in order to optimize the energy management by minimizing the energy
consumption from the battery, the electric motor also being used as a generator for the
charging of the battery when the vehicle decelerates or brakes.
Electric motor - is an electrical machine, with various operating modes that can be
employed, electric motor/generator, thus the relation between the mechanical power (Pm )
and the electric power (Pe ) to consider the sign convention is the following, relation 14:
Pe ¼ Pm Pp ð14Þ
where:
– Pm \0 for the operating mode electric motor;
– Pm [ 0 for the operating mode electric generator;
– Pp [ 0.
The user can deduct from the power loss (Pp ) efficiency corresponding to the motor
or the generator:
– for the operating mode electric motor, is used the relation 15:
Pe
g¼2 ð15Þ
Pm
– for the operating mode electric generator, is used the relation 16:
Pe
g¼ ð16Þ
Pm
From the perspective of the results of the simulation for the CO2 emissions we observe
an improvement of the fuel consumption by approximately 20–25 % for the configu-
ration HEVD on the cycle NMVEG (Table 1).
Table 1. CO2 emissions on the cycle NMVEG.

Cycle NMVEG Unit Diesel HEVD Variation (Diesel-HEVD)
Norm Euro 6b [g/km] – – –
Mileage [km] 11.058 11.14452 1%
Cumulated engine emissions [g] 1493.65 1155.55 −23 %
Global engine emissions [g/km] 135 104 −23 %
For the HC emissions (Non Burnt Hydrocarbons - Table 2) and CO (Carbon

Monoxide - Table 3), even if the norm is not respected not even in the reference
configuration (Diesel engine) we observe the maintaining of “ISO” level for HC
emissions and a significant reduction of the CO emissions for the HEVD configuration.
By observing the motor and exhaust emissions, for the CO emissions we observed that
even if for the motor emissions the difference is 8 % for the exhaust emissions we can
see a difference of more than 50 %. These is explained by weaker priming on the
catalyst for the HEVD configuration as a result of the deactivation of the combustion
engine and the dropping out of the thermal operation for it and for the DOC (see Figs. 5
and 6).
Table 2. HC emissions on the cycle NMVEG.

Norm Euro 6b [g/km] 0.05 0.05 0%
Mileage [km] 11.058 11.14452 1%
Cumulated engine emissions [mg] 5840.88 5283.972 −10 %
Global engine emissions [g/km] 0.52820 0.47413 −10 %
Cumulated exhaust emissions [mg] 1000.6 997.3836 0%
Global exhaust emissions [g/km] 0.09048 0.09009 0%
Table 3. CO emissions on the cycle NMVEG.

Norm Euro 6b [g/km] 0.5 0.5 0%
Mileage [km] 11.058 11.14452 1 %
Cumulated engine emissions [mg] 45742.4 49977.34 9 %
Global engine emissions [g/km] 4.13659 4.48448 8%
Cumulated exhaust emissions [mg] 7152.67 10930.23 53 %
Global exhaust emissions [g/km] 0.64696 0.98658 52 %
Fig. 5. Activating the combustion engine
Fig. 6. Evolution output temperatures engine

In the case of the NOx emissions (nitrogen oxides) - Table 4, the norm is respected
but we can see a degradation of approximately 36 % - motor emissions and 310 % -
exhaust emissions for the HEVD configuration. The engine emissions increase as a result
of the activation of the energy recovery strategy for part of the cycle with a high speed of
the 4 ECEs (ECE15 or UDC is Urban Driving Cycle), but especially for the EUDC
(Extra-Urban Driving Cycle), where the combustion engine has an additional loading in
order to work in the area of the economic pole (the interval from the motor field where the
specific fuel consumption is optimal). At the same time the difference of approximately
310 % found for the NOx exhaust emissions is the consequence of the existing from the
thermal state of the pollution removal systems from the exhaust line (catalyst, NOx-Trap)
as a result of the deactivation of the combustion engine on the cycle.
Table 4. NOx emissions on the cycle NMVEG.

Norm Euro 6b [g/km] 0.08 0.08 0%
Mileage [km] 11.058 11.14452 1%
Global engine emissions [g/km] 0.42184 0.57560 36 %
With regard to the mechanical particles (PM) - Table 5, even if the norm is
respected also we can see a degrading of *160 % both for engine emissions and for
the exhaust emissions for the HEVD configuration as a result of the activation of the
energy recovery strategy for the high speed areas of the 4 ECEs, but especially on the
EUDC cycle.
Table 5. PM emissions on the cycle NMVEG.

Norm Euro 6b [g/km] 0.005 0.005 0%
Mileage [km] 11.058 11.14452 1%
Global engine emissions [g/km] 0.08330 0.21534 159 %
4 Conclusions
The technology “Hybrid total” without loading from the domestic network associated
with the performing internal combustion engines, allows for a gain of CO2 of
approximately 20 %–25 % or approximately 20 g for a vehicle from the same segment
provided with an engine propulsion group with similar characteristics.
As a result of the deactivation of the internal combustion engine on the testing cycle
and the removal from the thermal state for both the motor and the catalyst the CO
exhaust emissions are smaller with approximately 50 % even if the engine emissions
increased with approximately 10 %.
Even if there is a notable improvement of the fuel consumption by using a hybrid
propulsion system, the impact on the NOx emissions is a negative one both for engine
emissions but especially for exhaust emissions, obtaining a degrading of more than
310 % for these emissions, new solutions being required to achieve the same level of
emissions.
The PM emissions were degraded with approximately 160 % as a result of an
additional load induced by the passing of the electric motor to the generator status on
the extra-urban cycle, which leads to an enrichment of the 0.6 threshold, threshold over
which the particle increase is exponential.
The using of a propulsion system using a GMP-DHEV without the rigorous
optimization of the energy management provides a significant gain for the CO2
emissions with the risk of degrading the other polluting emissions (a reduced fuel
consumption not being synonymous with low polluting emissions) [2].
References
1. Ioan Mircea Oprean: Modern Automobile. Romanian Academy Publishing House, Bucharest
(2003)
2. Michel, P., Charlet, A., Colin, G., Chamaillard, Y., Nouillant, C., Bloch, G.: Energy
management of HEV to optimize fuel consumption and pollutant emissions. In: 11th
International Symposium on Advanced Vehicule Control, AVEC 2012, Sep 2012, Seoul,
South Korea, p. CDROM (2012)
3. Help AMESim Rev12
4. http://uel.unisciel.fr/chimie/cinet/cinet_ch04/co/apprendre_ch4_05.html. Accessed 24 June
2016
5. https://fr.wikipedia.org/wiki/Ordre_de_réaction. Accessed 24 June 2016
Aspects of Experimental Research
on Hydrogen Fuelled Automotive
Diesel Engine
Ionel Mirica, Alexandru Cernat(&), Constantin Pana,


ionel_mirica@yahoo.com, cernatalex@yahoo.com,
constantinpana@yahoo.com,
niculae_negurescu@yahoo.com, cristi_cmt@yahoo.com
Abstract. The use of hydrogen at the diesel engines for improvement energetic
and ecologic their performance is the main objective of the paper. The research
work has been carried out on truck diesel engine at the 55 % load and 1400 rpm.
The engine was fuelled firstly with diesel fuel, then with diesel fuel and
hydrogen at different rates between 11–40 L/min. The minimum brake specific
energetic consumption decreases with *9 % and NOx emission decreases with
5.5 % due to combustion improvement at the hydrogen use. The engine ener-
getic and ecologic performances are improved.
Keywords: Combustion Hydrogen Diesel engine Pollutants emissions

Clean fuel
1 Introduction
Nowadays, in the global content regarding the impact on the environmental of the gases
emissions resulted from the fossil fuels combustion, an interest aspect discussed also on
the 21st Session of the Conference of the Parties from the 2015 Paris Climate Con-
ference in November 2015, and the gradual diminution of the worldwide oil reserves
contribute to the necessity of searching of alternative energy from durable and
renewable resources. The replacement of the classic fuels of oil origin by alternative
fuels was initiated many decades ago and represents one of the most actual preoccu-
pations linked by the possibility of pollutant emissions decrease and efficiency increase
of the diesel engine and necessity of fossil fuel saving [1]. These issues impulse the
research’s for developing and using new fuel sources which will allow the substitution,
even a partial one, of the classic fuels [1, 2]. Among alternative fuels, hydrogen
represents a viable alternative fuel due to its unlimited resources, being defined as the
cleanest fuel. Due to its physical-chemical properties that shows suitable characteristics
for a high efficiency burning process: a burning rate higher comparative to other
hydrocarbons fuels, wide flammability limits, very low ignition energy – hydrogen
provides a stable combustion process for lean and very lean mixtures, encouraging the
wide scale hydrogen use as fuel for engines. Extremely low cetane number and higher

428 I. Mirica et al.
autoignition temperature of hydrogen represent important issues of hydrogen use as

fuel at diesel engine, being necessary a diesel fuel pilot to initiate the combustion
process [1, 2]. The mole fraction of water vapors from the combustion products pro-
duced at the hydrogen combustion increases, the water vapors absorb an important heat
quantity from the energy released from combustion and thus the peak combustion
temperatures decreases [3, 4]. Bose [5] shows the increase in thermal efficiency, the
reduction of BSFC with 55 % for hydrogen fuelling, at full load [5]. Ashok [6] Senthil
Kumar et al. [7] use small quantities of hydrogen in a diesel engine primarily fuelled
with a vegetable oil (Jatropha oil). Results indicated an increase in the brake thermal
efficiency from 27.3 % to a maximum of 29.3 % with 7 % of hydrogen mass share at
maximum power output [6, 7]. Senthil Kumar [7] use hydrogen to fuel a direct
injection diesel engine at speed of 1500 rpm. The brake specific fuel consumption
(BSFC) decreases with 32 % when increase the hydrogen addition rate [7]. NOX
formation is higher with 25 % at 8 lpm compared to diesel fuelling [7]. Shekar [8]
shows the reduce of BSFC with 12.6 %, the decrease of smoke emission with 19.3 %
and an increase with 9.9 % in NO emission, for a diesel engine fuelled with bio diesel
and with hydrogen by induction [8]. Horng-Wen Wu [9] added hydrogen in the intake
manifold and designed an exhaust gas recirculation system for a direct injection diesel
engine. At different loads, for different EGR ratios (0 %…40 %) and hydrogen energy
share ratios of 0 %…20 %, the smoke emission decreases with 37.6 % and the NOX
with 59.5 % [9]. Paper target is the study of hydrogen use as fuel for a diesel engine,
study that shows new, important, considerable implication and effects of hydrogen use
at diesel engine, completing the general experimental investigation of engine running at
fuelling with alternative fuel for a clean environmental. The research work has been
carried out on truck diesel engine fuelled with diesel fuel and hydrogen by diesel gas
method.
2 Methodology of Experimental Investigation
The engine was mounted on a test bed equipped with adequate instrumentation, as
Fig. 1 shows. The investigated regimen was 55 % load, 1450 min−1 speed at normal
thermal regimen (80 °C cooling agent temperature) for different hydrogen flows. The
engine was fuelled only with fuel diesel to set up the reference and then with diesel fuel
and hydrogen in addition at different rate between 11.2 L/min and 40.4 L/min.
Hydrogen is injected into the engine inlet manifold at the pressure of 2.5 bar. For
measurement of the volumetric flow rate of hydrogen a special hydrogen flow meter
was used. The quantity of hydrogen in inlet air that substitutes the diesel fuel is
established by an energetic substitute ratio, xc [%], which takes into consideration the
energetically value of both fuels. At diesel fuelling the xc = 0. Authors developed a
thermodynamic model, zero-dimensional, two-zone for smoke quantity calculus, in
order to explain the tendency of variation for the smoke emission from the exhaust
gases at the dual fuelling.
To establish a correlation with the measured values of smoke emission, the ther-
modynamic model uses the difference between speeds of soot formation and oxidation,
the model being calibrated by in-cylinder pressure diagrams acquisitioned for 200
Aspects of Experimental Research 429
Fig. 1. Hydrogen diesel engine test bed schema. 1. hydrogen bottle 2.weight balance;
3. hydrogen valve; 4. diesel fuel valve; 5. diesel fuel tank; 6. pressure regulator; 7. diesel fuel
pump; 8. manometer high pressure; 9. diesel fuel filter; 10. low pressure manometer; 11. AVL
DiCom 4000 gas analyzer; 12. EGR valve; 13. air flowmeter; 14. supercharging pressure
manometer; 15-16. turbo-compressor; 17. dyno water cooling valve; 18. dyno water cooling
pump; 19. diesel fuel injection pump; 20. hydrogen pipe; 21. speed transducer; 22. fan 23.
radiator; 24. couple; 25. Hoffman eddy-current dyno; 26. D2156 MTN8 diesel engine; 27. water
pump for engine cooling; 28. Kistler piezoelectric pressure transducer; 29. injector; 30. charge
amplifier; 31. PC with AVL 617 Indimeter acquisition board; 32. dyno controller; 33. dyno
power cell; 34. diesel fuel pipe; Temperature indicators for: (a) inlet air; (b) exhaust gas; (c) oil;
(d) water cooling; (e) Oil pressure indicator.
consecutive cycles. At the same operating regime, the forming mass velocity of soot is
evaluated with the Kuo correlation [10–13]:

6290
mf ¼ Cf mc p 0:5
exp ð1Þ
T
where: coefficient Cf = (2.5…5.0)∙10−3, mc mass of fuel in gaseous state, p, T in-

cylinder pressure and temperature. The combustion of the soot particles is controlled by
chemical reactions that take into consideration the phenomena’s of adsorption of
oxygen at the particle surface and desorption of the reaction products from the particle
surface [10–13]. The global mass speed on surface unit for soot oxidation is established
by summarizing of desorption velocities of the considered reactive zones [10–13]:
v ¼ Mc
2 3
20 exp 15098 pO
T 7650
4 xA 3
þ 4:46 10 exp pO ð1 xAo Þ5
1 þ 21:3 exp 2063 p T
T O
ð2Þ
where coefficient Mc = 12 is the C molecular mass. The xA is a surface fraction related

with a very reactive zone and the 1 − xA fraction represent the less reactive zone. The
surface fraction filled by oxides xAo is defined by the equilibrium between oxygen
adsorption velocity from gaseous state for oxides forming and carbon monoxide des-
orption velocity from the droplet surface. With xAo is determinates the velocity of
carbon monoxide release after the oxidation into the very reactive zone takes place
[10–13]. The fraction xA decreases once with the temperature increasing and becomes
negligible comparative to coefficient of desorption velocity at high temperatures
[10–13]. With these relations the velocity of soot formation and oxidation are deter-
mined. By their integration the soot formation (SFL) and oxidation (SOL) laws are
determined. The difference between formation and oxidation of soot quantities estab-
lishes the smoke relative final quantity in the exhaust gases.
3 Results
Figure 2 shows the continuous decreasing in variation of brake specific energetic

consumption (BSEC) with substitute ratios. The BSEC decreases at the increase of
hydrogen addition in inlet air. The lowest BSEC of 9550 kJ/kWh is achieved for
xc = 3…3.9, corresponding to a hydrogen rate of 31.2,…, 40.2 L/min, versus to the
value registered only for diesel fuelling of 10550 kJ/kWh, due to better homogeneous
fuel-air mixture and the combustion improvement. The BSEC reduction is around
10 % versus diesel fuelling. Figure 3 shows the variation of NOX and smoke emissions
11000
BSEC [kJ/kWh]
10500
10000
9500
9000
8500
0 1 2 3 xc[%]4
Fig. 2. Variation of BSEC versus substitute ratio xc

500 0.4
450 0.3
NOx [ppm]
K [1/m]
400 0.2
350 0.1
300 0
0 1 2 3xc[%] 4 0 1 2 3 xc [%] 4
Fig. 3. NOX emission level and K Smoke Number versus substitute ratios xc
level with xc. Comparative to classic engine, in hydrogen-diesel dual fuelling operation
mode NOX emissions level decreases with 5.5 % at maximum xc. At the engine small
loads and hydrogen use the diesel-fuel cycle dose is reduced and the high effect of
combustion temperature on the NOX formation is reduced.
At the hydrogen adding, at the same engine load, the cycle dose of diesel fuel and
the combustion duration decreases, at same load regime, the heat release rate slightly
decreases lead to global temperature and NOX decreasing. In consecutive cycles, the
combustion duration with higher heat release rate comparative to diesel fuelling is
shorter. The maximum values of in-cylinder global temperature, calculated with the
thermodynamic model, decreases with *14 % comparative to reference engine, from
1600 K at xc = 0 down till 1375 K….1503 K at xc = 1.1,…, 3.9, below the values of
NOX thermal formation, fact that can explain the NOX emission reduction [3, 14, 15].
In case of fuelling with small hydrogen quantities in addition in inlet air, the explication
for NOX emissions level decrease is the fact that hydrogen burns fast, leading to the
avoiding of NOX formation due to a shorter duration of the combustion and high
temperatures (registered in some consecutive cycles) are registered only for short time
with values around 1.8,…,2 ms. Pechlivanoglou affirm that “the combustion is so rapid
that the high temperatures exist only for approx. 2 ms.” [16]. Lambe [1] and Talibi
et al. [3] explains NOX decrease at the hydrogen adding in air through the increase of
the mole fraction of water vapors in the combustion products produced at the hydrogen
combustion which absorb energy released from combustion and thus the peak com-
bustion temperatures decrease [3, 14, 15], which is in correlation with the anterior
mentioned reduction for the maximum values of gases temperature. Due to combustion
improvement and lower carbon content in the air-fuel mixture at hydrogen rate increase
the smoke emission level, evaluated by K smoke number, decreases comparative to
classic fuelling solution, Fig. 3. During combustion hydrogen forms a homogeneous
mixture versus heterogeneous mixture which is specific to diesel engine, fact that
contributes to the reduction of the smoke emission. Forward increase of hydrogen inlet
quantity, which affect the combustion temperature, leads to a slightly increase tendency
for smoke number with out exceeding the standard engine values. The smoke variation
tendency registered during experimental investigation, for a limited domain of
Fig. 4. Soot forming and oxidation laws for different substitute ratios xc
Fig. 5. Smoke relative emission for different substitute ratios xc
substitute ratio variation, can be explained thru the combined effect of in-cylinder
pressure and temperature. Soot forming (SFL) and oxidation (SOL) laws are presented
in the Fig. 4 for different xc.
For only diesel fuelling the soot formation is maximum (100 %) and the oxidation
also registered the maximum value, as Fig. 4 shows. At xc = 1.1 the soot formation
represents *60 %, initially being obtained with 34 % less soot inside the cylinder, but
the oxidation of formed soot being with 49 % less efficient comparative to diesel
fuelling. For xc = 2.2 soot formation represent 70 % and the soot quantity formed
inside the engine cylinder being with 30 % lower comparative to classic engine. The
soot oxidation is assured in the same terms of the previous lower value of xc. Actually,
for substitute ratios up till xc = 2.2, the soot oxidation is developed in a much efficient
way comparative to other substitute ratios. For xc = 3 the most lower value of soot
formation is obtained, comparative to diesel fuel, and the lowest value of soot oxidation
of 20 %, with 80 % less versus diesel fuel but in terms of a initial smoke emission with
53 % lower comparative to diesel fuel. At substitute ratio of xc = 3.9, the quantity of
soot is with 43 % lower comparative to diesel fuel, the oxidation of the soot being
reduced with 67 % versus diesel fuel. The increase of the substitute ratio from xc = 3 to
xc = 3.9 leads to the increase with 13 % of soot oxidation intensity, in terms of a 67 %

lower soot formation comparative to diesel fuel. The height oxidation rate is registered
for diesel fuelling, Fig. 4. The soot oxidation registered for dual fuelling is less
intensive comparative to diesel fuel, but the soot quantities formed inside the cylinder
are with 30 %…34 % lower for xc = 1,1…2,2 and with 43 %…53 % for xc = 3… 3,9.
Higher hydrogen rates influence more significant the oxidation process, also assuring
the formation of a lower quantities of soot, whereas the smaller hydrogen rate has a
higher influence on soot formation process. At xc = 3.9, the quantity of soot is with
43 % lower comparative to diesel fuel, the oxidation of the soot being reduced with
67 % versus diesel fuel. The increase of the substitute ratio from xc = 3 to xc = 3.9
leads to the increase with 13 % of soot oxidation intensity, in terms of a 67 % lower
soot formation comparative to diesel fuel. Figure 5 shows the variation of the smoke
relative emission (SFQ), theoretically determinate and compared with the experimental
values, presented in relative values of smoke number reported to the condition xc = 0.
Thru speeds and laws of formation and oxidation of the soot can be explained
according to the variation of the pressure and temperature regimes the reduction of the
smoke emission at dual fuelling, followed by a slightly increase tendency after a
specific xc value, without exceeding the values registered for xc = 2.2 or for diesel
fuelling.
4 Conclusions
The experimental study of dual fuelled diesel engine leads the following conclusions:
following the efficiency criteria, hydrogen use in dual fuel operation leads to a decrease
of almost 10 % in terms of brake specific energetic consumption (BSEC) comparative
to standard diesel engine. At the hydrogen-diesel oil dual fuelling operation mode NOX
emissions level decreases with 5.5 % comparative to diesel engine for a 3.9 % percent
of substitute ratio of diesel fuel by hydrogen. Comparative to classic engine, the smoke
emissions level decreases at the hydrogen addition increase. The tendency of variation
experimentally registered at the reduction of the smoke emission from the exhaust
gases is explained by the combined effect of pressure and temperature of in-cylinder
gases. Soot quantity formed inside the engine cylinder decreases at the increase of
hydrogen quantity, for all substitute ratios, the reduction being around 53 % at the
maximum hydrogen rate. At dual fuelling, higher hydrogen quantity accelerates the
soot oxidation process comparative to lower hydrogen quantities. Hydrogen in addition
at diesel fuel is a promising alternative fuel for diesel engines and the application of
diesel-gas method don’t require major modifications of the engine design.
Acknowledgements. The work has been funded by the Sectoral Operational Programme
Human Resources Development 2007–2013 of the Ministry of European Funds through the
Financial Agreement POSDRU/159/1.5/S/134398. The authors would like to address special
thanks to the AVL GMBH Graz Austria for providing the necessary equipments.
References
1. Lambe, S., Watson, H.: Optimizing the design of a hydrogen engine with pilot diesel fuel
ignition. Int. J. Veh. Des. 14(4), 370–389 (1993). Paper 14:370e89
2. Tomita, E., Kawahara, N., et al.: Hydrogen combustion and exhaust emissions ignited with
diesel oil in a dual fuel engine. SAE Pap 2001:2001e01e3503M (2001)
3. Talibi, M., Hellier, P., Balachandran, R., et al.: Effect of hydrogen-diesel fuel co-combustion
on exhaust emissions with verification using an in-cylinder gas sampling technique. Int.
J. Hydrogen Energy 39, 15088–15102 (2014)
4. Masood, M., et al.: Computational combustion and emission analysis of hydrogen diesel
blends with experimental verification. Int. J. Hydrogen Energy 32, 2539–2547 (2007)
5. Bose, P.K., Banerjee, R., et al.: Effect of hydrogen-diesel combustion on the performance
and combustion parameters of a dual fuelled diesel engine. Int. J. Energy Environ. 4(3),
497–510 (2013). ISSN 2076-2909
6. Ashok R.K.: A critical review on hydrogen-diesel fuel mode compression ignition engine.
Int. J. Appl. Innov. Eng. Manage. 5(3) (2016). ISSN 2319–4847
7. Senthil Kumar, R., et al.: Performance, emission and combustion characteristics of CI engine
fuelled with hydrogen. Int. J. Innov. Res. Sci. Eng. Technol. 4(2), 377–387 (2015)
8. Shekar, R., Purushothama, H.R.: Hydrogen induction to diesel engine working on bio diesel:
a review. Procedia Earth Planet. Sci. 11, 385–392 (2015)
9. Wu, H.-W., et al.: Investigation of the combustion characteristics and emissions of
diesel/hydrogen mixtures by using energy-share method in a diesel engine. Appl. Therm.
Eng. 42, 154–162 (2012)
10. Elkotb, M.M.: Fuel atomisation for spray modeling program. Energy Combust. Sci. 8, 61–91
(1982)
11. Kuo, T.W.: Evaluation of a phenomenological spray-combustion model for two open-
chamber diesel engines. SAE Paper no. 872057 (1978)
12. Appleton, J.P.: Soot oxidation kinetics at combustion temperatures. In: Propulsion and
Energetics Panel 41st Meeting Atmospheric Pollution by Aircraft Engines (1973)
13. Merker, G., Schwarz, C., Teichmann, R.: Internal Combustion Engines Fundamentals:
Running, Simulation, Measuring. Springer, Germany (2011)
14. Tomita, E., Kawahara, N., et al.: Hydrogen combustion and exhaust emissions ignited with
diesel oil in a dual fuel engine. SAE Paper: 2001-01-3503 (2001)
15. Christodoulou, F., Megaritis, A.: Experimental investigation of the effects of simultaneous
hydrogen and nitrogen addition on the emissions and combustion of the diesel engine. Int.
J. Hydrogen Energy 39, 2692–2702 (2014)
16. Pechlivanoglou, G.: Hydrogen Enhanced Combustion History, Applications and Hydrogen
Supply by Plasma Reforming. In: PPRE 2005–2007, University of Oldenburg (2007)
Researches on Combustion Quality for a Single
Cylinder Diesel Engine
Stelian Tarulescu(&) and Radu Tarulescu

{s.tarulescu,radu.tarulescu}@unitbv.ro
Abstract. The present paper presents researches on combustion quality for a

Diesel single cylinder engine. The tests ware made on an AVL single cylinder
test bed for gasoline and diesel engines. In this case, it was used an AVL 510 cc
Single Cylinder Engine Type 5402. The quality of the combustion for this
engine was analyzed and compared with the theoretical principles. During the
test was noticed that the main engines parameters are variable for 100 engine
cycles. The variation occurs in function of the air temperature, air/fuel mixture
formation, cooling system variation, combustion start moment and several
engine design parameters. In order to increase the combustion quality, some
measures can be applied: using different injection strategies; varying the
advance angle of the spark related to TDC; varying the quantity of fuel injected
per engine cycle; optimization of combustion chambers.
Keywords: Diesel Combustion Quality Engine Mixture
1 Introduction
In order to obtain data and optimize solutions for GDI engines, a research program was
developed at Transilvania University of Brasov, ICDT - Research & Development
Institute. The research stand is an AVL single cylinder test bed for gasoline and Diesel
engines. For the present paper, the tests was made on a AVL CR Diesel Single Cylinder
Research Engine 5402. The single cylinder can be setup in several configurations (with
multipoint injection, with direct injection, with turbocharger). The mixture formation
and combustion processes of the fuel can be monitored through the test bed component
software, AVL FIRE Commander 7.06c - IAV [2].
2 Combustion Quality Evaluation for Diesel Engines
In Fig. 1 is presented the typical data for cylinder pressure, fuel injector needle lift and
fuel pressure in the nozzle gallery through the compression and expansion strokes of a
direct injection Diesel engine. The rate of fuel injection can be obtained from the fuel
pressure, cylinder pressure, nozzle geometry, and needle lift profiles by considering the
injector as one or more flow restrictions. There is a delay between the start of injection and
start of combustion. The pressure rises rapidly for a few crank angle degrees, then more
slowly to a peak value. Injection continues after the start of combustion. The general

436 S. Tarulescu and R. Tarulescu
Fig. 1. Cylinder pressure p, injector needle lift I, and injection system fuel line pressure pi, as
functions of crank angle for small Dl diesel engines (Source: B. Heywood, Internal Combustion
Engine Fundamentals, p. 503)
Fig. 2. Cylinder pressure variation in function of combustion quality (Source: W. Pulkrabek)
shape of the rate of heat release curve is typical of this type of DI engine over its load and
speed range [5].
Estimates of the rate at which fuel and air mixture with composition within the
combustible limits is produced in Diesel sprays under typical engine conditions, based
on a variety of turbulent jet models of the spray show that mixing rates and burning
rates are comparable in magnitude. Estimates of characteristic times for the turbulent jet
Researches on Combustion Quality 437
mixing processes in diesel combustion chambers show these to be comparable to the

duration of the heat release process, and much longer than characteristic times for
evaporation and the combustion chemical kinetics (Fig. 2) [3, 4].
3 Used Equipments
The researches presented in the present paper ware made on a AVL CR Diesel Single
Cylinder Research Engine 5402 with following specifications:
– Bore: 85 mm; Stroke: 90 mm.
– Displacement: 510 cm3.
– Max. speed: 4200 rpm.
– Max. firing pressure: 180 bar.
– Max. BMEP: 14 bar at 2300 rpm supercharged.
– Max. output: 19 kW at 4200 rpm supercharged.
– Compression ratio: 17:1.
The test bed have some other components and systems: AVL Engine Control Unit
(AVL ETU 427); Coolant and conditioning Unit 577; AVL Fuel mass flow meter -
Type Flex Fuel; AVL Fuel temperature control; Intake Air Consumption Measurement
Device; Particle Evaluation - Micro Soot Sensor Continuous Measurement of Soot
Concentration; AVL PUMA Open Test bed Automation (Fig. 3) [1].
Fig. 3. Engine parameters read from the AVL PUMA Open Test bed Automation
The used software for intake and combustion process optimization is AVL FI2RE
Commander 7.06c – IAV. AVL FIRE was developed to solve the most demanding flow
problems in respect to geometric complexity and chemical and physical modelling.
FIRE offers a comprehensive computational fluid dynamics solution: a powerful set of
modules, features and capabilities, pre-and post-processing integrated in a common
environment and workflows and methods effectively supporting the use of the software
to solve any problem accurately [2].
The software used for engine parameters monitoring is AVL Indicom software.
During cycle based data acquisition the value of cycle time must vary all the time.
4 Research Methodology
The engine was tested in controlled laboratory conditions. The used fuel was Diesel
Petrom Standard. The atmospheric temperature was constant maintained at 18°C.
The test was made for 50 % load and for 1500, 2000, 2500, 3000 and 3500 rpm
engine speeds. The intake parameters are controlled by set the number of injections. In
this case it was used only two injection phases (first pilot injection and main injection)
from a maximum of for possible phases. The fuel mixture was adjusted by varying the
amount of fuel injected per cycle (injection period - μs). The ignition time was set in
crank angle degrees before top dead centre. The engine combustion quality was tested
for 100 engine cycles.
Fig. 4. The intake and combustion features for AVL CR Diesel Single Cylinder Research
Engine 5402, for 50 % Load and 1500 rpm
The parameters changes were made to obtain an optimal single cylinder pressure
curve and more optimal combustion (no detonations). In the Fig. 4 are presented the
intake and combustion features for AVL CR Diesel Single Cylinder Research Engine
5402. The upper window presents the parameters for intake and combustion quality:
combustion quality, combustions with more than 50 % efficiency and excess air
coefficient (mixture quality).
The main window presents the injection time (in function of engine crank angle and
TDC) and engine cylinder pressure.
The paper presents researches on combustion quality for an experimental single
cylinder Diesel engine. The tests ware made on a AVL single cylinder test bed for
single cylinder engines. In this case, it was used a AVL 510 cc Single Cylinder Engine
Type 5402. The quality of the combustion for this engine was analyzed and compared
with the theoretical principles. During the test was noticed that the main engines
parameters are variable for 100 engine cycles.
In the following diagrams are presented the variation of engine cylinder pressure
for 100 engine cycles.
Fig. 5. Variation of engine cylinder pressure for 100 cycles, for 1500, 2000, 2500, 3000 and
3500 rpm engine speeds
From Fig. 5 results that the cylinder pressure varies for all tested regimes for all 100
cycles. The biggest variations are registered when engine runs at 2000 and 2500 rpm.
We can connect this variations with air/fuel mixture inside of combustion chamber.
The mixture (lambda) is expressed as an electric parameter (an electric impulse value -
expressed in Volts).
Fig. 6. Variation of carburant mixture for 100 cycles, for 1500, 2000, 2500, 3000 and 3500 rpm
engine speeds
Figure 6 presents the variation of fuel/air mixture for all 100 engine cycles and for
the same engine regimes (50 % load and 1500, 2000, 2500, 3000 and 3500 rpm engine
speeds).
It results that the cylinder pressure varies for all tested regimes for all 100 cycles.
The biggest variations are registered when engine runs at 2500, 3000 and 3500 rpm.
In this condition we can conclude that the fuel/air mixture can influence the quality
of combustion for the analyzed engine.
5 Conclusions
In the research tests made on the experimental engine, a large set of data were collected
for the analysis of combustion quality and efficiency. The combustion quality can be
expressed by a set of graphical parameters, among which the most important is the
engine cylinder pressure. For functional conditions there were recorded 100 engine
cycles, at different engine regimes. It was observed a variation of cylinder pressure
curve in function of the excess air coefficient and indicated mean effective pressure.
In order to optimize engine operation and for a high quality of combustion we can
take the following measures:
– using an injection strategy with three or four phases;
– varying the advance angle of the spark related to TDC;
– varying the quantity of fuel injected per engine cycle;
– optimization of combustion chambers.
All these strategies can be tested using the software component AVL FI2RE
Commander 7.06c – IAV.
Another set of measures can be related to modelling the combustion chambers.
There has always been extensive debate over the optimum Diesel engine combustion
chamber design. There are a large number of options for cylinder head and piston
crown shape, spark plug location, size and number of valves, and intake port design.
The major combustion chamber design objectives which relate to engine perfor-
mance and emissions are:
– a fast combustion process, with low cycle-by-cycle variability, over the full engine
operating range;
– a high volumetric efficiency at wide-open throttle;
– minimum heat loss to the combustion chamber walls;
– a high fuel cetane requirement [6].
Acknowledgements. We hereby acknowledge the structural founds project PRO-DD (POS-CCE,

O.2.2.1., ID 123, SMIS 2637, ctr. No 11/2009) and Transilvania University of Brasov for providing
the infrastructure used in this work.
References
1. AVL - Tutorials and books for Engines Test Cell equipments usage (2013)
2. Fritz, O.: GDI engine development according EU 6, AVL Tutorials and books for Engines
Test Cell (2013)
3. Pulkrabek, W.: Engineering Fundamentals of the Internal Combustion Engine, University of
Wisconsin, Platteville (2004)
4. Gatowski, J.A., Balles, E.N., Chun, K.M., Nelson, F.E., Ehchian, J.A., Heywood, J.B.: Heat
release analysis of engine pressure data, SAE paper 841359, SAE Trans., 93 (1984)
5. Heywood, J.B.: Internal Combustion Engine Fundamentals, edn. McGraw-Hill, Inc.,
New York (2000). ISBN:0-07-028637-X
6. Țârulescu, S., Țârulescu, R., Șoica, A.: Researches on Combustion Quality for a GDI
Experimental Engine. In: AFASES Conference, Brasov (2016)
Urban Transportation Solutions for the CO2
Emissions Reduction Contributions
Stelian Tarulescu(&) and Radu Tarulescu

{s.tarulescu,radu.tarulescu}@unitbv.ro
Abstract. Consumption of electricity and greenhouse gas production repre-

sents the actual global issues. To reduce energy consumption and the amount of
CO2 specific local measures need to be adopted. For this study it was analyzed
the urban transportation system from Brasov Metropolitan Area. The study was
made for the transportation route that connects Brasov city and Săcele city. For
Săcele city it was made a prediction model in order to the decrease the energy
consumption and the CO2 amount resulted from road transportation. For this
study ware considered four optimization measures: vehicle fleet renewal;
increasing the average travel speed; making bicycle lanes; and students mobility
management. There were calculated the proportions of decrease for the fol-
lowing: final energy consumption [MWh] and total CO2 Emissions / CO2
equivalent for each implemented measure.
Keywords: Transportation CO2 Measures Mitigation
1 Introduction
Many countries around the world have adopted very stringent regulations to monitor
and control urban traffic emissions. Energy use and carbon emissions are influenced by
how electricity is produced, use of energy in buildings and to move around the city.
Urban density and spatial organization are the key factors that influence energy con-
sumption, especially in the transportation and building sectors. The expansion of
built-up areas through suburbanization has been particularly prominent among the
metropolitan areas. Climate action over the urban level happens through local regu-
lations, urban services, programme administration, city purchasing and property
management, and convening of local stakeholders [1].
Dealing with climate changes will require many instruments and strategies
implemented for local conditions. If we want to meet the objective of cutting emission
by at least 30 % by 2050, the level generally agreed on to prevent dangerous
anthropogenic interference, and minimize the costs involved, a number of strategies
must be implemented or expanded in order to diminish the overall pollution level [1].
In the last three decades, emissions of greenhouse gases have increased by an
average of 1.6 % CO2 in fossil fuel use, reaching 1.9 %/year. In the absence of policy
action, it is expected that these patterns of emissions growth to continue. For Romania,
green house gases (GHG) resulted from different sectors also highlights the major

Urban Transportation Solutions for the CO2 Emissions Reduction Contributions 443
contribution to energy and transport sector, which means that they are areas on which
are necessary to implement measures and actions to reduce CO2 emissions [2].
In terms of environmental impact, there is a wide range of factors influencing high
level of CO2 emissions from road transport, such as supply and demand for road
vehicles, individual mobility needs, availability/lack of availability of alternative public
transport services common and associated costs of having a private vehicle. Although
the energy efficiency of vehicles was and is still growing, this is offset by increased
average length of travel, the increase of the fleet, and other variables, such as driving
style, traffic jams etc., which translates into an increase in the intensity of emissions of
GHG. In Romania transport policy aims to align the national transport system con-
tinues to Community Transport Policy principles defined in the Transport White Paper
(with corresponding updates) and the requirements of sustainable development of
Romania [2].
2 The Analyzed Area
For this study it was analyzed the urban transportation system from Brasov
Metropolitan Area. The study was made for the transportation route that connects
Brasov city and Săcele city.
The city of Săcele is located in the south-eastern county of Brasov, the historic
region of Bârsa, on the river Tărlung. Being located to the southeast of the Brasov city,
Săcele is one of 18 municipalities that are part of Brasov Metropolitan Area (as we can
see in Fig. 1). The road connections with these areas are insured by the national road
network (DN1 and DN1A) and county roads (DJ 103A, 103B, 102I).
Fig. 1. The analyzed area - Brasov city and Săcele city (Source: Google Maps)
Statistical data for Săcele city:

- The city area is 32,000 ha.
- Population: 32 185 inhabitants.
- Population density: 122 inhabitants/square km.
Săcele city, the second highest city of Brasov Metropolitan Area in terms of
population, is also the only city after Brasov that have his own public passenger road
transportation system [5].
3 Energy Consumption and CO2 Mitigation Model for Săcele
3.1 Mathematical Model for Săcele Transportation System

For Săcele city it was made a prediction model in order to the decrease the energy
consumption and the CO2 amount resulted from road transportation. For this study
ware considered four optimization measures: vehicle fleet renewal; increasing the
average travel speed; making bicycle lanes; and students mobility management.
There were calculated the proportions of decrease for the following:
– Final energy consumption [MWh].
– Total CO2 Emissions / CO2 equivalent [6].
For this study we start from initial data: Săcele vehicle fleet composition during
2008–2013; values expressed in tones of CO2 produced in the 2008–2012 period.
There were grouped the Săcele traffic volumes data for period 2009–2013 in order
to predict the traffic values for the 2014–2020 period. There results values for Săcele
fleet evolution for all vehicle categories for 2014–2020 period.
Using existing data on vehicle fleet, the total energy consumption and CO2 emis-
sions for Sacele in the 2008–2013 period and the resulting approximating fleet evo-
lution data during 2014–2020 we can calculate the total energy and CO2 emissions for
2014–2020 period.
Following the mathematical modeling results the dependence of total energy
consumption and amount of emitted CO2 over the vehicle fleet evolution. Two equa-
tions are described by polynomial second degree regression curves (also in Fig. 2):
XEnergy ¼ 52; 822 VE2 145; 17 VE þ 22493 ð1Þ
Where: XEnergy is the approximate amount of the total energy consumed, and VE is
the number of etalon vehicles. For this function, the standard deviation (R2) is 0.9743.
XCO2 ¼ 14; 897 VE2 61; 213 VE þ 5725; 3 ð2Þ
Where: XCO2 is the approximate amount of CO2 emitted, and VE is the number of
etalon vehicles. For this function, the standard deviation (R2) is 0.9601 (also in Fig. 3).
In order to decrease energy consumption and CO2 emission values it is necessary to
implement the following measures: vehicle fleet renewal; increasing the average travel
speed; making bicycle lanes; students mobility management.
Energy consumption without any applied measures
31000
29000
27000
Total energie [MWh]
25000
23000
21000 2
y = 52,822x - 145,17x + 22493
2
R = 0,9743
19000
17000
15000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Year 1- 13 from 2008 in 2020
Fig. 2. Energy consumption prevision for 2020 without any transportation applied measures
CO2 production without any applied measures
8000
7500 y = 14,897x2 - 61,213x + 5725,3

7000 R2 = 0,9601
6500
CO2 [tones]
6000
5500
5000
4500
4000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Year 1-13 from 2008 in 2020
Fig. 3. CO2 resulted from transportation activities prevision for 2020 without any applied
measures
These measures can be grouped into measures that will have direct results in the
reduction of fuel consumption and measures that will have a direct impact over the
traffic flows in the urban area of Săcele.
The standard emissions factors applied in this study are:
– 0.249 [tones CO2 / MWh] - for spark ignition engines;
– 0.267 [tones CO2 / MWh] - for compression ignition engines;
– 0.000 [tones CO2 / MWh] - for biofuels.
For the transportation sector, data for each fuel and vehicle were calculated using
the formula:
XUsedFuel ¼ D Caverage Y ð3Þ
Where: XUsedFuel - Energy consumption for road transportation [kWh]; D - route

distance [km]; Caverage - average fuel consumption [liters/km]; Y - conversion factor for
each fuel [kWh/liters]. Y = 9.2 for gasoline and Y = 10 for Diesel.
3.2 Vehicle Fleet Renewal

This measure may be applied to the municipal and local public transport fleet. The
proposed solutions are based on actions for greening public transportation system
between Săcele and Brasov.
Using the data on the number of municipal vehicles fleet, public transportation
Săcele fleet and from related fuel consumptions, the following measures are proposed:
– Fleet renewal for municipal fleet and public transportation fleet will lead to a
consumption decrease from 11.25 L/100 km to 7 L/100 km for light vehicles and
from 27 L/100 km to 24 L/100 km for buses.
– By reevaluating and improving the transportation system management and fuel
consumption monitoring the fuel consumption can be redused by 5 %. If this
scenario is adopted, will result a reduction in fuel consumption of municipal fleet
from 11.25 L/100 km to 10.69 L/100 km, and for local public transportation fleet
from 27 L/100 km to 25.65 L/100 km.
– Replacing 50 % of public transportation fleet with green vehicles (using 100 %
biofuel or electric buses).
After applying these measures will have a reduction: 627.95 [MWh] and 540.71
[tones CO2].
3.3 Increasing the Average Travel Speed

The level of service of the road network (expressed as average travel speeds) is gen-
erally low. Average travel speeds are 66 km/h for the national road network, 34 % less
than the target of 100 km/h. Similarly, the average speed of traffic in urban areas is
between 15 km/h and 40 km/h.
Measurements on test cycles that are close to actual traffic conditions showed a
slight influence of travel speed limits over the pollutant emissions. In general, the
lowers speeds increase pollutant emissions, but due to turbulence generated behind the
vehicle quickly dispersed into the atmosphere pollutants are avoiding their local
concentration.
In the case of the Municipality of Săcele was made a dependency calculating
average fuel consumption in function of average speed of the vehicles. It has been
determined fuel consumption equation depending on the vehicle speed. The graphical
representation is given by Fig. 4 [3].
It has been determined that the rate of decrease of the fuel consumption with the
increase of average speed of vehicle travel is 7.35 % for a speed increase from 35 to
40 km/h and 13.45 % for a speed increase from 35 to 45 km/h.
It was established an average speed for Săcele by making measurements at peak
morning and evening hours. The analyzed route was from Brasov Boulevard to George
Moroianu Avenue, in both directions, from 8.00–9.00 and 17.00–18.00.
It results that an increasing of average travel speed (on 7.8 km route) from 35 km/h
to 45 km/h will result a decrease in fuel consumption of 13.45 %. Thus we have a
reduction: 3106.33 [MWh] and 785.59 [tones CO2].
Fig. 4. Fuel consumption equation depending on the vehicle speed (Source: [3] Mitran Gabriela,
a.a, A model for determining the average speed of light duty vehicles in urban traffic in terms of
minimum pollution and fuel consumption)
3.4 Bicycle Lanes for Săcele

Cycling has a major role in the metropolitan urban transport. It helps reduce conges-
tions, local air pollution and emissions of greenhouse gases. About 20 % of car
journeys are less than 3 km, a distance that can be traveled easily by bike in less than
15 min. If people choose to make some of those trips by bike, we could have a
considerable impact on local congestion and road pollution. The study is based on
approximate traffic volumes decrease by creating bicycle tracks in Săcele. According to
the analyzed data in this study for cyclists routes, resulted a decrease of 7.8 % road
traffic volumes after the construction of a bicycle lane in Săcele. After applying this
percentage reduction for total vehicle we have a reduction of: 2455.28 [MWh] and
651.86 [tones CO2] [4].
3.5 Students Mobility Management

The target group of this study are students in primary and secondary schools. If they
would use much often the public transportation (existing system or special routes),
from the approximately 8,000 etalon vehicles of Săcele fleet, we will have about 1000
etalon vehicles less in the first year of implementation at peak morning and evening
rush hours. It is anticipated that for each year the number of vehicles will be reduced by
500 cars. After applying this reduction we will have a reduction in energy consumption
and CO2 emissions: 570.63 [MWh] and 178.07 [tones CO2].
4 Conclusions
For this study were analyzed four measures that can be adopted in order reduce energy
consumption and CO2 emissions: vehicle fleet renewal; increasing the average travel
speed; making bicycle lanes; and students mobility management. The most effective
measure, increasing average travel speed of 7.8 km route from about 35 km/h to
45 km/h will result in a decrease in fuel consumption of 13.45 % and a decrease in
Fig. 5. The evolution of energy consumption and CO2 production by building a bicycle track
Fig. 6. The evolution of energy consumption and CO2 production if is applied students mobility
management program
energy consumption and CO2 emissions by 13.50 %. Making bicycle lanes will lead to
lower energy consumption and CO2 emissions by 11 %.
In Fig. 5 is presented the evolution of energy consumption and CO2 production by
building a bicycle track. Municipal and public transportation fleet renewal, fuel con-
sumption monitoring program and replacement of 50 % of a fleet of public transport
vehicles with green vehicles will lead to a 3 % lower energy consumption and 9.3 %
less CO2 emissions. Mobility management program for students in Săcele will lead to a
decrease in energy consumption and CO2 emissions by 3 %. Figure 6 shows the
evolution of energy consumption and CO2 production by students mobility manage-
ment program.
Acknowledgements. We hereby acknowledge the structural founds project “Plan de acțiune

privind energia durabilă pentru Municipiul Săcele”, 2964/13.03.2015 and Transilvania Univer-
sity of Brasov for providing the infrastructure used in this work.
References
1. Kamal-Chaoui, L., Robert, A.: Competitive Cities and Climate Change, OECD Regional
Development Working Papers No. 2, OECD Publishing, © OECD (2009)
2. Vasile, E., Balan, M., Grabara, I., Balan, S.: Measures to reduce transportation greenhouse gas
emissions in Romania. Pol. J. Manage. Stud. (2012)
3. Gabriela, M., Ion, T., Florian, I., Sorin, I., Viorel, N., Alexandru, B.: A model for determining
the average speed of light duty vehicles in urban traffic in terms of minimum pollution and
fuel consumption, CAR Pitesti (2011)
4. Protected Bicycle Lanes, New York City Department of Transportation – Polly Trottenberg,
Commissioner, September 2014
5. Stadiul implementării măsurilor din responsabilitatea R.A.T. Braşov din PLANUL DE
ACŢIUNE PENTRU ENERGIE DURABILĂ 2010–2020
6. Strategia de dezvoltare a Municipiului Săcele 2014–2020
Mission CO2 Reduction
Roland Seebacher(&)
LuK GmbH & Co. KG, Buehl, Germany

roland.seebacher@schaeffler.com
Abstract. In the race to achieve global CO2 targets, longer gear ratios helps to
reduce fuel consumption in significant extent. The operating point of the internal
combustion engine then shifts to lower speeds and specific torque is increased.
Both measures lead to more pronounced rotational irregularity, however. The
resulting higher design requirements for vibration isolation will nevertheless be
reliably met by current technology as it is incorporated into today’s engines and
those targeted for the next evolution stage. The problem revolves around the
next generation of engines, which will require even more capable systems.
Although the technology offered by the dual-mass flywheel in conjunction with
a centrifugal pendulum absorber is a prime candidate, the summation damper is
also worth considering if a way can be found to extend its high potential at low
operating speeds to mid-range and higher speeds. Schaeffler continues to
investigate both concepts with a great deal of interest. The key to developing a
more responsive summation damper lies in the ability to vary one relevant
parameter with respect to engine speed. A solution that is robust, affordable, and
can be deployed on a large scale has not yet crystallized, however.
Keywords: Fuel consumption Downspeeding Vibration Damper systems

Pendulum absorber Summation damper
1 Introduction
The internal combustion engine will continue to be the dominating force behind
individual mobility for some time to come. The biggest challenge in this context,
however, revolves around lowering fuel consumption in line with ever more stringent
legal requirements while at the same time maintaining driving comfort and pleasure.
Fuel consumption could then be further reduced by opting for longer gear ratios, for
example. At the same time maintaining the driving pleasure, combustion engines with
higher torque provided at lower speeds, are necessary.
The end result – “extreme” down speeding – has disadvantages, however, espe-
cially when it comes to future engines, where few cylinders and/or feature cylinder
deactivation will be widely used. In order to realize the comfort and convenience
expected by end customers, ever better systems for isolating, or dampening, vibrations,
must be developed [1]. Although the centrifugal pendulum-type absorber (CPA) de-
veloped by LuK offers good potential for the coming years, in the long term, even more
capable systems will need to be integrated.

452 R. Seebacher
2 Improving the Efficiency of the Powertrain

and the Challenges to Be Overcome
As mentioned above, shifting the operating point of an engine to lower operating

speeds (downspeeding), fuel consumption can be significantly reduced. For example,
when the mean operating speeds of a current 2.0-L diesel engine are reduced by 10 %,
it is possible to consume 5.6 % less fuel under NEDC testing conditions. This potential
can only be tapped, however, if doing so does not lead to any drawbacks in driving
dynamics or comfort. Thus, to ensure that these driving dynamics remain fairly con-
sistent and comparable maximum torque must also be increased by approximately
10 % (Fig. 1).
Fig. 1. Operating point shifting and potential reduction in consumption with downspeeding
In addition, it is foreseeable that usable speeds will be expanded much further down
in the rev range. Some engines in the future will even reach their peak torque at below
1,000 rpm! Compared to today’s engines, this will allow these power units to theo-
retically reduce their consumption by 11 % under NEDC testing conditions. Such
engine developments ultimately lead to considerably higher vibrations from the pow-
ertrain. This initially becomes evident in the rotational irregularity that increases pro-
portionately to an increase in torque or a drop in engine speed. Adding to this is the fact
that as engine speed goes down, the excitation frequency becomes more closely aligned
with the natural frequency of the rest of the powertrain. Figure 2 summarizes the effects
on the rotational irregularity in the powertrain. Relative to a current engine, the
oscillation range at the transmission input doubles for the same damper technology
when engine speed is reduced by 10 %. This marks the starting point at which target
comfort levels can no longer be attained. Some drivers would even intentionally avoid
Mission CO2 Reduction 453
low engine speeds for this reason and thereby not profit from the lower fuel con-
sumption otherwise possible [5, 6]. Further downspeeding amplifies the situation dis-
proportionally. When maximum torque is available below 1,000 rpm, the comfort
target at this speed is undershot by more than 600 %. In order to achieve an acceptable
comfort level with these engines, performance-oriented damper systems must be fitted
and are critical to ensuring that the consumption benefits afforded by downspeeding
can, in fact, be realized.
Fig. 2. Rotational irregularity at the engine and transmission input for current and future engines
3 Vibration Isolation – State of the Art
Some 20 years ago, the requirements placed on damper technology dramatically rose as
a result of the direct-injected diesel engines then offered for passenger cars (Fig. 3).
This shift in engine technology presented the developers of these systems with
entirely new challenges [2–4]. The resulting rotational irregularity could not be suffi-
ciently counteracted using the available torsion-damped clutch disks. Although the
principle of the low-pass filter was known, it was not regarded as being technically
feasible until the dual-mass flywheel (DMF) was introduced in passenger-car appli-
cations. By leveraging its comprehensive knowledge of the operating principles of
passive damping systems, LuK systematically started investigating the underlying
correlations early on and was consequently able to offer a compatible solution that met
the emerging challenges in good time. Many years of know-how in metalworking then
finally led to a robust product.
In the years that have passed, specific torque outputs have more than doubled in
comparison to the first turbocharged, direct-injection diesel engines. The resulting
effect is that even today, some engines experience torsional vibrations that cannot be
454 R. Seebacher
Fig. 3. Dramatic increase in performance requirements for vibration-dampening systems
counteracted with a DMF alone [8]. The answer to these increased requirements is the
centrifugal pendulum absorber (CPA), which is a damper assembly that introduces
additional mass external to the power flow. The dual-mass flywheel and centrifugal
pendulum absorber have been continually refined and advanced and will meet the
requirements associated with the upcoming evolutionary stages set for the current
generation of engines [7]. The next engine generation, however, which is currently
under development, will call for vibration isolation measures that are even more
capable, which is why Schaeffler is not only investigating the possibilities and con-
straints of today’s technology, but is also looking at alternative solutions [9].
4 Alternative Solutions – Options and the Operating

Principles That Define Them
Before implementation concepts are considered at product level, the operating prin-
ciples that govern them must be thoroughly evaluated with respect to future require-
ments. It is in this context that the method that uses simple, linearized models to
investigate the operating principles has proven successful. Not only the technical
potential of the different approaches must be factored into the overall assessment, but
also their cost-benefit ratio, whereby the objective must always be to find approaches
that offer equal, uniform performance across an engine’s entire operating speed range.
Improvements made at very low engine speeds are not optimal if they compromise the
progress already achieved in the mid and high-speed ranges. In addition, only those
solutions that comply with the restrictions for installation space and weight and are just
as robust as current systems when it comes to friction, wear, and manufacturing tol-
erances are promising candidates.
The following systems will be investigated to determine whether (and under which
conditions) their physical potential is capable of isolating the torsional vibrations of a
motor that utilizes an extreme downspeeding concept so that comfortable driving is
possible from 800 rpm.
5 Spring-mass System – Principle of the Dual-mass Flywheel
The basic operating principle of this arrangement is that two masses connected to each
other by a spring-damping system oscillate against one another. In terms of the
operating range used today and the excitations that are encountered, dampers
demonstrate overcritical performance, and provide better isolation as frequency
increases. When frequency drops, the resonance frequency is more closely aligned with
these excitations and torsional vibrations become more prevalent. Theoretically, it is
also possible to use a spring-mass system to reach the required target even in extreme
downspeeding scenarios. This, however, would require the mass to be increased by a
factor of 3.5 or the spring rate to be reduced by a factor of 17 compared to the base
construction (Fig. 4). Neither is realistic. Arguments not in favor of increasing mass are
the increased installation space required, added weight, and worse driving dynamics.
Reducing the spring rate by an extreme amount is also not possible as a result of the
installation space problem and the compromised driving experience that would result.
Fig. 4. Isolation capacity and limitations of the spring-mass system

456 R. Seebacher
6 Anti-resonance – Principle of Interference
The following describes two concepts for generating anti-resonance: The spring-mass
absorber and the summation damper. Although both concepts use a different operating
principle, they produce similar results under the same conditions.
7 The Spring-mass Absorber
The spring-mass absorber is based on a second spring-mass system. When this system
is excited at its resonance frequency, an opposing oscillation is generated that ideally
annihilates the original excitation. With a conventional absorber connected via a spring,
this effect occurs at exactly one frequency – the resonance frequency of the absorber.
The drawback is an additional resonance point above the absorber resonance
frequency.
A conventional absorber (Fig. 5) is therefore not a suitable means of reducing
torsional vibrations in the powertrain. What is required is an absorber whose damp-
ening frequency corresponds to the ignition frequency of the engine at all times. This
property is fulfilled by the centrifugal pendulum absorber (Fig. 6), which restoring
force is dominated by the centrifugal force of the absorber mass. Since the centrifugal
force changes quadrically in relation to the engine speed, the centrifugal pendulum
absorber has a absorber frequency that is proportional to this speed. This is the ideal
property or attribute for reducing torsional vibrations in the powertrain, since a fixed
excitation order can be dampened.
Fig. 5. Principle and isolation effect of a conventional absorber

Fig. 6. Centrifugal pendulum absorber (CPA) as a speed-dependent absorber
Figure 7 shows how effective the mass of a centrifugal pendulum absorber is. The
graph depicts, in relation to the engine speed, by what factor the secondary mass would
have to be increased for similar performance – e.g. by a factor of 3 at a speed of
1,000 rpm or a factor of 9 at 1,500 rpm. With a CPA, a vibration isolation figure of
100 % could theoretically be achieved up on a defined frequency. In demonstrator
vehicles, a decoupling performance rating of up to 99 % was already demonstrated in
conjunction with a DMF. This, in turn, makes it easy to meet the requirements of
today’s engines and their upcoming evolution stages. Current systems are even capable
of fulfilling the requirements of two-cylinder engines. The potential offered by the CPA
is described in an additional article in this book [7].
When engine speed drops, the centrifugal pendulum absorber must absorb more
energy. The ability of this pendulum to respond depends on the mass involved and the
vibration angle, whereby the latter is intrinsic. The mass of the pendulum can also only
be increased to a certain extent due to the installation space available. Whether the CPA
can produce a vibration isolation that is also compatible with the excitations of the next
generation of engines is not entirely clear at present. Recent improvements made to the
system support this working hypothesis, however. Nevertheless, Schaeffler also con-
tinues to search for alternative approaches. Using mass intelligently is the key to
implementing future solutions.
458 R. Seebacher
Fig. 7. Equivalent effective mass inertia of a centrifugal pendulum absorber
8 The Summation Damper
Another way of dampening vibrations with anti-resonance is to add two vibration paths
together. Figure 8 charts this principle. Vibrations are transferred via a spring-mass
system along the one path and directly to a lever on the other. The pivot point of the
lever (summation unit) is void of force and motion from a dynamic vibration
perspective.
As in the case with a conventional absorber, a summation damper can also
decouple 100 % of vibrations but only for a single frequency. The summation damper
therefore has an advantage over the absorber in that no additional natural frequency is
generated. Annoying vibrations above and below the anti-resonance frequency remain
present, however.
The frequency to be isolated, or targeted, can theoretically be selected as required.
When coordinating the system, the summation damper provides one additional
parameter not available with the conventional absorber – the lever ratio in addition to
the spring rate and the rotary mass (J). Another benefit is that the system can also be
configured so that a dampening effect is achieved on the primary side (engine side).
Further arrangements are possible in addition to the summation damper characterized in
Fig. 8. For example, the spring can be positioned at any point required (Fig. 9).
Comprehensive testing has revealed that the same basic laws and principles apply
irrespective of the positional arrangement of the spring. The anti-resonance frequency
can even be calculated for all concepts using a single formula. Assuming that the lever
ratio, spring capacity, and mass J do not change, not only is the same anti-resonance
frequency yielded for all of the concepts, but also an identical transfer response. When
the transfer response for design concepts with different anti-resonance points is
Fig. 8. Principle and isolation effect of a summation damper
Fig. 9. Variations in spring arrangement for the summation damper
considered, the typical properties of a summation damper become apparent.

Anti-resonance frequencies can theoretically be shifted to any low engine speed. Doing
this, however, not only reduces the absorption width, but also the isolating properties
above the anti-resonance frequency (Fig. 10). This, in turn, means that a summation
damper configured for very low anti-resonance responds sensitively to fluctuating
parameters. A satisfactory solution can only be achieved if at least one of the three
relevant parameters is variable with respect to engine speed. In a direct comparison, the
460 R. Seebacher
Fig. 10. Influence of the anti-resonance frequency on the absorption width
Fig. 11. Evolution of the absorber and summation damper
summation damper has a slightly higher theoretical potential for dampening vibrations
than the conventional damper (Fig. 11). Having said this, the advent of the centrifugal
pendulum absorber has already provided a solution for realizing a variable-speed
damper and is currently being used in volume production applications. Variable-speed
summation dampers, on the other hand, have yet to be integrated.
References
1. Gutzmer, P.: Individuality and variety. In: 10th Schaeffler Symposium (2014)
2. Kremmling, B., Fischer, R.: The automated clutch. In: 5th LuK Symposium (1994)
3. Fischer, R., Berger, R.: Automation of manual transmissions. In: 6th LuK Symposium (1998)
4. Welter, R., Herrmann, T., Honselmann, S., Keller, J.: Clutch release systems for the future. In:
10th Schaeffler Symposium (2014)
5. Müller, M., Kneissler, M., Gramann, M., Esly, N., Daikeler, R., Agner, I.: Components for
double clutch transmissions. In: 9th Schaeffler Symposium (2010)
6. Müller, B., Rathke, G., Grethel, M., Man, L.: Transmission actuators. In: 10th Schaeffler
Symposium (2014)
7. Kooy, A.: The evolution of the centrifugal pendulum absorber. In: 10th Schaeffler
Symposium (2014)
8. Fidlin, A., Seebacher, R.: DMF simulation techniques. In: 8th LuK Symposium (2006)
9. Kroll, J., Kooy, A., Seebacher, R.: Torsional vibration damping for future engines. In: 9th
Schaeffler Symposium (2010)
Experimental Research Concerning
the Influence of the Electrochemical Reactions
Induced by Products of an Electrolytic Cell
Over the Emission Pollutants in ICE
Marius Nicolae Moldoveanu1, Cheng Lin Tang1,

and Cornel Aramă2(&)
1
Fujian Worldwide Automotive Co. Ltd., Quanzhou, Fujian, China
moldoveanu.marius@qq.com
2
“Henri Coandă” Air Force Academy, Brașov, Romania
aramis5791@gmail.com
Abstract. Any lean mixture – favorable to fuel economy and low CO2 emis-
sion – is prone to produce NOx . Catalytic convertors are used as post com-
bustion oxidation of unburned HC and CO but are not suitable to reduce NOx .
There are only few solutions for NOx reduction based on post combustion
treatment of exhausted gases. The recently concerns about the pollution from
diesel engines by the inherent NOx formation in lean combustion and the
economical motivation to avoid established post combustion treatments com-
bined with the necessity of reduction of CO2 resulted from even ideal com-
bustion have driven us to search the possibility of mixed procedures by using an
electrolytic cell as source of reactants for a relatively simple and well known
reaction that may solve the problems of lean combustion.
Keywords: Electrochemical reactions Pollution Nitrogen Nitrogen

Oxides Electrolytic cell Carbon Cleaning Machine “ECO” maintenance
system
1 Introduction
It is well known that pollution caused by internal combustion engines has become a
major issue. The international efforts to reduce the pollution started in 1960 and
continue today. Engineering requires today qualification in various other fields
apparently not related to the field in order to control the combustion and adjust the
parameters to achieve the mechanical work desired but to maintain the pollutants
within stringent levels imposed. Besides the electronics, automation etc. the role of the
specialists in chemical field has increased, as well as the extension to fuels
compositions.
At European level mandatory EURO norms become more and more stringent and
car manufacturers have had to invest in optimization of various specific processes
aiming ideal combustion equation.

Experimental Research Concerning the Influence of the Electrochemical Reactions 463
The reality is more complex than theories and experimental laboratory conditions.
Considering the fact that air contains Nitrogen, the ideal combustion process is
impossible to be achieved.
Exhausted gases from Internal Combustion Engines [1] contain the following main
pollutants:
CO – Carbon Monoxide
HC – Hydrocarbon (unburned residues)
NOx – an instable combination of Nitrogen Oxides NO and NO2
PM – particulate matter composed of incomplete burned fuel and various oxides of
fuel composition
SOx – Sulphur Oxides.
Theoretically CO, HC and PM can be reduced to CO2 and water by proper com-
bustion inside the combustion chamber. SOx are reduced by optimization of fuel
composition [2].
NOx are considered one of the major toxic pollutants in exhaust gases due to the
proportion in the summarized index of toxicity and due to the negative effect over
human bodies. Nitrogen Oxide and Nitrogen Dioxide are produced inside the com-
bustion chamber according to equations:
N2 þ O2 ! 2NO ð1Þ
2NO þ O2 $ 2NO2 : ð2Þ
The last equation is reversible and this makes the flue concentration analyses more
difficult. Nitrogen Dioxide NO2 is approximately four times more toxic than CO and in
presence of water has the tendency to convert into Nitric Acid HNO3 (a substance that
attacks the fine elements of the lungs). The reaction is:
3NO2 þ H2 O ! 2HNO3 þ NO: ð3Þ
This affirmation emphasizes the importance of reducing NOx in exhaust gases of

ICE.
There are major difficulties regarding this achievement due to the fact that NOx
emissions show a dependency of parameters of the cycle different to specific fuel
consumption. It means that reduction of NOx emissions will trigger an increase of fuel
consumption and reverse. According to these observations, the correlation of those two
aspects can be achieved only by the structure of fuel/air mixture in a correlation that
may obtain an average temporary temperature level of the mass of gases in the com-
bustion chamber and a minimal uniformity of temperature field.
2 Content
Recent research in the field shows that optimization of combustion inside the com-
bustion chamber reaches an optimal peak and now is the time that imposes the presence
of auxiliary equipment’s for the improvement of specific parameters of toxicity and fuel
464 M.N. Moldoveanu et al.
economy. Despite the effectiveness of the NOx reduction these solutions do not con-
tribute to fuel economy and increase the operational costs.
Modern traffic combines two major conditions for improper combustion and makes
pollution more evident:
1. Irregular working conditions
Irregular working conditions (sudden acceleration followed by prolonged idling)
induce imperfect combustion leading to formation of carbon deposits accumulation
inside the combustion chamber. These are a source of even worse combustion
conditions. Ignition timing is altered to prevent knocking, air/fuel ratio is altered to
sustain required torque and further more solid deposits and sots are produced. EGR
is not used in transitional regimes. Due to the improper combustion conditions the
functionality of TWC, SCR is reduced and DPF are more easily clogged.
2. Concentration of pollutants
Increased pollutants volumes concentrated in urban areas without a continuous
circulation of air masses and average temperature higher facilitate further more
re-composition of even more dangerous pollutants. The chemical reactions of NOx
in the environment convert the gas in acid aqueous mixtures.
Considering to aspect (1) virtually there is no single one vehicle in traffic that can
sustain the original emission level to its designed values according to current standards.
Considering the aspect (2) the natural processes can be used and the negative effects
can be turned into an asset in the goal to achieve a better efficiency of ICE.
Starting from this point our team have been testing the effect of an electrolytic cell
and the presumed Hydrogen role in positive results observed by various experiments
but focused on the electrochemical aspect of combustion and the real composition of
the products of electrolysis in the positive results experiments. Our technology imply
carbon deposits cleaning procedure followed by a chemical enhancing of catalytic
converter from CO and HC reducer to NOx reducer.
We have presented the role of alkaline ions produced by an on-board oxyhydrogen
generator combined with EGR containing NOx in a previous paper [6].
The study of electrochemical processes allows us to record a Particulate Matter
reduction to 51.85 % on a diesel engine without any additional anti-pollution equip-
ment. We have noticed the prolonged positive effect even though the exposure to the
electrolytic cell products was of a short period of time but were concerned with
possibly increased NOx pollutants as was reported by most experiments conducted by
others with similar systems.
Usually such electrolytic cells are used as presumptive on-board fuel savers (aka
HHO) or off-board Carbon Cleaning Machines with limited but positive effects over the
emissions. Electrolytic cell produces gases: Hydrogen, Oxygen and (depend on design
and operational parameters) water vapors. These water vapors are contaminated by
alkaline electrolyte (commonly KOHðÞ or NaOHðÞ ). Such systems are designed to
eliminate the alkaline ions from the stream of gases produced in order to prevent the
corrosion of various parts of the engine.
Carbon Cleaning Machines – a little more sophisticated electrolytic cell in fact – are
used with good results for restoring the engines original emission values, power and
fuel consumption. The carbon cleaning procedure is supposed to be achieved by
admission of a certain volume of Oxyhydrogen stoichiometric mixture or pure
Hydrogen. The general accepted theory is that the Oxyhydrogen mixture or Hydrogen
gas burns faster and with higher temperature than fuel and releases water and the
overheated steam is the cleaning agent. Previous carbon deposits accumulated on
cylinder head, valves and spark plugs are removed on exhaust pipe. This is true for
clean Hydrogen gas or Oxyhydrogen mixture used as cleaning agent. In order to
achieve a substantial result large volumes of dry (scrubbed) electrolytic gases are used.
Systems running at several kW able to generate 10 or even more LPM of gas are used.
3 Hypothesis
We have concluded that the traces of Kð þ Þ or Nað þ Þ ions present in water vapors are
susceptible to induce further reaction [3, 4, 5]. The extreme conditions from com-
bustion chamber may induce the migration of alkaline ions and lead to generation of
even more Hydrogen by chemical reaction with water:
K ð þ Þ þ H2 O ¼ KOH ðÞ þ H ðmonatomic ProtiumÞ: ð4Þ

Furthermore, ionized water (due to heat and static charges), overheated Nitrogen
and Nitrogen Oxides produced by the engine become “active” by reactions:
H2 O ¼ H3 Oð þ Þ þ OHðÞ ð5Þ
N2 þ O2 ¼ 2NOx ð6Þ
NOx þ KOH ¼ KNOx þ H2 O ð7Þ
HC þ O ¼ CO2 þ H2 O. ð8Þ
Those reactions take place inside the piston chamber during the combustion and
also inside the fitted catalytic converter as part of the entire power train assembly.
In the catalytic converter porous body, the flue of exhausted gases is cooled and
even the salts of Nitrogen previously formed may be decomposed in presence of noble
metal catalytic element by reactions:
KNOx ¼ N þ O þ KOH ð9Þ
N þ N ¼ N2 ð10Þ
O þ O ¼ O2 ð11Þ
CO2 þ KOH ¼ KCO3 : ð12Þ

Furthermore, the exhausted gases and water vapors with all chemical byproducts
are passed through the TWC. The porous ceramic impregnated with catalytic materials
is not “poisoned” with alkali hydroxides that are trapped inside the pores. In this
moment the function of a TWC is enhanced. It became a NOx trap as well.
Potassium ion Kð þ Þ prevents the NOx formation inside the piston chamber and is
expelled and accumulated inside the catalytic converter most probable in form of KCO3
and KNO3 salts. In this situation the Pt (Platinum) catalyst is “poisoned” with free
Kð þ Þ ions “released” from salt forms due high temperature of HC post combustion
oxidation process and the ionized water vapor of exhaust gases. Under the substrate of
Pt catalyst the constituent element form a plasma and due lower pressure of the process
the NOx and CO are reduced to N2 and turned into CO2 . The Platinum catalyst not only
thermally oxidizes HC and CO to H2 O and CO2 but decompose NOx to N2 and O2 in
series of reactions. The result is a clean exhaust flux with lesser CO2 and water vapors
free of CO, HC and NOx with harmless salts.
4 The Experimental Unit
To prove this hypothesis, we redesigned a Carbon Cleaning Machine of low gas output
– considered improper for serious and sustained effects – to neutralize but not to
remove the alkali ions and proceeded to tests. The neutralization may also improve the
lubrication as well but we do not have the equipment to verify it.
The system description (Fig. 1):
An electrolytic cell (1) filled with electrolyte (K) and a piezo-electric fog generator
(2) combined into a functional unit correlated to vehicle’s ICE (5) and additional
antipollution equipment (6). The electrolytic cell (1) has a power control unit that
convert the alternating current electricity into direct current electricity by means of a
high efficiency AC/DC inverter (3) with voltage output in a range of 24–27 V
adjustable by a potentiometer (4). The piezo-electric fog generator (2) is enclosed in a
case (2a) that contains water up to a certain level (WL). The case (2a) is connected by a
Fig. 1. The system description of experimental unit

pipe (1a) with electrolytic cell (1) and has a tube (2b) with the end lower than pipe (1a).
A flexible pipe (2c) connects the device to the engine (5) intake manifold (5a).
5 Test Results
After a series of positive effects, we proceed to Chinese accredited emission inspection

with a first vehicle – Audi A6L, gasoline engine 1.8T AT (gasoline turbo engine with
automatic transmission).
We tested the vehicle and in original status could not pass the annual emission test
due to values CO – 38,4 g/km and HC + NOx – 8,6 g/km (Fig. 2). The accepted
values are CO – 10,8 g/km and HC + NOx – 4,0 g/km. We proceeded with our
technology and obtained the results: CO – 0,1 g/km and HC + NOx – 0,1 g/km
(Fig. 3).
Fig. 2. The original status emission test
Fig. 3. The emission test after treatment
The positive results are sustained for a 4-month period of time when the vehicle is
driven more than 10.000 km from the base line odometer readings of 205.405 km to
216.380 km when the second test shows the values: CO – 0,85 g/km and HC + NOx –
0,52 g/km (Fig. 4).
Fig. 4. The emission test after 11075 km and 4 months

The conclusive diagrams for the tests results are presented in Figs. 5, 6 and 7.
Fig. 5. The test HC + NOx results
Fig. 6. The CO test results
Fig. 7. The test results after 11075 km and 4 months
Considering the positive results in such radical range recorded with more others
vehicles we start measuring the parameters during the procedure in order to establish
the correlation between the actual engine conditions and the system working param-
eters for optimal results.
The emission values before, during and after treatment are measured with a Gas
Analyzer model SV-YQ Tianjin Shengwei Development of Science Co. Ltd. (com-
ponents: Analyser SV-5Q and Opacimeter SV-5Y).
Please notice the water and carbon residues eliminated by exhaust pipe in Fig. 8
and the emission tester records at the beginning and at the end of procedure (Table 1).
Fig. 8. The water and carbon residues eliminated by exhaust pipe
Table 1. The emission tester records at the beginning and at the end of procedure
The moment of values CO [%] HC [ppm] NOx [ppm] Date/Time
The original values 1.46 128 68 16.05.2016/10.11 AM
After treatment values 0.00 0 4 16.05.2016/11.23 AM
6 Conclusions: “ECO” Maintenance System and Beyond
In actual paper we present the tests with a maintenance system that include generically
an electrolytic cell and a water mist machine combined in order to prove the hypothesis
that by chemical reactions the electrolysis products can neutralize the pollutants
emission in a ICE (in particularly the lean mixture emission of NOx ). The goal was
achieved. We proved that may be combined the Carbon Cleaning Machine action with
the effect of alkaline ions with the positive effect over the PM, CO, HC and NOx .
According to above presented experimental results the hypothesis elaborated by us
was proved and due electrochemical effect of electrolytic cell products (a) virtually any
on road vehicle can be “restored” to near optimal combustion conditions by the effect
of Carbon Deposits Cleaning by Oxyhydrogen gas and (b) the pollution control aux-
iliary equipment already existent can be restored or even more, upgraded, being able to
reduce NOx at levels beyond the original parameters. The dual effect is possible by a
simple and cost-effective maintenance procedure using the device developed by our
team.
The mandatory annual emissions test conducted by accredited tests facilities in R.
P. China proved more than 50 % reduction (in stage one test – simplified system) of
PM in case of a diesel engine and more than 90 % reduction of NOx in case of gasoline
engines (in a stage two – more elaborated system).
The working parameters of the maintenance system elaborated by our team were
observed with a calibrated exhausted gases analyzer and a working procedure was
elaborated.
This will allow us to develop the stage three – complex system able to respond to
acute demand of automotive market situated under a high pressure of strict emission
regulation and low fuel consumption.
Our hypothesis was confirmed and is possible that - with proper research and
development support - to achieve a stage three – an onboard device able to use water
based harmless solutions in order to totally replace combustion of hydrocarbon fuels
and create a fuel flex system that will have a positive impact over on road vehicles
contribution to environmental pollution.
References
1. Bobescu Gh. ș.a.: Motoare pentru automobile și tractoare, vol. III-Economie și ecologie. Surse
energetice alternative. “Tehnica-Info” Publishing House, Chișinău (2000)
2. Chiriac R. ș.a.: A study of combustion of hydrogen-oxygen gas mixture enriched gasoline in a
spark ignition engine. In: International Conference of Scientific Paper, CAR 2005, Pitești, 02–
04 November 2005
3. Amminex Emissions Technology A/S, New world standard for NOx reduction. http://
amminex.com. Accessed June 2016
4. Chiriac R. ș.a.: Effects of Gasoline-Air Enrichment with HRG Gas on Efficiency and
Emissions of a SI Engine, SAE Paper 2006-01-3431 (2006)
5. Conte, E., Boulouchos, K.: Influence of hydrogen – rich – gas addition on combustion,
pollutant formation and efficiency of a IC–SI engine, SAE Paper 2004-01-0972 (2004)
6. Moldoveanu M., Aramă C.: Experimental research concerning the influence of the on-board
hydrogen supply equipment on the engines combustion. In: International Conference of
Scientific Paper, AFASES 2014, Brașov, 22–24 May 2014
Energy-Based Approach to Model a Hybrid
Electric Vehicle and Design Its Powertrain
Controller and Energy Management Strategy
Nicole Salloum, Serge Francis, and Charbel Mansour(&)
Industrial and Mechanical Engineering Department, Lebanese American

University, Byblos, Lebanon
{nicole.salloum,serge.francis,
charbel.mansour}@lau.edu.lb
Abstract. This paper presents a methodology to develop an energy manage-

ment strategy that would quasi-optimize the fuel consumption of a parallel
Hybrid Electric Vehicle in real-time operations over the entire trip. Hence, the
main highlight of this study is the presentation of a coherent energetic modeling
technique to simulate a parallel hybrid electric vehicle, by providing an
energy-based representation of the powertrain components; presenting a sys-
tematic technique to design the vehicle controller structure and designing
(through a comprehensive global-optimization approach) a rule-based EMS
showing close-to-optimal consumption results over the entire scheduled trip.
Keywords: Energy-based modeling Energetic macroscopic representation

Hybrid electric vehicle Vehicle controller Energy management strategy
1 Introduction
Hybrid electric vehicles consist of the combination of two power sources: an internal
combustion engine (ICE) and an electric motor. Thus, the complexity of the powertrain
architecture requires extensive modeling and simulation. Several efforts have been made
to model such vehicles that resulted in different modeling techniques. For instance, the
physical system modeling technique provides the benefit of a realistic modeling of
physical components so that it can be easily implemented on hardware [1, 5]. However,
it requires expertise to implement new models. Another illustration of a modeling
approach is the mathematical modeling technique: a process of using mathematical
structures such as graphs, equation, and diagrams to represent a particular model. Its
main advantage is that it allows finding out the essential characteristics of a component
while removing the gaps in knowledge about the component [8]. Another modeling
method presented in this paper is the Energetic Macroscopic Representation (EMR),
very advantageous for modeling HEVs as it allows tracking the energy flow throughout
the various powertrain components, it provides a simple control over the powertrain
components, and ensures optimal control by combining the model to a global opti-
mization energy management strategy [2]. In this paper, the EMR is used to model a
parallel HEV, along with its powertrain controller and energy management strategy.

472 N. Salloum et al.
2 Powertrain Architecture
The vehicle studied corresponds to a front-wheel drive parallel hybrid architecture that
combines a mono-directional thermal powertrain to a bidirectional electric powertrain.
The thermal powertrain consists of a 1.2 L downsized-turbocharged engine and a 5
gears automated manual gearbox. The electric powertrain consists of a 27 kW motor, a
power electronics module and a 6.5 Ah NiMH battery. A pre-transmission parallel
hybrid configuration is considered as illustrated in Fig. 1, with direct coupling between
the engine and the motor, separated by a clutch system in order to reduce the engine
drag losses during electric drive mode. The vehicle and components specifications
needed for modeling are summarized in Table 1.
Fig. 1. Pre-transmission parallel hybrid configuration considered in this study.
Table 1. Vehicle and components specifications

Vehicle specifications Unit Value
Vehicle Mass Kg 1560
Frontal area/Drag coefficient m2/- 1.2/0.3
Wheel radius m 0.32
Gearbox efficiency - 0.95
Gearbox ratios - [3.64 1.95 1.28 0.97 0.77]
Differential ratio - 3.76
Auxiliaries’ consumption W 294
Battery voltage/capacity V/Ah 240/6.5
3 Powertrain Design and Model
3.1 Energetic Macroscopic Representation

EMR is the synthetic energy-based graphical tool used to model the considered vehicle.
Its main objective is to simulate the energy distribution throughout the powertrain [2].
Each component in the latter is defined as a single subsystem, known as element
illustrated in standard shapes and colors depending on their physical properties. Ele-
ments are grouped according to their physical nature: energy sources and sinks, con-
version elements, accumulation elements, and coupling elements. They are represented
Energy-Based Approach to Model a Hybrid Electric Vehicle 473
by pictograms, illustrated in Appendix A [7]. All elements are connected by two

opposing vectors that represent their action and reaction. The product of the variables
on the corresponding vectors determines the instantaneous power exchange, thus
defining the integral causality principle. For example, the internal combustion engine
(ICE) converts chemical energy to mechanical energy. Being a multi-physical con-
version element, it is represented by an orange circle in the model. Its output is the
crankshaft torque (action) whilst its input is the crankshaft rotational speed (reaction).
Therefore, the instantaneous mechanical power produced by the engine can be obtained
by multiplying the action and reaction variables.
3.2 EMR of the Studied Model

The model is divided into subsystems that include the engine, the transmission, the
motor, the battery and the vehicle chassis as stated in the first section. These different
components of the vehicle and their corresponding relationships are introduced in this
section.
Engine
The internal combustion engine (ICE) represents a 4 stroke turbo charged engine. It
produces the engine torque TICE, which is controlled by a reference torque TICE,ref
acquired from the MCS. In EMR, the ICE is represented by a controllable power source
with the torque TICE as output and the torque reference TICE,ref as tuning input. ICE is
thus considered a conversion element. The engine drag torque is given by Eq. 1:
Te0 ¼ ae JICE þ Tf ð1Þ
where αe is the acceleration required at the crankshaft, JICE is the inertia of the engine
and Tf is the friction torque function of the engine speed. The fuel consumption of the
ICE is obtained from the engine specific fuel consumption (sfc) map as function the
engine torque and rotational speed.
Electric Motor
The motor is powered through a power electronics module by a DC voltage (Ubat) and
is directly controlled by the reference electric motor torque (TEM,ref) generated by the
MCS. Transient behavior of the motor is not included in the model for simplicity. This
machine is considered as an electromechanical conversion (circle) with the voltage Ubat
and the rotational speed wm as inputs, the current Ibat and the torque TEM as outputs and
the reference torque TEM,ref as tuning input. The motor drag torque Tm0 is given by
Eq. 2.
Tm0 ¼ am Jm ð2Þ
where αm and Jm are the acceleration and the inertia of the electric motor respectively.
Hence, the motor electric power is expressed in Eq. 3.

(
TEM wm
gm for TEM 0
PEM ¼ ð3Þ
TEM wm gm for TEM \0
Battery and auxiliaries

The total power required from or supplied to the battery is given by: Pbat = PEM + Paux
where the auxiliaries’ power is 250 W and the efficiency of the DC-DC converter is
85 %. Therefore, the power consumption due to auxiliaries is 294 W. The current Ibat
accounts for the motor load Imot and the auxiliaries load Iaux (Eq. 4). Equation 5
expresses the equation of Imot which results from the battery power exchange, Pbat. It is
expressed as function of the open circuit voltage of the battery Voc and the battery
internal resistance R, both function of the State of charge (SOC). The battery is con-
sidered a source element. The battery state of charge SOC is given by Eq. 6.
Ibat ¼ Imot þ Iaux ð4Þ

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Voc ðSOC Þ Voc2 ðSOC Þ 4 P RðSOC Þ
Imot ¼ bat
ð5Þ
2RðSOC Þ
Z t
SOC ¼ Cini þ I dt Cmax ð6Þ
t0
Transmission System
In EMR, it is represented by a mechanical conversion element (square). It represents
the driveline of the powertrain which consists of a 5 speed manual gearbox along with
the differential. The gearbox torque and speed are expressed in Eqs. 7 and 9 respec-
tively. Note that both thermal and electrical mechanisms are mounted on the same
shaft. The mechanical coupling element sums the two torques TICE and TEM as shown
in Eq. 8.
TGB ¼ Tdriveline KGB gGB þ TGB losses ð7Þ
Tdriveline ¼ TICE þ TEM þ Tdrag ð8Þ
wICE ¼ wwheel KGB ð9Þ
And the traction force and speed on the wheels as expressed in Eqs. 10 & 11
respectively:
1
Fwheel ¼ TGB ð10Þ
Rwheel
1
wwheel ¼ vveh ð11Þ
Rwheel
Vehicle Chassis
The chassis accumulates kinetic energy in the mass of the vehicle. The used vehicle has
a small size with an initial mass (mi) of 1474 kg. Adding the masses of the electric
motor and the battery, the total vehicle mass is:
mveh ¼ mi þ mEM þ mbat ð12Þ
where the mEM and mbat are function of the electrification ratio ER.
Equation (13) represents the kinetic energy that builds up in the mass of the vehicle
in function of the traction force Ftraction and the resistive road forces Froad.
d
mveh vveh ¼ Ftraction Froad ðð13ÞÞ
dt
Froad is calculated in Eq. 14.
Froad ¼ f0 þ f1 vveh þ f2 v2veh ð14Þ
where f0, f1 and f2 are provided from an experimental test bench.
4 Powertrain Controller
4.1 Maximum Control Structure
Once the vehicle is modeled, it is self-evident to design the powertrain controller. An
Inversion-Based technique, also known as Maximum Control Structure (MCS), is used
for its simplicity and rapid deduction of the control strategy. In order to determine the
variables that will be controlled, a tuning path is developed to link the tuning inputs to
their outputs. Thus, the powertrain elements that are represented by arrows in the
tuning path are the elements that are inverted, constituting the backward tuning path.
The inversion of the determined elements consists of a set of standard rules that
abide by the EMR integral causality principle. Consequently, time-independent vari-
ables such as the rotational speed of the crankshaft are directly inverted using the same
formulas of the EMR model whilst time-dependent variables are inverted using a
proportional integrator (PI) controller. Therefore, this control strategy defines the right
inputs to apply to the system from the desired output. For a better illustrative repre-
sentation, the control blocks offered by the EMR library differ from the defining blocks
by their blue color and tilted orientation, as illustrated in Appendix A.
4.2 MCS of the Studied Model

As stated previously, the MCS consists of identifying first the tuning path and the
control path, and then invert step-by-step the time-independent elements of the tuning
paths. All variables obtained in this section are referred to as reference variables since
they are controlled by the MCS. Figure 2 illustrates the identified tuning path and its
three tuning inputs: TEM_ref and TICE_ref as reference control torques for the two power
sources (ICE and the electric motor) and Fbrake_ref as the reference control braking
force.
As a second step, the MCS of the tuning path is deduced by inverting the relevant
elements equations, and the generated control path is represented in Fig. 3.
Fig. 2. Tuning path of the powertrain.
Fig. 3. Control path of the powertrain.
Inversion of the Transmission

The driveline torque is split between the engine and electric motor torques as shown in
Eq. 9. The split ratio of the electric motor torque to driveline torque is known as the
torque split factor and denoted uTSF. The latter represents the operating driving mode.
Hence, a value 1 of uTSF corresponds to electric drive mode, 0 to thermal drive mode and
–1 to brake energy recovery mode. Values between 0 and 1 indicate that both the engine
and the electric machine provide torque for traction, and values between –1 and 0
indicate that the engine provides excess power to drive the vehicle and recharge the
battery through the electric machine (operating in generator mode). Values of uTSF are
obtained from the strategy block that will be further discussed in the next section.
The equations used in the MCS are:

TEM ¼ Tdriveline ref þ Tdrag uTSF ð15Þ
TICE ¼ ðTdriveline ref þ Tdrag Þ ð1 uTSF Þ ð16Þ
As for the reference driveline torque, the torque losses due to the gearbox ineffi-
ciency, and the reference torque of the gearbox, are obtained by directly inverting
Eqs. 7 and 8 to get the following equations:
TGB ref
Tdriveline ref ¼ ð17Þ
KGB gGB
TGB ref
TGB losses ¼ Tdriveline ref þ ð18Þ
KGB
TGB ref ¼ Fwheel Rwheel ð19Þ
Inversion of the Chassis

The chassis, being an accumulation element, cannot be directly inverted as it would
violate the integral causality principle of EMR. The control of the actual velocity of the
vehicle requires a PI controller in order to avoid having a derivative operator.
The EMR of the model, along with its inversion based control, is shown in the
following figure (Fig. 4).
Fig. 4. EMR and MCS of the studied hybrid electric vehicle.
5 Energy Management Strategy
In order to achieve the control strategy in MCS, another control level is added to the
EMR which is the vehicle main control that includes the Energy Management Strategy.
The latter is usually implemented to optimize several objective functions such as the
fuel consumption, the gear-shifting, the driver’s comfort, and the emissions. This study
focuses on the minimization of energy consumption only, while the other objectives are
considered in future works.
To optimize the fuel consumption, several energy management methods can be

found in the literature. One of them is the Dynamic Programming (DP) which leads to
the global optimal solution. However, DP is a method that cannot be used in real-time
because of the heavy computation time required. The adopted technique for real-time
control is the Rule-Based EMS, which generates a local instantaneous optimum [3, 4].
The aim of this study is to combine the advantages of both strategies to ensure
close-to-optimal fuel savings. This new strategy is known as optimized RB EMS [6, 9]
and will be elaborated in this section.
Building the optimized RB EMS consists of 3 steps:
1. Run the DP algorithm on a pre-defined driving cycle.
2. Cross-compare the load power results with the obtained torque split ratio in order to
identify the different driving modes under optimal control (Fig. 5).
3. Deduce the rules that define the operating modes based as function of the vehicle
load power.
6 Results
The EMR model, the MCS, and the optimized RB EMS have been tested on the New
European Drive Cycle (NEDC), and obtained results are compared to the DP model
results.
The following figure shows the torque split factor as a function of the power load
which serves as a basis to determine the rules of the respective driving modes. For
instance, a negative power load corresponds to the Brake Energy Recovery (BER) mode.
A positive power load smaller than the threshold PEV corresponds to the Electric Motor
(EM) mode. A power value ranging between the two threshold values PEV and PPS
determines the power Peaking Source (PS) mode, and a value greater than PPS corre-
sponds to the Engine (ICE) mode.
Fig. 5. Optimized rule-based EMS based on DP optimization results for the NEDC.
The parameters used to validate the model and its control are (1) the velocity profile
to make sure the actual velocity follows the input velocity (Fig. 6), (2) the battery SOC
that should follow the battery sustainability constraint of maintained initial and final
SOC of 60 % (Fig. 7), and (3) the fuel consumption to make sure it is close to the
optimum resulted from DP. A 2.3 % fuel consumption difference was observed when
comparing the DP results (3.82 l/100 km) to the optimized RB EMS results
(3.91 l/100 km).
Fig. 6. Predefined and controlled velocity profiles for the NEDC
Fig. 7. DP and RB results of the battery SOC
7 Conclusion
This paper presents the design and simulation of a hybrid electric vehicle using an
energy-based modeling technique, the Energetic Macroscopic Representation. This
approach presents two major advantages namely the ability to track the power flow
through the sub-elements of the powertrain and the easiness to deduce the control level
by a step-by-step inversion of the components. A new energy management strategy,
optimized Rule-Based strategy, is introduced which yields to a close-to-optimal fuel
consumption in a real time framework.
Appendix A: Synoptic of Energetic Macroscopic Representation
References
1. Patil, K., Molla, S.K., Schulze, T.: Hybrid vehicle model development using
ASM-AMESim-simscape co-simulation for real-time HIL applications. SAE Technical Paper
Series (2012). doi:10.4271/2012-01-0932
2. Letrouve, T., Delarue, P., Bouscayrol, A. (2009). Modelling and control of a double parallel
hybrid electric vehicle using energetic macroscopic representation. In: 2009 8th International
Symposium on Advanced Electromechanical Motion Systems & Electric Drives Joint
Symposium (2009). doi:10.1109/electromotion.2009.5259137
3. Salmasi, F.R.: Control strategies for hybrid electric vehicles: evolution, classification,
comparison, and future trends. IEEE Trans. Veh. Technol. 56(5), 2393–2404 (2007). doi:10.
1109/tvt.2007.899933
4. Serrao, L., Onori, S., Rizzoni, G.: A comparative analysis of energy management strategies
for hybrid electric vehicles. J. Dyn. Syst. Meas. Control 133(3), 031012 (2011). doi:10.1115/
1.4003267
5. Singh, R., Dugaya, N.: Design and simulation of a 25/250 V DC-DC boost converter fed by a
PV pannel in SIMSCAPE library of MATLAB. Int. J. Enhanced Res. Sci. Technol. Eng. 4(4),
12–17 (2014)
6. Cheng, Y., Horrein, L., Mansour, C., Bouscayrol, A., Dumand, C.: Modeling and control
optimization for Peugeot 3008 Hybrid4. Paper presented at Proceedings of the SIA
International Powertrain Conference: The low CO2 spark ignition engine of the future and its
hybridization, Versailles, France (2015)
7. Barrade, P., Bouscayrol, A.: Energetic Macroscopic Representation–An energy-flow based
methodology dedicated for the control of multiphysics systems (2011)
8. Novoseltsev, V.N., Novoseltseva, J.A.: Mathematical modelling - advantages and shortages
(2004)
9. Mansour, C.J.: Trip-based optimization methodology for a rule-based energy management
strategy using a global optimization routine: the case of the Prius plug-in hybrid electric
vehicle. In: Proceedings of the Institution of Mechanical Engineers, Part D: Journal of
Automobile Engineering (2015). doi:10.1177/0954407015616272
Dynamic Control of an Electric Vehicle
with Traction Induction Motor
Marius Constantin Georgescu(&)
Transilvania University of Brasov, Brașov, Romania

mgeorg@unitbv.ro
Abstract. During recent years, to solve as soon as possible the environment

pollution problem, considerable improvements have been made in research and
in international legislation. As research and manufacturing results are mentioned
the new types of hybrid and electric vehicles and as legislation improvements
have started to view carbon dioxide as a greenhouse gas that traps the earth’s
heat and contributes to global warming. The present paper has, as a main target,
the description of a modern and environmentally friendly solution regarding the
dynamic control of an electric vehicle with traction induction motor. The
dynamic control proposed by the author provides better traffic safety and an
efficient management of electrical energy for the vehicle (reduced consumption
by approximately 12 %).
Keywords: Electric vehicle Induction motor Dynamic control
1 Introduction
Automobile releases many liquid or/and gaseous wastage which are the cause of
environment pollution. These components are: nitrogen oxide (NOx), hydrocarbon
(HC), carbon monoxide (CO), volatile organic compounds (VOCs) and sulphur di-and
tri-oxide (SOx). All of them are depending on the content of fuel-based engine. To avoid
the environment pollution, the development politics of an eco-friendly and sustainable
transport promote more and more electric vehicles use. Requirements for electrical
motors, as main vehicle propulsion sources, are getting more and more demanding. In
this aim, passenger and freight vehicles must accomplish a large autonomy, short
re-fueling time, acceleration of 10–15 s to reach speeds of 0–100 km/h, drive uphill on
5–10 % ramps at legal top speeds and without environmental hazards. To have a good
selection of electrical traction motors, must be respected the main following require-
ments: high torque at starting, good power/gauge ratio, flexible traction characteristics,
energy efficiency, extended speed range ability, good thermal cooling conditions,
safety/reliability and acceptable costs. The traction characteristics help to optimal
evaluate the vehicle dynamics in acceptable economic conditions.

Dynamic Control of an Electric Vehicle 483
2 Vehicle Propulsion Dynamics

2.1 Motion Forces
In order to respect the vehicle’s operating constraints and to have good dynamics, it
requires an extended speed range and a constant power operating flexible region for
different hill climbing conditions. Figure 1 presents a schematic representation of the
forces acting on the vehicle in motion [1].
Fig. 1. Forces during vehicle motion
From the figure results the expression of the vehicle movement equation, as
follows:
d
mv vðtÞ ¼ FT ðtÞ ½FA ðtÞ þ FRF ðtÞ þ FHC ðtÞ þ FD ðtÞ; ð1Þ
dt
where mv is the loaded vehicle mass [kg], v the vehicle longitudinal speed [m/s], FT the
traction force [N], FA the aerodynamic friction force [N], FRF the rolling friction force
[N], FHC the hill climbing force [N] and FD the disturbance force [N] that includes all
other not yet specified effects. The traction force is negative, for instance during the
vehicle braking. For the electrical vehicles case, the mathematical expressions for all
these main forces are the followings [1]:
• traction force FT, is calculated as
Tem ir Pem
FT ðtÞ ¼ gtr ¼ 2700 gtr ; ð2Þ
R v
where Tem is the electric motor torque [Nm], ir the gear ratio, ηtr the transmission
efficiency, R the tire radius [m], Pem the electric motor power expressed in [HP]. The
vehicle longitudinal speed v, expressed in [km/h] is calculated as:
ir
v ¼ 0; 377 nm ; ð3Þ
R
where nm is the motor rotation in [rpm].
484 M.C. Georgescu
• aerodynamic friction force FA, is calculated as
qa Af Cd 2
FA ¼ v ; ð4Þ
2
where ρa is the air density [kg/m3], Af the vehicle frontal area [m2] and Cd the aero-
dynamic drag coefficient estimated by experiments or by computer programs.
• rolling friction force FRF, is calculated as
FRF ¼ lrf mv g cos a; ð5Þ
where μrf is the rolling friction coefficient, g the gravity acceleration [m/s2] and α the
uphill driving angle. The rolling friction coefficient depends on vehicle speed, tires
pressure and road surface conditions.
• hill climbing force FHC, due to the gravity, is calculated as
FHC ¼ mv g sin a: ð6Þ
• disturbance force FD [N], that summarizes all other not yet specified effects [1].
To have a complete estimation of traction characteristics, one must calculate the
dependence between the vehicle power Pv and its speed within traffic, as follows:

q Af Cd 2
Pv ¼ mv g Cr þ a v v; ð7Þ
2
where Cr is the wheel rolling coefficient.
2.2 Traction Induction Motor

Presently, brushed dc motor, brushless dc motor, ac induction motor, permanent
magnet synchronous motor and switched reluctance motor are the main motors used for
Electric Vehicle (EV) driving. The choice of traction electric motor for a specific
electric vehicle is dependent on main factors, such as: utility of the EV, easy, smooth
and accurate control, efficient output power at shaft, conversion efficiency of 90 %,
constructive simplicity, high torque when the vehicle slows down or stops, no extra
gears for higher power curves and good power/weight ratio. As consequence, induction
motors are among the top candidates in electric vehicle propulsion [3, 4]. They are
widely used in modern electric vehicles. Some research even concludes that induction
motor provides better overall performance compared to other motors [4]. An induction
squirrel-cage motor has some advantages as: low-cost for maintenance, high efficiency,
high reliability, easy for cooling and firm structure, making it very competitive in EV
driving. In this paper the traction induction motor is considered a time-varying
multi-variable non-linear system, hence the modeling task is not easy. The induction
motor fed by a PWM voltage inverter is field oriented controlled [2].
Adjusting of vehicle running speed must be completed by limiting the

acceleration/deceleration torque at suitable values, for passengers. To easy reproduces
the acceleration pedal function, the direct torque control method can be also used [3].
As traction motor it has been used an induction motor with the following technical
data: rated power Pn = 90 kW (120 HP), rated line voltage Un = 400 V, rated current
In = 153,5 A, pole pairs p = 1, rated rotor speed n = 2975 rpm, rated mechanical
torque Tm = 288,8 Nm, stator resistance R1 = 0.1113 Ω, rotor resistance R2 = 0.01685
Ω, stator inductance L1 = 0.0002373 H, rotor inductance L2 = 0.0003075 H, mutual
inductance Lm = 0.014 H and efficiency η = 94.1 %. The induction motor is fed at
rated voltage by a PWM inverter with IGBT power transistors, having 10 kHz com-
mutation frequency, efficiency of 95 % and power density of 3.5 W/cm3.
3 Dynamic Control of the Vehicle
3.1 Traction Control

The traction control system is required to be fast in response and low-ripple. Electric
vehicles require that traction induction motor has a wide range of speed regulation. In
this aim, two requirements must be fulfilled:
– to guarantee the speed-up time, the induction motor must have large torque output
under low speed and high over-load capability;
– to operate at high speed, the driving motor must have certain power output at
high-speed operation.
For EV one of the most advantages is the quick and precise torque response of the
electric motor. Besides, important information including wheel angular velocity and
torque, can be achieved much easier by measuring the electric current passing through
the motor. Based on these remarkable advantages, a couple of advanced motion con-
trollers are developed, in order to improve the handling and stability of EV [4]. To
develop a Traction Control System (TCS) with high performances, the control algo-
rithm must be independent on the identification of the road adhesion coefficient. Also it
must have an excellent robustness to estimate the vehicle velocity error. In this aim, has
been developed an anti-slip controller for an EV based on Variable Structure Control
(VSC) method. In order to have a VSC with good robustness, the control law adopted
is an equivalent control with switching control [5]. In this case, the output torque of the
induction motor is expressed as [4]
Tm ¼ Tmeq DTsgnðsÞ; ð8Þ
where Tmeq is the induction motor equivalent torque, ΔT the hitting torque of control
drive and sgn(s) = {−1,0,1} is the system switching function.
The sliding motion includes two processes:
– approaching motion which makes the system at any time in any position approach
to the sliding face in limited time;
– sliding motion which occurs when the EV reaches sliding surfaces.
486 M.C. Georgescu
The control process is based on a correct estimation of the wheel slip ratio λ
calculated with the following relation:
xR v
k¼ ; ð9Þ
xR
where ω is the wheel angular velocity [rad/s], R the tire radius [m] and v the vehicle
longitudinal speed [m/s]. In relation 9 are using the standards SAE wheel slip defini-
tion. In this case, ωR represents the longitudinal speed of the wheel center. The TCS
based on VSC strategy is presented in Fig. 2.
Controller
Hitting control torque ΔT

calculation
-1, 0, 1
Switching function
Equivalent torque of electrical Tmeq - Electrical Tm

motor calculation
+ Vehicle
motor
ω, v, Tm
Fig. 2. Torque control system based on VSC strategy
Regarding the high-frequency-chattering on sliding surface, is better to combine the

VSC advantages with the Model Following Control (MFC), [4], [6].
The MFC strategy is recommended in order to decrease the fluctuation to the motor
torque and slip ratio of the tire. It only requires the induction motor’s output driving
torque and the tire’s angular speed signal to put TCS into practice. This combined strategy
between VSC with MFC is practical because the longitudinal velocity estimation and the
optimal slip ratio identification can be ignored. Thus, the high-frequency-chattering near
the sliding surface can be avoided [4]. The author estimates that with such combined
system the induction motor consumption can be reduced by approximately 12 % and the
efficiency of energy recovery from braking increases by about 5 %, which lead to the
expansion of distance traveled on a single charge by more than 15 %. These estimations
take into account the standard European driving cycles.
3.2 Induction Motor Control

There are different control methods for induction motor, recommended by some
authors [2–4]. To control in speed and torque the induction motor, the Field Oriented
Control (FOC) method is used [2]. The drive system simulation is provided in
Simulink, as shown in Fig. 3. As shown in this figure, the block with induction motor,
contains the following subsystems: speed controller, drive control with FOC,
three-phase voltage inverter and induction motor. This block drive was created using
the block AC3 from the SimPowerSistems/Electric drives library and is intended to
simulate the FOC of the induction motor.
To use the original block AC3 for traction purpose, some modifications are set to
use it in a three-phase system with speed-control. The Asynchronous machine block
implements a squirrel cage induction machine that works as a motor and as a regen-
erative brake.
Fig. 3. Simulink diagram of control drive simulation
3.3 Vehicle Dynamics

The vehicle dynamics describes the EV behavior based on its parameters and
depending on external factors which influence motion. This block is depicted in the
Fig. 4.
The main vehicle technical data are: traction induction motor rated power Pn =
90 kW, maximum longitudinal speed v = 160 km/h, acceleration time of 10 s
(0-100 km/h), autonomy away of 470 km, total mass mv = 1200 kg, no tire wheel
radius rw = 0.3 m.
In the vehicle dynamics block diagram the main subsystems are: the Longitudinal
Vehicle Dynamics and the Drive Wheels (Tire, Tire1). These blocks are interconnected
with each other, the differential mechanism beeing also connected to the transmission
shaft which is driven by the traction induction motor.
488 M.C. Georgescu
Fig. 4. Vehicle dynamics block diagram
4 Practical Results
After conducting real-life experiments the following waveforms represent the behavior
of the modelled system. In Fig. 5 are depicted the waveforms representing the
parameters evolution of traction induction motor.
400
300
stator current, A
200
100
0
-100
-200
-300
-400
(a)
5000
motor speed, rpm
4000
3000
2000
1000
0
-1000
-2000
(b)
electromagnetic torque, Nm
400
Measured
300
Reference
200
100
-100
-200
-300
-400
0 5 10 15 20 25 30 35 40 45
(c)
time, s
Fig. 5. Parameters evolution of traction induction motor

Refering to this figure, some comments are made:

– during starting (0–12 s), the stator current is around 300 A at constant magnitude
and its frequency increases with increasing in speed. The electromagnetic torque is
quite constant around 300 Nm;
– at time t = 12 s the rotor speed becomes almost constant at 4000 rpm and main-
tained in a time of 8 s. During this time the stator current decreases at a third of the
current value of the acceleration and electromagnetic torque becomes constant
having 60–70 Nm. These are the values of current and torque to keep speed rotor
shaft that drives the vehicle at a constant speed, but over rated motor speed.
– at time t = 20 s the speed is controlled to rise, so it is a slower acceleration. There is
a low constant torque region sufficient to achieve a slower acceleration.
In Fig. 6 a simulation of the vehicle acceleration is presented, during 45 s until the
maximum speed of 160 km/h.
400
stator current, A
300
200
100
0
-100
-200
-300
(a)
6000
motor speed, rpm
5000
4000
3000
2000
1000
0
-1000
(b)
electromagnetic torque, Nm
350
Measured
300
Reference
250
200
150
100
50
0
-50
0 5 10 15 20 25 30 35 40 45
(c)
time, s
Fig. 6. Simulation of vehicle acceleration
5 Conclusions
This paper presents computer modeling and simulations related to a system drive for
vehicle propulsion with an induction motor control system based on VSC strategy. As
future research the author follows the implementation of a soft controller which
combines the VSC advantages with the MFC.
490 M.C. Georgescu
References
1. Calin, M.D., Georgescu, M.C., Helerea, E.: Magnetic Materials for Electrical Machines used
in Transportation. Lambert Academic Press, Saarbrücken (2015)
2. Leonhard, W.: Control of Electrical Drives. Springer, Heidelberg (1985)
3. Gyuláné, V., Gergely, G.B.: Electric Vehicles. BMEVIVEM263, Budapest (2012)
4. Soylu, S.: Urban Transport and Hybrid Vehicles. Sciyo, Croatia (2010)
5. Buckholtz, K.R.: Reference input wheel slip tracking using sliding mode control. In: SAE
2002 World Congress (2002)
6. Hori, Y.: Future vehicle driven by electricity and control-research on four wheel motored.
UOT Electric March II 51, 954–962 (2004). IEEE
A Propulsion System for Means of Transport
with Non-autonomous Electrical Traction
Tudor Mitran(&), Nicolae Chioreanu, Catalin Chioreanu,

Alexandru Rus, and Horia Beles

tudor_mitran@yahoo.com,
{nchioreanu,alrus}@uoradea.ro,
catalin.chioreanu@gmail.com, horia.beles@gmail.com
Abstract. The paper presents a new, innovative concept regarding a propulsion

system for non-autonomous electric vehicles (trolley buses, trams, subway
frames). The system’s main part is the electric-hydraulic generator. The gen-
erator converts electrical energy into hydrostatic energy that’s stored in an
accumulator. By using a hydraulic motor, this hydrostatic energy is converted in
motor torque in form of rotational motion. Because the electric motor operates in
a single regime, the electric consumption is reduced. Also the new system
ensures better dynamic performances and operational safety.
Keywords: Electric traction Conversion of energy Storage of energy

Electric-Hydraulic Generator
1 Introduction
The current trend in the field of the propulsion systems for means of transport with
electrical traction is to conceive and to realize models with reduce electric energy
consumption and with electromagnetically emissions that would be within the rules
imposed by the legislation. In this paper is analyzed a new, innovative propulsion
system, based on a patented invention [1] that by the concept point of view exceeds by
far the current stage of development in this field.
The last generation propulsion systems for non-autonomous electric vehicles
(trolleybuses, frames and subway frames supplied by DC contact lines with voltages of
600 V or 750 V, electric locomotives supplied by AC contact lines with voltages of
25 kV and 50 Hz frequency) are powered by rotary asynchronous three-phase motors
equipped with a three-phase inverter in the case of DC contact lines or with a static
frequency convertor in the case of the supply with AC contact lines [2].
The three-phase motor transforms the electric energy into mechanical energy in the
form of rotational motion that is transmitted through a reduction gear to the vehicle’s
wheels by direct coupling [3]. Due to this coupling mode, the engine has a stable
operation only if the produced power is equal at any moment with the power necessary
(the resistance power) to operate the vehicle. Generally, the resistance power is variable
during the operation. Due to this, the current motors must operate in variable regime,

the produces the reduction of the efficiency. The reduction of the efficiency depends on
the intensity of the motor’s operation regime variation.
In conclusion, the main disadvantage of the electric motor’s coupling (through a
reduction gear) direct to the vehicle’s wheels is the operation in a variable regime that
produces an increased consumption of electric energy. At the induction electric engines
the variation of the operating regime is made automatically, without any special device,
the amperage of the electric current in the windings is modified depending on the value
of the resistance torque.
The asynchronous motors have the following disadvantages: they consume reactive
power, at the startup they produce the overstrain of the power supply network (the
amperage at the startup is 2 to 7 times greater than the value of the rated current); for
the single phase motors a startup system is necessary (consisting of auxiliary winding
and capacitor). The main disadvantage of the synchronous motors is the need of a
startup system and a synchronization one.
Also, the induction electric motors have a constant rotational speed (quasi constant
at the asynchronous motors), and for it’s modification special devices called converters
are necessary, to modify the supply electric current parameters (voltage, amperage).
The converters have the main disadvantage that they alter the variation of the voltage
and the amperage wave of the electric current. Due to this altering (altered operating
regime), the efficiency and the power factor of the motors decrease, and produce
disruptive electromagnetic emissions. Although the electric motors are reversible
machines, the recovery of the braking energy is possible only at the actuation systems
with three phase asynchronous motors by using a complex electronic system that
produces great disruptive electromagnetic emissions.
2 The Operational Scheme of the Proposed Propulsion

System
The new propulsion system, which is the subject of the analysis, consists of three main
parts: the rotary electric-hydraulic generator GH, the hydro-pneumatic accumulator AH
and the hydraulic motor (Fig. 1). The system includes also a motor command system
SCM, for the hydraulic motor.
The electric-hydraulic generator GH, converts the electric energy in hydrostatic
energy, which is transmitted and stored in the hydraulic accumulator AH. Also, the
accumulator AH has the role to store the recovered braking energy.
Fig. 1. The operational scheme of the propulsion system

A Propulsion System for Means of Transport 493
The hydraulic motor MH takes over and converts the hydrostatic energy from the
accumulator into mechanical energy, usually under the form of rotational motion, with
the following parameters: M1 [Nm] is the motor torque and n1 [RPM] is the rotational
speed. The parameters of the rotational motion can be adjusted by the motor command
system SCM that modifies the volumetric capacity and the torque of the hydraulic
motor. Also, the motor command system can realize: the recovery of the braking
energy (by passing the hydraulic motor, which is a reversible machine, in a pump
regime) and the reversing of the rotation sense.
The operation of the propulsion system is as follows: the electric-hydraulic gen-
erator GH produces and transmits the hydrostatic energy to the accumulator AC until
the pressure of the hydraulic liquid in the accumulator reaches the maximum admis-
sible value. In this moment, automatically, the generator stops it’s operation until the
pressure in the accumulator drops to the minimum admissible value. In this way, the
electric-hydraulic generator operates in a single regime, with maximum efficiency and
automatically, on the start-stop principle. The hydraulic motor takes over the energy
from the accumulator and converts it in mechanical energy at the parameters necessary
to drive the vehicle’s wheel. At braking, the hydraulic motor is passed in a pump
regime and, in this way, the braking kinetically energy is converted into hydrostatic
energy that is stored in the accumulator AH.
In Fig. 2 is presented the operation scheme of the rotary electric-hydraulic gener-
ator. The main elements of the generator are: the single phased motor with a
short-circuited rotor ME, without startup winding and the hydraulic pump PH. The
pump PH is reversible and with constant volumetric capacity. The pump’s discharge
pipe is put alternately in connection with the accumulator (pump regime) or with the
tank Rz (motor regime) by the one sense unlock valves Sd1 and Sd2 and by the one
sense valves S1 and S2. The unlock of the valves Sd1 and Sd2 is done through the
distributor Dc which is with electromagnetic command.
The start-stop of the generator is made depending on the pressure of the hydraulic
liquid inside the accumulator AG, measured by the pressure transducer SP. The start of
the generator is made by passing the pump PH in motor regime. So, the electric motor
will be put in operation until it reaches the nominal rotational speed. In this moment,
Fig. 2. The operational scheme of the rotary electric-hydraulic generator

the electric motor ME is connected by the electronic contactor CE to the electric current
supply network, and the pump PH is passed into a pump regime. The information
regarding the moment in which the system pump-electric motor reaches the nominal
rotational speed is provided by the rotational speed transducer Th. The stop of the
generator is made by disconnecting the electric motor ME from the electric current
network by the contactor CE, when the pressure in the accumulator AG reaches the
maximum admissible value.
The electronic contactor CE, the command distributor Dc, the pressure transducer
SP, the rotational speed transducer Th and the electronic command and supply circuits
form the automatic command system of the generator that, depending on the pressure
inside the accumulator AG, realizes the start and the stop of the generator.
The operation of the generator is as follows: when the value of the hydraulic liquid
inside the accumulator AH is the minimum admissible one, the signal provided by the
transducer SP produces the switch of the distributor Dc from position I to position II.
The distributor Dc produces the unlock of the valves Sd1 and Sd2 that determines the
entry of the pump PH in a motor regime. The pump PH in motor regime drives the
electric motor ME until the synchronous rotational speed is reached. In this moment,
the signal provided by the rotational speed transducer Th produces the opening of the
contactor CE and the switch of the distributor from position II to position I. The
contactor CE connects the electric motor ME to the electric current network, and the
distributor Dc blocks the valve Sd1 and Sd2. The valves Sd1 and Sd2 in position
blocked determine the entry of the pump PH in a pump regime. In this moment, the
pump PH, driven by the electric motor ME, discharges hydraulic liquid under pressure
in the accumulator AG and in the accumulator of the propulsion system AH through a
one sense valve. In the moment in which the value of the pressure inside the accu-
mulator AG is the maximum admissible one, the signal provided by the transducer SP
produces the closure of the contactor, CE that disconnects the motor from the electric
current network.
3 An Experimental Model to Demonstrate the Advantages

of the Proposed Concept
In order to demonstrate and to validate the parameters of the proposed propulsion

system the authors have thought about an experimental model. The validation of the
functional parameters of the system will be made on a test bench for dynamics tests.
The operational parameters are: The nominal power P = 50 kW and the maximum
velocity 90 km/h.
The stand will permit the testing and the determination of the characteristic
dimensions of the auctioning system and the measuring of the following parameters:
the power, the torque, the speed, the traction force and the electric energy consumption.
The scheme of this experimental model is presented in Fig. 3.
The electric-hydraulic propulsion system SAE is coupled to the drive axle PM
consisting mostly of central transmission and the differential TC + D, the planetary
axles and the drive wheels.
A Propulsion System for Means of Transport 495
Fig. 3. The scheme of the test bed for dynamic tests
The test bed consists of: the treads BR, the brake F and the fly-wheels V. The brake
F develops the resistant torque and the fly-wheels develop the inertia torque (it rep-
resents the reduced mass of the vehicle). In this way one can simulate the operation of
the propulsion system in conditions similar to the real ones, and with the possibility to
determine the dynamic performances. Also, the stand will be equipped with a complete
set of measurement and control devices, necessary to carry out the measurements and
the tests.
The demonstration and the validation of the operational parameters of the proposed
system will be done in laboratory conditions on this test bed for dynamic experiments.
The stand will allow the testing and the determination of the dynamic characteristics of
the propulsion system and the measurement of the following parameters: the power, the
torque, the speed, the traction force and the electric energy consumption. During the
experiments it will be determined the acceleration and the braking qualities.
4 Conclusions
The new propulsion system for non-autonomous vehicles with electric traction has the
following advantages:
a. Reduce electric consumption. The generator’s electric motor operates only in a
single regime, at maximum efficiency (the amperage of the electric current in the
windings is not variable) and without idle (in this case, the consumption of reactive
power is significantly reduced).
b. It realizes all the function necessary for the propulsion (the modification of the
torque, of the rotational speed and of the rotation sense), including the braking one,
with the recovery of the braking energy. The propulsion system, no matter it’s
complexity, is formed by only one electric motor (of high power, that have a better
efficiency than the ones of smaller power), obviously with one or more hydraulic
motors that can be direct coupled to the vehicle’s wheels.
c. Improved performances and operational safety. At the same power, the vehicles
equipped with the new propulsion system will have superior dynamic qualities,
because the maximum power of the system is instantaneously available.
Generally, the non-autonomous vehicles with electric traction are used in urban
passenger transport, where the start and the stop have a high frequency. At stop,
because the recovery of the braking energy, the hydrostatic energy in the accumulator
rises and can be used at the start for acceleration (the dynamic performances are
increased).
As mentioned in the abstract, the paper’s subject is the presentation of a theoretical
innovative concept corresponding to a technological maturity level TRL 2 (the for-
mulation of the technological concept) regarding a propulsion system for
non-autonomous electric vehicles. So no experimental results are available to validate
the technical efficiency. But taking into consideration the recovery of the braking
energy and the fact that the electric motor runs in a single regime, one can assume a
20-25 % drop in electric energy consumption. The validation of the assumption
regarding the electric energy consumption can be made only on the proposed test bed
presented in chapter 3 (TRL 3). This is a subject for a future project. Best regards,
References
1. Chioreanu, N.L.: Single regime running engine (Motor monoregim). Invention patent (Brevet
de inventie) BI 127362 B1, CBI a 2011 00903/14.09.2011, RO-BOPI 8/2015, OSIM
Bucharest (2015). www.osim.ro/publicatii/brevete/bopi_2015/bopi_inv_08_2015.pdf
2. Popovici, O., Popovici, D.: Electric Traction (Tractiune Electrica). Mediamira Publisching,
Cluj Napoca (2009). ISBN 978-973-713-230-7
3. http://www.clean-auto.com
Validation
of a Human-and-Hardware-in-the-Loop
Control Algorithm Using Real Time
Simulation
Ionut Alexandru Stoica(&), Marius Valentin Bataus,

and Ioan Mircea Oprean
Department of Automotive Engineering,

a.i.stoica@gmail.com, mvbataus@yahoo.com,
mircea_oprean@yahoo.com
Abstract. One of the key issues of modern passenger cars is the customer-
perceived performance. This paper aims to investigate the development of an
electric vehicle test platform which can be used as a human-and-hardware-in-the-
loop simulator for drivability studies, with emphasis on the validation of the
control algorithm using real time simulation. The evolution of the electric vehicle
model is presented during stages of testing of real time simulation equipment,
describing the transition of the control algorithm from variable to fixed step
simulation with emphasis on how to convert an offline model in a real-time one.
Keywords: Electric vehicle Powertrain Modeling Drivability Real time

simulation H2iL
1 Introduction
Customer-perceived performance is a key issue in modern passenger car development.

While trying to keep a delicate equilibrium between the fuel economy, emission reg-
ulations, comfort and dynamic behavior, automotive and OE manufacturers go to great
lengths in order to investigate the drivability issues. If a current or future production
vehicle requires a change with direct influence on the dynamic behavior, the modifi-
cation can be made at the cost of many prototype vehicles, over long periods of time.
Furthermore, they have no direct way of conducting a subjective investigation of a
vehicle’s dynamic behavior during the early stages of the development process using a
real driver in real traffic conditions.
Modern passenger cars have some of the most complex multidisciplinary systems.
Their development involves the comprehension of a wide variety of levels of perceptual
abilities, physical skills and technological understanding of automobile drivers, all of
these attributes increasing the difficulty of transferring standard control theories directly
to the automotive field. For example, the project of integrating a 6 speed automatic
gearbox in an A-segment vehicle involved the handling of more than 600 calibration
values in over 20 vehicle variants [3]. Present day automotive manufacturers have
498 I.A. Stoica et al.
adopted modeling and simulations in the development and tuning of the automotive
control. These techniques are becoming preferred due to lower costs and development
times, but they offer no or low subjective feedback on the dynamic behavior. Some
proposed solutions to overcome this are adjustable chassis and H2iL (human-
and-hardware-in-the-loop) simulators.
The UNIKAT vehicle from Automotive Testing Papenburg (ATP) is a universal
chassis with adjustable kinematics and adjustable masses, developed by ATP for the
investigation on the effects of kinematics, of mass distribution and centers of gravity, of
tires and brakes on vehicle behavior and comfort. The chassis is also the core of a
dedicated training program for testing engineers, chassis designer engineers and people
involved in vehicle dynamics simulations [4].
The “inDrive Simulator” project from Ingenieurgesellschaft Auto und Verkehr
(IAV) was developed as a testing platform for future cars before the first prototype is
even built. The purpose of the project is to create a simulator capable of being drive in
real-life road traffic which can be used by virtually any individual after appropriate
training. All that is required are mathematical models and a base vehicle, not neces-
sarily the target vehicle, but as similar as possible [2].
This paper aims to investigate the development of an electric vehicle test platform
which can be used as a H2iL simulator for drivability studies, with emphasis on the
validation of the control algorithm using real time simulation. The evolution of the
electric vehicle model is presented during stages of testing of real time simulation
equipment, describing the transition of the control algorithm from variable to fixed step
simulation.
The electric vehicle model is developed using the LMS Imagine.Lab AMESim
simulation platform and tested using Matlab/Simulink and AMESim for standalone
simulations and in co-simulation. The complex two dimensional model has integrated
vehicle suspension and an electric powertrain located on elastic mounts. The solver
steps are analyzed with the target of implementing the real time simulation. Before
validating the H2iL control algorithm using real time simulation platforms, the
Simulink version of the control algorithm is tested in co-simulation with the AMESim
plant model. Some details on how to convert an offline model in a real-time one are also
presented.
Using the xPC Target and dSpace real time simulation platforms, the control
algorithm is tested during a full throttle take-off maneuver, covering the launch and the
first two gear changes. This allows the analyzing of the simulator’s capability to follow
an imposed longitudinal acceleration profile.
2 Electric Vehicle Model
There is a diversity of possible electric vehicle models, given by their level of com-
plexity, which can be used for qualitative or quantitative studies of the dynamic
behavior. Considering that the paper focuses on the development of an electric vehicle
test platform based on a B-class segment vehicle, having the electric drive packaged
and supported in a similar manner with the conventional powertrain, the model to be
Validation of a Human-and-Hardware-in-the-Loop Control Algorithm 499
Fig. 1. Simplified electric vehicle model
used for the validation for the control algorithm needs to be a two dimensional model
which considers the mounting stiffness and geometry.
A simplified model was first developed in Simulink. The simplified model (Fig. 1.)
is built using dedicated submodels (subsystems) for the vehicle, the electric motor, the
inertia, the reducer and the equivalent drive shaft. This model is used for the first tests
of the control using real time simulation.
The model, developed using LMS.ImagineLab AMESim, is obtained using sub-
models from Mechanical, Powertrains and IFP Drive libraries. These models are used
for performance and energy consumption studies. It is started with a simple unidi-
mensional vehicle model that is evolved in stages to the final configuration (Fig. 2.).
In order to use the model for drivability studies, the first step is to use a two
dimensional planar model and to introduce the stiffness of the drivetrain, by separating
the motor inertia and adding a rotary spring-damper that considers the equivalent
drivetrain stiffness [9].
The inertia of the rotor is separated by adding a rotary load submodel at the motor
shaft. The equivalent drivetrain stiffness is computed with the formula:
sdt ¼ sis shs e =ðsis þ shs e Þ ð1Þ
Where sdt is the equivalent drivetrain stiffness, sis is the transmission input shaft
stiffness and shs:: e is the half-shaft’s equivalent stiffness.
The formula used to compute the equivalent stiffness of the half-shafts was deduced
taking into account the influence of the differential.
shs e ¼ 4slhs srhs =ðslhs þ srhs Þ ð2Þ
Where slhs is the left half-shaft stiffness and srhs is the right half-shaft stiffness.
The model is further developed by adding the vehicle suspension. For this, a 2D
vehicle model is necessary. The sprung mass is modeled by an infinitely rigid body. In
fact, the stiffness of the carbody is considered high enough not to impact the dynamics
of the vehicle located below 20 Hz. The rigid body is subjected to external mechanical
actions which may cause the movements along the degrees of freedom considered: roll,
pitch and yaw. It is therefore subjected to the forces coming from the vehicle sus-
pension and the restoring forces coming from the engine mounts. Furthermore, the
carbody is also subject to mechanical actions of its weight and aerodynamic drag
forces. Outgoing causalities of the body are its movements and travel speeds in the
three degrees of freedom studied; these variables are inputs to the suspension of the
Fig. 2. Two dimensional vehicle model
vehicle and the junctions for the connections of the engine mounts with the body. The
variables describing the interaction between the carbody and the suspension and engine
mounts are modeled in different frames:
• The absolute frame: it’s origin is located in the plan containing the contact points
between the wheels and the road;
• The vehicle body frame: it’s origin is located in middle of the front axle;
• The dynamic behavior frame: it’s origin is located in the center of mass.
The basic principal of dynamics is applied in the center of mass and in the absolute
frame [1]:
8
< Mv €xcar ¼ Fxs!c=f þ Fxs!c=r Fdragx þ FGMPx
Mv €zcar ¼ Fzs!c=f þ Fzs!c=r Mv g Fdragx þ FGMPZ ð3Þ
:
Jv €hcar ¼ Tfs þ Trs Tdrag þ TGMP
Where Mv is the vehicle mass, Fxs!c=f and Fxs!c=r are the longitudinal forces from
the front and rear axles, Fxz!c=f and Fxz!c=r are the vertical forces from the front and
rear axles, Fdragx;dragz are the longitudinal and vertical aerodynamic drag forces, Tdrag is
the pitch torque of the aerodynamic drag, FGMPX and FGMPZ are the longitudinal and
vertical forces from the powertrain, TGMP is the pitch torque of the powertrain, Jv is the
vehicle’s pitch inertia in G, hcar is the carbody’s pitch angle in G, x; zcar are the
coordinates of the center of mass – G, Tfs and Trs are the front and rear axles’ pitch
torque.
The unsprung masses of front and rear axle are subtracted from the vehicle mass
and introduced in suspension submodels. The suspension damping and stiffness as well
as spindle mass are modeled in two planar degrees of freedom: longitudinal and vertical
translations. Suspension forces are calculated in the carbody dynamic behavior frame.
Therefore, input displacements from the carbody and forces from the tyre which are
expressed in the absolute frame are transformed to be expressed in the carbody
dynamic behavior frame. The submodel computes suspension forces and torque related
to longitudinal, vertical directions and self-rotating axis as follows [5]:

Fx ¼ Kx ex hx ð_ex Þ þ m €ex
ð4Þ
Fz ¼ Kz ez hz ð_ez Þ þ m €ez
Where m is the spindle mass, Kx is the longitudinal suspension stiffness, Kz is the

vertical suspension stiffness, hx is the longitudinal suspension damping, hz is the
vertical suspension damping, ex is the longitudinal suspension compression and ez is
the vertical suspension compression.
At this stage, it is necessary to first bring the vehicle in equilibrium by setting the
initial preload of the suspensions. The equilibrium state is obtained by running a
simulation with 0 Nm torque request from the driver and storing the suspension
parameter values from the end of the simulation, after the model has self-stabilized
(Fig. 3.).
The final step is to support the electric drive in the engine compartment by means of
elastic mounts. The motor is now connected with the transmission through a motor
block unit that includes also his rotational inertia. This is the same submodel as that
Fig. 3. Vertical acceleration of 2D vehicle model before and after stabilization

used for conventional powertrains. It provides the powertrain 6 degrees of freedom and
ensures the coupling between powertrain mass, vehicle mass and engine mounts.
The motor block is connected to the frame of the body by three engine mounts
which are implemented in a similar manner with a conventional powertrain. Studies
have shown that the engine mounts have to meet a compromise between the various
functions they perform. Indeed, the trade-off can be presented as follows: firstly, the
engine mount must support the weight of the powertrain when the vehicle is stationary
and limit movements in a tip-in, which implies a high static stiffness. Secondly, the
engine mount should also be able to filter the vibrations to which the powertrain unit is
subjected during various driving maneuvers, which requires high levels of damping and
a low dynamic stiffness in the frequency domain impacted. The engine mounts are
modeled using Kelvin-Voigt cells. The Kelvin-Voigt model is suitable for the study as
it is both sufficiently representative of the low frequency and the implementation
process is less complex [1].
A stabilization process is necessary in order to obtain the equilibrium position (the
electric motor on the mounts and the vehicle body on the suspension). This process is
made in a similar manner as the previous step, by running a simulation with 0 Nm
torque request from the driver, keeping the preload values elastic mount values from
the end of the stabilization process (Fig. 4.). By balancing the vehicle, an almost total
reduction of the vertical acceleration is obtained.
This model was used to validate offline the control strategy and to determine some
mechanical specifications needed to make the H2iL simulator more robust [9].
Fig. 4. Movement of the 6 dof : a. unbalanced; b. balanced

3 The Control Algorithm
The control of the vehicle acceleration is based on a PID block that has a cycle input
with data measured and stored under the form of time dependent acceleration (Fig. 5.).
The signal from the source is filtered using a general transfer function block. It is linked
to the vehicle’s electronic control unit with braking and acceleration controls. Using
positive (0 to +1) and negative (-1 to 0) saturation blocks for separating the acceleration
and braking signals, it adjusts the vehicle’s speed by decreasing the error between the
model’s actual speed and the input speed from the cycle, in meters per second. The
vehicle’s velocity is obtained by integrating its longitudinal acceleration measured
using high sampling frequency. The measured signal is first filtered using a 3rd order
Butterworth low-pass filter, using a cutoff frequency of 10 Hz. It has been proven [8]
that this model is robust enough, insuring a good correlation between the imposed
longitudinal acceleration profile and the actual one.
Fig. 5. The control algorithm
4 Validation of the Control Algorithm Using Real Time

Simulation
Two simulation platforms are chosen to order to verify the robustness and the quality of
the control algorithm: an Intel Pentium III (801 MHz) that was running xPC Target
(MathWorks) and a dSpace 1006 HiL. To set the models for the platforms, a two stage
simulation scheme is used: one with a variable step and one with a fixed step. The
variable step solver continuously adjusts integration step size in order to shorten
simulation time, reducing integration step size when model states are changing rapidly,
to maintain accuracy and increasing the step size when the states changes are slow and
there is a need to avoid unnecessary steps. The fixed step solver uses the same size
throughout the simulation and is the representative algorithm that indicates whether or
not the model can be run in real time [6, 7].
The control algorithm is firstly tested on the xPC Target platform using the sim-
plified Simulink version of the vehicle model. After performing the simulation with
variable step, the performance of the control algorithm control is analyzed (Fig. 6.). It
is noticeable how the algorithm with variable step size changes the step size in order to
meet the required tolerance for error, the step size having peaks of 10−15 to end the
simulation. The peaks are present due to accumulation error resulting from the inte-
gration of speed and the inability to achieve the level model imposed acceleration. For
an experienced user, the number of discontinuities and recovery speed are useful for an
overall assessment of capacity to run the model with a fixed step. This analysis was
Fig. 6. Evolution of the integration step - Simulink
Fig. 7. Evolution of the longitudinal accelerations for the Simulink model
used to determine coarse step size that was used for simulating fixed step: the interval is
0.1 ms… 1 ms and a fixed step size of 0.2 ms is chosen.
The simplified Simulink model is tested using a 0.2 ms fixed step size and an ode1
(Euler) integration algorithm and compared to a variable step simulation. The electric
vehicle model is subjected to a take-off maneuver, followed by two gear changes
(Fig. 7.).
The performances of the control algorithm in real time are assessed using xPC
Target (MathWorks), version 3.4. Three cases were simulated based on recorded data: a
cycle consisting of two gear changes, a fast take-off and a tip-in maneuver. For each
case stability and proper functioning of the control algorithm were monitored.
The minimum, maximum and average Task Execution Times (TET) for each
maneuver have been recorded and stored (Table 1). They offer a good indication of the
control algorithms functionality.
Table 1. Task Execution Times

Average TET Minimum TET Maximum TET
[s] [s] [s]
Two gear changes 5.29110−6 4.01710−6 22910−6
Take-off 5.16210−6 3.97710−6 22910−6
Tip-in 5.26110−6 4.02410−6 22910−6
The testing of the control algorithm on the dSpace 1006 HiL is made using the
complex vehicle model, in two steps. The first step is to validate the robustness and the
stability of the algorithm using fixed step simulation; the second step is to test the
control algorithm on the HiL station.
In the case of the variable step simulation, the performance of the control algorithm
can be evaluated using the run stats block from the Simulation library. This tool is used
to analyze performance integration model because the step value indicates links
between specific events and drastic reductions of integration steps, which can slow the
simulation. By analyzing the evolution of the current integration step, it can be seen
that its size remains constant, although there are some notable reductions below the
main variation interval 0.3 … 3 ms (Fig. 8.). Given the literature [6, 7] recommending
an interval of 0.5 ms… 1 ms for the integration step, for fixed step simulations and
range of variation obtained, an integration step of 0.5 ms is chosen.
The complex AMESim model is tested using an Euler integration method with a fix
step of 0.5 ms on a 14 s long cycle, consisting of a fast launch and two gear changes.
No notable differences can be observed between the two acceleration profiles (Fig. 9.).
Fig. 8. Evolution of the integration step – AMESim
Fig. 9. Evolution of the longitudinal accelerations for the AMESim model

The validation of the control algorithm on the HiL station is made by combining the
advantages of two simulation environments: the control algorithm modeled in Simulink
has been merged with the complex vehicle model developed in AMESim. This solution
was chosen because it combines the advantages of the both modern simulation envi-
ronments. Simulink offers the advantage of mathematical modeling based on signals
that can be applied to multiple systems from different fields, as long as the equations
describing systems are known. However, Simulink is difficult to use effectively,
requiring permanent and substantial changes to improve or evolve studied systems. On
the other hand, AMESim uses a physical simulation environment in which complex
components stored in ordered libraries, can be easily interconnected to create a com-
plete model for the simulation and analysis of mechanical, electrical, hydraulic or
pneumatic systems.
The merging of the two simulation environments has been made by adding a
Simulink co-simulation block to the complex AMESim electric vehicle model by using
a specific interface icon. The new control block has one input, the electric vehicle
speed, and two outputs, the commands for acceleration and brake. In order to connect
the Simulink control with the AMESim model, an AME2SLCoSim block from the
AMESim Interfaces library has been added to the control algorithm in the Simulink
sketch. Being the correspondent of the AMESim block, it has two inputs, the accel-
eration and brake commands, and one output: the electric vehicle speed. In order to
avoid an algebraic loop, a Unit delay block has been added (Fig. 10.).
Fig. 10. Control algorithm with the cosimulation block
The validation of the control algorithm using the Simulink - AMESim cosimulation
is made in two stages, by comparing the acceleration profiles obtained using a variable
step ode45 (Dormand-Prince) integration method and the acceleration profiles obtained
using an Euler integration method with a fixed step of 0.5 ms (Fig. 11.).
The validation of the control algorithm on the dSpace 1006 HiL is made in the
same configuration; the only difference is the replacing of the cosimulation block with
of an S-function block. The configuration of the S-function parameters is set so that it
allow the model to be subjected to a fixed step simulation. After the model is compiled
into and *.sdf file, it is uploaded on the HiL station and it is subjected to the complex
maneuver. A ControlDesk custom built layout is used as interface. The results are
positive with no overruns and good correlation with the offline simulation.
Fig. 11. Evolution of the longitudinal accelerations for the cosimulation method
5 Conclusions
It has been proven that a fixed step of 0.5 ms can be successfully used to run the
algorithm on two real time simulation platforms (xPC Target and dSpace) while per-
forming typical maneuvers used for vehicle’s comfort and dynamic behavior assess-
ment (a full throttle take-off, covering the launch and the first two gear changes).
Running the complex electric vehicle model on the real time simulation platform
has showed that the electric vehicle is capable of following an imposed acceleration and
the solution can be advanced to further development (HiL and H2iL testing). This also
demonstrate that is possible to use similar complex models of the developed
vehicle-powertrain configurations to pilot in real time the H2iL simulator by generating
the acceleration signal in accordance with the driver demands.
Acknowledgements. This project was supported by POSDRU/107/1.5/S/76903 contract num-

ber 817/01.03.2011.
References
1. Eller, B.: Contribution a l’Etude Amont de l’Agrement de Conduite de Systemes de
Propulsion de Nouvelle Generation, These de doctorate, Nantes (2011)
2. Jaensch, D.: Driving a Car That Does Not Even Exist Yet, Automotion, 02nd edn, p. 4.
Berlin, Germany (2008)
3. Kirschstein, S., Remelhe, F., Stolze, B.: Efficient transmission application by using modern
offline tools. In: Proceedings of 8th CTI International Symposium of Innovative Automotive
Transmission, Berlin (2009)
4. Knobloch W.: Vehicles Dynamics Seminars, Automotive Testing Technology International,
p. 506. Stuttgart, Germany, May 2007
5. LMS. ImagineLab AMESim Powertrain Library Rev 12
6. Miller, S., Wendlandt, J.: Real-Time Simulation of Physical Systems Using Simscape,
MATLAB News and Notes, January 2010
7. Schuette, H., Waeltermann, P.: Hardware-in-the-Loop Testing of Vehicle Dynamics
Controllers – A Technical Survey, SAE paper no. 2005-01-1660, SAE 2005 World Congress
& Exhibition, Detroit, USA, April 2005
8. Stoica, I., Bataus, M.V., Negrus, M.: Validation of a human-and-hardware-in-the-loop control
algorithm. U.P.B. Sci. Bull. Ser. D 76(4), 69–78 (2014)
9. Stoica, I., Bataus, M.V., Oprean, I.M.: Modeling of electric vehicles for driveability control
applications. The European Automotive Congress 2015 - EAEC-ESFA. Springer, Heidelberg
(2015)
Electronic Control Systems of E-Smart Vehicle
Radu Tarulescu(&) and Stelian Tarulescu

{radu.tarulescu,s.tarulescu}@unitbv.ro
Abstract. In this paper the authors present the electronic control systems used
to manage the driving performances and optimize the energy supply for E-Smart
electric vehicle. E-Smart is an electric vehicle obtained through conversion, of a
Smart ForTwo City vehicle, from the internal combustion propulsion system to
a system that uses a three-phase asynchronous motor. The requirement of the
electric vehicle entire control system is to have a controller able to accomplish
the submission tasks from the driver.
Keywords: Electric vehicle Curtis controller Sensor
1 Introduction
Referring to electric vehicles, the global automotive industry is undergoing a significant

transformation. Electric vehicles are an emerging technology that manufacturers are
increasingly turning towards. Increasingly more vehicle manufacturers are at various
different stages of electric vehicle development and commitment. The environmental
benefit of electric vehicles over fossil fuel powered vehicles is that they don’t produce
any polluting emissions.
Driving an electric vehicle produces no carbon dioxide or any other environmen-
tally damaging emissions at the source [1]. The average price of an electric vehicle is
higher than a conventional fossil fuel powered vehicle, because of the high production
costs of the electric vehicle’s battery, but savings are made from electricity prices being
cheaper than gasoline or Diesel fuels [2].
In this article will be presented E-Smart electric vehicle’s electronic systems.
E-Smart (Fig. 1) is a vehicle obtained by converting a Smart ForTwo City Coupe
Fig. 1. From 1999 Smart ForTwo City Coupe to 2016 E-Smart electric vehicle

510 R. Tarulescu and S. Tarulescu
vehicle manufactured in 1999 and fueled by gasoline, with an electric propulsion

system. The old gasoline engine of 33 kW power was replaced with an electric
three-phase asynchronous motor of 22 kW power, model N50D2.
2 E-Smart Vehicle Electronic Devices
2.1 Electronic Control Module

The engine control unit (ECU) of the old gasoline engine was replaced with an elec-
tronic control unit E-Car-Box based on a Curtis 1238 controller (Fig. 2). Al of the
commands from the vehicle’s driver are processed in this controller and given to the
electric motor.
Fig. 2. E-Car-Box electronic control unit and Curtis 1238 controller
Curtis 1238 controller has advanced pulse width modulation technology for effi-
cient use of battery voltage, low motor harmonics, low torque ripple, and minimized
switching losses.
Fig. 3. The 1314 PC programming station and software

Electronic Control Systems of E-Smart Vehicle 511
Also, power limiting maps allow performance customization for reduced motor
heating and consistent performance over varying battery state of charge. Powerful
operating system of this controller allows parallel processing of vehicle control tasks
and motor control tasks. Curtis 1238 controller is easily programmable through the
1314 PC programming station (Fig. 3).
Fig. 4. Curtis 840 display for main parameters
Curtis 840 display (Fig. 4) show all important information from the Curtis con-
troller. It can show motor power and temperature, battery status, motor speed in rpm,
motor torque in Nm, vehicle speed in km/h and total voltage at the controller.
Fig. 5. Emus Evgui application for smart phone
That parameters can be displayed also on the smart phone using an android free
application named Emus Evgui (Fig. 5). In addition, it can display more information
about the state of battery, discharge voltages of individual battery cells and the highest,
lower and medium cell voltage. Curtis 1238 controller communicate with mobile phone
through a Bluetooth device.
2.2 System Electronic Interface (SEI)

The System Electronic Interface (SEI) is an interface between E-Smart vehicle and
Curtis controller. SEI (Fig. 6) is a board which generate a signal of type “charger on”
and a signal of type “neutral/drive” for the Curtis 1238 controller.
Fig. 6. Inputs and outputs signals for System Electronic Interface (SEI)
The SEI board gives to the Curtis controller a signal that there is a power of 230 V
available because the charging cable is plugged into the vehicle. Having this infor-
mation, the Curtis controller inhibits drive mode, showing “Charger” on the Curtis 840
display and the driver is not able to drive away with the charging cable plugged. This is
the safety function of the SEI board. The other function is to enable vehicle’s driving,
being generate a signal “neutral/drive” to the Curtis controller. If the signal is off, the
Curtis controller is in the neutral state and the main contactor being not pulled. The
power bank of the controller is not powered and the motor is stopped. If the signal is
on, the main contactor pulls and the Curtis controller is ready to drive the motor.
2.3 Battery Management System

The battery management system of EMUS type (Fig. 7) is used for batteries with more
cells (like the battery cells mounted on the E-Smart vehicle) connected in series to form
a higher voltage battery pack.
Fig. 7. The battery management system EMUS

Each cell of the battery has an electronic module which monitors cell’s parameters,
controls the balancing current and communicates with central control unit of the sys-
tem. The control unit of EMUS battery management system is connected to cells
electronic modules via double twisted-pair optically isolated interface over which it
communicates and collects cell characteristic information and manages the battery [3].
The control unit of EMUS also has a USB interface for configuration, which is used to
connect it to a PC to configure the battery management system via specific free
downloadable software.
3 E-Smart Vehicle Functionality
3.1 E-Smart Driving Modes

E-Smart vehicle works just like any full electric vehicle. The electric current is drawn
from the batteries pack and applied, under the management of the electronic control
module E-Car-Box, to the electric motor. Inside the motor, this current is passed
through a magnetic field, which results in a torque that is used to turn the vehicle’s
wheels [4].
Fig. 8. The switches for enable and disable the regenerative or economic driving mode
This process can be reversed during braking of the vehicle. Effectively, the elec-
tronic control module converts the motor to a generator and the regenerative braking
system can recapture kinetic energy and store it as electrical energy [5].
The regenerative driving mode can be enabled by switching the red button in 1
position (Fig. 8). Also, the regenerative driving mode can be disabled by switching the
red button in 0 position. Another driving mode can be selected, the economic driving
mode. In this mode the battery pack discharge is limited even if the accelerator pedal is
fully pressed to save the energy and the vehicle speed is limited to 36 km/h. The
economic driving mode can be enabled or disabled by switching the green button
(Fig. 8).
3.2 E-Smart Motor Parameters

Using Curtis 840 display and Emus Evgui application were determined power, torque
and motor speed of E-Smart vehicle. Measurements were made for both driving modes,
regenerative and economic from vehicle starting, on straight road without slope
(Fig. 9).
Fig. 9. The chosen road for tests
Table 1. Measured parameters for the economic and regenerative driving modes.
Motor speed [rpm] Torque [Nm] Power [kW]
Economic Regenerative Economic Regenerative Economic Regenerative
drive drive drive drive drive drive
35 257 18 62 3.18 5.34
297 636 66 112 4.27 18.02
702 1214 68 93 10.34 17.58
907 1722 69 70 12.07 17.42
1257 1909 63 60 10.91 17.36
1451 2235 47 51 10.68 17.34
1672 2489 39 45 10.52 16.91
1798 2645 37 40 10.43 16.23
1880 2773 35 37 9.96 15.37
1929 2889 32 33 9.45 13.42
2154 2993 31 29 6.36 13.15
2224 3065 28 28 6.22 13.08
2310 3123 26 27 5.92 13.03
2375 3220 24 26 5.68 13.03
2415 3315 23 26 5.57 13.03
2472 3386 22 25 5.41 13.01
2522 3460 19 24 5.38 12.76
2565 3484 16 23 5.24 12.29
In Table 1 are displayed the consumed power and torque values depending on the
motor speed for the both driving modes.
It should be mentioned that the E-Smart vehicle’s acceleration pedal was fully
pressed for all the road in both cases.
In Figs. 10 and 11 are showed the diagrams of power and torque in function with
motor speed for the boat driving modes, economic and regenerative.
Fig. 10. The power and torque diagram for economic driving mode
Fig. 11. The power and torque diagram for regenerative driving mode
The values obtained for the economic driving mode are lower because the limiting
controlled from the E-Car-Box electronic control module. Instead, the traveled distance
using this economic mode was nearly 15 km longer than the normal (approximatively
65 km starting with batteries fully loaded). With the regenerative mode the distance
traveled with batteries pack fully loaded was approximatively 50 km, but with a top
speed of 80 km/h.
4 Conclusions
Like all electric vehicles, the good functioning of E-Smart depends solely on the correct
functioning of electronic control systems. The main role in this case is held by the
controller Curtis 1238 controller, which is actually the “brain” of the entire system.
Actually E-Smart vehicle, on operating mode, is very much like mobile robots [6]. This
paper presents an overview of E-Smart, an old vehicle, pollutant, transformed into a
modern vehicle, friendly with the environment, which should be implemented in a
large city like Brasov.
References
1. Leitman, S., Brant, B.: Build Your Own Electric Vehicle, 3rd edn. McGraw-Hill Education
Publishing, New York (2009)
2. Larminie, J., Lowry, J.: Electric Vehicle Technology Explained, 2nd edn. John Wiley & Sons,
Ltd., Hoboken (2012)
3. EMUS Battery management system, User’s manual, JSC Elektromotus (2011)
4. Soylu, S.: Electric Vehicles - the Benefits and Barriers. InTech Publishing (2011)
5. Stevic, Z.: New Generation of Electric Vehicles. InTech Publishing (2012)
6. Tarulescu, R.: Optimizarea Configuratiei Senzorilor Utilizati la Robotii Mobili. GlobeEdit
Publishing (2016)
Battery Management System of E-Smart
Vehicle
Radu Tarulescu(&) and Stelian Tarulescu
Transilvania University of Brasov,

Brasov, Romania
{radu.tarulescu,s.tarulescu}@unitbv.ro
Abstract. In this paper the authors present the battery management system
(BMS) used for monitoring, evaluation, charge control, and battery cells bal-
ancing of E-Smart electric vehicle. E-Smart is an electric vehicle obtained
through conversion, of a Smart ForTwo City vehicle, from the internal com-
bustion propulsion system to a system that uses a three-phase asynchronous
motor. The purpose of the battery management system is to guarantee safe and
reliable operation for all battery cells used to supply the engine electric energy.
Keywords: Electric vehicle Battery LiFePO4
1 Introduction
E-Smart (Fig. 1) is a vehicle obtained by converting a Smart ForTwo City Coupe

vehicle manufactured in 1999 and fueled by gasoline, with an electric propulsion
system. The old gasoline engine was replaced with an electric three-phase asyn-
chronous motor power supplied from a pack of 32 batteries of LiFePO4 type. A simple
failure in a battery management system can lead to battery damage or safety issues for
the driver. Therefore advanced battery technologies require a Battery Management
Systems (BMS) to ensure safety and long life performance of the battery packs [1].
One of the BMS functions is to protect the battery cells, responding to external fault
conditions by isolating the any cell as well as addressing the cause of the fault. Another
function of the BMS is to determine the State of Charge (SOC) of the batteries for
providing the fuel gauge indication [2]. The BMS monitors and calculates the SOC of
each individual battery cell to check for uniform charge in all of the cells in order to
verify that individual cells do not become overstressed. BMS also monitors the con-
ditions of individual battery cells, maintain all the cells within their operating limits,
sets the battery operating point to allow regenerative braking charges to be absorbed
without overcharging and provide information for driver displays and alarms predict
the range possible with the remaining charge in the battery [3].

Fig. 1. E-Smart electric vehicle
2 E-Smart Vehicle Battery Management Components
2.1 Batteries
To supply with energy, the electric motor has been adopted a solution using LiFePO4
(Lithium Iron Phosphate) rechargeable batteries. LiFePO4 batteries have more constant
discharge voltage and are considered to offer better safety than other batteries [4]. Also
that type of batteries have the ability of a much faster recharge and higher discharge
rates than other and usually higher number of recharge cycles (above 2000), meaning
longer life when are not fully discharged. LiFePO4 batteries have the operating voltage
between 3.2 V and 3.3 V, the minimum discharge voltage being 2.8 V and the max-
imum charged voltage 3.6 V.
Fig. 2. The battery pack from E-Smart vehicle with 32 LiFePO4 cells
For E-Smart, the chosen LiFePO4 batteries model was SP-LFP-66AHA with
nominal capacity of 66Ah, max charging current of 132A (optimal 22A) and max
discharge current of 198A (optimal 22A). There was placed 32 batteries in the pack
(Fig. 2).
Battery Management System of E-Smart Vehicle 519
2.2 Batteries Charger

For charging batteries in optimal conditions, on the E-Smart was installed a charger of
Elcon TC type (Fig. 3), specific for the LiFePO4 type of battery cells. That charger
have a wide input voltage range, between 85 V (AC) and 265 V (AC), being conve-
nient for electric networks fluctuation. Elcon TC charger is also fully sealed and
waterproof, having an IP46 protection class. The charger have thermal self-protection,
so when its internal temperature exceeds 75°C the charging current will reduce auto-
matically. If exceeds 85°C, the charger will shut down, and when the internal tem-
perature drops, it will resume charging automatically.
Fig. 3. The Elcon TC battery charger mounted on the E-Smart vehicle
2.3 Battery Management System

For the batteries pack of E-Smart vehicle is used a battery management system of EMUS
type (Fig. 4). The control unit of EMUS BMS is connected to cells electronic modules
via double twisted-pair optically isolated interface over which it communicates and
collects cell characteristic information and manages the battery [5]. The control unit of
Fig. 4. The battery management system EMUS

EMUS also has a USB interface for configuration, which is used to connect it to a PC to
configure the battery management system via specific free downloadable software.
Each cell of the battery has an electronic module which monitors cell’s parameters,
controls the balancing current and communicates with central control unit of the sys-
tem. There are four types of the EMUS battery management system cell modules to
allow the installation on the battery pack: a Top type module, a Bottom type module,
15 of A and 15 of B module types. The type of the module is printed on the each
module. Bottom cell module is dedicated to be installed on the cell which has negative
(−) terminal of the whole battery pack [5]. Top module is installed on the cell which
has positive (+) terminal of the pack. Type A and B modules are installed on the cells in
between (Fig. 5).
Fig. 5. The batteries electronic module types and connections
With these modules, battery management system can balance each cell voltage
during charging, and during vehicle’s running. Also, the management system measures
and calibrates temperatures of each cell, each of EMUS cell modules having temper-
ature probes (Figs. 6, 8 and 10).
Fig. 6. The electronic modules mounted on the E-Smart vehicle batteries

2.4 Parameters Display System

The batteries parameters will be displayed on the smart phone using an android free
application named Emus Evgui (Fig. 7). Also, that application can display more
information about the state of battery pack, discharge voltages of individual battery
cells and the highest, lower and medium cell voltage.
The electronic control module of E-Smart communicate with mobile phone through
a Bluetooth device connected to the EMUS battery management system.
Fig. 7. Displayed parameters with Emus Evgui application
3 E-Smart Battery Management Functionality

3.1 E-Smart Test Procedures
For tracking the battery discharge process of E-Smart vehicle it was chosen a route
from Brasov city because here can be found the majority of traffic conditions. On
departure the batteries were fully charged, the road being traveled by the driver without
any passengers [6] (Fig. 8).
During the trip have been used the both driving modes available to the E-Smart
vehicle, the regenerative driving mode and the economic driving mode.
3.2 E-Smart Batteries Pack Parameters

Using the Emus Evgui application were determined the minimum, maximum and
average values for batteries voltage and temperature at every five kms traveled.
Measurements were made with the vehicle at a standstill state, the values being free
of influence by the mode of accelerator pedal action or other factors. The average value
of the ambient temperature was 12°C. In Table 1 are displayed the measured values.
The values obtained for voltage are common, LiFePO4 battery discharge mode
being similar to what is seen in the chart (Fig. 9). Maximum discharge have place in
Fig. 8. Brasov city locations from the chosen route for testing
Table 1. Measured parameters for the E-Smart battery cells

Distance traveled [km] Voltage [V] Temperature [oC]
Max Min Average Max Min Average
0 3.44 3.31 3.41 14 11 13
5 3.30 3.28 3.28 22 15 18
10 3.28 3.26 3.27 24 16 19
15 3.27 3.25 3.26 24 17 19
20 3.28 3.26 3.26 24 17 19
25 3.28 3.26 3.26 25 19 21
30 3.26 3.25 3.26 26 22 22
35 3.26 3.24 3.25 29 21 23
40 3.25 3.23 3.24 28 22 24
45 3.25 3.23 3.24 30 22 25
50 3.23 3.20 3.21 35 24 28
55 3.21 3.18 3.19 34 24 28
60 3.18 3.12 3.03 35 25 29
65 3.10 2.77 2.96 33 24 28
the first and last traveled kilometers. Considering that the electric motor is used by
E-Smart vehicle is air-cooled, its temperature has not increased excessively the entire
distance traveled, being under 30°C (Fig. 10).
Fig. 9. The average values for batteries voltage and temperature
Fig. 10. E-Smart vehicle parked during measurements
4 Conclusions
The good functioning of E-Smart batteries pack depends on the correct functioning of
battery management system. The main role in this case is EMUS management battery
system, without that the charging, functioning and discharging of battery calls would
be difficult to achieve. This paper presents a view of the E-Smart vehicle in terms of
energy management because on full electric vehicles the energetic autonomy it’s the
most important thing.
References
1. Lyman, E.W., Tredeau, F.P.: Battery Evaluation and Battery Management System, Proquest.
Umi Dissertation Publishing (2011)
2. Rahn, C.D., Wang, C.Y.: Battery Systems Engineering, 1st edn. John Wiley & Sons, Ltd.,
Hoboken (2013)
3. Shekhawat, D., Spivey, J.J., Berry, D.: Fuel Cells: Technologies for Fuel Processing, 1st edn.
Elsevier Science Publishing, Amsterdam (2011)
4. Scrosati, B., Garche, J., Tillmetz, W.: Advances in Battery Technologies for Electric Vehicles.
Woodhead Publishing Series in Energy, 1st edn. (2015)
5. EMUS Battery management system, User’s manual, JSC Elektromotus (2011)
6. Tarulescu, R.: Optimizarea Configuratiei Senzorilor Utilizati la Robotii Mobili. GlobeEdit
Publishing (2016)
New Materials, Manufacturing
Technologies and Logistics
Worldwide Diversity Management
of “Global Access” Renault/Dacia’s Vehicle
Range by Using a Modern PLM System
Bogdan Bădescu and Pascal Candau(&)
Renault Technologie Roumanie, Bucharest, Romania

{bogdan.badescu,pascal.candau}@renault.com
Abstract. The “Global Access” Range of Dacia/Renault has to handle a huge

diversity in order to manage a global production for our customers worldwide.
It’s about conceiving 16 vehicles in 4 interconnected R&D centers and
manufacturing in 12 plants on four continents in the context of multi-sourcing.
This is possible by using a strong and versatile Product Lifecycle Manage-
ment (PLM) system which permits a real-time coordination for all R&D oper-
ations involving updating, sharing, transferring and archiving of all type of
current data (3D, Working Breakdown Structure/WBS or standard documents).
The paper shows briefly the main characteristics of the Renault/Dacia’s PLM
System and why this solution is optimum.
Keywords: Renault Diversity Product lifecycle management Product data

management PLM PDM
1 Context
Globalization imposes carmakers to search for alternatives to consolidate their plat-

forms and then reducing the cost of designing and manufacturing cars for the global
markets i.e. to implement standardized components and manufacturing processes that
can be efficiently deployed across global locations.
Carmakers are also coping with an increasing competitive pressure to reduce their
development time (time-to-market). This is possible only if the conception convergence
time and rework time are drastically reduced by fast digital process feasibility and
validations.
2 Complexity of “Global Access” Range of Renault/Dacia
2.1 About Renault Technologie Roumanie (RTR)

RTR is the only complete automotive engineering center in Eastern Europe and the
Mediterranean region and the largest Renault engineering center outside France, with
approximately 2300 engineers.

528 B. Bădescu and P. Candau
Part of the engineering network of Renault, RTR’s main areas of expertise is the
conception and improvement of vehicles, as well as adapting engines and gearboxes for
the “Global Access” range models.
For all its products, RTR is also in charge to develop the related manufacturing
technologies for the vehicles and powertrain plants.
In its three main locations, RTR brings together all the activities needed in the
development of an automotive project:
• Bucharest: the engineering offices develop and adapt vehicle projects based on the
Global Access platform and mechanical projects for this range.
• Titu: the technical center performs tests for the vehicles and the powertrain parts
developed by the engineering offices.
• Mioveni: RTR provides the technical support to the Dacia plant and to its
suppliers.
RTR represent the engineering hub for all vehicles developed for the “Global
Access” Range which represents one third of vehicles sold worldwide by Renault.
Fig. 1. RTR: a first-order key-role in Renault’s global engineering (Source: the authors)
The Renault’s Engineering Centers

• TCR - Technocentre (France), global engineering headquarter
• RTR - Renault Technologie Roumanie (entities in Romania, Turkey and Morocco)
• RTA - Renault Technologie Americas (entities in Brazil, Argentina and Colombia)
• RTS - Renault Technologie Spain
• RTK - Renault Samsung Technical Center in South Korea
Worldwide Diversity Management of “Global Access” 529
In line with its ambition to increase sales outside Europe, the Renault Group designs
and develops its vehicles at a global level. Local development of vehicles ensures they
respond closely to market requirements, while advance planning of necessary modi-
fications allows them to be marketed internationally.
2.2 Managing Diversity

Cultural and Geographical Diversity
In the present context when the process of conception and production goes invariably at
a global scale, an effective and mandatory cooperation involves a close interaction
between engineers and technicians from different places and cultures with diverse
scales of values. (See Fig. 1.)
A powerful PLM system and a well standardized interface between projects’
contributors with well-defined rules and deliverables could avoid the most of misun-
derstandings (cultural and technical).
Besides the collaboration inter-engineering (between R&D Centers) there is the
interaction with the world of manufacturers. Each R&D Center has to work together
with the plants in the launching phase from the first milestone which involves the plant
until the Manufacturing Approval at least.
Handling the Diversity Given by the Mix between Product Diversity/

Manufacturing Sites/Coexistence of Vehicles Generations
The Global Access Range has to manage a multitude of vehicles from the same
generation but also in the same time with some from old generations (as a result of
certain local commercial strategy).
For instance: at present moment, Logan 1st generation is still produced and sold in
Iran, Russia or South Africa.
Simultaneously, Logan 2nd generation is sold widely over the world. For the
management of their serial life these products are obviously present in the
Renault PLM both in Bill Of Material (BOM) and 3D database.
In the same time, we are deeply involved in designing the facelift of the 2nd
generation and in parallel we have just started the development of Logan 3rd
generation.
Consequently, all the references in BOM/3D database which cover the versions of
Logan mentioned above have to coexist in a full coherence and ready to be updated
from different stages in order to fit the market demand and the regulation changes in an
easy and fast manner (Fig. 2).
Handling of 3D Models Diversity
In order to build the complete digital mock-up an important quantity of 3D parts are
inherited from other existing vehicles from same or different range
(carry-over/carry-across parts). Concerning the specific parts, some 3D designs are
created by suppliers and a significant quantity is made “in house” by RTR engineering
teams.
An example of the circuit of 3D information for Global Access Range “results of
crash numerical simulation”:
Fig. 2. Worldwide diversity bodies/plants for Global Access Range (Source: internal Renault
document)
Stage 1: All the parts are collected in RTR by the Digital Mock-up manager who is
responsible that all 3D arrive in time and at the expected level of maturity with respect
to the project milestones. Deliverable: complete vehicle digital mock-up. Tool:
NewPDM
Stage 2: The digital mock-up is analyzed by architects and corrected by designers,
every on his specialty. Deliverable: validated digital mock-up for a specific milestone.
Tool: NewPDM
Stage 3: The digital mock-up is made available in PDM system for RNTBCI (India
R&D Center) for model meshing and assembling. Deliverable: FE model available in
PDM system. Tool: NewPDM
Stage 4: Starting from model, the CAE engineers of RTR performs all crash simu-
lations and present to the conception project engineers the results and the correction
proposal after one or several iterations. Deliverable: Crash numerical simulation
report. Tool: WAC/DocING
Stage 5: The digital mock-up is corrected to fulfill the milestone parameters. Deliv-
erables: Digital Mock-up validated accordingly with the milestone requirements. Tool:
NewPDM
Off-Stage: All conclusions/corrections reports are available for the related further
projects (in RTR) or for the further re-industrializations of same project carried on by
others R&D Centers (other RTx). Tool: NPDM/WAC/DocING
3 The Renault’s Modern PLM Platform

3.1 Why PLM?
PLM (Product Lifecycle Management) is an enterprise tool which allow an effective
management of entire amount of digital product data (3D models, WBS or standard
documents) related to whole design activity that accompanies the product to the
beginning of the upstream phase onto its end of life (“from cradle to grave”- concept).
In order to achieve the goal of “Manufacture the good product and manufacture the
product good/well”, in the industrial practice it is translated as deliver the precise
information to the right person at the right time, this is the challenge of a PLM solution.
In industry, the Product Lifecycle Management (PLM) is the process of managing
the entire lifecycle of a product from inception, through engineering design and
manufacture, to service and disposal of manufactured products [1, 2]. PLM integrates
people, data, processes and business systems and provides a product information
backbone for companies and their extended enterprise [3].
3.2 Choosing the Right and Tailored PLM Platform

In June 2009, Renault has selected Dassault Systèmes’ V6 PLM as its new global
product development solution, in order to reduce the time-to-market (development
time), standardizing the components and improve productivity and product quality for
its global automotive engineering processes to support an internal program called
NewPDM. Renault is a pioneer and the largest user of the version V6 of Catia in the
automotive world. Today (July 2016) there are more than 10.000 Renault users which
means five times more than next company (carmaker or automotive supplier).
From more than 7 years ago, Renault acted as a main partner of Dassault Systèmes
for defining the technical specification from client’s point of view in searching as a
trendsetter the PLM of near future.
During the weekly sessions, Renault has also its own agenda with its specific
requirements. The volume of changes required by Renault during the last 7 years is
around 10.000.
As a result, to fit the Renault’s needs, the V6 PLM platform “powered by Dassault”
has a level of exclusive customization is around 10 % (Catia & Enovia).
Since 2014, an initiative of members of the governance of PFA France (Plateforme
Filière Automobile) was launched a program called VALdriv “Value driven Product
Lifecycle Management” lead by Renault and PSA Peugeot Citroen. This research
group includes French Automotive OEM’s and suppliers like RENAULT, PSA
Peugeot-Citroen, Michelin, Valéo, Plastic Omnium and others with the declared stake
to improve “the added value by the digital tools throughout the automotive product
lifecycle” [5].
In the present, Renault has advanced in the deployment of the Catia/ENOVIA
V6-based collaborative platform in all geographies and throughout all brands to enable
the company to collaborate on the creation of new product designs in real time.
The online access to the digital mock-up (DMU) will lead to a simplification of the
collaboration between engineering sites. This use of a unique, collaborative interface
for all developers’ worldwide supports simultaneous product/process (assembly)
engineering to get it ‘right the first time’.
Renault’s original engineering processes use now a unified collaborative platform
using a single, standardized data model, and solutions that are deployable for all
engineering divisions. At the heart of Renault’s strategy to transform its product
development is the V6 platform’s virtualization of the entire product lifecycle within a
truly collaborative environment that is both flexible and precise. Other requirements
were the need to reinforce collaboration with extended enterprise partners and suppliers
and to ensure iso-performance and data consistency between globally dispersed teams
(Fig. 3).
Fig. 3. The New PDM’s big modules (Source: internal Renault document)
3.3 Advantages – Why DS’ V6?

In Dassault Systèmes’ V6 solution, Renault found an integrated and collaborative PLM
environment allowing more operational transparency and the possibility to validate
scenarios through virtual simulation and online management of the DMU. This will
significantly ease decision making throughout the stakeholder community and across
the entire product lifecycle of Renault automobiles, from conception through design,
compliance, simulation, and manufacturing.
Related to Renault’s complex needs, the chosen PLM solution is optimum versus
other “best in class” PLM solutions.
For instance, in comparison with his main competitor the NewPDM of Renault
offers:
• A tolerance to latency which allow an efficient work between an R&D Center and
its oversea remote site. In fact, Dassault is the solely to offer an integrated 3D
collaboration solution, who takes into account latency and security of each of the
end-users. This solution sends through the network only the position matrix, while
the alternative solutions based on screen sharing/remote monitor are relying on

video streaming that is much heavier.
• Solution providing a full coherence between CAD and DMU (since both of them
are fully integrated)
• Capacity to manage the dynamic links between a part reference in the Working
Breakdown Structure of BOM system which is replicated in NewPDM (logical tree)
and its unique correspondent in the product structure (3D physical tree). In the same
time, the reference is linked at a container with the related documents of the part
(product requirements, product specs, material characteristics, etc.).
“In the coming years, we will develop collaborative engineering. The partnership
with Dassault Systèmes will contribute to this strategy. We chose the V6 platform
because its integrated PLM environment brings our global teams together through
real-time collaboration and online-enabled design anywhere. We have chosen to deploy
the full V6 solutions portfolio, simply because it offers a remarkable collaborative,
scalable, open and easy-to-use solution.” said Odile Desforges, EVP engineering and
quality, Renault at that time [4].
“Dassault Systèmes and Renault share the same vision of sustainable innovation
through global collaborative product development. The automotive industry needs to
consider economic and social aspects more than ever to make sure they create the
vehicle that perfectly meets the customer needs and values. All stakeholders must be
actors of the process of sustainable innovation”, explains Bernard Charles, president
and CEO, Dassault Systèmes [4].
4 Case Study: The Worldwide Administration/Tracking

of a BIW Common Part: Rear Skirt (End-Panel) of Sandero
The further example concerns a common part of Sandero: rear skirt (or rear end-panel)
which consists in a simple sheetmetal part or as an assembly if two nuts are added by
welding.
The difference seems insignificant at a first look, but could be a great deal for the
supply chain if the oversea transport is made as simple part instead of an assembly. The
stack of simple sheetmetal part is obviously much more compact and therefore it
generates a significant lower supply cost by unit (it is well known that any logistician
hates “to transport air”).
Our example shows the main possibilities of sourcing of the same part in respect
with the geographical position of the Renault Plant acting as customer of the part:
Case 1: Dacia Pitesti Plant (Romania) produces “in house” the assembly part for using
it on its own assembly line
Case 2: Dacia Pitesti Plant produces as a supplier the assembly part and sends it to the
Renault/Nissan/Lada plant in Togliatti (Russia)
Nb Sites 15
Nb Bodies 16
MOSCOU
Nb Réf. KPI 1 148 BL90 H79 IZHEVSK
Parteners IL 56% L52
TOGLIATTI
PITESTI RF90 LB1A
VALLADOLID H79 BLK52 BL52 FK67 BMH
TANGER IL 68%
J92 FK67 B52 BURSA
IL 80% TEHERAN
OUED TLELAT L90 BU90
CASABLANCA
L52 IL 90%
BL52
IL 15% NASHIK
L90
CHENNAÏ
ENVIGADO H79 J92
BL90 H79 BL52 IL 26%
IL 3%
CURITIBA ROSSLYN
CORDOBA BL90 H79 BL52 U79 BU90
L52 IL 72% IL 12%
Fig. 4. Complexity of Supply Chain for the BIW parts in Global Access Range (Source: internal
Renault document)
Case 3: Our supplier in Tangier produces the assembly part and sends it to the Renault
Plant in Tangier (Morocco)
Case 4: Our Supplier in Tangier produces the assembly part and sends it to the Renault
Plant in Casablanca (Morocco)
Case 5: Our supplier in Tangier produces the simple sheetmetal part and sends to
Renault Plant in Curitiba (Brazil) which makes the assembly and uses it on its line
Case 6: Our Supplier in Tangier produces the simple sheetmetal part and sends to a
local supplier in Argentina, which makes the assembly and sends it to the Renault plant
in Cordoba (Argentina) (Fig. 5).
For all this cases, the BOM system (with its working breakdown structure included
in PLM platform) has to identify every part in respect with the couple: supplier/
plant-client, but also with his affiliation at the same geometry and material (inter-
changeable part).
More than that, the number of references are multiplied by the fact of having two
different class of materials (common and galvanized steel) each with several properties.
Also, the PLM platform provides a link onto the 3D/2D of the part, the sheet of
material properties.
There is a functional/bidirectional link with the Enterprise Resource Planning
(ERP) systems of the company (for ordering/supplying etc.)
Fig. 5. Complexity of Supply Chain for the BIW parts in Global Access Range (Source: the
authors)
Renault’s best-in-class solution for Product Lifecycle Management system helps to

bring together people, processes, content and systems to achieve a compelling com-
petitive advantage by substantially accelerating innovation, time-to-market and revenue
generation by collaboratively developing, building and managing products.
References
1. Kurkin, O., Januška, M.: Product life cycle in digital factory. In: Knowledge Management and
Innovation: A Business Competitive Edge Perspective. Cairo: International Business Informa-
tion Management Association (IBIMA), pp. 1881–1886 (2010). ISBN 978-0-9821489-4-5
2. “About PLM”. CIMdata. Accessed 25 Feb 2012
3. “What is PLM?”. PLM Technology Guide. Accessed 25 Feb 2012
4. “Le programme VALdriv PLM” - “Value driven Product Lifecycle Management”. Accessed
22 Jan 2014
5. Renault chooses Dassault Systèmes Full V6 PLM to Improve the Company’s Productivity and
Product Quality. Accessed 29 Jun 2009
Potential Capabilities of Shape Memory Driven
Automotive Devices
Joachim Strittmatter1,2(&), Paul Gümpel1,2, and Oskar Lattner1

1
Hochschule Konstanz, University of Applied Sciences, Constance, Germany
{joachim.strittmatter,paul.guempel,
oskar.lattner}@htwg-konstanz.de
2
WITg, Tägerwilen, Switzerland
{j.strittmatter,p.guempel}@witg.ch
Abstract. In the automotive industry a strong effort has been undertaken to

reduce the weight of modern vehicles. In order to reduce the energy con-
sumption and to improve the environmental sustainability, the importance of
weight reduction activities is even growing faster. As lightweight designing is
becoming more and more expensive and show less potential savings, new
approaches are needed. One promising technology could be the use of shape
memory elements. In the last years a lot of potential application possibilities
were presented, demonstrating the benefit of these functional elements in
automotive design solutions: they often reduce complexity, weight and design
space of an actuation device and enable new functions. In addition they work
silently and are therefore ideally suitable for comfort applications in the pas-
senger cabin. Because of the current trend to electric vehicle the hitherto existing
drawback of a high electrical energy consumption of shape memory actuators in
some design proposals is not given any more.
Keywords: Shape memory alloys Actuators Weight reduction
1 Introduction
The shape memory effect was discovered for the first time in the year 1949 in some
experiments upon CuZn- and CuAl-alloys [1]. In the following decades this phase
change effect was further investigated and an extensive research work was carried out
especially for the NiTi-alloy system [2–4]. Nowadays SMAs have become a wide field
of research on the one hand and also part of an industrial market with real products and
applications in different technical areas on the other hand. In the first place medical
applications have to be mentioned, for example stents and guide wires [5, 6]. But also
in the field of domestic appliance the number of household applications using shape
memory elements is rapidly growing [7].
In the automotive industry, in contrast, some growing interest in these functional
materials can be mentioned, but the activities were mostly limited upon pure research,
patent filings and the development of some first demonstrators. The reasons for this
restriction are complex and multilayered: for a long time construction algorithms and
designing tools were missing, reliable material producers and suppliers were hard to

Potential Capabilities of Shape Memory Driven Automotive Devices 537
find, the phase change temperatures were not high enough and the effect stability was
sometimes too small. Working with shape memory elements very often was a time and
cost intensive trial-and-error process. Fortunately a lot of these drawbacks have been
overcome in the last years and therefore the automotive industry is more and more
interested in the development of new applications, especially when SMAs are used as
actuators.
Considering the automotive industry it can be asserted that in recent years a strong
effort has been carried out to reduce the weight of modern vehicles. In order to reduce
the energy consumption on the one side and to improve the environmental sustain-
ability of any car on the other side, the importance of weight reduction activities is even
growing faster. As lightweight designing is becoming more and more expensive and
show less potential savings, new approaches are needed. One promising technology
could be the use of shape memory elements. In the last years a lot of potential
application possibilities were presented, demonstrating the benefit of these functional
elements in automotive design solutions: they often reduce complexity of an actuation
device and enable new functions, thereby they reduce the weight and the required
space. In addition they work silently without any sound and are therefore ideally
suitable for comfort applications in the passenger cabin of an automotive vehicle.
Because of the current trend to electric vehicle the hitherto existing drawback of a high
electrical energy consumption of shape memory actuators in some design proposals is
not given any more.
Although the SMAs are described in a multitude of papers and books, this paper
gives a short overview of commercially available SMAs, explaining the unusual
behavior, the fundamental material mechanisms and the limitations of its use in
automotive applications. In the second part of the paper diverse examples are given to
demonstrate the benefit of shape memory driven automotive devices, some of them
developed by the authors.
2 State of Art of Shape Memory Alloys
2.1 Physical Effects and Fundamental Material Mechanisms

The most important property of SMAs is the reversible change of its shape by means of
thermal influences or interactions with external mechanical stress. During this change of
shape very high resetting forces may occur, for which reason this material is considered
to be very interesting for a lot of different applications. The shape memory effect is based
upon a reversible and diffusionless phase change between a high temperature phase
called austenite and a low temperature phase called martensite. By means of an appro-
priate change of temperature or mechanical stress this lattice structure may be provoked.
This change of shape or phase change, respectively, occurs in the solid state, far below
the melting temperature of the alloy. The phase change of SMAs is showing a hysteresis
and is generally characterized by four so-called phase change temperatures. In the case of
the austenitization these are the austenite-start-temperature, AS, and the austenite-finish-
temperature, AF. For the building of the martensite structure two temperatures are used:
the martensite-start-temperature, MS, and the martensite-finish-temperature, MF.
538 J. Strittmatter et al.
Fig. 1. Hysteresis and phase change of a shape memory material (Source: Liu [8])
Figure 1 shows schematically this hysteresis during the phase change of shape memory
materials.
The reversible change of shape provoked by the shape memory effect may occur in
different forms, which are generally named as:
• one way shape memory effect (OWSME)
• two way shape memory effect (TWSME)
• pseudo-elasticity (PE) or super-elasticity (SE)
Generally every SMA is able to conduct every mentioned shape memory effect;
which type of shape memory effect is executed depends upon the material, its pro-
duction and its thermomechanical treatment and is reliant on the temperature, defor-
mation and the mechanical stress [9].
One Way Shape Memory Effect (OWSME). The OWSME describes the ability of a
SMA that was deformed in the martensite condition to reset this deformation (shape
change) only by heating. The percentage of this reversible deformation can be repre-
sented up to 8 % [3].
Figure 2 shows as a practical example for the OWSME a wire or a clamp that
consists of shape memory material. First the clamp will be deformed by applying an
external force. Afterwards the clamp will be heated and the original shape will be
restored. A further cooling will not cause any visible change of shape.
Fig. 2. Principle of the one way effect (OWSME)

This effect of a reverse transformation into the original shape by means of the
austenitization only occurs after a previous deformation in the martensite condition and
is therefore called OWSME. Colloquially it can be said that the SMA “remembers” its
original shape in the high temperature phase (austenite). The phase change from
austenite to martensite does not change the shape.
Figure 3 shows the course of mechanical stress (σ), temperature (T) and elongation
(ε) during the OWSME. At the beginning the SMA is in the martensite condition, no
mechanical stress is applied and therefore no elongation can be observed. When the
SMA is mechanically loaded first an elastic area is run through, where the mechanical
stress rises at low elongation. This area of elastic elongation is followed by an area
where the elongation rises at nearly constant mechanical stress. This area is the
martensite plateau, where a SMA can be pseudo-plastically elongated. A further
loading after this reversible area would cause real-plastically deformation till fracture
[8]. If the SMA is unloaded within the pseudo-plastic area, a certain elongation remains
(ε1W). If a SMA in this condition is heated, the elongation returns back to zero.
Fig. 3. One way effect (OWSME or also ε1W) (Source: Liu [8])
Two Way Shape Memory Effect (TWSME). SMAs with a TWSME are able to
“remember” two shapes: one in the austenite at high temperature (in the same way as
the OWSME) and additionally another shape at lower temperature and martensite
structure. The percentage of reversible deformation of the TWSME is lower compared
to the OWSME and can be represented up to 4 % in the case of NiTi alloys [3].
The shape in the lower temperature condition has to be adjusted by thermome-
chanical treatments. No new mechanical stress is necessary to change between these two
shapes: the shape of the high temperature condition is formed by means of the austen-
itization and the shape of the low temperature condition by means of the martensitization.
The full effect is achieved when heated to the austenite-finish-temperature and when
cooled to the martensite-finish-temperature, respectively (see Fig. 4).
According to [10] the TWSME may occur extrinsically or intrinsically:
• An extrinsic TWSME effect is defined if a force that deforms the actuator element in
the cold condition is integrated into the setup
• An intrinsic TWSME effect is defined if the shape memory element is able to
change (“switch”) between the two shapes only by changing its temperature
Fig. 4. Two way effect (TWSME or also ε2 W) (Source: Liu [8])
Pseudo-Elasticity (PE) or Super-Elasticity (SE). PE describes a mechanical SME

where the alloy exhibits a “rubber-like” behavior. By application of an external force a
reversible shape change can be observed that overpasses the reversible shape change of
conventional spring steels up to ten times and can be practiced up to 8 % [3].
The reason for this effect is based upon a mechanically enforced phase change
within the material. For this purpose the material has to present 100 % austenite in the
beginning of the loading. Through the mechanical loading stress-induced martensite is
formed at a critical stress level. Subsequent unloading causes a reverse transformation
into the initial austenite structure (see Fig. 5).
Fig. 5. Pseudo-elasticity or super-elasticity (PE, SE or also εPE, εSE) (Source: Liu [8])
The pseudo-elastic effect only occurs at temperatures between the austenite-

finish-temperature, AF, and the martensite-desist-temperature, Md. Md represents the
upper limit for the formation of stress-induced martensite. At higher temperatures the
austenite structure remains in a stable condition without formation of stress-induced
martensite [11].
2.2 Available Alloys: Characteristics and Limitations

The thermal SME was observed in noble metal, Cu-, Fe- and NiTi-based alloy systems
and in non-metal shape memory materials [12]. It can be said that with the availability
of the NiTi alloys, discovered around 1960 in the Naval Ordnance Laboratory
(NOL) and therefore called “nitinol”, the commercial use of the shape memory effect
has started. Nowadays the NiTi- and Cu-based alloy systems are mainly used and
present the highest availability. Although an increasing interest and development of
Fe-based SMAs can be observed within the last years, Table 1 shows the most
important properties of NiTi-, CuZnAl- and CuAlNi-SMAs.
Table 1. Characteristic values of thermal SMAs

(Source: Janocha [12], p. 159, and own data)
Property Unit NiTi CuZnAl CuAlNi
One way effect, max. % 8 4…5 6
Two way effect, max. % 4 1 1,2
N = 10 % 6 1 1,2
N = 1000 % 2 0,8 0,8
N = 100000 % 0,5 ≥104 ≥103
Max. number of cycles ≥10 6
Phase change temperatures °C −100 … +120 −200 … +120 −200 … +170

Hysteresis °C 30 15 20
Overheating, max °C 400 150 … 200 300
Young’s modulus, martensite kN/mm2 28 … 60 70 80
Young’s modulus, austenite kN/mm2 83 … 100 70 … 100 80 … 100
Tensile strength, martensite kN/mm2 700 … 1100 700 … 800 900 … 1200
Tensile strength, austenite kN/mm2 800 … 1500 400 … 800 500 … 800
Strain to rupture, martensite % 40 … 60 40 … 50 8 … 10
Strain to rupture, austenite % 12 … 30 10 … 15 5…7
Thermal conductivity W/mK 10 … 18 120 75
Thermal expansion coefficient 10−6 1/K 6,6 … 10 17 17
Specific electric resistivity 10−6 Ωm 0,5 … 1,1 0,07 … 0,12 0,1 .. 0,14
Density 103 kg/m3 6,45 7,9 7,15
Corrosion resistance Very good Sufficient Good
Biocompatibility Very good Bad Bad
In most of the applications NiTi-alloys are used because of their advantageous

combination of the material properties. They exhibit the highest values of OWSME and
TWSME, the best effect stability, the smallest risk of overheating, excellent corrosion
resistance and biocompatibility and moreover a very high specific electrical resistivity
[12]. The most important disadvantage seems to be their high price level.
2.3 Potential Capabilities for SMAs in Automotive Applications

Concerning the application of SMAs in automotive vehicles the use of actuator ele-
ments seems to be most promising. Practical examples are being presented in the next
chapter. Table 2 shows the general advantages and disadvantages using NiTi-SMAs as
actuators.
Table 2. Important properties of SMAs in actuator applications (Source: Janocha [12], p.160)
Pros Cons
High energy density (high working capacity Effect stability is strongly dependent on the
per volume) quality of the alloy
Almost immediate shape change High price (NiTi-alloys)
Multiple forms of shape change (elongation, Application requires competent and skillful
contraction, bending, rotation) consulting (manufacturers, experts)
Simple mechanisms are possible (low Minor energetic efficiency
complexity of the actuator design)
Possible limitation of the effect for certain Limitation of application bandwidth because
element areas of heating/cooling restriction
Clean, noiseless and nonarcing operation Risk of degradation and fatigue
conditions
In addition to the design of classical shape memory actuator elements in automotive

vehicles it is also possible to use SMAs as crash elements, utilizing the energy dissi-
pating property during the phase change [13]. Based on the different electrical resis-
tivity of the austenite and stress-induced martensite structure the application of SMAs
as sensor elements is possible as well.
3 Examples for the Use of SMA in Automotive Devices

3.1 Overview of Potential Applications
Figure 6 shows an overview of potential applications of SMAs in car systems. Several
of them are still under development; some of them are even shown in certain
demonstrators [14]. One application that already passed to a standard-production
application in luxury class vehicles of some car manufacturers is presented in the next
chapter.
3.2 First Serial Automotive Application: SMA Pneumatic Valves

for Lumbar Support in Car’s Seats
The first high-volume product (>5 Mio actuators/year) is an automotive valve used to
control low pressure pneumatic bladders in a car seat that adjust the contour of the
lumbar support/bolsters. The overall benefits of SMA compared to traditionally-used
Fig. 6. Overview of potential applications of SMAs in car systems (Source: Mercedes-Benz)
solenoids in this car seat application (e.g.: lower noise, electromagnetic compatibility,
lightweight, required space, power efficiency/consumption) are the crucial factors in the
decision to replace the old standard technology with SMA [15].
According to a big SMA manufacturer [16] the SMA actuators are a perfect fit for
valves because the micron thin SMA wire allows for numerous valve form factors
enabling various options for attachment. In addition, the included proportional control
(without any additional sensor) is the most suitable option for every valve supplier [16].
The silent and minimalistic design specification of the SMA air and gas valves offers
benefits for a wide use in the automotive industry. This has contributed to winning
significant production volume in the automotive sector [17]. This actuator is designed
to fit into a customized valve or valve block offering 2 × 2/2 way, replacing two single
magnets that would have to be used otherwise (see Fig. 7). Some of the impressive
Fig. 7. SMA actuator with SmartFlex wires and electronic control board (left side). The SMA
wire is conducted along the plastic part and activates the valve against the bias spring when
heated (right side) (Sources: [18, 19])
actuator characteristics [18] are listed below: Voltage: 9–16 V; Temperature: −40°C–
80°C; Stroke: 0,4 mm; Closing force: 0,8 N; Full open/close < 120 ms; Weight <
20 g; Noise < 25 dBa.
One of the leading SMA companies is producing every year 10 million of these SM
actuators that are installed in all main vehicles’ platforms of Daimler, BMW, GM,
Hyundai, Ford and Porsche/VW [19].
3.3 Further Potential Applications

The Konstanz University of Applied Sciences and the WITg are working in the field of
SMAs for more than 20 years and various automotive applications were investigated,
some of them still are under development. In these applications a strong focus was set to
quick changing actuators in order to use NiTi wires in automotive pedestrian safety
systems [20, 21]. Another field of activity is located in the energy dissipation of a crash
impulse by means of shape memory elements [22]. Recently the authors are working
upon new automotive systems with SMAs that can be used on the one side for comfort
tasks in the passenger cabin and on the other side also for crash situations, presenting
new functions in order to prepare the passengers to the particular accident situation [23].
In the Conference up-to-date information of these projects will be given; for lack of
space in this paper only one actuator type for automotive applications will be presented.
In some applications, the stroke of a SMA based on the two way effect is not
sufficient. Due to this, a structure has been developed, in which stroke as well as lifting
force are variably expandable. The function to a specific installation length to increase
the stroke of one or more SMA actuators is realized with a linear mechanical cascading
mechanism. This is based on a connection of the actuator elements with the superposed
cascade element on which in turn are mounted further SMA actuators. For the desired
contraction the number of levels of the cascade serves as summation of each stroke
path. Depending on the lifting capacity required, the number of used SMA wires can be
increased or a larger cross-section could be used. Figure 8 shows the first prototype on
the left side and its further development on the right side.
Continue setting options with respect to the movement can be achieved by the
connection of the electrical activation. With a series circuit of each layer the cascade
Fig. 8. Cascade systems for SMAs (left: first design, right: further developed design)
performs a slow and uniform motion. In contrast to this, with a parallel circuit it results
a rapid contraction.
4 Summary and Outlook
In the automotive industry a strong effort has been undertaken to reduce the weight of
modern vehicles. In order to reduce the energy consumption and to improve the
environmental sustainability, the importance of weight reduction activities is even
growing faster. As lightweight designing is becoming more and more expensive and
show less potential savings, new approaches are needed. One promising technology
could be the use of shape memory elements. In the last years a lot of potential
application possibilities were presented, demonstrating the benefit of these functional
elements in automotive design solutions: they often reduce complexity, weight and
required space of an actuation device and enable new functions. In addition they work
silently and are therefore ideally suitable for comfort applications in the passenger
cabin. Because of the current trend to electric vehicle the hitherto existing drawback of
a high electrical energy consumption of shape memory actuators in some design
proposals is not given any more.
References
1. Lagoudas, D.C.: Shape Memory Alloys – Modeling and Engineering Applications. Springer
(2008)
2. Saburi, T.: Shape memory materials. In: Proceedings of the International Symposium
SMM99, Kanazawa, Japan (2000). Trans Tech Publications Ltd.
3. Gümpel, P., Gläser, S., Jost, N., Mertmann, M., Seitz, N., Strittmatter, J.: Formgedächt-
nislegierungen – Einsatzmöglichkeiten in Maschinenbau, Medizintechnik und Aktuatorik;
Band 655, expert Verlag Renningen 2. Aufl (2016)
4. Otzuka, K., Wayman, C.M: Shape Memory Materials. University Press, Cambridge (1999)
5. Duerig, T.M.: An overview of nitinol medical applications. Mater. Sci. Eng. A(273–275),
149–160 (1999)
6. Morgan, N.B.: Medical shape memory alloy applications — The market and its products.
Mater. Sci. Eng. 378, 16–23 (2004)
7. Furukawa Techno Material Co., Ltd. https://www.furukawa-ftm.com/english/nt-e/appli/
juutaku.htm
8. Liu, Y.: Formgedächtnis – Mikroventile mit hoher Energiedichte. Universität Karlsruhe,
Dissertation (2003)
9. Fink, R.: Superelastische Formgedächtnislegierung als Sensorelement. HTWG Konstanz,
Master’s Thesis (2004)
10. Büttgenbach, S.: Mikrosystemtechnik. Institut für Mikrotechnik, Technische Universität
Braunschweig (2008)
11. Stoeckel, D., Yu, W.: Superelastic NiTi Wire. Wire J. Int., 45–50 (1991). presented at the
60th Annual Convention of the Wire Association International, Boston, MA, October 1990
12. Janocha, H.: Aktoren mit thermischen Formgedächtnis-Legierungen. In: Janocha, H.:
Unkonventionelle Aktoren – Eine Einführung, Oldenbourg Wissenschaftsverlag GmbH,
München 2. Aufl., pp. 155–173 (2013)
13. Clipa, V., Gheroghita, V., Gümpel, P., Strittmatter, J.: Characterization of NiTi shape
memory damping elements designed for automotive safety systems. J. Mater. Eng. Perform.
23(5), 2696–2703 (2014). Springer
14. Butera, F.: Shape memory actuators for automotive applications. In: Auricchio, F. (eds.)
Shape Memory Alloys, Advances in Modelling and Applications, International Center of
Numerical Methods in Engineering CIMNE, Barcelona, Spain, pp. 405–426 (2001)
15. Shape-memory alloy, Wikipedia: https://en.wikipedia.org/wiki/Shape-memory_alloy
16. Actuator Solutions GmbH. http://www.actuatorsolutions.de/products/valves/
17. Actuator Solutions GmbH. http://www.actuatorsolutions.de/products/valves/air-gas/
18. Actuator Solutions GmbH. http://www.actuatorsolutions.de/products/valves/air-gas/33-
actuator/
19. Butera, F.: SMAs for Industrial Applications – From first NiTiNOL to a Big Success, https://
www.saesgetters.com/sites/default/files/Presentation%20by%20F.%20Butera%20SAES%
20Group.pdf
20. Gläser, S., Gümpel, P., Kilpert, H., Strittmatter, J.: Quick changing actuators for safety
systems in automobiles. In: Proceedings of the 10th International Congress on Automotive
and Future Technologies CONAT 2004, Braşov, Romania (2004)
21. Chiru, A., Gheroghita, V., Gümpel, P., Strittmatter, J.: Future applications of Ni-Ti alloys in
automotive safety systems. IJAT 15(3), 469–474 (2014)
22. Gheorghita, V., Strittmatter, J., Gümpel, P., Akarasanon, P.: New Testing machine for
measuring the damping properties of shape memory wires in dynamic conditions, Annals of
Faculty Engineering Hunedoara. Int. J. Eng. Fascicule 2, 119–122 (2013). Tome XI
23. Gheorghita, V., Guempel, P., Ceron, A.E., Strittmatter, J.: shape memory alloy actuators in
steering columns. In: Proceedings of the XXVI microCAD International Scientific
Conference, Miskolc, Hungary (2012)
Applicability of Composite Forming
Technologies for Automotive Components
Lucian Rad(&) and Anghel Chiru
Transilvania University Brasov, Brasov, Romania

{lucian.rad,achiru}@unitbv.ro
Abstract. The purpose of this paper is to present the proper technologies

available today to create automotive components in composite materials based
on carbon fiber. In 2016, major manufacturers have focused on composite
materials suitable choice of material for the purposes component products,
design and development of opportunities for improving the technology for mass
production.
Keywords: Composite materials Technologies
1 Introduction
The purpose of the research is to find the proper technology for obtain composite
material components in mass production. It will be made a comparison between three
of most usual technologies in order to compare the parts obtained.
Given the US CAFE standards requirements (Corporate Average Fuel Economy)
for vehicle manufacturers, to meet the thresholds of consumption of 3.8 l/100 km by
2025, in conjunction with requests EU (European Union) regarding CO2 emissions of
75 g CO2/km by 2025 lightweight materials have become a necessity [3].
A weight reduction of 100 kg leads to a reduction in average fuel consumption of
0.6 l/100 km.
Composite materials based on glass fiber and those based on carbon fiber, ther-
moplastics, thermos-settings but new aluminum alloys may be what specialists call
“multi-material vehicle” [2].
There is particular interest to develop technologies for the production of small
structural components with complex geometry, such as the steering column bracket.
The chart below shows the situation of forming technologies, the most used, depending
on the time of production, geometric complexity and length of carbon fiber (Fig. 1).
2 Technologies Used for Steering Column Components
Two components are found in the majority of steering columns, namely, steering
column bracket, with complex geometry shape and steering column housing with a
relatively simple geometric shape.

548 L. Rad and A. Chiru
Fig. 1. Forming technologies function of parts complexity and fiber length
These two components are representative in terms of tension/forces and geometric

complexity.
Steering column housing: it makes the connection between tubular shaft and
steering wheel. Steering shaft tubular is used both for transfer torque and transfer forces
driven by the driver through the steering wheel. To achieve this part have been con-
sidered three forming technologies: autoclave (Prepreg), RTM and space winding
(Fig. 2).
Fig. 2. Forming technologies for steering column housing
Steering column bracket: For this part the main purpose liaising between the steering
column and vehicle transfer driving forces acting on the casing column bracket on the
vehicle. For this part it was used the same three technologies, considering that the
shape is more complicate comparing with housing.
The comparison between forming processes was made taking into account the
reduction of weight, stiffness obtained and the natural frequency (Fig. 3).
Applicability of Composite Forming Technologies 549
Fig. 3. Forming technologies for steering column bracket
2.1 Autoclave Technology

The process includes manually cutting and hand lay-up Prepreg layers that make up the
voluminous workmanship of all components examined and post-production processing
(Fig. 4).
Fig. 4. Autoclave process-general presentation
Prepreg’s are produced using a cutting tables run by computer, so-called “cutter”,
according to guidelines laid fiber so that losses are minimal and optimal use of layers.
Prepreg’s tailored forms are placed manually in one form or tool. For hollow
interior components, such as casing column, the volume of labor is higher because it’s
working with the halves tools for deposited layers.
For prototyping housing and bracket steering column were used Sigratex Prepreg’s
titled “2 × 2 fabric’’ and epoxy resin matrix [1].
FEM analysis results show that the variant that uses the steering column bracket
and housing made of composite material obtained by the autoclave process, does not
meet the specifications on the stiffness and natural frequency. After the study param-
eters MEF was established the optimal structure of layers and wall thickening from
1.64 mm to 2.08 mm. The results were examined in comparison with the previous tests
and the values obtained by products of the series.
The process is ideally applicable for small series up to 1’000 parts/year and pro-
duction costs and the components are comparable to those of steel prototypes or alu-
minium for the same number of pieces. CAD model also shows the possible reduction of
both parts, using CFRP components with 992 g relative to the steel components.
For FEM analysis it was used the ANSYS software for composite materials. The
constraint points were set in bracket fixing points on the vehicle body and it was
simulated the vertical and horizontal loads of 500 N on the drive wheel point.
Replacing the composite components obtained in series steering column, it was
performed experimental measurements and the results were compared with FEM results.
The values obtain for FEM simulation and experimental measurements are shown
in Fig. 7. The same methodology it was used for all forming processes describe below
(Fig. 5).
Fig. 5. Weight saving using autoclave CFRP components (Source: T. Heitz Doctoral Thesis)
FEM analysis and experimental results are different due to data entry that do not
express the real behavior of the material. There is some calculation model which could
be applied to this case but none of them describe the real characteristic for composite
material (Fig. 6).
Fig. 6. FEM analyze for CFRP parts by autoclave
Fig. 7. Natural frequency and displacement for autoclave composite parts

Production structural components of the steering column with autoclave process

allows a variety of degrees of freedom and it is ideal for small series production. Due to
manual execution are facilitated the introduction of UD tapes, sandwich materials and
metal inserts in almost any geometric configuration.
The process can serve as a starting point for further developments.
2.2 RTM Technology

The process consists of combining both components braiding process and RTM pro-
cess. Overview on the steps of obtaining the two components is made in the image
below (Fig. 8).
Fig. 8. RTM process-general presentation
Description of the housing steering column process: To achieve Preform it was used
the process of basic textile (weaving) which is modernized, automated and equipped
with appropriate tools. These elements are processed into final product form through a
resin injection process, RTM.
Description of the bracket steering column process: The basic construction of the
lateral side pieces consisting UD multiaxial fiber orientation 0°, –45° and +45°. This
process is very important and to assess the suitability preform on the tool and the
possibility to execute a process RTM. The 4 preform’s (2 × sides, front axle, rear axle)
were placed in the tool and tool parts closed (Fig. 9).
Fig. 9. FEM analyze for CFRP parts by RTM process

The FEM results show that the steering column use components obtained by RTM
are fulfilled requirements of the specification on its own frequency. The values of input
parameters are the same with previous process. The output values are shown in Fig. 10.
CAD models show the possibility to decrease the weight with 670 g, using CFRP.
On the RTM production process for FEM calculations were made on the following
marginal conditions:
– Mass Substitute wheel weight is set at 3.58 kg
– Inner tubes steering column are steel
– The volume of fiber is 50 % against 60 % for sour safety factor in achieving target
Fig. 10. Natural frequency and displacement for RTM composite parts
Fig. 11. Weight saving using RTM-CFRP components (Source: T. Heitz Doctoral Thesis)
Overall investigations conducted had very positive results because they fulfilled the
main requirements of the specification. MEF calculations proved to be a helpful means
of great importance for fast and accurate dimensioning constructive. The process has a
high automation potential, can achieve tactical production < 3 min/part. The process is
proper for small series of up to 10’000 pieces/year (Fig. 11).
2.3 Winding Space Technology

The two manufacturing technologies for housing pultrusion and space winding for
bracket are fundamentally different. For this reason the components are separately
treated (Fig. 12).
Fig. 12. Space winding process-general presentation
Space winding process description: This new process use 2 robots, handling robot
who support the tool and winding robot who make the winding process.
On-device port inserts are fixed by means of cover plates. After winding, the device
is removed from the robot and check possible defects of winded bracket and fix the
pressing plates with screws. Subsequently, the polymerization takes place console in
closed form, in a furnace pre-heated to 80°C. Five minutes after introduction into the
furnace, its temperature reaches 100°C and the device is left in the oven for 25 min
followed by cooling in air to the forming console for one hour and then extracted from
the mold (Figs. 13, 14 and 15).
Fig. 13. FEM analyze for CFRP parts by space winding process
Fig. 14. Natural frequency and displacement for space winding composite parts
Fig. 15. Weight saving using space winding-CFRP components (Source: T. Heitz Doctoral
Thesis)
In this case, saving weight is 639 g using CFRP components comparing with steel
and measurements fulfilled the specification requirements (Table 1).
Table 1. Weight saving in function of manufacture process
3 Conclusion
Comparing the three technologies, it could be conclude that for most weight saving
case, the requirements are not fulfilled and if we add the fact that it’s need production
time smaller than 3 min/part, we can develop further the space winding process as
series process for small and complex geometry parts in composite materials.
Comparing the FEM results with experimental results, it can be seen the differences
caused by mathematical model used for simulation. The challenge is to obtain the
proper mathematical model for each forming process in order to have the real behavior
of part created.
Regarding the technologies used for this study, it can be conclude, based on
measurements results, that space winding technology is the most proper for develop
considering the automation potential and the possibility to change the fiber orientation
on force direction.
This process must be improved regarding the impregnation system, tensioning

device and tools in order to obtain a minimum production time for big series production
up to 50.000 parts/year.
References
1. Heitz, T.: The physical-mechanical properties of structural components made of fiber
composite materials in the application of steering columns in cars. Doctoral thesis, Brasov,
Romania (2013)
2. Rad, L.E., Chiru, A., Leahu, C.: Calculation of the steering column bracket made of
composite materials reinforced with continuous fibers. In: Proceedings of the European
Automotive Congress EAEC-ESFA, pp. 567–575 (2015)
3. Lucintel, L.L.C.: Assessment of Global Carbon Fiber and Carbon Composites Market (2015)
Constructive Optimization of Composite
Materials Automotive Components
Anghel Chiru and Lucian Rad(&)
Transilvania University, Brasov, Romania

{achiru,lucian.rad}@unitbv.ro
Abstract. This paper presents the steps made for develop a new steering col-
umn console in composite material, based on requirements of VW. Develop-
ment of lightweight construction, will be analyzed, considering design concepts
and new manufacturing process.
Keywords: Composite materials Technologies
1 Introduction
Steering column bracket is fixed to the steering column of the car through the four
attachment points and along with the casing column, it connect the steering wheel and
steering shaft. Through adjustment lever has the ability to adjust the position of the
steering wheel up/down, forward/backward. The console contact with its components
of the steering column is shown in Fig. 1.
Fig. 1. General components of steering column [Source: Thyssenkrupp]
The starting point is the standard steel steering column. The target objective is to
develop cost effective steering column components that render maximum weight
reduction to the overall system and can be produced in high volume series.
To restore the design of structural components for the steering column is important
to make considerations of both integral and differential methods of design, in terms of
economical and manufacturing process [1].

Constructive Optimization of Composite Materials Automotive Components 557
Following the product development methodology below, it was created two new
designs for the steering column VW Golf 6, using Catia and Solidworks software and
considering the space winding manufacturing process (Figs. 2, 3, 4, 10, 13 and 14).
Fig. 2. Product development methodology
2 Concept Evolution for Steering Column Console Using

Space Winding Technology
Space winding technology is a new process for manufacturing small parts with com-
plex geometry. The objective is to obtain a composite console steering column which
could replace the steel steering column.
In this study, two versions of the console CFRP will be presented, considering new
tools and equipment, new metallic inserts and new raw material.
We chose the steering column bracket VW Golf 6 as replace component. The
requirements needed to be met by the steering column were extracted from this vehicle
specifications:
• The natural frequency: vertical >50 Hz and horizontal >50 Hz
• Rigidity: vertical ≤1.6 mm and ≤1 mm horizontal
• Control climatic conditions –30/+80° C
2.1 Version 1
To create steering column brackets in composite materials, is needed the following
components and steps: carbon fiber roving Tenax HTS 40 24 K, resin system (matrix)
Neukadur 986 + 246, tool with pressure plates and the oven for polymerisation pro-
cess.After carbon fiber impregnation, hand winding it was made on the aluminum tool.
After winding process, the pressure plates close the tool, creating the shape of bracket.
The curing was made using the oven according to resin specification. The final step is
the demoulding process, obtaining the bracket in composite materials.
558 A. Chiru and L. Rad
Fig. 3. Winding process version 1
Fig. 4. Metallic inserts for version 1
The inserts. The need of using these inserts has resulted in the need to ensure the
required rigidity of the bracket and to provide a means of mounting the bracket on the
vehicle structure.
Axial inserts and the support inserts are made out of aluminum in order to provide
an optimal weight bracket and ensure stiffness. Also, due to the stiffness, inserts made
of steel sheets were implemented where the transverse force was bigger.
Tools. It was made of aluminum (Fig. 5), is equipped with a main board on which
plates are mounted where the fiber bundle is placed through winding process. Pressure
plates are fixed by screws and assure the tightening and give the desired shape of the
console.
For this case, the winding was made by hand in order to experiment the process and
observe possible problems.
Fig. 5. Tools for obtaining the composite steering column for space winding process
Results. On the beam structure it was applied a force (F = 500 N) vertically and
horizontally. FEM analysis showed in horizontal plane, perpendicular to the fibers, the
maximum displacement of 0.53 mm to 500 N load and vertical displacement reaches
the maximum value of 0.3 mm. FEM simulation also does not capture the real behavior
of the console. In terms of requirements specifications, horizontal characteristic fre-
quency was not met, although the value obtained is not critical (Fig. 6).
Fig. 6. FEM simulation, deformation (stiffness) and natural frequency tests results for version 1
Problems encountered in version 1 of the steering column bracket composite were

derived related to surface quality, mold geometry, the winding planes, non-parallelism
of the two lateral sides. The sharp edges of the mold were cutting carbon fibers which
led to a premature breakage of the console and to preload of the whole structure. Due to
lack of constant pressure, there were air gaps and fiber and matrix excess affecting
tensile and compressive strength.
2.2 Version 2
For this version it was required to optimize metal inserts, the mold used and automating
the entire process in order to achieve production time of 2 min/part.
For this version it was developed devices in order to obtain a full automatic process.
The carbon fiber is impregnated with resin in new impregnation bath, then pass
through guidance system (winding robot) which wind the fiber on the tool (tool support
robot). The pressure plates close the tool and the oven cures the part according to a
cycle track given by resin supplier. After demoulding it was obtained the final shape of
the bracket without any other post-production processes (Fig. 7).
The mathematical model and experimental tests for each version are developed and
described in different paper (reference [3]).
Fig. 7. Winding process version 2
Inserts. For this version, aluminum support inserts were designed for a better fit and
considering that it’s necessary to avoid sharp edges, based on experimental results of
version 1.
Fig. 8. Metallic inserts for version 2
Tools. For this version two tools were presented in Fig. 9, one equipped with rein-
forcement and one without reinforcement. The console without reinforcement it’s an
optimized shape of the console obtaining a smaller weight and follow the
specifications.
Fig. 9. Tools, w/o reinforcement for obtaining the composite steering column
Winding was carried out using two robots, one winding and one for the tool
support.For this, new devices have been used as impregnation bath, guide system, each
with multiple versions.
Three versions of the guiding device, illustrated in Fig. 8, have been designed and
developed. The aim is to guide and ensure the orientation of the impregnated fiber.
The challenge was that the hardened resin on the device, makes it impossible to be
reused repetitively. The direction of the fiber, as well as the size of the outlet are
important factors.
Fig. 10. Different guiding device for space winding process
Impregnating bath (Fig. 11) plays an important role, when it impregnates the
carbon fiber with matrix and squeezes the excess resin carbon fiber. The wetting
ensures the quality of the composite material in its new format and also plays an
important role in the polymerization of the material. If the fiber is impregnated in
excess, large amounts of resin could accumulate, harden, and destroy the tool. If too
little fiber is impregnated, the quality of the material will be poor since the layers will
not adhere to each other and after polymerization, and the final piece will be easily
unwound. The ideal proportion is 60 % fiber and 40 % matrix.
Fig. 11. Impregnation bath for space winding process

Results. In order to conclude if there is needed reinforcement or not, it was made a

FEM comparison and the results are similar. It means that is it possible to use, further,
the variant without reinforcement winning in weight 10 grams/part.
Regarding the requirements in terms of natural frequency and deformation (stiff-
ness), FEM analyses results and experimental results it’s showed that new bracket
shape met the specifications (Fig. 12).
Fig. 12. Deformation and natural frequency tests results for version 2,w/o reinforcement
To analyze the feasibility of new console obtained, it took a combination of ver-

sions of the casing column assembly serial console CFRP. The results shows that we
can successfully use CFRP console.
Fig. 13. Feasibility study for steering column console made of composite materials
The study showed that tension fiber has great influence on the stiffness and limited
natural frequency. Also the clamping force must be large enough to evacuate the air in
the piece and small enough not to damage the fibers. To have constant tension, a fiber
tensioning system that comes out of the impregnation bath and a tension coil voltage to
secure a fiber before impregnation is required.
Fig. 14. Steering column bracket steel (left) and carbon fiber bracket (right)
Throughout the development process we were used a series of epoxy resins and
carbon fibers. The characteristics of these materials have a decisive impact on pro-
duction time and on winding itself. Weight bracket obtained, using composite mate-
rials, is lower by 48 % compared to steel bracket.
3 Conclusion
In order to obtain full automatization of the process, it was set up the usage of two
robots, one that supports fiber guidance system and other support for the mold that is
winding.
The trajectory of the robot is important in order to avoid the mold collision. Also
winding speed plays an important role in getting a low winding time.
The winding speed could be increased up to the maximum (1 m/sec). In that case
the impregnation of the fiber it’s made without spreading the resin. Also the tensioning
system after impregnating fiber brake coil voltage gives a constant thread, a uniform
impregnation and a placement on tools as desired geometry.
One of the great advantages of this newly developed process is the possibility to be
executed using roving bobbins sites with good price and the fibers of the roving can be
submitted in accordance with the flow paths defined and lines of force.
Irregularities in the roving layer deposition can be tolerated within certain limits so
that in the event of winding techniques irregularities in laying the fibers free permit
ensuring a proper quality. The new process, developed in this work allows additional
degrees of freedom winding space by using a robot. With production assembly
described we can achieve a continuous production with a production cycle of under
2 min/part. Developing new components for composites technology, we can create
parts with complex geometries, light-weight and with strong material properties.
The next step is to improve the space winding process devices in order to obtain
continuous and repetitively process for large production series > 50.000 parts/year,
respecting the customer specification requirements, without affecting steering column
assembly functionality.
References
1. Chiru, A., Rad, L.E.: New technology for composite materials parts. Proceedings of the
European Automotive Congress EAEC-ESFA 2015, pp. 209–217. Springer, Heidelberg
(2016)
2. Rad, L.E., Chiru, A.: Calculation of the steering column bracket made of composite materials
reinforced with continuous fibers. Proceedings of the European Automotive Congress
EAEC-ESFA 2015. Springer, Heidelberg (2016)
The Evaluation of Rheological Properties
of Composites Reinforced with Hemp,
Subjected to Photo and Thermal Degradation
Mariana Domnica Stanciu1(&), Ioan Curtu1, Mihai Groza1,

and Adriana Savin2
1
Department of Mechanical Engineering,
{mariana.stanciu,curtui}@unitbv.ro,
mihai.groza90@yahoo.ro
2
National Institute of Research and Development
for Technical Physics, Iasi, Romania
asavin@phys-iasi.ro
Abstract. The cars of the future are designed so that to replace partial or total
the fossil fuels. For this reason, the structural modifications of the materials used
in the cars structure are researched being studied the lightweight materials,
ecological and environmentally friendly, reusable and with improved perfor-
mance in terms of mechanical strength (static and dynamic), thermal insulation
and acoustic, with interface aesthetic. The paper aims to assess the structural
modifications and visco-elastic properties of the composites reinforced with
hemp mat subjected to artificial aging by exposure to UV radiation. Thus,
samples having a percentage of reinforcing fiber hemp about 50 % were sub-
jected to a dynamic mechanical analysis (DMA), being determined visco-elastic
behavior at different frequencies loading. Then, samples were exposed for 168 h
at the UV-A radiations, after that their rheological behavior was re-evaluated. In
terms of structural and morphological changes, the samples were analyzed with
the atomic force microscope (AFM) and method of contact angle both before
and after accelerated aging. It was found that after UV radiation, on the surface
of the composite structure reinforced with hemp mat occurred rearrangements of
chemical chains and increases of the surface roughness.
Keywords: Bio-composites reinforced with hemp mat Complex modulus

Photo-Degradation (UV) Dynamical Mechanical Analysis (DMA) Atomic
Force Microscope (AFM)
1 Introduction
Due to the environmental problems caused by the large volume of waste, the world is
moving towards composite materials reinforced with natural fibers. Although natural
fibers lower mechanical properties than synthetic fibers, they have advantages such as
low density, low price of the raw material and biodegradability [1, 2]. Thus, composites
reinforced with natural fibers are popular in areas where the weight of materials plays
an important role. One such area is that of automotive interior components.

566 M.D. Stanciu et al.
Given the close link between fuel consumption and vehicle weight, we can deduce
that the weight of materials used in the construction of a vehicle affects the amount of
CO2 emitted to the atmosphere with adverse effects on the environment. The interior
automotive parts industry, composite materials based on natural fibers reinforcement
are found mainly in the mat fibers or woven fabrics [3, 4]. A big challenge is getting
some components of composite materials reinforced with natural fibers showing visible
surfaces, with no need to cover them with other materials. Using these materials may be
obtained parts of the panels or panels with simple geometry without large indentations
as shown in Fig. 1. The coating materials are applied as in the case of flat panels, and
the panel fixing metal structure door is done by applying sticking coupling systems [3].
Fig. 1. Types of EcoCor panels based on hemp fibers
Unlike the metal or ceramic materials, due to the visco-elastic polymer, there is a
strong dependence between the properties of the polymers, time, temperature and other
environmental factors (UV) [5, 6]. In general, the polymers have an elastic behavior in
response to an applied force quickly and a viscous behavior in response to an applied
force slowly. These two states denoted the complex behavior of polymers. The tests to
determine flow and relaxation phenomena provide important information about the
dimensional stability of the polymer [7, 8]. The microscopic investigation with SEM or
AFM provides information about the modification of topology and morphology of
composite surface [8, 9]. The use of polymeric materials in the environment with UV
radiation can cause photo-degradation of the material as the breaking of chemical
bonds of polymer chains, which can lead to a discoloration of the material, oxidation of
its surface, to a weakening of the material and micro-cracking of the matrix its. Also,
exposure to UV radiation can cause stress concentrators on the surface of the polymeric
material (which can lead to a premature failure), and can make the material more
sensitive to the action of other factors (i.e. humidity) [10]. The paper evaluates the
changes in physical and mechanical properties of composites reinforced with hemp
fabric as a result of thermal and ultraviolet degradation.
The Evaluation of Rheological Properties of Composites 567
2 Materials and Experimental Method

2.1 Materials
The analyzed composite material contain hemp mat and polyurethane resin (RAI-
GITHANE 8274/RAIGIDUR CREM), with 50 % percentage of reinforcing natural
fibers. To evaluate the rheological behavior of composite reinforced with hemp fibers,
from the plate were cut 5 samples coded with the letter C (Fig. 2) whose geometric and
physical characteristics are shown in Table 1. This material is a commercial product,
currently used in car body structure. Initial, the characteristic curve (σ-ε) and elastic
modulus were determined at tensile test. So, the average values of elastic modulus are:
E1 = 2773 MPa (in longitudinal direction) and E2 = 3770 MPa (in the transverse
direction).
Fig. 2. Types of specimens and their geometry
Table 1. The physical characteristics of samples before UV exposure

Sample Length L [mm] Width/[mm] Distance d [mm] Thickness Mass
code hi [mm] mi [g]
C1 50 40 10,18 1,98 0,95
C2 50 40 10,22 1,99 0,92
C3 50 40 10,26 1,94 0,94
C4 50 40 10,27 1,96 0,91
C5 50 40 10,25 1,93 0,82
2.2 Experimental Methods

To study the rheological behavior of the composite materials reinforced with mat hemp
and polyurethane resin in the first stage was examined the roughness of surfaces speci-
mens using atomic force microscope (AFM - Atomic Force Microscope) NTEGRA Probe
Nanolaboratory, in the way semi-contact, using a cantilever type NSG 10 with constant
force from the Institute for Research and Development of Transilvania University of
Brasov. AFM measurements were processed using NOVA 1.0.26 RC1 software that
enables roughness measurement and calculation of average surface roughness.
Then the dynamic modulus E * and damping coefficient tanδ it were determined using the
equipment Netzsch DMA 242 C, endowment National Institute of Research and
Development for Technical Physics, Iași, Romania. The samples were exposed to UV
radiation for 168 h in artificial aging cabin placed on the Institute for Research and
Development of Transilvania University of Brașov. Subsequently, physical and
mechanical properties were revalued by the aforementioned methods and equipment
(DMA and AFM).
3.1 The Determination of the Surface Roughness Before and After

Exposure to UV Radiation by Means AFM
In tests conducted with AFM, were obtained 3D images (Fig. 3) and images of phase
contrast (Fig. 4) to the surface of each specimen. Based on these images the root mean
square of the roughness surface was determined before and after UV exposure with
value of 23.2 nm before exposure to UV and 32.26 nm after artificial aging. It is noted
that the surface roughness of the composite reinforced with hemp mat (C3) increased
after exposure to UV radiation, which shows an oxidative degradation of the respective
areas. This phenomenon can be linked to the composite capacity to absorb water, so to
modify the hydrophilic/hydrophobic properties [13, 14]. If hemp mat-reinforced
material, artificial aging has increased the surface energy, the material becomes
hydrophilic. These chemical modifications, morphological and structural effects after
artificial aging lead to a decline of the mechanical properties. Moreover, the color of
samples after UV exposure tends to be dark.
In the analysis of phase contrast images (Fig. 4), it has been found that before the
UV exposure, the surface of the test specimen is homogeneous; after exposure to UV the
migration of chemical compounds to the surface occurs which indicates that the surface
becomes more porous, pore size varying depending on the type of reinforcement
a b
Fig. 3. Topographical scan of a composite reinforced with hemp fibers obtained with AFM,
before (a) și after UV exposure (b)
a b
Fig. 4. The phase contrast micrographs before (a) and after UV exposure (b)
elements. However, after UV exposure, the specimens are observed on the surface of
black spots, thus the composition becomes uneven. Photo-degradation is initiated by
solar radiation, which results in the absorption of UV radiation by chromophores and in
the activation of excited states in macromolecules. Durability is a serious issue from
both a health and safety aspect and in terms of economic costs of automotive parts and
automotive products [11].
3.2 The Evaluation of the Dynamic Modulus and Damping Coefficient

for Different Frequencies of Dynamic Loads
In mechanical dynamic analysis were performed two types of tests. The first group
consisted of varying the frequency of load, keeping the temperature constant at the
value of 30° C. One by one, samples were placed onto the DMA equipment and were
subjected to a 3-point bending (support spacing: 40 mm). Every 30 min, the frequency
of the application force (in the amount of 6 N) was set to the following values: 1 Hz,
5 Hz, 10 Hz and 50 Hz. Every 0.25 s was recorded movement, which through the
acquisition system and data processing has been translated into a storage mode (E’) and
loss modulus (E’) under which it was calculated the complex modulus E*, the damping
tanδ as the ratio of E’’/E’. In Fig. 5, it is noticed that with increasing load frequency
increase the value of damping factor tan δ.
After artificial aging it is observed a nonlinear behavior of damping factor compare
to behavior of composite before UV exposure. This can be explained by structural
changes in resin, the phenomenon highlighted mainly by AFM image analysis with
consistent degradation of the polyurethane resin surface.
In the case of complex modulus E* (Fig. 6), the values are higher by about 26 %
after exposure to ultra-violet, which means that the material became more rigid. The
trend of increasing the modulus of elasticity depends on the frequencies of load.
At the beginning of the test, the material exhibits a nonlinear behavior; after about
10 min were observed a stabilization of the samples response (the complex modulus
0.090 1 Hz
0.085 C4
0.080 1 Hz UV
0.075
Damping tanδ
5 Hz
0.070
0.065 5 Hz UV
0.060 10 Hz
0.055
10 Hz
0.050
UV
0.045 50 Hz
0.040
0 10 20 30 50 Hz
UV
Time t , [min]
Fig. 5. The variation of damping tanδ for different frequencies before and after UV exposure
1900
C4
Complex Modulus E* , [MPa]
1800
1 Hz
1700
1 Hz UV
1600 5 Hz
1500 5 Hz UV
1400 10 Hz
1300 10 Hz UV
50 Hz
1200
50 Hz UV
1100
0 10 20 30
Time t , [min]
Fig. 6. The variation of complex modulus E*
recording almost the same value with increasing of time test). However, it is noted that
in the first 10 min of test, the internal friction is “drop by about 17 % after the loss
modulus is approximately constant over time, both before and after exposure to UV.
The value of loss modulus E” after artificial aging increases by approximately 22 %
regardless of the frequency of application.
3.3 The Assessment of Dynamic (Complex) Modulus and Damping

Coefficient with Increasing the Temperature
A second category of tests consisted in varying the temperature in DMA oven from 30
to 120°C with a heating rate of 2-3°C/min. One by one, samples were placed onto the
analyzer and for 44 min a force of 6 N were applied on specimens with frequency by
1 Hz. Every one seconds was recorded deflection, which through the acquisition
system and data processing has been translated into a storage mode (E’) and a loss
modulus E”. In Fig. 7 is observed that before UV exposure, the damping factor tanδ
aging tends to decrease from the 30°C, stabilizing at around 60°C. After UV exposure,
the damping is approximately constant up to a temperature of 50°C, then recording a
continuous increase, so that around 100°C the temperature where the two values of
damping (one before and one after aging) getting to be approximately equal. This
phenomenon can be explained by the fact that, after UV exposure of chemically bonds
of the matrix are stabilized, while increasing the transition temperature from 45 to 50°
C. Comparative values of glass transition temperature Tg were obtained by Baghaei
[13] in case of composite with hemp fibers/PLA, the Tg recorded around 57-61°C
which corresponds to the α relaxation of the amorphous regions in matrix.
0.075
Before UV C2
0.070
0.065
Damping tg δ
0.060
0.055
0.050
0.045
After UV
0.040
0.035
30 40 50 60 70 80 90 100
Temperature T , [ºC]
Fig. 7. The variation of the damping tanδ in relation with temperature during the test
In terms of complex modulus E*, the two curves of variation (one before and one
after aging) had a similar trend, except that after exposure to UV radiation, the module
has increased by about 8 % from pre-exposure (Fig. 8). The similarly results were
obtained by [12].In both cases, the dynamic modulus decreased in value at a rate of
about 12 MPa/°C. [13, 14] show that the storage modulus of the composites reinforced
with hemp showed remarkable dependence on fibre orientation. The storage modulus
was highest for the hemp/PLA (0) yarn composite, followed by the hemp/PLA (45) yarn
composite and the hemp/PLA (90) yarn composite. The increase in the stiffness of fibre
containing samples, revealing effective stress transfer from the fibre to the matrix at the
interface, can be interpreted as good adhesion between the fibres and the matrices [13].
2900
Complex Modulus E* , [MPa]
After UV C2
2700
2500
Before UV
2300
2100
1900
1700
30 40 50 60 70 80 90 100
Fig. 8. The variation of the complex modulus E* with temperature
After exposure to UV radiation, the energy losses undergo a stabilization

throughout the temperature range 30-100°C compared with the downward trend con-
tinuing of the loss modulus recorded before aging (Fig. 9). Figure 10 shows that the
deflection is increasing in both cases; but after exposure to UV are recorded lower
values by about 9 %, which shows a decrease in the elastic properties of the poly-
urethane composite reinforced with hemp mat.
190
Before UV
Loss Modulus E'' , [MPa]
180 C2
170
160
150
140
130
120 After UV
110
100
30 40 50 60 70 80 90 100
Fig. 9. The variation of the loss modulus with increasing the temperature
0.70
C2
Deflection vmax , [mm]

0.65
0.60
0.55
Before UV
0.50
0.45
After UV
0.40
30 40 50 60 70 80 90 100
Fig. 10. Variation of maximum deflection with increasing the temperature
4 Conclusion
In conclusion, the paper highlighted artificial aging effects on mechanical properties of

composites reinforced with hemp mat with 50 % hemp fiber content and polyurethane
resin. Thermal aging and photo-degradation increases the rigidity of the composite
which is manifested by the loss of capacity to absorb deformation energy, the
bio-composite structures becoming brittle. The polyurethane composites reinforced
with hemp mat undergoes structural changes, morphological and mechanical degra-
dation as a result of exposure to UV radiation and temperatures up to 100°C. Increased
temperature accelerates most of the degradation processes that occur in composites
reinforced with hemp such as oxidation, chemical attack and mechanical creep. Surface
degradation due to chemical changes of the matrix lead to decreased of the visco-elastic
properties and hence the service life of structures made of composites reinforced with
hemp mat and polyurethane resin. Solutions to prevent or delay the aging can be the
surface coating with UV varnishes or their use as support structure of automobile
interior parts. The importance of all protective measures is due to the fact that the
degradation from one agent can also reduce resistance to other agents.
Acknowledgement. This paper was supported by Program partnership in priority domains -

PNII under the aegis of MECS -UEFISCDI, project No.PN-II-PT-PCCA-2013-4-0656.
References
1. Peijs, T.: Composites for recyclability. Mater. Today 6(4), 30–35 (2003)
2. Bismarck, A., Baltazar-Y-Jimenez, A., Sarlkakis, K.: Green composites as Panacea?
socio-economic aspects of green materials. Env. Dev. Sustain. 8(3), 445–463 (2006)
3. Terciu, O.M., Curtu, I., Stan, G., Cerbu, C.: Lignocellulosic composites for automotive
industry. In: The 11th International Congress on Automotive and Transport Engineering
CONAT2010, pp. 117–124 (2010)
4. Stanciu, M.D., Terciu, M.O., Curtu, I.: Lignocellulosic Composites. Compozite lignocelu-
lozice. Application in Automotive Industry (in Romanian language. Editura Universității
Transilvania din Brașov (2014)
5. Fernandes, E.M., Correlo, V.M., Mano, J.F., Reis, R.L.: Novel cork–polymer composites
reinforced with short natural coconut fibres: effect of fibre loading and coupling agent
addition. Compos. Sci. Technol. 78, 56–62 (2013)
6. Ayrilmis, N., Jarusombuti, S., Fueangvivat, V., Bauchongkol, P., White, R.H.: Coir fiber
reinforced polypropylene composite panel for automotive interior applications. Fiber Polym.
12(7), 19–26 (2011)
7. Chauhan, A., Kaith, B.: Recycling the cellulosic biomass to competent material. J. Text. Sci.
Eng. 1, 1–5 (2011)
8. Cherizol, R., Sain, M., Tjong, J.: Modeling the rheological characteristics of flexible
high-yield pulp-fibre-reinforced bio-based nylon 11 bio-composite. J. Encapsul. Adsorpt.
Sci. 5, 1–10 (2015)
9. Peng, Y., Liu, R., Cao, J.: Effects of antioxidants on photodegradation of wood
flour/polypropylene composites during artificial weathering. BioResources 9(4), 5817–
5830 (2013)
10. Essabir, H., Bensalah, M.O., Rodrigue, D., Bouhfid, R., Qaiss, A.: Structural, mechanical
and thermal properties of bio-based hybrid composites from waste coir residues: fibers and
shell particles. Mech. Mater. 93, 134–144 (2016)
11. Maxwell, A.S., Broughton, W.R., Dean G., Sims, G.D.: Review of accelerated ageing
methods and lifetime prediction techniques for polymeric materials – NPL Report (2005)
12. Placet, V.: Characterization of the thermo-mechanical behaviour of Hemp fibres intended for
the manufacturing of high performance composites. https://arxiv.org/ftp/arxiv/papers/0906/
0906.3597.pdf
13. Baghaei, B., Skrifvars, M., Salehi, M., Bashir, T., Rissanen, M., Nousiainen, P.: Novel
aligned hemp fibre reinforcement for structural biocomposites: porosity, water absorption,
mechanical performances and viscoelastic behaviour. Compos. Part A 61, 1–12 (2014)
14. Shahzad, A.: A study in physical and mechanical properties of hemp fibres. Adv. Mater. Sci.
Eng. 2013, 9 (2013). Hindawi Publishing Corporation, ID 325085
Optimization of the Plastic Injection Molding
Parameters by Using the Taguchi Method
to Prevent Jammed Components
in the Assembly Process of Mechatronic
Devices
Ion Cristian Braga(&) and Anisor Nedelcu

jbc69mc@gmail.com, a.nedelcu@unitbv.ro
Abstract. Mechatronic devices are more and more present in the automobile
industry, with roles, to support the control of the car or to interact with the entire
system to make the driving safe and to increase passengers comfort. If in the
past these devices were simple, used for tasks such as actuation of the door
opening or other similar servo-mechanisms, nowadays, they consist of a com-
bination of a printed circuit board, including microprocessors and software, but
also mechanical and plastic parts working in complement with each other to
reach a common purpose. Often, the small plastic parts are designed to be used
in devices for translation, rotation or both movements, and subsequently the
haptic feedback feeling associated to these is important to be kept within the
requested range. This paper presents the optimization of the plastic injection
molding parameters, by using Taguchi method in order to reduce risks in the
serial production parts with dimensional precision. Also, the optimization
method is targeting prevention of jammed components in the assembly process,
and parts with changed roughness of the surface that could lead to a friction or a
bad haptic feeling. The study is performed on a plastic material component
obtained from polypropylene (PP) and the chosen parameters are modeled with
an orthogonal array (OA), signal-to-noise (S/N) ratio and analysis of variance
(ANOVA). They will be applied to analyze the shrinkage effect on the molded
parts. The Minitab software will be used to support the design of the experi-
ments based on the aforementioned factors, determine the number of necessary
trials, and analyze variance used to determine the parameters that affect the
quality of the mentioned characteristics. The results will show the best combi-
nation of parameters with optimal interaction to be used in injection molding in
order to supply plastic parts for a reliable assembly process of the mechatronic
devices.
Keywords: Injection molding Shrinkage Warpage Taguchi method

ANOVA Risks

576 I.C. Braga and A. Nedelcu
1 Introduction
Recently, Automotive 4.0 has increased industry awareness of automobile future

production by 2020. Based on the project Industry 4.01 this is one revolution in the
industry which combines production methods with state-of-the-art information and
communication technology, a project meaning the future of manufacturing and work in
Germany [1]. This near future will increase the challenge of the automotive manu-
facturers in their scope to produce parts with an increased quality and decreased cost.
Even today, automakers are challenged to find solutions together with customers, to
find robust and reliable materials reduced weight, and to decrease the overall mass of
the car. In Europe, statistics suggest that by 2020, the average vehicle weight will
shrink to a little more than a ton, its 1970 level, after peaking at nearly 1.5 tons in 2010
[2] (Table 1).
Table 1. Plastics will account for 18 % of the average vehicle’s weight (Source: AT Kearney:
Plastics. The Future for Automakers and Chemical Companies, June 2012)
For a mechatronic device, the combination of the quality of the subassemblies is

very important, and a small variation in the surface quality or dimension of part can
generate either noise or jamming.
Therefore, as plastic materials are increasingly used in vehicles, the plastic injection
molding processes used must be optimized to conserve the characteristics of these
parts. Also, the manufacturer must take into account the parameters used for setting the
machine, as these can greatly influence the final quality of the desired product In the
1
AUTOMOTIVE 4.0, Sensing the road ahead for tier 1 suppliers, December 2015.
Optimization of the Plastic Injection Molding Parameters by Using the Taguchi Method 577
end, this can influent the next assembly processes because of combination between two
or more parts or plastic parts versus metallic, rubber or other materials.
The plastic injection molding processes use raw material plastic as granule and in
the beginning the drying of the material must be considered in order to obtain quality
parts. This is assured using dryers. Following the dryer stage the material is inserted
into an injection unit to be heated and melted. In the final stage this is injected into the
mold to be cooled and solidified in order to obtain the desired shape of the part.
The parameters for adjusting the process such as cooling temperature, demolding
temperature, molding temperature, melting temperature, injection time, holding time,
mold close or open time, injection pressure, back pressure or holding pressure are
parameters to be optimized, at the end the parts injected will be produced at the quality
level required.
This paper presents the optimization of the process parameters in order to reduce
the shrinkage of housing, made from polypropylene (PP), as a part of a mechatronic
device which uses multiple parts that must slide together. A very important charac-
teristic for the mechatronic devices used in the automotive is the haptic feeling. This
concept is referring to the confirmation feedback felt by the user (driver or passenger)
when a button from the mechatronic device is pressed, so, the dedicated function was
activated [3]. And the haptic feeling is described in detail by the customers, transposed
in measurable parameters as travel way and force, both must to be within specific
limits.
The issue in the current manufacturing processes of these mechatronic devices is
that due to very small field of the tolerance of the component parts which must be
assembled together, the channel of the housing could become thinner due to shrinking,
then the other part which must slide inside the channel can have some friction, and the
haptic feeling will be out of customer requirements. Even this haptic will be accepted
on a limit, there could be occurred the noise. Or, the opposite, the parts are not slide and
will be jammed.
The planning of the manufacturing processes in the plastic molding injection area
uses the batches which are delivered in warehouse, then to assembly lines. And the
process validation of these batches is based on the measuring of the dedicated
important dimensions of the injected parts, at start of production and certain period, the
samples being taken, more or less, immediate after injection molding (and not kept a
while to be sure that are measured after shrinking).
The solution is to find the combination of the parameters which will reduce at
minimum the shrinkage of the housing, in that way, the process validation described
above will be relevant.
The optimization of the parameters will help to prevent either the noise in case of
play between the sliding parts neither the jamming of these parts. The study was
conducted using the Taguchi method and the data is processed using the Minitab
software. The goal was to find the factor-levels combination that reduces performance
variability and to adjust the factor-level that brings performance closer to the target [4].
2 Taguchi Method
Genichi Taguci was born in 1924 in Japan, where he worked as an electrical engineer
who worked during 1950’s to improve telephone communication system of Japan’s
post. Taguchi defines the level of quality of a product as the total loss incurred by
society due to a failure of a product to perform as desired when it deviates from the
delivered target performance levels, the well-known loss function. He saw the quality
improvement as a continual effort to reduce the deviation of a characteristic from the
nominal value which will lead to this loss, and he demonstrates that the noise (meaning
uncontrollable factors) is responsible for this deviation. The proposed tables known as
orthogonal arrays (OAs) made the design of experiments very easy [7, 8]. For instance,
if we take three factors A, B, and C, with two levels of each, a full factorial experiment
will require 23 = 8 trials, while OA L4 will require only 4 experiments [11]. This is
more useful when the number of factors increased, as the number of levels increase in
complexity as well.
Therefore, to achieve the desired quality of a product, based on Taguchi’s rec-
ommendation, the study must follow three stages: system design, parameter design and
tolerance design [4, 9].
The experiments will run and their results will be observed with the signal-to-noise
(S/N) ratio to determine the optimal level of the process parameters [10]. In the end, the
statistical analysis of variance (ANOVA) will show the influence of the factors to each
other.
3 Study Regarding the Optimization of the Plastic Injection

Molding Parameters
3.1 Material, Machine and Parameter Selection
The study was performed in a system which uses the injection molding machine Engel
Victory 100 tons. The material used for producing the housing, is a polypropylene
(PP) polymer. As the shrinking characteristic is an important quality factor, the fol-
lowing parameters were selected to be optimized: melting temperature, in the range
between 220 and 240°C, holding pressure, in the range between 85 and 105 bars,
injection pressure, in the range between 120 and 140 bars, and holding time in the
range between 5 and 7 s.
Table 2. Selection of the parameters and their levels (Source: authors’ development)
Code Parameter Unit Level 1 Level 2 Level 3
a Melt Temperature °C 220 230 240
b Holding pressure bar 85 95 105
c Injection pressure bar 120 130 140
d Holding time sec 5 6 7
3.2 Selection of Orthogonal Array

The selection of the orthogonal array depends on the degrees of freedom, where
formula 1 was used to calculate the minimum number of necessary experiments:
Xnv
Nexperiments ¼ 1 þ j¼1
ðLi 1Þ ð1Þ
Where nv represents the number of variables (factors) and Li represents levels of the
factors.
Because this experiment takes into account four factors with three levels, the OA
L9 is a minimum to be chosen, but the L27 is more appropriate.
Following the calculation, Minitab outputs the proposal of combinations for each
experiment. In this investigation, 27 trials were proposed – see Table 3. The a, b, c and
d are the code of the parameters from Table 2, and each level is represented by the
numbers 1, 2 and 3. The orthogonal array shows the combinations to be used for the
trials, from 1 to 27.
Table 3. Selection of orthogonal array (Source: authors’ development)

Trial a b c d
No.
1 1 1 1 1
2 1 1 2 2
3 1 1 3 3
4 1 2 1 2
5 1 2 2 3
6 1 2 3 1
7 1 3 1 3
8 1 3 2 1
9 1 3 3 2
10 2 1 3 1
11 2 1 1 2
12 2 1 2 3
13 2 2 3 2
14 2 2 1 3
15 2 2 2 1
16 2 3 3 3
17 2 3 1 1
18 2 3 2 2
19 3 1 2 1
20 3 1 3 2
21 3 1 1 3
22 3 2 2 2
(Continued)
Table 3. (Continued)
Trial a b c d
No.
23 3 2 3 3
24 3 2 1 1
25 3 3 2 3
26 3 3 3 1
27 3 3 1 2
3.3 Shrinkage Measurements

The relative shrinkage was calculated as the difference between the size of the mold
cavity and the size of measured product divided by the size of the mold cavity,
multiplied with one hundred (percentage). Formula 2 (shown below) was applied:
S = ðMd-PdÞ=Md 100 ð%Þ ð2Þ
Where: S represents shrinkage, Md represents mold dimension of the measured

characteristic, and Pd represents product dimension of the measured characteristic [5].
This study uses three trials for each experiment and the signal-to-noise (S/N) ratio
was calculated based on these trials. The S/N ratio calculation depends on the char-
acteristic to be measured or optimized. Because a variety of S/N ratios exists, we
considered three as generally applicable [4]:
– Bigger-is-the best quality characteristics;
– Nominal-is-the best quality characteristics;
– Smaller-is-the best quality characteristics.
As our target characteristic is shrinkage of the part, the signal-to-noise (S/N) ratio
appropriate is smaller-is-the best and is calculated with formula 3:
1 Xn
g ¼ 10log yi2 ð3Þ
n i¼1
where: g represents ratio, yi represents value of the quality characteristic of the ith trials,
and n represents number of the observations.
3.4 Shrinkage Values and S/N Ratio

For each of the 27 experiments, three trials were run. The shrinkage for each element
was calculated using Formula 2. The data for each trial with measured and calculated
Table 4. Measured and calculated shrinkage for each experiments

Melt Holding Injection Holding Trial Trial Trial S/N Mean
temp. pressure pressure time 1 2 3 ratio
220 85 120 5 1.333 1.338 1.343 −2.5292 1.338
220 85 130 6 1.672 1.675 1.677 −4.4786 1.67467
220 85 140 7 1.24 1.24 1.244 −1.8778 1.24133
220 95 120 6 1.333 1.339 1.339 −2.5227 1.337
220 95 130 7 1.494 1.497 1.499 −3.5025 1.49667
220 95 140 5 1.768 1.773 1.773 −4.966 1.77133
220 105 120 7 1.386 1.391 1.392 −2.8582 1.38967
220 105 130 5 1.415 1.416 1.417 −3.0213 1.416
220 105 140 6 1.471 1.474 1.48 −3.3759 1.475
230 85 140 5 1.995 2.001 2.001 −6.0163 1.999
230 85 120 6 1.637 1.637 1.638 −4.2827 1.63733
230 85 130 7 2.072 2.075 2.08 −6.3432 2.07567
230 95 140 6 0.985 0.988 0.988 0.11365 0.987
230 95 120 7 1.131 1.133 1.139 −1.0949 1.13433
230 95 130 5 0.87 0.871 0.871 1.20296 0.87067
230 105 140 7 1.073 1.078 1.079 −0.6417 1.07667
230 105 120 5 1.512 1.515 1.518 −3.6083 1.515
230 105 130 6 1.418 1.421 1.423 −3.0499 1.42067
240 85 130 5 2.095 2.1 2.105 −6.4444 2.1
240 85 140 6 1.988 1.989 1.99 −5.9727 1.989
240 85 120 7 1.396 1.396 1.398 −2.9019 1.39667
240 95 130 6 1.485 1.49 1.495 −3.4638 1.49
240 95 140 7 1.099 1.102 1.104 −0.841 1.10167
240 95 120 5 1.22 1.221 1.222 −1.7343 1.221
240 105 130 7 1.79 1.794 1.796 −5.0732 1.79333
240 105 140 5 1.66 1.661 1.663 −4.4091 1.66133
240 105 120 6 1.707 1.71 1.712 −4.6582 1.70967
shrinkage can be seen in Table 4. Using Minitab [6], the S/N ratio, as well as the Mean
was computed. In order to properly compute all the desired values, the analysis using
the Taguchi design should be activated. Under the Start dropdown menu, choose DOE,
followed by Taguchi and Analyze Taguchi Design. Additional functions such as
Standard Deviation and Coefficient of Variation (R) can be calculated, but the Smaller
is Better option must first be picked under the Options drop menu.
4 Results and Interpretation
The response for S/N ratio smaller-is-better is presented in the Table 5. It can be seen
that the optimal combination of parameters used to reduce the shrinkage and to prevent
the jammed component in the assembly process chosen based on the level with a
highest value for each factor. As the code of parameters given in Table 2, the optimal
combination is a2, b2, c1, d3. Delta represents the difference between min-max and
gives the influence of the significant factor for shrinkage. In the presented case, as in
the ranking calculation, the holding pressure gives the first influence, followed by the
melting temperature, injection pressure, and finally the holding time.
Table 5. The response for S/N ratio smaller-is-better

Level Melt temp. Holding press. Injection press. Holding time
1 −3.237 −4.539 −2.91 −3.503
2 −2.636 −1.868 −3.797 −3.521
3 −3.944 −3.411 −3.11 −2.793
Delta 1.309 2.671 0.887 0.728
Rank 2 1 3 4
Fig. 1. Main effects plot for S/N ratios
The graphs with the main effects plot for S/N ratios are presented in Fig. 1, and the
highest values can be seen: for the parameter melting temperature, code “a” in Table 2,
the highest value is the one for level 2, so, a2 will be chosen. The same for the other
three parameters, for the holding pressure is b2, for the injection pressure is c1 and for
the holding time is d3.
4.1 ANOVA Analysis

The Minitab 17 software was used also to perform the analysis of variance, ANOVA,
for the balanced data from experimental design, based on the same number of obser-
vation. The computed data is presented in Table 6:
Table 6. ANOVA analysis

Factors Degree of Adjusted Adjusted F-value Percentage Rank
freedom sum of mean
squares squares
Melt 2 7.724 3.862 1.284 7.629 2
temp.
Holding 2 32.36 16.18 5.378 31.964 1
press.
Injection 2 3.898 1.949 0.648 3.850 3
press.
Holding 2 3.106 1.553 0.516 3.068 4
time
Pooled 18 54.151 3.008 53.488
error
Total 26 101.239 100
5 Conclusions
The research study aim was to find a solution for the problem in the manufacturing
processes of the mechatronic devices where is used plastic injection molding parts in
combination to slide and to create the requested haptic feeling by the customer. Due to
variation in manufacturing and field of tolerance, there can be occurred a bigger play
between the components, and this can lead to a noise and bad haptic feeling or even
smaller play between the components, and this can lead to the jamming and bad haptic
feeling.
The Taguchi method was used in order to find the minimum relevant experiments to
run and to choose the optimal combination to achieve the minimal value of the
shrinkage. Based on the range of adjusting for each parameter, the combination found is:
melting temperature at 230°C, holding pressure at 95 bars, injection pressure at 120 bars
and holding time at 7 s.
The study shows that the influence of the holding pressure is noticeable in com-
parison to the other factors. And this ranking was also supported by the analysis of
variance study which validates the influence of each factor in the form of a percentage,
at the end, the ranking factors show after holding pressure, the melting temperature, the
injection pressure and holding time, but with less influence.
The result helps to prevent bad assemblies in the mechatronic devices which use the
plastic injected molding parts, due to shrinking of the components. And the validation
of the injection molding processes could be made with less effort, less cost and quick
measurements by taken samples from the line at the start of the production and at the
established period, as soon the parts are out of the mold.
References
1. Platform Industrie 4.0. http://www.plattform-i40.de/I40/Navigation/EN/Industrie40/
WhatIsIndustrie40/what-is-industrie40.html
2. AT Kearney. http://www.atkearney.ro/sustainability/ideas-insights/article/-/asset_publisher/
LCcgOeS4t85g/content/plastics-the-future-for-automakers-and-chemical-companies/10192
3. Confirmation Haptics for Automotive Interfaces. http://www.immersion.com/wp-content/
uploads/2015/10/CUE-whitepaper_jun13v1.pdf
4. Ranjit, K.R.: A Primer on the Taguchi Method, 2nd edn., pp. 5–14. SME, Michigan (1990)
5. Singh, T, Singh, M.P., Alam, M.M.: Taguchi and ANOVA analysis of shrinkage of injection
moulded polypropylene Component. IJSER (2014)
6. Minitab. https://www.minitab.com/en-us/products/minitab/
7. Taguchi, G.: Introduction to Quality Engineering. Mc Graw Hill, New York (1990)
8. Roy, R.K.: Design of Experiments Using the Taguchi Approach: 16 Steps to Product and
Process Improvement. John Wiley & Sons, Inc. (2001)
9. Lal, S.K., Vasudevan, H.: Optimization of Injection Moulding Process Parameters in the
Moulding of Low Density Polyethylene (LDPE). IJSER (2013)
10. Ionescu, C.N., Nicoara, I., Gruescu, C.: The Integration of Classic and Taguchi Experimental
Design into the Projection and Optimization of the Process Parameters in the Controlled
Laser Processing of Polycarbonate Cogwheels, UEM, Resita (2008)
11. Unal, R.: Taguchi approach to design optimization for quality and cost: an overview. In:
Annual Conference ISPA (1991)
Key Characteristics of the World Class
Manufacturing Concept in the Production
of Chassis for Buses Industry
Luminiţa Ştirbu1(&), Cosmin Știrbu2, Nadia Belu2,

and Viorel Nicolae2
1
Piroux Industrie Romania, Mioveni, Romania
luminita.stirbu@piroux.com
2
{cosmin.stirbu,nadia.belu,viorel.nicolae}@upit.ro
Abstract. This paper presents the approach of the implementation of WCM in

the chassis production in the bus industry.
To believe in the merits of WCM (and thus for WCM to be a part of their
work), shop floor operators need to experience the benefits (one – time on site
experience, not from any books). The effects of WCM are best felt when
everyone is involved. Not working harder, but working smarter, not being
obstructive but co-operative, not the customer has to choose what we produce,
but “we produce what the customer wants”. In addition, the customer becomes
more and more demanding. He wants more options, the lowest price and he
wants it now. That is why we cannot survive without WCM, because only if we
make the most of the best methodologies and techniques at our disposal in
balanced coherence, are we able to approach the ideal as much as possible.
Which manufacturing company would not want that?
For chassis plant from the bus industry, the latest WCM strategy is a strong
focus on education.
Keywords: World Class Manufacturing (WCM) Chassis production Buses

industry TPM
1 Introduction
World Class Manufacturing [WCM] is a global term for the most effective method-
ologies and techniques to give to our clients what they want: quality, delivery and cost
effectiveness [1–6].
The WCM methodologies and techniques that are known throughout the world are
TPM, EFQM, Kaizen, TQM, Six Sigma, JIT, and Lean Manufacturing.
Hence, WCM refers permanently to the ideal situation: ownership and the
non-acceptance of losses. Is WCM promising us an EL DORADO? Concern, on this
point we have to go back to reality from our theoretical ideal. World Class Manu-
facturing is just as physically strenuous, complicated, and demanding as any other form
of production. However, it is much more effective because of its completeness,
transparency and harmonious coherence between people and processes, between

586 L. Ştirbu et al.
methodologies and techniques. Therefore, WCM is more successful, more enjoyable,

and healthier for human beings and companies [1–6].
We do not have to work harder; we have to work smarter. Nowadays it’s mandatory
to be co-operative, and we cannot waste the time of others, nor block them. We
produce what the customer demands. In addition, the customer becomes more and
more exigent. He wants more options, the lowest price, and he wants it now. An old
dictum says: “There’s no time to rest content with half measures”. That is reality at the
beginning of the 21st century. That is why we cannot survive without WCM, because
only if we make the most of the best methodologies and techniques at our disposal in
balanced coherence, are we able to approach the ideal as much as possible. That is what
the customer wants. World Class Manufacturing; which manufacturing company
would not want that? [1–6].
WCM begins with an ideal situation; this means that involved employees always
have the production processes run without losses. Then, one assesses the ideal vs. real
situation and second focuses on differences between the two processes. This difference
is called loss. WCM aims to eliminate this loss. So, in conclusion based on WCM:
• One learns not to accept losses
• Creation of ownership, meaning that people carrying out a certain production
process feel ownership of this process and initiate improvements, as well as
implementing them.
Thus the four main characteristics of WCM are successively:
• Making losses visible
• Improving in team format
• Organizing process-oriented
• Standardizing working methods.
The most important characteristics of WCM are [1–6]:
• WCM focuses on people who are part of the operation. Together with them losses
are traced and made visible.
• WCM forms teams to find and eliminate the cause of a certain loss, for only by
removing the cause of a problem, will the problem stay away forever. Solutions
achieved by curing the symptoms and not removing the cause are not real solutions.
WCM is focused on product organization. Instead of dividing the work over many
specialized departments, WCM organized in a process stream oriented manner.
There is, ideally, a process organization every product-market-combination of your
company knows its own team, to which activities adding to the total value are
assigned, as well as the support services directly and exclusively related to them.
• The fourth and last main feature is the permanent standardization. In addition, the
‘process owners’ should take this step so that we can verify that the step is under
control [1–6].
Key Characteristics of the WCM Concept in the Production of Chassis 587
2 World Class Business Process Management
We are used to saying that the most boring and repetitive work is done in factories.
This is not always true. It is also true that in offices monotony kills enthusiasm and
damages productivity. Office productivity can be measured as a component of client’s
satisfaction, either externally or internally [1–6].
Many companies waste a lot of time filing useless information or obsolete data.
How much time is wasted when a 10-page report can be easily summarized in 3 lines?
How many information shuttles from one office to the other for “competence conflict”
with overlapping roles and responsibilities? [1–6].
The system to avoid all this exists. There is a methodology which improves effi-
ciency and effectiveness, working simultaneously on 5 key-factors:
• work quality
• enhancement of employee skills
• improved information flows and management
• clarified roles and procedures
• more effective information technology.
Through 3 intervention phases (Identify-Quantify-Rectify) – without creating
friction, or conflicts, or dissatisfaction in the office – it is possible to:
• eliminate useless or extremely repetitive tasks, reducing stress and routine
• define processing standards that employees can gradually improve
• increase employee skills thereby enhancing their abilities and talents
• win against bureaucracy (“we always have done it this way…”), promoting inno-
vation and rewarding those employees that are most creative [1–6].
Applied with rigor and thoroughness, this method enables the Human Resources
function to express all their skills, using the best information technology in a friendlier
and enjoyable environment [1–6].
3 Technical and Managerial Pillars of WCM
The WCM system is made up of ten technical and ten managerial pillars, illustrated as a
temple (see below). The ten technical pillars are as follows:
• Safety (Occupational safety)
• Cost Deployment (Distribution of Costs)
• Focused Improvement
• Autonomous Maintenance
• Workplace Organization
• Professional maintenance
• Quality Control
• Logistics & Customer Service
• Early equipment Management
• People Development
• Environment (and Energy) [1–6].
The ten managerial pillars are:

• Management Commitment
• Clarity of Objectives
• Route map to WCM
• Allocation of Highly Qualified People to Model Areas
• Commitment of the Organization
• Competence of Organization towards Improvement
• Time and Budget
• Level of Detail
• Level of Expansion
• Motivation of Operators [1–6].
The Cost Deployment pillar is one of the most important among the discussed
concepts. To arrive at a proper Cost Deployment, you need a team from accounting,
finance, and operations. Cost Deployment is an accounting technique for assigning
actual costs to each loss and waste that happens in a factory. Based on this cost
deployment, the prioritization of which loss to attack first can be made with economical
reasoning. A supplementary advantage of Cost Deployment is that all improvement
work in the organization is assigned an equivalent saving potential. This motivates
further improvements and is the best argument for convincing people who have a lot of
reservation to change and prefer to stay suspicious and not confident [1–6].
Another important concept of WCM is that change always starts with a model area.
The model areas are pilots for the implementation of the principles. For example, the
plant typically chooses the worst performing machine as a model machine for the
Autonomous Maintenance pillar. Through a dedicated project, using WCM tools and
techniques, this model machine is “brought back to basic condition” and made the best
performing machine in the plant. The learning points and good practices are thereafter
shared with the rest of the plant. This is, however, a challenging way to implement an
“excellence area” that leads to small improvement for the total performance of the
plant. The good practices are spread overall [7–9].
The “concept of zero” is also a very important notion in WCM. The target of WCM
is zero waste, zero defects, zero breakdowns and zero inventory. The model areas
should prove achievement of zero for several weeks before solutions are spread [7–9].
WCM is not a static never-changing improvement program. In 2010, an Energy
sub-pillar was introduced in the Environment pillar “to improve the ability to identify
and implement measures to reduce waste and achieve greater energy efficiency” [7–9].
4 Appling WCM in the Buses Industry
Achieving world class manufacturing status is a major attempt (or milestone) to most
organizations. This requires companies to do exceptionally well (or at least performing
better than their competitors) in most competitive advantages - quality, time, cost,
flexibility, technology, innovation, etc. The following framework outlines the required
techniques and the roadmap for achieving world class manufacturing in a plant in the
chassis industry (Fig. 1).
Fig. 1. Roadmap for achieving world class manufacturing in a plant in the chassis industry
In the past, Manufacturing has been perceived as a pure cost generator. The only
task of manufacturing management was to be able to reduce production costs by 2,5 %
per year. Nobody is arguing now that manufacturing is a value creating core of the
Company.
This is due to the fact that world leaders are able to compete thanks to reliable and
adaptable processes.
For example, you have a route to transform a plant into a source of competitive
advantage following 3 stages:
• Restore basic conditions of a reliable production flow through reduction of
variability
• Improve competitiveness through process cost reduction
• Achieve competitive advantage through a global performance improvement of the
manufacturing
The first step is focused on an intensive “Loss Intelligence” phase: by linking strategic
goals to manufacturing performance, it is possible to define the improvement priorities to
be attacked. In fact the deployment of Volume, Quality, Cost, Delivery, Human losses
defines, quantifies, and prioritizes improvement opportunities. The second step is devoted
to learning and test the methodology by attacking losses with pilot teams. Pilot teams
work on a detailed “road map”, and thanks to Fast Track Team routes they have an
excellent payback [10–12].
The third step is designed to create the “Learning Organization” in order to support
the expansion of the approach through-out the Company. We involve management in
“pillars” structure to replicate loss intelligence and train, support, audit team activities.
For our Chassis plant from buses industry, the latest WCM strategy are based on:
• TPM, originally the floor was dirty with metallic powder, chemicals and nobody
understood where the cause defects came from. So operators started cleaning the
floor and developed a bottom-up cleaning practice, which gave the management an
incentive to deploy a plan involving QRQC (Quick Response Quality Control)
meetings and then TPM;
• strong focus on education by empowering shop floor staff with training to solve
problems on the line, that, this bottom –up approach was complementary to a top –
down deployment of Kaizen and TPM across the company;
• job rotation at shop floor operator level, or when there is a shortage of workers on a
given large projects, others will be assigned to help; we minimize operators mistake
in the repetitive operations, if there are persistent problems, they will rotate
operators;
• quality is a key, but productivity and efficiency comes from operators who have
fine-tuned senses to detect defects; the key factors is the operator’s experience, so
we are attaching a lot of experience in operator training.
5 Conclusions
This paper describes the beginning of the implementation of WCM in a chassis plant
from the bus industry and demonstrate that the effects of WCM are best felt when
everyone is involved.
The benefits of WCM are, by nature, synergistic (the more people are involved, the
stronger the benefit, which is why Japanese –style teamwork is so important). The
purpose of WCM allow to a dream to become reality: not working harder, but working
smarter, not being obstructive but co-operative, not heaving the customer choose what
we produce, but “we produce what the customer wants”.
The WCM practice is an integral part of the operator’s job (not an “extra” nor a
separate chore to complain about).
The acquisition, understanding and application of WCM knowledge is an empirical
process.
The WCM practice works best when employees are self-taught and self-driven
(what is the role of the foreman/section manager/WCM champion?). Thus repeating the
WCM process will lead to continuous improvements.
To believe in the merits of WCM (and thus for WCM to be part of their work), shop
floor operators need to experience the benefits (one-time on site experience, not from
any book).
A top down approach to the implementation of WCM does not work well.
References
1. http://www.makigami.info/world-class-manufacturing/
2. http://www.docfoc.com/what-is-world-class-manufacturing
3. https://www.linkedin.com/pulse/20141210005634-47919295-world-class-manufacturing
4. https://www.linkedin.com/pulse/world-class-manufacturing-wcm-sachin-dhawale?
forceNoSplash=true
5. http://better-operations.com/2013/05/22/world-class-manufacturing-at-chrysler-and-fiat/
6. http://net1.ist.psu.edu/chu/wcm/frame.htm
7. Rother, M., Shook, J.: Learning to See: Value Stream Mapping to Add Value and
Eliminate MUDA. The Lean Enterprise Institute, Inc. One Cambridge Center, Cambridge
(2003)
8. International Trade and Japan: Annual White Paper on International Trade and Foreign
Direct Investment by JETRO. http://www.jetro.go.jp/en/stats/white_paper/
9. Japanese Market Trends: In-depth Japan market reports covering a wide range of products,
services and industries. http://www.jetro.go.jp/en/market/reports
10. Murase, Y.: Challenge towards World Class Manufacturing Course, Japan Institute of Plant
Maintenance (2012)
11. Terada, A.: Fundamental Secret of LEAN Implementation JMAC Consultants Inc.
12. Shingo, S.: A study of the Toyota Production System. Productivity Press (1989)
Advanced Automotive Assembly Line Trends
as Tools in Optimizing Production Line
Performance
Hariton Poparad(&)

hariton.poparad@unitbv.ro
Abstract. This work proposes an overview of current and future automotive

assembly line methods, highlighting challenges facing manufacturing executives
to increase efficiency and optimize the assembly line output. Increasingly,
assembly line manufacturing is characterized by “concurrent processes”—
multiple parallel activities that feed into a final assembly stage, rather than a
single large system. These processes require sophisticated communication sys-
tems, material flow plans, and production schedules. This paper is focused on
the methods and processes implemented to increase the speed and accuracy
needed to meet competition challenges and strengthen the overall manufacturing
process. The conclusions underline innovative present trends on automotive
assembly line methods which reduce waste, speed up automated processes, and
allow for high quality products.
Keywords: Manufacturing process Assembly line Production line
1 Introduction
The technological advances of the 21st century have brought major changes to the
practice of automobile assembly, due to the need to comply with environmental
restrictions and infrastructure regulations. The imposed regulations lead to the demand
in innovative solutions relative to vehicle design, validation, and mass production
activities. Such changes have led to the development of the electric car and the self
driving car concepts, as viable alternatives to the combustion engine.
For example, the electrical automobile incorporates a multitude of traditional
engineering elements of design and build, as well as previously unconventional tech-
nologies such as aeronautic elements of steering done through extensive wiring,
regenerative brake, satellite communication, and radar surveillance. Such leaps in
technological advances have been made possible through the extensive research,
development and investments into the transformation and optimization of assembly
lines to meet business and consumer demands.
Assembly factories perform the final stages in the production process of the
automobile. The needed components are previously designed, purchased, and shipped
from external suppliers or produced internally by specialized company divisions; these
sub-assemblies and systems can add up to more than 4,000 components. The next step

Advanced Automotive Assembly Line Trends as Tools 593
involves the assembly of the major structural components (underbody, body side, roof),
building a semi-assembled state of the product.
From there, the assembly line is seen as a progressive repetitive action, in which
standardized, interchangeable components are added to the product in a
semi-assembled state, which is then transported in specific sequences from one station
to another in order to obtain the finished product.
2 Configurative Variations of the Assembly Line
The diversity of the assembly line procedures is the direct result of technological
improvements, careful planning of production, and specific assembly conditions and
strategies established by each company, considering [3, 6, 8]:
– Limited resources and investment capital;
– International competition;
– New legislation dictating operating procedures;
– Human resources and supply base availability.
Through this technological process, the automobile in its finished form is assem-
bled more efficiently by cutting down time, lowering cost, and exerting less effort
compared to an assembly process in which operators transport the components to be
assembled in a single cell, static work station.
The assembly lines are designed for specific operations, and the workers trained to
execute tasks in technological sequence. Efficiency is established by specialized con-
veyors, which transport the components in a semi assembled state to specific stations,
thus eliminating the need for manual operation and worker travel.
The new assembly line methods have been introduced to optimize productivity and
efficiency in the assembly process. The progress made in the advancement of such
techniques has additive benefits, which leads to and end product that is more refined
and complex from a production standpoint. The assembly is sustained through con-
current processes, which are parallel lines that supply the final line of assembly [2, 5].
The assembly line process needs complex communication systems that can pro-
gram and monitor the components as they move from station to station. The assembly
line can be categorized as a single, complex mega-system, and a single stop in the
process results in halts and major disruption of production [7].
Maximum coordination is required to keep the entire assembly line system and its
components moving smoothly. The use of computers has allowed the system to be
monitored, controlled, and to perform more sophisticated functions - actions which
have made the reduction in stock cost of components associated with production
possible.
Figure 1 shows the generic organizational flowchart of automobile assembly. The
method in which assembly line operates to produce the final product in a set time is also
shown. The major assembly concepts are individually highlighted, and we can see the
activities outside suppliers are responsible for in clear, while the activities done right on
the assembly line are depicted in grey.
594 H. Poparad
Fig. 1. Generic organizational flowchart of automobile assembly
Just in time assembly has become a popular technique due to the necessity of
accurately planning and cost reduction of components needed during production [2].
This technique allows the assembly plants to have the stocks of semi-assembled
components needed for production just for a couple of days, creating a continuous flux
of production and assembly. This is possible due to the flexibility in production that
suppliers have and their higher involvement in adapting to the assembly plant’s daily
needs and prognosis, thus making the entire process more efficient and dynamic.
Future development trends include components produced by suppliers directly inside
the assembly plant, and improving electronic communication between the plant and its
suppliers, in efforts to more efficiently troubleshoot any potential problems that arise.
The current practices of major automobile companies, based on work principle,
(GM, Ford, Daimler, Fiat-Chrysler, Toyota, etc.), are [1, 5, 6, 9].
2.1 Team Based Assembly

Team based assembly is an advanced technological process that allows flexibility for
the operators who work alone or in a team and perform repetitive tasks. Through team
based assembly concepts, the operators have the possibility to move along the line and
perform additional tasks such as identifying, adjusting, and even correcting possible
problems that may arise during the final inspection.
This method requires increased operator responsibility and knowledge of assembly

process details, which is a useful quality control tool.
2.2 Modular Assembly

The scope of the modular assembly line is to improve productivity, and it does so by
using auxiliary parallel lines as shown in Fig. 2. This system is applied in automobile
industry, and it requires the assembly of complex modules (i.e. chassis structure,
suspension, steering, and body) on independent lines, while their complete assembly
occurs on the final assembly line.
Fig. 2. Suspension modular assembly line [1, 8]
2.3 Cell Assembly

The concept of cell assembly is the result of the development of automation that is both
functional and technological. Through this method, robots manipulate parts and per-
form the welding or fastening operations, while the operators perform 3 or 4 different
auxiliary tasks. The cells can be operated by one or multiple workers, based on their
complexity.
2.4 “U” Shape Assembly Line

Arranging all the assembly stations in a continuous line, is not always the most efficient
way of organizing the workflow form a technological standpoint. However, by
arranging the assembly line in a “U” shape Fig. 3, allows station operators to perform
their tasks within it. Additionally, the operators can watch the assembly process and
ensure its smooth progression by having an easy view of the upcoming components.
The “U” shape assembly is flexible enough to allow operators to prepare additional
operations for the assembly stage.
596 H. Poparad
Fig. 3. Steering module, “U” shape assembly line [2, 6]
This concept allows performing technological operations with ease, improve

communication between assembly line operators, and shorten the time needed to set up
the lines for operation.
In current practice, most assembly lines in the automobile industry are semi-
automatic systems, along which the product moves along to be assembled. At every work
station there are specific assembly operations, with a stationary operator and tools.
Every automobile producing company has developed its own techniques of design
and organization assembly lines, to produce competitive, high quality, cost saving
products.
3 Advanced Automotive Assembly Line Trends
The most intense work phase in the assembly line process is the final stage assembly,
demanding nearly 50 % of the effort needed to assemble the whole automobile [2].
Since the final assembly is the critical point of the entire assembly process, the oper-
ators and union leaders can have a major influence upon it, and use it as a bargaining
tool. Consequently, there is a growing desire to eliminate the human factor through a
complete automation, an unrealistic model at the moment. Thus far, an organized blend
between manual and automatic operations renders the best results.
Another key assembly trend is to further optimize logistics, since their role in
sustaining the assembly process by compiling data and record keeping of activities of
the suppliers as well as those of the assembly plant is integral. The importance of
optimizing logistics is shown directly in the company’s immediate performance results,
and taking in account the following objectives provides a more controlled handling of
the system:
– Reducing unused stock and the capital that is tied in it;
– Improve the quantity of available components as planned for production;
– Decrease of door-to-door time (time needed to get components from suppliers);
– Improve the working conditions and interactions between assembly line operators.
The measurement of the ratio between the components needed to sustain the
assembly process and the finished product is an important tool aiming to improve the
system’s logistics by providing information about just-in-time part supply and that of
optimizing the assembly through lean production [3, 4, 9].
Trends for logistics and supply chain work optimization include:
– The suppliers of components and subsystems will continue to receive responsibility,
as the demand for ready to be assembled components increases.
– The use of sensors and image processing systems are extensively employed, as
efforts to eliminate subjective decisions making by human operators and to assure
the outmost quality in the assembly process.
– The use of automated guided vehicles with adjustable height that allow for the
adjustment in the work envelope during the assembly process. These automated
vehicles serve as both assembly and transport platforms.
– The use of long continuous assembly lines has been replaced with shorter segments.
This makes the assembly process easier to be supervised, as operators can identify
possible problems from other stations.
– Systems with flexible mechanics are needed due to the large numbers of automobile
models variants produced and the fast changes that need to be made throughout the
assembly process.
– In current practice, the large automobile manufacturing companies instate a 2–4
week assembly plant shut down to prepare the transitions to the new and improved
versions of the models. During the shut down process, assembly line maintenance
occurs with tool recalibration, and integration of new work stations. This process
requires careful planning and synchronization between the suppliers of components
and assembly line production managers.
4 Conclusion
– The extensive use of computers to conduct the planning, controlling, and moni-
toring processes of the assembly line plays a critical role in the optimization of
resources, time allocation, and the communication of information and trou-
bleshooting of the system.
– The concepts of smart factories and robotic factories are gaining popularity, and
some companies are opting for complete robotization in their factories. Such fac-
tories require human oversight of overall operations only, since most work pro-
cesses are controlled and operated by robots entirely.
– Utilizing robots equipped with integrated vision technology which automatically
halts the robot’s operation on physical contact with human operators is a key feature
for the future that ensures a safe working environment.
598 H. Poparad
– The assembly line is a critical component of the production process, and maintain it
to run continuously impose maximum coordination between the system and its
components.
– The balance between human factors that may affect production and automation is a
major player in the optimization of the assembly line and overall success of the
company.
– Research projects are in progress to incorporate cloud technology into robotics,
which will allow for automatic real-time transfer of data into cloud platforms.
References
1. Croci, F., Perona, M., Pozzetti, A.: Work force management in automated assembly systems.
Int. J. Prod. Econ. 64, 243–255 (2000)
2. Dove, R.: Custom lines - built just in time. Automot. Manuf. Prod. (2010)
3. Hounshell, D.: From the American System to Mass Production. Johns Hopkins University
Press, Baltimore (1994)
4. Kalpakjian, S., Schmid, S.: Manufacturing Engineerig and Technology, 7th edn. Prentice
Hall, New Jersey (2013)
5. Maloney, D.: New Roots, Modern Materials Handling, October 2003
6. Nieble, B., Freivalds, A.: Methods, Standards, and Work Design, July 2002
7. Whitfield, K.: Assembly: how standard can you get? Automot. Des. Prod., March 2004
8. Umble, M., Gray, V., Umble, E.: Improving Production Line Performance IIE Solutions,
November 2000
9. Zylstra, K.: Lean Distribution, Applying Lean Manufacturing to Distribution, Logistic, and
Supply Chain. John Wiley & Sons, New Jersey (2012)
Cleaning Methods for Flux Pollution
Measurement in Automotive Coolant
Loop Components
Hervé Bourgeais and Claire Demarcq(&)
Valeo Thermal Systems, La Suze-sur-Sarthe, France

{herve.bourgeais,claire.demarcq}@valeo.com
Abstract. For the worldwide technology of heat exchangers, Nocolok® flux is

used to remove aluminium oxide during brazing under controlled atmosphere.
This flux is known to react with corrosion inhibitors from coolant by creating
silicate paste. Coolant flows in heat exchangers: several temperature cycles,
coolant ageing, inhibitors consumption will enhance this phenomenon. By this
way, heat exchangers are losing thermal performance. That is the reason why car
manufacturers ask to reduce residual flux amount. This presentation deals with
different existing methods to evaluate heat exchangers cleanliness: how can be
measured the residual flux amount? Different processes used by Valeo to reduce
the initial flux amount will be also presented. The main difficulty of this study is
to minimize the flux quantity while maintaining a good brazing quality. Flux
amount reduction cannot be done by sacrificing heat exchangers robustness and
reliability. Customer’s specifications are staying at a high requirement level.
Keywords: Heat exchangers Aluminium alloys Flux reduction

Cleanliness Flux deposit
1 Introduction
Valeo Thermal Systems is dedicated to produce heat exchangers for automotive

industry. In La Suze sur Sarthe, we develop and produce heater cores for the
cabin heating. Flux (mixture of K, Al, F form) is used for Controlled Atmosphere
Brazing1 process to:
• Penetrate into the oxide layer (as it has extremely low surface tension)
• Dissolve the natural oxide (Al2O3 or Al(OH)3) present on aluminium surface
• Enhance surface wettability
• Enable filler metal to move
• Protect surfaces from oxidation during brazing cycle
As flux is non corrosive, internal and external residues are staying on the surfaces
after the CAB process. Residues seen with Scanning Electron Microscope2 are
1
CAB (Controlled Atmosphere Brazing) will be used in the rest of the document.
2
SEM (Scanning Electron Microscope) will be used in the rest of the document.

600 H. Bourgeais and C. Demarcq
Fig. 1. Flux residues with SEM at 1010 and 4020 magnification [1]
presented on Fig. 1. They can have needles or square shape (needles shape for post
brazed residues: KMgF3, and square shape for K, Al, F post brazed residues) [1].
During brazing, flux is submitted to different reactions as explained in the Fig. 2,
due to the temperature change from ambient to 600°C. At the end of brazing process,
the formation of the compound K3AlF6 is responsible for the solubility of the flux into
liquids.
Fig. 2. Reactions of flux during brazing process
Disadvantages of internal flux residues are that they can react to coolant additives,
such as corrosion inhibitors, creating gel and hydroxides formation. Corrosion inhi-
bitors consumption will also increase the cylinder block aluminium corrosion creating
aluminium oxide. This will lead to channels plugging, as seen in Fig. 3 [2], then
performance decreasing of heater core or radiator and warranty return as this reaction
appears generally at least one year after the first circulation date.
To avoid heater core plugging and to protect the coolant loop, car’s manufacturers
request to decrease more and more the residual flux amount. In the Table 1, are
presented some requests from our customers.
Cleaning Methods for Flux Pollution Measurement 601
Fig. 3. Plugging channels in a heater core
Table 1. Request of car’s manufacturer for flux residue rate

Car manufacturer Rate according to year
BMW 1 g of K+ 0,7 g of K+ in 2012
in 2008
VW-Audi No request For MQBa 7,3 mg/l * 3,5 Wish in 2016: 25 %
in 2011 l = 26 mg of F– in 2016 decrease versus
2014
Daimler-Chrysler No request before 2016 0,3 g/l of K+ in 2016
Ford 0,3 g/l of K+
a
MQB is a vehicle platform for Volkswagen (VW).
2 Cleaning Methods
Some car’s manufacturers have their own methods to clean exchangers, others do not
have any. That is the reason why, Valeo has developed its method: PDTNVX15029
[3]. The goal to develop one single method for all exchangers is to be able to compare
different technologies of exchangers between them, to compare different processes and
also to compare the improvement of one process.
In the Table 2, different cleaning methods are summarized, and the Figs. 4, 5 and 6
show the principle of the cleaning to collect flux residue.
Volume of test bench for Valeo method depends on the size of the exchanger. 8 L
is for the smallest one, and is particularly used for heater cores.
Table 2. Synthesis of cleaning methods to collect flux residue

Method VW [4] Valeo [3] MBNa [5]
Product rinsing time (hours) 24 17 20
Rinsing volume Demineralized water (or
desalinated water)
Flow rate (l/h) 480 400 – 5000 No flow
Pressure (bar) Patm. 1,1 ± 0,5 Patm.
Temperature (°C) 80 90 ± 2 90 ± 2
Test bench volume (l) 3,5 8 – 15 ≈0,5
a
MBN for Mercedes-Benz Norm.
1. Test bench
2. Container
4 3. Pump
2 5 4. Piping
6 5. Heater core
1 6. Thermometer
Fig. 4. Valeo test method for heater core cleaning to collect flux residue
The complete volume of deionized water circulates for 17 h in the heater core at the
temperature of 90°C by using a pump and a themostatic bath. The temperature is
automatically controlled by a digital thermometer. After 17 h, the heater core is
emptied in the container and the rinsing solution is removed from the test bench and
naturally cooled at room temperature. Then it is shaked to be well homogenized. The
cleanliness of the test bench is insured by measuring the conductivity of the “residual
water” which may remain in the bench. This one must be lower than 20 µS/cm and a
blank is kept to measure K+ content before performing a new test. Moreover, the
quality of deionised water must insure a conductivity lower than 5 µS/cm.
The duration of 17 h has been demonstrated by our colleagues from Skawina on
radiators, with a safety coefficient of 2,5. On the Fig. 5 they have recorded the con-
ductivity of the rinsing solution versus time. After 7 h, the conductivity remains stable
and is directly linked to the ions concentration, and particularly K+.
In VW method, as presented in Fig. 6, the rinsing liquid (desalinated water) passes
through a thermostatic bath at 80°C before going inside the heat exchanger. Then the
rinsing liquid is collected into a polyethylene reservoir which is unique for each test.
Fig. 5. Record of one radiator measurement: λ = f(t)

Fig. 6. VW test method for exchanger cleaning to collect flux residue
After 24 h of circulating test, the whole system is emptied in the reservoir and cooled
naturally to room temperature. The main disadvantage of this cleaning method is its
duration and that the system is not submitted to some pressure to allow collecting
“sticky” residue present on internal surfaces.
In MBN method (see Fig. 7), the heat exchanger is completely filled up with
demineralized water and an expansion reservoir is put on it to collect liquid in case of
volume expansion during the test. It is useful to have a small hole on the top of the
reservoir so that the test is performed at ambient pressure. This complete assembly is
put in an oven at 80°C for 20 h.
Fig. 7. MBN test method for exchanger cleaning to collect flux residue
The main disadvantage of this method is that there is no circulation of the rinsing
liquid, this is less efficient, because complex internal surfaces cannot be reached.
The main advantage of Valeo method is to standardize the residual flux quantifi-
cation method for CAB heat exchangers. It enables to compare different technologies
and different products, without taking into account which customer the product is
dedicated to.
3 Measuring Methods
After cleaning procedure, a sample of rinsing solution is collected. Generally the sam-
pling corresponds to 250 ml of rinsing solution. According to methods used and pre-
sented in Table 3, potassium is determined by ICP3 and fluoride is determined by IC4.
Table 3. Methods used to measure flux residue

Method VW [3] Valeo [2] MBN [4]
Measurement IC on F– ion ICP on K+ ion ICP on K+ ion
Both assumptions are made:

• the value of flux residue is well below the flux solubility limit (around 4 g/l)
• post brazed flux residue are made of 85 % KAlF4 and 15 % K3AlF6
The main choice done to measure internal flux residue is to determine the K+
content into the rinsing solution and then to approach the value of flux residue. It is not
recommended to use the fluorides concentration to estimate the residual flux level in the
heat exchanger, due to the fact that this element is not completely dissociated in water.
In general it forms some complexes that hide the presence of some fluorides. By this
fact, the analysis of the rinsing water does not show the real concentration of this
element.
4 Flux Deposit Processes
Valeo uses different processes to deposit the flux linked to history of heater cores and
improvements done on new product or process developments:
• Dipping process: Parts are assembled and totally dip into flux slurry at a deter-
mined percentage of dry flux. The process is presented in the Fig. 8. After dipping,
parts are blown to control flux residues quantity on the core and then dried in an
oven to remove water. The main advantages of this process is to access to complex
surfaces.
• Spraying process: Assembled parts or components go under a shower of flux slurry
at a determined percentage of dry flux. At the end of the process, parts are heated to
be dried and to remove completely water. An example of spraying process is
presented in Fig. 9. The main advantage of this process is to reach hidden areas and
the application is easy to set up.
• Paint fluxing process: Flux is put down in a liquid format on components through
nozzles by spray technique. The process is presented on the Fig. 10. The solution is
3
ICP: Inductively Coupled Plasma, method used to measure element concentration in liquid with
optical spectrometry or mass spectrometry detector. [6].
4
IC: Ion Chromatography [7].
Heater core
Fluxing tank
Fig. 8. Dipping process to apply flux in and on exchangers
Nozzles
Exchangers
Fig. 9. Spraying process to apply flux on external surfaces of exchangers
Fig. 10. Paint fluxing process mainly used for components before heat exchangers assembly
composed of water, flux and a binder with organic additives to insure the viscosity
of the liquid. After paint fluxing, parts are dried in an oven (continuously process) to
evaporate water. It enables to have parts which are easy to handle as flux is well
fixed on the surface thanks to the binder.
Fig. 11. In situ fluxing process on TAHC (TAHC: Thin Aluminium Heater Core) tube
• In situ fluxing process: Flux is put down in a paste format (flux with solvent basis)
on components with a wheel by rolling deposit. As it is presented in Fig. 11, the
wheel under the aluminium coil in U tube shape is dipped in the flux paste and by
rolling and advance of the aluminium coil, the flux is deposited on the aluminium
tube and forms a “flux ribbon”. This process enables to apply flux where it is
necessary (areas to braze). After the deposit, there is a natural drying at ambient
temperature to let the solvent evaporate.
5 Results
First results presented in Table 4 show K+ and F− content obtained on two standard
products: CORION5 heater core (manufactured since 1999) and TAHC (manufactured
since 2005). Fluxing process for CORION is dipping. For TAHC different fluxing
processes are used: in situ for the tube, paint fluxing for the header and spraying for the
external surface. K+ is evaluated according to Valeo method and F− is evaluated
according to VW method.
Table 4. Actual situation of series products

Product K rate average (mg) F– rate average (mg) VW Specification F– rate (mg)
+
CORION 195 89 22a

TAHC 94 38 26
a
The original specification from VW is expressed in mg/l per meter of coolant-side brazed
seam length, corresponding to the tube length for a simple folded tube. VW has the most
severe request for internal flux residue.
For both products, the specification is not respected. That is the reason why
improvements have been done on TAHC serial process by decreasing the rate of
internal flux in the tube. These new results are presented on the Table 5.
5
CORION: Cross counter flow One row Return Integrated One shot Nocolok®.
Table 5. Comparison between serial and improved situations

Product K+ rate F– rate VW Specification F–
average (mg) average (mg) rate (mg)
TAHC serial state (38 % of flux 94 38 26
in solvent)
TAHC improved state (20 % of 46 17
flux in solvent)
Another way to estimate the internal flux residue is to measure directly the internal
quantity of dry flux present in the tube and before assembly and brazing (this simple
test is performed by scratching and weighing). Several measurements have been done
on tubes according to different ratios of flux in solvent and the results average is
presented in the Fig. 12.
Fig. 12. Dry flux density related to ratio of flux present in the solvent (flux slurry)
On the curve, it is clearly obvious that the point at 14 % is not stable. This ratio is
not mastered on the tube manufacturing machine. If this point is removed, the curve
changes and a linear relationship between dry flux density and ratio of flux deposit is
obtained with a correlation coefficient of 0,9986 as it is presented on Fig. 13.
Fig. 13. Dry flux density related to ratio of flux present in the solvent without point at 14 %
6 Discussion
In addition to measurements of internal flux residual, of dry flux density, mechanical

tests and brazing quality of heater cores have been evaluated according to Valeo
functional specifications. For metallographic analysis, brazing quality is conform for
38 % of flux in solvent and for 18 ± 2 % of flux in solvent. For mechanical resistance,
evaluated with burst test and cycled pressure test, both of these flux ratios are conform.
The improvement way in this article is clearly dedicated to the internal fluxing
process. Different studies, made in this domain, have shown that the external fluxing
contributed only for 10 % of the internal flux residue; the other 90 % come from
internal flux deposit.
7 Conclusion
The main advantage of Valeo cleaning and measuring method to evaluate the internal
flux residue is to propose a common method for all heat exchangers and all tech-
nologies. By measuring the potassium rate instead of fluoride rate, the real level of
internal flux residue can be reached Moreover the method is independent from cus-
tomer functional specifications. It enables to compare all products between them and to
determine improvements done at process level.
According to different results presented, Valeo has made the choice to decrease the
flux ratio in tubes from 38 % in solvent to 18 ± 2 %. This new ratio enables to be in
conformity with the request of one of the most demanding customers for the internal
flux residue; this ratio is also mastered by the process and evaluated for all other
characteristics of the product.
This study will be continued by the establishment of a relationship between the dry
flux density present in the tube and the real internal flux residue present in heater cores
and measured according to Valeo method PDTNVX15029. As the reaction between
flux and aluminium surface is more physical than chemical during brazing process, all
flux deposits on the aluminium surface, will be present after brazing as flux residues.
So far the results are not numerous enough to be presented.
The next process improvement way is the study of fluxes with a low solubility in
water. The first difficulty of this solution is the application of such flux, as water is the
main vehicle of the flux deposit. The expected advantage is to obtain less dissociation
of the flux in the water/coolant to avoid reactions between flux residue and coolant and
warranty returns.
References
1. Gray, A.: Influence of Residual Flux Level on the Corrosion Behavior of CAB Alloys (2001)
2. Lanceleur, J.: Internal test report: 14WR074-LM1480 bis_TAHC-MQB (2014)
3. Zaldivar, G.: Flux Residues Evaluation on CAB Aluminum Heat, PDTNVX15029 (2015)
4. Brüschke, Günzl, Götz: Cleanliness Requirements for Components that Convey Coolant,
VW01137 (2014). Strassmann: Residual Foreign Matter in the Heat Exchanger, PV 3678
(2010)
5. Hesse: Determination of the Flux Content Readily Soluble in Water in Brazed Heat
Exchangers Made of Aluminum Materials, MBN 10479 (2014)
6. Water Quality – Determination of Selected Elements by Inductively Coupled Plasma Optical
Emission Spectrometry (ICP-OES), ISO 11885 (2009)
7. Qualité de l’eau - Dosage des anions dissous par chromatographie des ions en phase liquide -
Partie 1 : dosage du bromure, chlorure, fluorure, nitrate, nitrite, phosphate et sulfate, NF EN
ISO 10304-1 (2009)
Design and Development of Chassis Along
with Novel Under Run Protection Devices
for Commercial Vehicles with the Use
of an Innovative Approach Resulting
in Lighter Weight of the Vehicle
Ashesh Anil Shah(&) and Himanshu Arora
VE Commercial Vehicles Ltd., Pithampur, India

{aashah,harora}@vecv.in
Abstract. The paper presents design and development of chassis along with the
novel under run protection devices for commercial vehicles. The concept was
experimentally validated as per govt. regulations.
In today’s commercial market, no customer wants to lose payload for his
vehicle, and ends up overloading it. It not only leads to accidents, but also
deteriorates the life of different vehicle components which are designed for
specific loading.
We have addressed this concern by converting cross members of frame and
under-run protection devices for front as well as rear of the vehicle into alu-
minum materials with patented cross sections [1]. We have avoided the welding
process and used high strength rivets. The riveting process used here is very
suitable for the closed cross sections, as it is fastening operation where only the
single side access is available.
Finally with the introduction of aluminum cross members, Al FUPD and
Al RUPD we have reduced the vehicle weight significantly resulting in 276 kg
payload increment, which will come as an additional benefit to the customer. It
has been tested digitally as well as physically with satisfactory results as per
VECV durability and endurance testing standard.
Keywords: Aluminum frame Aluminum cross members Payload Al

FUPD Al RUPD Durability and endurance testing
1 Introduction
In today’s market, payload, in addition to fuel efficiency, is one of the major factors
contributing to the success of any commercial vehicle. The payload carrying capacity is
directly governed by the unladen weight and vehicle GVW. Since GVW is restricted by
govt. regulations as per the vehicle platform level, only the vehicle aggregates are
subject for improvements in order to provide enhanced payload carrying capacity and
satisfy the vehicle owner.
A heavy unloaded vehicle is not only a setback for profitability, but it also has
negative effects on the environment. Less payload capacity means increased vehicle

Design and Development of Chassis Along with Novel Under Run Protection Devices 611
trips for a constant load, and subsequently increased fuel consumption. This not only
affects the rate of consumption of natural resources, but also deteriorates the air quality
and increases emissions.
We have based our research on the aluminization of structural members of the
vehicle frame to result in the same level of performance and lighter structure as
opposed to the conventional steel frame. The changes we are going to discuss ahead are
related to aluminum cross members and aluminum underrun protection devices.
1.1 Frame Overview

The frame is considered as the backbone of a vehicle, especially in commercial vehicles
with ladder type structure. It serves the dual purpose, as support for assembling other
components as well as taking various static and dynamic loads under vehicle running
conditions. A ladder frame can primarily be divided into two structural sub categories:
side member assemblies and cross member assemblies.
The side member assembly consists of a long member, stiffeners and welded
hardware. The long member section can be categorized into two major surfaces: the
web and the flange. The web is the portion that is aligned to the sides of the assembly at
which all major components are bolted to the frame. The web is responsible for
transferring maximum load of the vehicle to axles. Flanges are dimensionally smaller
than web, but act as a mounting surface for critical components like suspension
brackets. Flanges also resists to deformation of the twisting frame during turning.
The cross member assembly consists of gussets and cross members along with any
additional components such as legs or CM stiffeners. The cross member is responsible
for the transfer of the load from LH to RH chassis side member assembly. It also
provides stiffness to frame for lateral, torsional and vertical loads (Fig. 1).
Fig. 1. Frame schematic
1.2 Under Run Protection Devices

Under run protection devices are provided on the truck as passive safety features for the
surroundings, in case of a crash condition. These are structural members that reduce the
chances of any foreign object getting under the vehicle. FUPD and RUPD commonly
612 A.A. Shah and H. Arora
comprise of two brackets connected to a long bar that takes crash load. The assembly is
connected to the frame with two additional brackets. Without UPD, the smaller vehicle
passenger compartment is likely to interact with stiff commercial vehicle chassis frame
structures. However, with UPD, the smaller vehicle front-end structure gets involved in
the crash helping to control the energy absorption and safe-guarding the passenger
compartment.
We have designed new Aluminum FUPD and RUPD with a patented cross section
[1], serving the purpose of providing under run protection and is lighter by 24 kg than
the conventional steel UPDs (Fig. 2).
Fig. 2. FUPD/RUPD with bar cross-sections
2 Development Phases with Virtual Validations
The frame, FUPD and RUPD all were designed and validated for an 8 × 2 vehicle with
GVW of 31T.
2.1 Frame
The first and most important objective of a durable frame is to withstand to reliability
and performance loads: 2.5 g Vertical Load, Articulation (125 mm vertical displace-
ment), Racking Load, Lateral Load, Tramp Load on First Front Axle, Tramp Load on
Second Front Axle, Jounce Load on First Front Axle, Jounce Load on Second Front
Axle, and Braking.
Static Loads on a frame are classified in two categories:
1. Uniformly distributed loads for entities like cargo body, cargo and frame;
2. Point loads applied onto the frame at the CG of individual components and
assemblies.
Design Initiation Phase: The first concept design was based on the virtual simulation
optistruct frame model [2] as shown in Fig. 3. This model identified weight reduction
opportunities for front cross member modules estimated to 57 kg. Later, an engineering
model was developed in the first iteration as one illustrated in Fig. 4.
Stress concentrations captured by Optistruct
Fig. 3. Optimization result for aluminum frame
Cross member assemblies replicating optistruct
Fig. 4. First iteration design of aluminum frame
In the third iteration, the proposed frame with aluminum cross members was
matured with engineering and manufacturability inputs as well as addressing weak
areas observed in the CAE analysis (Fig. 5).
Flanges added to C section cross members to improve stiffness
Fig. 5. Third iteration design of aluminum frame
In the fifth iteration, the design satisfied the vertical, lateral and torsional stiffness
requirement of the frame. Complexity was reduced to a great extent and modularity
was also maintained. All manufacturability feedbacks were taken care of, in the final
design. Final weight reduction of 252 kg was observed compared to the existing frame.
The cross member design was finalized as two modules: front and rear, placed
strategically to transfer major loads between the side members. Four front modules and
two rear modules were used to get desired frame stiffness.
The front C section at top and bottom of the two gussets is riveted in the horizontal
orientation. The gussets are bolted to the Web of the frame. Similarly, in the rear, the
two C sections are riveted in the vertical orientation as shown in Fig. 6. Longer gusset
Cross members Long gusset for rear CM
Cross members
Front CM gusset
Fig. 6. Aluminium cross members (front and rear)
is required in rear cross member because rear axle reaction is greater than front axle
reaction. Longer gussets distribute the load over a longer span of side member.
Cross members were placed to support the point loads from left to right on the
frame side member. The final frame configuration looked as shown below (Fig. 7).
Fig. 7. Final iteration design for aluminum frame
Initially, a box structure was selected for the cross member assembly manufactured
as bent sheet metal. However bending aluminum gussets proved to be a challenge for
the manufacturer, thus, extruded sections were used for gussets development. Fol-
lowing similar observations, the extrusion process was selected for the development of
all cross members, gussets and stiffeners.
CAE analysis for torsional, vertical and lateral loading frame stiffness was per-
formed, and the achieved results were compared to the existing steel frame, as shown in
Table 1.
Table 1. Aluminum frame stresses compared to the validated product baseline
Stresses developed in all the components of the frame were compared with the base
material Yield stress. The frame was declared safe for implementation only after every
single component registered as FOS ≥ 1.
2.2 The Underrun Protection Devices

During the first iteration, the following FUPD an RUPD sections were proposed
(Fig. 8):
According to CMVR norms, UPD must follow certain dimensional and load
requirements. AIS-069 [3] is followed for FUPD and IS-14812 [4] is followed for
RUPD design.
RUPD
FUPD
Fig. 8. FUPD/RUPD cross sections in first iteration

Fig. 9. FUPD/RUPD load-cases
The force equal to 50 % of vehicle weight was applied on P1, and equal to 100 %
of vehicle weight on P2 in the horizontal direction of FUPD, Fig. 9.
The crash test CAE validation was carried out on LS Dyna [5] elements. The
aluminum section was finalized based on bar deformation results for both FUPD and
RUPD. Profile sections were validated accordingly to the manufacturing process
capability. The FUPD as well as RUPD bar are executed using the extrusion process.
We have used special Hydrabom rivets which can be assembled in conditions when
only single side is accessible for riveting. It has high tensile strength of 13000lbs; hence
welding can be eliminated from the assembly process.
The final section of the FUPD and RUPD is shown in Fig. 10:
FUPD RUPD
Fig. 10. FUPD/RUPD final cross sections
FUPD and RUPD load case used for the CAE simulation were validated as per
AIS-069 and IS-14812 [3, 4] respectively, as shown in Figs. 11(a), (b) and 12(a), (b).
Deformation resulted in the RUPD and FUPD simulations were well within the
standard limits. The total deformation allowed in the structure from the rearmost point
of the vehicle was also within the limits set as per AIS-069 & IS-14812 [3, 4].
Fig. 11. (a) Aluminum RUPD simulation at P1 (b) Aluminum RUPD simulation at P2
Fig. 12. (a) Aluminum FUPD simulation at P1 (b) Aluminum FUPD simulation at P2
3 Physical Validation
Rig level validation proved the FUPD and RUPD concept. Vehicle level testing is
currently under validation passing 1500 endurance cycle without any failure. Also, no
bolt loosening and wear of holes on aluminum gussets has been observed till now.
Physical aluminum frame picture is shown in Fig. 11.
Future scope of work includes testing of the vehicle for ride comfort and handling
(Figs. 13 and 14).
Fig. 13. Physical Aluminum frame picture

Fig. 14. FUPD test rig
4 Conclusion
Aluminum and its alloys are rising as one of the major substituent of steel in auto-
mobile industry. Light weighting of many structural and non-structural parts can be
done by using aluminum designs.
One of the major concerns of working with aluminum sheet metal is the cracking of
sheets observed while bending operation. So we came up with Aluminum extrusion,
which is good alternative for manufacturing.
Virtual validation followed by rig testing has proved that aluminum design is safe
for implementation in FUPD and RUPD bars and aluminum extrusion offers excellent
weight saving opportunity in many structural components like frame side members,
underrun protection devices and cross members. Physical validation of FUPD &RUPD
along with Al Cross member has been completed successfully on vehicle level. Hence
it is safe for implementation to replace steel for structural members and offer weight
reduction as compared to conventional designs.
The vehicle build was completed with the proposed aluminum components and
offered a net weight reduction of 276 kg per vehicle.
In this study, we found that Aluminum can be used to design structural members
and capable to take same load as compared to steel and offer weight reduction in range
of 25–30 %.
Acknowledgments. The authors would like to express their gratitude to Mr. Rajinder Singh
Sachdeva, Executive Vice-President, Technology Development, VE Commercial Vehicles Ltd.,
for granting permission to publish this work. Special thanks to Mr. Vimal Mulchandani, Vice
President, CVD and Mr. Vishal Gupta, Dy. General Manager, CVD and Mr. Umashanker Gupta,
Dy. General Manager, CAE for their guidance and encouragement to develop present work.
Also, the contributions and support of all CAD, CAE engineers and respective testing engineers
are highly appreciated.
References
1. India Patent by Volvo Eicher commercial vehicle, Application Number 201621001073 Under
run protection device for heavy vehicle
2. Altair – OptiStruct Help
3. Automotive industry standard for Front under run protection device, AIS-069, doi, September
2006, by The Automotive Research Association of India (ARAI), Pune
4. IS 14812: Automotive Vehicles - Rear Underrun Protective Device - General Requirements
(2005)
5. LS Dyna - Livermore Software Technology Corporation (LSTC)
The Quality Management Principles and Their
Incidence Within ISO 9001:2015
Laurentiu Aurel Mihail(&)

l.mihail@unitbv.ro
Abstract. The study that is the base of the present paper intends to underline
the prevalence of the new set of quality management principles within the new
form of the ISO 9001. For this scope, the ISO 9001:2015 standard was
exhaustively analyzed and, for each chapter the corresponding quality man-
agement system principles were extracted. Therefore, the methods used are
analytical; of most importance being the absolute mastering of the new and of
the old approach of quality standard analyzed and of the application of quality
management principles with their correct meaning. The results are presented in a
graphical mode, depicting the incidence of the quality principles within each
chapter. This is important for implementation and certification scopes. Because
one of the reasons of the changes within the ISO 9001 standard was for bal-
ancing its structure, the main objective of this study must underline the validity
of this target.
Keywords: ISO 9001:2015 Quality management principles Leadership

Evidence based decision-making ISO/TS 16949
1 Introduction
A multitude of factors promoted the change of the standard ISO 9001. First of all, it
needed to adapt to an economically, politically and societal changing world, in order to
reflect the increasingly complex business environments in which organizations cur-
rently operate. Also, it needed to provide a consistent foundation for the future of the
business environment and to ensure the needs of all relevant interested parties of the
organization. Finally, it needed to ensure an optimal alignment with other management
system standards (environment, HACCP - Hazard Analysis and Critical Control Point,
OHSAS - Occupational Health and Safety, social accountability, information security,
sectorial quality management systems, etc.).
According to [1] the main considered changes of the new version of ISO 9001 are:
• 10-clause structure and core text for all management system standards according
with the annex SL [2];
• Increased compatibility with services and non-manufacturing sector users;
• Clearer understanding of the organization’s context;
• The “process approach” is strengthened and more explicit;

The Quality Management Principles and Their Incidence Within ISO 9001:2015 621
• Addressing the concept of preventive action by introducing the risk identification

and mitigation plan [3];
• The term “documented information” replaced by the terms “document” and
“record” [4];
• Control of externally provided products and services replaces purchasing/
outsourcing;
• Increased emphasis on seeking opportunities for improvement.
2 The Quality Management Principles Within the ISO 9001
2.1 The Structure of the New ISO 9001:2015

The structure of the standard ISO 9001 has been changed, especially for its alignment it
with the structure of the famous “Annex SL” [2]. Therefore, the ISO 9001 standard is
composed of 10 chapters, with the chapters 4 to 10 containing the certification
requirements.
Those chapters are aligned with the same Deming’s P.D.C.A. (Plan - Do - Check -
Act) cycle (Fig. 1), as ISO 9001:2008. As new aspects, it appears the “Organization
and its context” and “Needs and expectations of relevant interested parties”, as inputs
beyond the “Customer requirements”. Within the quality management system model
PDCA cycle, the central role is occupied by the “Leadership”, required with a sup-
plemental accent by the ISO 9001:2015.
Fig. 1. The new model of the PDCA approach of the quality management system according
with the ISO 9001:2015 (Source ISO 9001:2015, p. 7)
622 L.A. Mihail
2.2 The New Set of Quality Management Principles

Causing controversies in the beginning of the new version of the ISO 9001 proposal
[5, 6], the new set of the quality management system principles has only 7 principles
[7] (Table 1). This is one less than the old set [8]. The actual “process approach”
principle is more emphasized, underlining, the old principle of “system approach to
management”. Therefore, the ISO 9001:2015 strengthens the process approach of the
quality management system, as a better way to manage the organization.
Table 1. The old and new quality management principles

The quality management system principles The new quality management system
according with ISO 9000:2005 principles according with ISO 9000:2015
Customer Focus (P1) Customer focus (NP1)
Leadership (P2) Leadership (NP2)
Involvement of people (P3) Engagement of people (NP3)
Process approach (P4) Process approach (NP4)
System approach to management (P5)
Continual improvement (P6) Improvement (NP5)
Factual approach to decision making (P7) Evidence-based decision making (NP6)
Mutually beneficial supplier relationships (P8) Relationship management (NP7)
The relative importance of each principle will vary from organization to organi-
zation and can be expected to change over time, this being the approach for imple-
menting it within the quality organizational quality management systems [7].
By comparing the new edition of the standard to the previous model, the main
target of the present study is to underline the spread of the quality management
principles within the body of the new edition.
3 Data Analysis for the Researched Study
By analytical investigation of ISO 9001:2015 [9], taking into account of the quality
management principles definitions (Fig. 2), and the own practical and theoretical body
of knowledge, the incidence of the quality management principles within each chapter
was underlined.
Within the chapter 4 of the ISO 9001:2015 the more spread principle is NP 6
(evidence based decision making), followed by NP 7 (relationship management). Also,
NP1 and NP2 (customer focus and leadership) are present (Fig. 3). By this evidence, it
can be considered that when analyzing the context of the organization there must be
kept documented information for making the decisions. Additionally, it is important to
note the way in which the relationship between the relevant interested parties is taken
into consideration. This relationship must be managed with leadership and customer
focus as its primary target.
Fig. 2. The new set of quality management principles and their new approaches & meanings
Fig. 3. The incidence of the quality management principles in the chapter 4
Chapter 5 of the ISO 9001:2015 specifically takes into consideration the evidence
based decision making and leadership for the “Leadership” chapter (Fig. 4). Also,
importance is placed on customer focus, and on the relationship between management
and the interested parties.
Chapter 6 (“Planning”) of the ISO 9001:2015 also takes into consideration the
evidence based decision making and leadership principles (Fig. 5). Additionally, the
chapter focuses on customer orientation, the process approach, and improvement.
Chapter 7 (“Support”) primarily takes in consideration the leadership principles
(Fig. 6). Relative importance is also held by the same evidence based decision -
624 L.A. Mihail

making followed by people engagement principle. Chapter 8 (“Operation”) explores

the leadership and evidence based decision – making principles. Relative importance is
also held by the customer focus and relationship management (Fig. 7). Chapter 9
(“Performance evaluation”) takes in consideration the leadership and evidence based
decision making principles. Importance is held also by the same customer focus and
improvement principle (Fig. 8). Chapter 10 (“Improvement”) takes in consideration the
leadership and customer focus & improvement principles, the engagement of people &
evidence based decision making principle (Fig. 9).

626 L.A. Mihail
Fig. 10. The absolute value incidence of the quality management principles within the ISO
9001:2015 for underlining the balancing of the quality management principles incidence
4 Conclusions
The actual form of the ISO 9001 is quite similar as the old approach and, also, it
proposes important modifications (see Fig. 10). These modifications will help orga-
nizations (especially the ones with mature and effective quality management systems)
to grow and it will provide the auditors a field to mastering their activities in a more
appropriate manner. As a result, this new edition of the standard will give the
opportunity to improve a company’s weaknesses by means of the organizational
quality management systems. Comparing the present study with previous ones [10–12]
it can be concluded that between the two versions of ISO 9001, the changes are
important, but not fundamental. The main reason of the formal changes it was reached,
for balancing the spread of the quality management principles within the standard.
The presented study is an own and original idea applied previously for studying the
previous editions of quality management standards [10, 11]. The results of these past
studies are encountered within the present form of the ISO 9001 standard, as major
modifications operated by the technical committee ISO TC 176. Because of the
magnitude of this last study, and because of the editorial space at disposition, it was
presented in a very condensed manner.
5 New Developments for ISO/TS 16949 According

with the ISO 9001 Family of Standards Changes
Strictly for the quality within automotive supply chain, starting from the ISO
9001:2015 the IATF [13] has approved the first draft of the new automotive quality
standard to include additional stakeholder feedback. The revision work team has
organized a conference, also, to obtain further stakeholder input, and for presenting a
final transition strategy from ISO/TS 16949:2009 to the new automotive industry
quality standard [13]. Feedback from gap assessments was used to further refine the
new automotive quality standard and the associated rules for achieving and maintaining
IATF recognition.
Some of the new enhancements to the ISO TS 16949 include the following [13]:
• Requirements for safety-related parts and processes;
• Enhanced product traceability requirements to support regulatory changes;
• Requirements for products with embedded software;
• Warranty management process including addressing NTF (No Trouble Found) and
use of automotive industry guidance;
• Clarification of sub-tier supplier management/development requirements;
• Addition of corporate responsibility requirements.
References
1. ISO/TC 176/SC 2/N1282: ISO 9001:2008 to ISO 9001:2015 Summary of Changes. http://
isotc.iso.org/livelink/livelink/open/tc176SC2public
2. ISO/IEC Directives, Part 1 Consolidated ISO Supplement — Procedures specific to ISO -
Annex SL (normative) Proposals for management system standards (2015)
3. ISO/TC 176/SC 2/N1283: ISO 9001:2015 Risk-based thinking. http://isotc.iso.org/livelink/
livelink/open/tc176SC2public
4. ISO 9000:2015 - Quality management systems - Fundamentals and vocabulary (2015)
5. ISO/CD 9001 - Quality Management systems - Requirements (2013)
6. ISO/FDIS 9001 - Quality management systems - Requirements (2015)
7. QMP ISBN 978-92-67-10650-2. http://isotc.iso.org/livelink/livelink/open/tc176SC2public
8. ISO 9001:2008 - Quality management systems - Requirements (2008)
9. ISO 9001:2015 - Quality management systems - Requirements (2015)
628 L.A. Mihail
10. Mihail, L.: The correlation between the quality management principles and the ISO TS
16949:2002 requirements. In: Proceeding of the International Conference on COSME 2008,
pp. 347–352. Transilvania University of Brasov, Romania (2008)
11. Mihail, L.A., Negrus, A.: The incidence of the quality management principles in the
automotive supplier’s management systems. In: Proceeding of the Congress, CONAT 2010,
Romania, pp. 125–132. SAE, FISITA, EAEC, Transilvania University (2010)
12. Hoyle, D.: Automotive Quality Systems Handbook, 2nd edn., Elsevier Ltd., Amsterdam
(2005)
13. International Automotive Task Force. http://www.iatfglobaloversight.org/default.aspx
Product Quality Control Optimization
for Selection of Measurement and Control
Devices
Constantin Buzatu and Iulian Alexandru Orzan(&)

{cobuzatu,orzan.iulian}@unitbv.ro
Abstract. One of the major decisions that a manufacturing process designer

must take, knowing annual production volume, average size of manufacturing
batches and their frequency, is the selection of the measurement and control
devices used for product quality control.
The approach of scientific criteria for the cost optimizations necessary for
product quality verification based on different influence parameter types: devices
value, actual working time, additional maintenance costs, maintenance and
energy consumption, are analyzed in this paper by determining the critical
number of verified pieces from which one variant is more economically than the
other one.
Keywords: Quality Measurement Control Optimization Efficiency
1 General Information
A constant trend in material production for the past few decades was orientation over
consumer needs which became more and more diversified. The effect of this was the
that the gravity center of production processes moved to small and medium batches
organized in similar products for constructive point of view, but different in sizes.
The effect for production processes and for product quality verification is that the
technologies and the equipment developed in order to maintain this high flexibility rate,
which can respond back quickly and efficient from economic point of view to a
competitive and very dynamic market, with high quality and distinct demands.
All of these have directly influenced the conception and the adaption way on
efficient technical-economic criteria and also on methods and equipment’s used for
product quality verification and analysis.
For economic point of view, the product quality verification is a non-productive tie,
which drives an increase in production costs; for this reason quality verification must be
limited to an absolute necessary for each process from the entire production steps
necessary for one product.
In this way, a considerable influence over the production processes efficiency is the
optimization of selection of measurement and control devices for product quality on
technical and economic criteria.

630 C. Buzatu and I.A. Orzan
2 Actual Principles for Methods and Devices for Product

Quality Control and Measurement Selection
Methods and devices for quality measurement and control selection is dependent on
different factors, from which the most important are [1]:
a. tolerances (dimensional, geometrical, micro geometrical) [1];
b. production batch (unique, serial, mass production) [2];
c. safety of the device, automation level reducing the risk of injury [6, 7];
d. quality operation efficiency [8];
e. quality costs [3, 5, 8];
f. qualification of the quality assurance inspector [6];
g. verified dimensions with specific parameter characteristics [1, 3, 4];
h. the division value from the analogical calibrated dial, for analogical devices, or the
last number that can be read on the digital or numerical system dial [1];
i. error limit of the devices or the methods used [1, 5];
Next lines presents a methodology for analyzing and optimizing the selection of
measurement and control devices used by the production facility on economic criteria.
3 Optimized Selection of Measurement and Control Devices

from the Point of Time in Service Cost of the Equipment’s
Used for Product Quality Verification
Because at the selection of the optimal variant for the production process (design
phase), an important balance has the time for evaluation of product quality, with
measurement or control, it is necessary to determine the costs over the process, divided
for each product, independent of other spends, grouped in manufacturing and assembly.
Vu
Cv ¼ þ CS ð1Þ
Nam Nu 60
Where, Vu is the equipment value, purchased or designed and made by the owner
or by a specialized company, in Euro, Nam is the working hours in one year or another
period of time, CS are the additional costs necessary for maintenance of the equipment
and the energy consumption of this.
Parameter Nu can be calculated with the relation:
Nu ¼ Nt=v np ð2Þ
Where, Nt/v is the time, in hours, for the product quality verification with mea-
surement or control and np is the number of products for verification. From relations (1)
and (2) can be obtained:
Product Quality Control Optimization 631
Vu
Cv ¼ þ CS ð3Þ
Nam Nt=v np 60
For parameter CS the next relation can be used:
CS ¼ Cit þ Cmt þ Cee ; in Euro ð4Þ
Where all this costs are calculated like this [2, 3]:
1. Attendance costs, Cit, for the product quality control devices and can be calculated
with:
pffiffiffiffi Nvt
Cit ¼ 3 þ 0:065k 3 G ; Euro=piece ð5Þ
nps
– k is the constructive complexity coefficient of the device, with values from

k = 2 ≤ 3.6, for electric drive and k = 0.75 ≤ 0.9, for electromechanical drive;
– G is the device mass in kilograms;
– nps is the possible number of simultaneous products verification;
2. Maintenance costs of the device, Cmt, can be calculated with the relation:
pffiffiffiffi Nt=v
Cmt ¼ 0:228k 3 G ; Euro=piece ð6Þ
nps
3. Costs for electrical energy used for driving the device are calculated with relation:
cme
Cee ¼ 1:2Pef dbv þ aPn dauvv ; Euro=piece ð7Þ
0:6
– Pef is the effective power used for verification of product quality, in kW;
– δbv is the base time used for the quality verification, in min/piece;
– α is the ration between nominal power and the idle power of the motor, value given
in Table 1;
– PN is the nominal power of the electric motor, in kW;
– δauv*v auxiliary time consumed in the verification activity, in min/piece;
– cme average cost of the electrical energy; in Euro/min;
Table 1. Value of α
P, [kW] 1.6 0.8 1.1 1.5 2.2 3 4 4.5 7.5 10
α 0.155 0.14 0.133 0.119 0.1 0.094 0.084 0.07 0.065 0.06
With the help from relation (3) it can be observed that it is taken in consideration
also the costs for using high performance product quality verification devices, but
expensive, also the operation productivity on this devices, with the help of effective
working time, respectively the numbers of pieces that can verified simultaneously.
At the same time, it can be observed, from analyze of relation (3), that if Nam and
Nu will be identical, the more efficient verification method is the one with the less
expensive device.
Between 2 devices, which have the same value Vu, it is recommended to select the
one that assurance the maximal productivity for the quality verification. If it is needed
to analyze which method has more advantages, between measurement, m, and control,
c, can be written, for the same number of verified pieces, np, using m, c indices,
associated with relation (3) parameters, next relations:
Fig. 1. a. Product quality verification costs dependencies by npcr. b. Product quality verification
costs dependencies by npcr
Product Quality Control Optimization 633
Vum
C vm ¼ þ CSm ð8Þ
Namm Nt=vm np 60
Vuc
Cvc ¼ þ CSc ð9Þ
Namc Nt=vc np 60
Adding the condition that both costs categories to be equal, meaning Cvm = Cvc,
results critical number of pieces npcr, from which one or the other method is more
advantageous:
60ðCSm CSc Þ
npcr ¼ ½ 1 ð10Þ
t Vu N
Vum Namm Nvm c amc N t
vc
Graphically presented the 2 relations (8) and (9), for the 2 methods of verification,
measuring and control, of product quality, can found in practice next variants, Fig. 1a, b.
Fig. 2. a, b. Variation of product quality verification total costs by number of verified pieces np
If np ≤ npcr, then the efficient method used for product verification, for the case in
Fig. 1a, is the one by measuring, and if np > npcr, then the best to use is the control
method, in case of CSm > CSc.
If np < npcr, then the efficient method used for product verification, for the case in
Fig. 1b, is the one by control, and if np > npcr, then the best to use is the measuring
method, in case of CSc > CSm.
If we represent graphically the relations (8) and (9) it can obtain the next variants,
where the hashed domain is the economic efficiency of one or the other verification
method (Fig. 2).
4 Conclusions
Taking in consideration all presented in this paper, it can come off next conclusions and
recommendations:
• Every measurement or control process of product quality verification must be
designed and analyzed from the point of technical and economic performance
optimization;
• Measurement and control variants of the product quality are depending from a wide
range of factors, which can’t be taken in consideration with subjective approach,
only by applying mathematical scientific analysis optimization with costs, or pro-
ductivity, or if possible both of them performance;
• The results of product quality verification can influence the management and
technological decision, by being a reaction cycle for each production process for
quality assurance.
References
1. Buzatu, C., Lepadatescu, B.: Echipamente si tehnologii moderne de masurare si control a
calitatii produselor (Modern Equipment’s and Technologies for Product Quality Measurement
and Control). MATRIX ROM, Bucharest (2013)
2. Buzatu, C.: Technologii de fabricatie in contrusctia de masini (Manufacturing Technologies in
Auto Industry). Universitatea Transilvania, Brasov (2003)
3. Buzatu, C.: Elemente de proiectare tehnologica si management in fabricatia produselor din
cconstructia de masini (Technological Design and Management Elements for Product
Manufacturing in Auto Industry). Matrix-Rom, Bucuresti (2012)
4. Olariu, C.: Costul si calculatia costurilor (Cost and Costs Calcultion). Editura Didactica si
Pedagocica, Bucuresti (1978)
5. Bralla, J.G.: Handbook of Product Desging for Manufacturing. Mcgraw Hill, New York
(1986)
6. Mcmahon, C., Browne, J.: CAD/CAM Principle, Practice and Manufactuing Management.
Addison-Wesley, England (1998)
7. Walken, J.M.: Handbook of Manufacturing Engineering. Marcel Dekken Inc., New York
(1996)
8. Lane, C.: Modular Fixturing Handbook. Carr Lane Mfg Co., St. Louis (1991)
Improvement of the 8D Analysis Through
a System Based on the “Internet of Things”
Concept Applied in Automotive Industry
Viorel Nicolae1, Laurenţiu Mihai Ionescu2(&), Nadia Belu1,

and Știrbu Luminița Elena3
1
Faculty of Mechanics and Technology, University of Piteşti, Pitești, România
{viorel.nicolae,nadia.belu}@upit.ro
2
Faculty of Electronics, Communcations and Computers,
University of Pitesti, Pitești, România
laurentiu.ionescu@upit.ro
3
Piroux Industrie Romania SRL, Mioveni, România
luminita.stirbu@piroux.com
Abstract. The Eight Disciplines Problem Solving - 8D analyses represents, for

the automotive industry companies, the main working tool with which the defect
analysis is made following a complaint received from an (internal or external)
customer, as well as the means used to communicate the results of this analysis,
all these supplemented by required the corrective and preventive actions. 8D
purpose is to identify, correct, and eliminate recurring problems and when it’s
being used the right way can help the companies come to a solution for the most
complex of problems. This paper presents a solution to decrease the response
time for the 8D methodology: by warning of the workstations and the warehouse
to the appearance of a customer complaint and by using a software for the
computerized management of some documents needed for the 8D analysis.
Keywords: 8D methodology Customer complaints Radio network Zigbee
1 Introduction
The 8D method (also called G8D, Global 8D, TOPS 8D) is one of the most widely used
problem-solving tools related to nonconformities reoccurrence prevention in the
manufacturing process, commonly used for complaints management in automotive
industry. The 8D problem –solving methodology was developed in Ford Motor
Company in the mid – 1980 s to be used by their suppliers to improve the resolution of
problems [1]. Presently, 8D method is a tool of ISO/TS 16949:2009 and has been
widely applied in automotive industries for service or products problem solving,
including the problems with supplier qualification confirmation, manufacturing process
deviations, defects, maintenance, customer complaints, returned purchases etc. [2]. The
purpose of the Technical Specification ISO/TS 16949 is the development of a quality
management sytem that provides continuous improvement, with a particular emphasis
on the prevention of errors and reducing variation and reducing the amount of waste in
the supply chain [3]. The 8D methodology involves teams working together in order to
636 V. Nicolae et al.
solve problems, using a structured 8 step approach to help focus on facts [4]. 8D
methodology are described as follows:
D1. Problem description. This step identifies the problem to be solved and detail spec-
ified by quantifiable parameters. The quality of the information received from the
client is very important for the definition of the problem in the most accurate manner.
D2. Reference to similar products. This step is performed with the purpose to detect
other products with the same nonconformity risk.
D3. The first analysis. This step is used to identify the causes of the failure to detect
the problem: in what place should the nonconformity be detected? why wasn’t the
nonconformity detected?
D4. Immediate action plans. The purpose of this step consists in setting immediate
action plans to protect the client.
D5. Final analysis. This step is carried out with the purpose of obtaining a real and
complete of the situation in order to find the root causes and to decide the
optimum actions required for the treatment of the causes.
D6. Final action plans. The purpose of this step is to decide an action plan for the
elimination of the root causes identified in stages 3 and 5. Permanent actions are
analysed and applied in order to prevent the recurrence of the problem permanently.
D7. Confirmation of the action plans. The effectiveness of the final action plans is
checked. It is a key stage meant to avoid the recurrence of the quality problem.
D8. Preventing a recurrence of the problem. This step involves modifying specifica-
tions, update training, review work flow, and improve practices and procedures in
order to prevent recurrence of the given and similar problems.
There are several research areas for quality improvement in industrial processes.
These involve increasing the efficiency and response to the appearance of the customer
complains. One direction involves the intervention in the production chain so as to
provide a slight revision to the notification of complains [5]. Another way involves
using the computers and the information technologies to analyze 8D. There are several
specialized software in this area, presented in [6].
On the other hand, electronic communication techniques have evolved in the last
time. Therefore they are implemented more often in the industrial sector. Thus, some
important improvements to the 3G communications (cellular networks) access proce-
dure allowing a significantly performance boost for alarm messages in manufacturing
area [7]. New architectures of networks are proposed as the advanced wireless access
network paradigm, where cloud computing is used to fulfill the centralized large-scale
cooperative processing for suppressing co-channel interferences [8]. Using ad-hoc band
free communication protocols, such as Zig-Bee, it is present more than ever in the
industry. ZigBee is encountered in industrial radio networks because that is low cost
and low power compared to other wireless technologies [9]. It can be used in the sensor
networks with other solutions [10].
This article presents a complete solution that includes information technologies and
radio networks, which will join the chain analysis 8D profile in an automotive man-
ufacturing company. The solution integrates electronic devices made to operate in
industrial environments, indoor radio communication systems and client – server
applications.
Improvement of the 8D Analysis Through a System 637
2 The Solution Presentation
The solution consists of three distinct modules: embedded systems for monitoring /
warning of the workstations, embedded systems for warehouse and software application
for the integrated systems management. The solution is based on the using of the “Internet
of Things” concept and its implementation in the manufacturing processes. Thus,
regardless of its type, each workstation can be equipped with an embedded system that, on
the one hand, allows monitoring some activities, their automated registration in an activity
report and on the other hand, reception through an internal communications network of
the eventual complaints from the customer. In the moment of complaint reception, the
embedded system warns the workstations about the occurrence of non-conformity, then
they select and submit the activity reports related to the non-conformity.
In the same time, the embedded systems present in the different warehouse divisions
are informed about non conformities in the case when there are similar parts in the stock. It
will inform staff about the imminence of control to identify non-conformities parts.
Embedded systems are managed by a computer application. The quality manager can send
messages to the embedded systems involved in the manufacturing process of the com-
plained parts. The embedded systems presented in workstations can respond automatically
with the activities reports regarding manufacturing process of the complained parts.
After the receptions of the complaints (by email) the flowchart is: the quality
manager uses the software application to receive the indications from the workstations
involved in the manufacturing process of the non-conformity parts also about the
location in warehouse where this kinds of parts are stored. In the same time, the quality
manager sends the warnings to the workstations and to the divisions of the warehouse.
The warnings messages are received by the embedded systems. The warehouse and the
workstations are prevented about a quality control. Instead, the workstations return a
report of activity to manager and control commission.
The application of our solution involves the use of software technology running on
servers and database. The application server is responsible for selecting of the work-
stations involved in the manufacturing process and the communication with embedded
systems using intranet radio network while the database server is responsible for
storing information related to the workstations and the warehouse divisions.
Embedded systems are based on the local processing units and radio transceivers to
communicate data to and from the access point that communicates with the computer
where running the application. Integrated systems that are at workstations may contain
contact type sensors for monitoring activities. This system can significantly reduce the
time of the 8D analysis. This is because the start-up sequence in response to a customer
complaint can sometimes be quite slow. The steps in 8D analysis assume the answer to
client timeframes imposed.
The system was implemented in a manufacturing process of components parts from
the automotive industry. The solution proposed shows advantages compared to clas-
sical 8D methodology: the response time reduction in startup sequence of 8D analysis
and by creating a database with reports activities in manufacturing process which can
be used for future 8D analysis and for prevention of future customers complains.
A system block diagram is shown in the Fig. 1:
Fig. 1. System block diagram
Central station receives from the quality manager, via the client application,
information on a non-compliance reported by the client for a particular part. The central
station automatically decides on workstations involved in the sectors of manufacturing
and warehouse where there are stored the parts (indicated by the customer in complain).
Through the network radio signals are transmitted to the terminal points where are the
workstations and the stored parts. The staff at the endpoints is informed about the
problem and about the possibility of control. Thus, the first step is immediately
informed of the sites responsible for the occurrence of conformity. Thus, the first step is
the immediate information of the workstations responsible for the occurrence of the
faults. Further here can take preparatory actions for detecting the factors that led to the
problem. Also, there are informed storage points that owns pieces that may have
non-conformities. Endpoints will send reports to the central station on actions in the
manufacture and storage of the parts with problems. The reports will be stored in a
database associated with the notification of the problem and will be readable by the
personnel involved in the quality management and manufacturing organization. This
staff is involved in 8D analysis.
In the following sections will be presented, in detail, the components involved in
the solution proposed in this paper. The solution involves the use of interdisciplinary
technologies in the service of optimizing the management of industrial processes:
information technology, client server systems, radio communication networks in indoor
conditions, embedded systems and Internet of Things concept.
3 The Central Station and Client Application
In this section we present the structure of the central station, client application and
information technologies involved in its design. The Fig. 2 illustrates a block diagram
of the central station and client. As can be seen in the figure, the central station involves
both server applications running on server computer and client application that is
displayed on client workstations from the management and quality staff.
Fig. 2. Central station – client application architecture
It used Java and MySQL solution as freeware technologies. On the other hand,
given the importance of data management, it used a professional portal, Liferay, cus-
tomized to support specific requirements of the 8D analysis. Thus, it has a robust
authentication tool, based on user groups with different access rights. In this way can
enable reporting for three types of users: the administrative and quality staff, the
personnel involved in coordinating the activities of production and the client itself.
The portal enables the quality staff to introduce information about the complaint
and posting reports on the activities involved in the manufacturing and storage. Per-
forming reports was done by using a professional environment dedicated to reports:
Jasper server. Along with the reports editing server allows easy management of the
reports and to display reports. Client application interface is based on the use of
PrimeFaces, a framework with specialized tools for client applications.
4 The Radio Network
Data communications between the central station and the end-points is the key of the
proposed system. All these are provided by a radio network based on ZigBee PRO
protocol. The network structure is shown in the Fig. 3. ZigBee PRO protocol provides
data communication under indoor condition for distances between 500 m – 1 km
between two points.
As you will see there is the possibility of extending coverage using ZigBee wireless
routers. Communication is in 2.4 GHz band at a power which is not harmful to living
beings and, most importantly, communication is free.
Fig. 3. Radio network architecture
There are a gateway to the central station, which acts as a coordinator to the
network - one who builds and manages network. The network management is auto-
matically. The coordinator periodically sends requests to endpoints and routers to
detect their presence and to determine the data routes. An end point can transmit data to
the coordinator directly if there are in the coverage area or through a router. The role of
the router can be dynamically allocated to certain endpoints.
5 The Endpoints from Workstations and Warehouse
As a basic structure, the end points are similar whether there are to workstations or to
warehouse. The difference between them do inputs and inputs management. An
end-point of a workstation block diagram is shown in the Fig. 4:
As you can see, we have several types of inputs: analog and digital. Digital inputs
collect data from contact type sensors. They can be attached to the mechanical drive
units for the workstation. The analogue sensors can take various types of non -electrical
signals as: ambient temperature or workstation temperature for different level, pressure,
light. From the inputs signals are converted in numbers. There are stored in memory
along with time data from real-time clock (integrated inside processing unit).
When requesting information about a specific activity, the processing unit do a
search in the local database, stored in memory and extract the required information.
The processing unit is based on an AVR microcontroller (Atmel) and the storage
memory has capacity of 4 GB. A recording has a total of 32 bytes so it can store
records per second for about 4 years.
Fig. 4. End-point for workstation
Fig. 5. End-point for warehouse
In the Fig. 5 is presented end-point for warehouse. Unlike end point from work-
station, here input consists of a bar code reader. In this way it can determine the
warehouse sector location where it was stored each piece.
6 Conclusions
The innovations and originality that brings the work are:

– decreased the communication time between the quality team and the staff involved
in the manufacturing process from notification client complains;
– increased the speed of collecting information on manufacturing processes during 8D
analysis;
– increased useful information that could lead to resolution of non-compliance;
– decreased overall 8D analysis time especially in the first phase;
– the possibility of introducing an interface for customer information about the
problem solving steps and integration with the standard means of reporting in 8D
analysis. The application and benefits of this study can be further explored in the
automotive industry.
References
1. Langley, G.J., et al.: The Improvement Guide: A Practical Approach to Enhancing
Organizational Performance, 2nd edn. Jossey-Bass Publishers, San Francisco (2009). ISBN
978 - 0 - 470 - 54903 - 2
2. Riesenberger, C.A., Sousa, S.D.: The 8D methodology: an effective way to reduce
recurrence of customer complaints? In: Proceedings of the World Congress on Engineering
2010, Vol III WCE 2010, June 30 - July 2, 2010, London, U.K. (2010) ISBN:
978-988-18210-8-9
3. Misztal, A.: Technical determinants of success the quality management systems implemen-
tation in automotive industry. Int. Sch. Sci. Res. Innov. 9(8), 1928–1933 (2015)
4. Chen, H.R., Cheng, B.W.: A case study in solving customer complaints based on the 8Ds
method and Kano model. J. Chin. Inst. Ind. Eng. 27(5), 339–350 (2010)
5. Malik, J., Rathee, R.: Evaluation of MRR in CNC drilling through optimization of En-8D.
Int. J. Enhanced Res. Sci. Technol. Eng. 4(6), 317–323 (2015)
6. Kifor, V.C., Tudor, N.: Quality system for production software as tool for monitoring and
improving organization KPIs. Int. J. Comput. Commun. 8(2), 235–246 (2013)
7. Condoluci, M., Dohler, M., Araniti, G., Molinaro, A., Sachs, J.: Enhanced radio access and
data transmission procedures facilitating industry-compliant machine-type communications
over LTE-based 5G networks. IEEE Wireless Commun. 23(1), 56–63 (2016)
8. Peng, M., Li, Y., Zhao, Z., Wang, C.: System architecture and key technologies for 5G
heterogeneous cloud radio access networks. IEEE Netw. 29(2), 6–14 (2015)
9. Afifi, G., Halawa, H.H., Daoud R.M., Amer, H.H.: A ZigBee-based Industrial WLAN. In:
2015 Tenth International Conference on Computer Engineering & Systems (ICCES),
pp. 55–59 (2015)
10. Qiao, B., Ma, K.: An enhancement of the ZigBee wireless sensor network using bluetooth
for industrial field measurement. In: 2015 IEEE MTT-S International Microwave Workshop
Series on Advanced Materials and Processes for RF and THz Applications (IMWS-AMP),
pp. 1–3 (2015)
A Case Study Regarding the Implementation
of Six Sigma in an Assembly Process
for the Automotive Parts
Dorin-Ion Dumitrascu, Adela-Eliza Dumitrascu(&), and Anghel Chiru
“Transilvania” University of Brasov, Brasov, Romania

{d.dumitrascu,dumitrascu_a,achiru}@unitbv.ro
Abstract. This paper aimed to investigate the process of six-sigma imple-

mentation in an assembly process for vehicle parts. The Six Sigma objectives
consists in reducing the occurrence of specific defects, scratches and deforma-
tion of the assembled doors, saving material and reducing costs for internal
repairs. By implementation of Six Sigma, it was possible to improve the DPMO
indices (Defects per Million Opportunities) with 70 percent and to analyze the
quality characteristics considered critical for the monitored process, in order to
determine the performance level of the product. The results indicate that with
proper implementation of Six Sigma methodology it can be achieved a positive
impact on the quality parts by reducing defects, improving the flexibility of the
assembly process and customers’ satisfaction, too.
Keywords: Assembly process Six Sigma methodology Defects per Million

Opportunities (DPMO) Capability analysis
1 Introduction
Six Sigma is a strategy of continuous quality improvement for an organization that is

used in many fields. In general, Six Sigma is a methodology to improve the process by
reducing the defects of products, minimizing the process variation and improving the
capability of manufacturing processes. The Six Sigma objectives are focused on
increasing the profit and reducing the costs, by minimizing the defective rate of per-
formed products [1, 2].
Due to the diversity of industrial products and competitive environment, organi-
zations have searched strategies to improve the processes and products quality such as
total quality management (TQM), ISO certification, Six Sigma, etc.
Six Sigma is the new statistical approach for total quality management, a global
approach aimed to improve customer satisfaction, which is not the same as quality
improvement. Based on this improved customer satisfaction, Six Sigma is increasing
source of profitability for the organization cumulating the following effects [3]:
• a decrease of scraps, work correction and generally the non-quality costs;
• an improvement of machine availability and synthetic yield rate;
• better market segments, considering product quality improvement.

644 D.-I. Dumitrascu et al.
One of the basic principles of Six Sigma is to reduce the process variability, which
derives in particular from:
• variability of materials;
• procedures variability;
• variability of the condition where production process evolves.
Six Sigma has two major perspectives. One is a statistical perspective, the other is
the business perspective. From the statistical point of view, the term Six Sigma is
defined as having less than 3.4 defects per a million of produced items or a success rate
of 99.9997 %, where sigma is a term used to represent the average of process variation.
From a commercial perspective, Six Sigma is defined as a business strategy used to
improve business profitability, to increase effectiveness and efficiency of all operations
in order to meet the customer needs and expectations or to exceed them [4, 5].
Six Sigma promotes a philosophy of excellence and it is based on using two
methods: DMAIC and DMADV [6, 7].
DMAIC method consists of the following steps: definition phase, measurement
phase, analysis stage, stage and phase control improvement; it is applied when you
want to improve products/services or existing processes.
DMADV method consists of the following steps: defining stage, measurement
phase, analysis phase, design phase/stage and verification phase useful when you want
to design new products or processes for a specific imposed capability.
In this paper is presented a study consists on implementation of Six Sigma
methodology using DMAIC method in order to improve the assembly process. By
implementation of Six Sigma methodology it allows us to identify the effective way to
find out the critical points of the analyzed process. For analyzed process, which initially
is carried out with open doors in assembled state on the vehicle, there are a number of
defects (deformation of the assembled doors and paint scratches) that need repairing or
even the parts are rejected. The main objective is to improve the assembly process by
decreasing the number of defective units with a percent of minimum 70 %.
2 Implementation of Six Sigma. Case Study

2.1 Define Phase
The objective of Six Sigma is to reduce the number of rejects due to specific problems
that appear during the doors assembly process on the vehicle, by reducing the occur-
rence of specific defects (door scratches and sheet shocks).
Thus, the following targets are defined with related indicators:
• Improving the number of DPMO (Defects per Million Opportunities) by 70 %, so
reduction with the same percentage of the number of defective units, reported to a
million opportunities of developing these defects;
• Increasing the number of positive results at the car body and paint;
• Saving materials;
• Cost savings due to internal repairs.
A Case Study Regarding the Implementation of Six Sigma 645
2.2 Measure Phase

In the second phase are presented measurements undertaken for further analysis of how
internal failure occurs, reported proportional for the four doors assembled motor
vehicles (for two models of the car body). Thus, for the first model body is observed
that most failures occur on the back door on the left side, followed by the right front
door and then by the other two doors, in the same proportion (Fig. 1). It can be
concluded that failures occur approximately at the same proportion on each door,
resulting the idea that the corrective actions that will be reached after applying opti-
mization of the entire cycle will be applied to all four doors.
Fig. 1. Defects distribution for assembled doors of the analyzed type I car body
For the second examined car body model, it is observed that most failures occur at
the front door on the right side, followed by the back door on the right side and the
other two doors on the left side (Fig. 2). From here comes the same conclusion,
according to which the optimal corrective solution will be applied to all four doors.
In order to determine the performance of the products, there are monitored the
quality characteristics considered critical for the analyzed process, defects caused
during assembling process: paint scratches and sheet shocks per 1000 products.
Fig. 2. Defects distribution for assembled doors of analyzed type II car body
2.3 Analyze Phase

Due to previously recorded data, there are identified the possible causes that generated
the unfavorable recorded situation. It is analyzed the defect by drawing up a
cause-effect diagram and as generating potential sources are included (Fig. 3):
• environment - workplace design, inadequate lightening;
• measurements - Cpk process capability indices;
• material - processed material strength (in a state of semi and finished parts),
resistance of the paint layer, the protective elements of the door;
• labor force - improper assembly and usage of tools, lack of training, failure
procedures;
• methods - improper design of the workplace, how to adjust the edges;
• machine tools - door handles transporters, improper maintenance of devices, faulty
design of the conveyor.
The main causes that produce the irregular flow are faulty assembling process and
inadequate equipment usage (as causes related to the labor force) and the doors con-
veying and transfer systems (as causes related to machine tools).
2.4 Improvement Phase

Process improvement consists in the use of an optimized transfer device to ensure
disassembly of the vehicle doors, in order to avoid the appearance of defects related to
Fig. 3. Case-effects diagram of the assembly process
the previous mode of organization of assembling the inside components. This device
allows not only a fence door, but its proper grip due to its overall dimensions and
weight.
As an effect of applied corrective measures, the process performances by product
quality point of view it was monitored for three months. Figure 4 shows that the defects
percent records a significant decrease.
Monitoring the occurrence of defects on the assembly line, it is seen predominantly
low levels, with a steep decrease slope, corresponding in time, immediately after
implementation of the corrective solutions. The proposed and implemented optimiza-
tion has led to a record level of DPMO by 1098, related to a defective fraction of 4.4.
Compared to the initial state, it results an improvement of more than 85 % of the
process, for the monitored critical quality characteristics.
2.5 Control Phase

The analysis of the samples dependency for means of monitored defects specific to
initial and optimized stages it is detailed in Fig. 5.
The t test for determining the samples dependence is a statistical technique that is
used to compare two populations if the two samples are correlated. Data were recorded
before (def b) and after implementation of improvement actions (def a).
The analysis of process capability for monitored defects specific to initial and
optimized stages are presented in Figs. 6 and 7.
25
20
Defects [%]
15
10
0
Month 1 Month 2 Month 3
Fig. 4. Monitored process performance
Paired t Test for the Mean of def b and def a

Summary Report
Mean Test Paired

Is def b less than def a? Statistics Differences *
0 0 .0 5 0 .1 > 0.5 Sample size 3
Yes No Mean 237.33
90% CI (-595.49, 1070.2)
P = 0.754 Standard deviation 494.00
The mean of def b is not significantly less than the mean of * The difference is defined as def b - def a.
def a (p > 0.05).
def b def a
Mean 345.67 108.33

Standard deviation 379.91 150.41
Distribution of the Differences
Where are the differences relative to zero?
0 Comments
-- Test: T here is not enough evidence to conclude that the

mean of def b is less than def a at the 0.05 level of
significance.
-- CI: Quantifies the uncertainty associated with estimating
the difference from sample data. You can be 90% confident
that the true difference is between -595.49 and 1070.2.
-- Distribution of differences: Compare the location of the
differences to zero. Look for unusual differences before
interpreting the results of the test.
-500 0 500 1000
Fig. 5. Distribution for means differences applying t test

Fig. 6. Poisson capability analysis for def b
Fig. 7. Poisson capability analysis for def a

3 Conclusions
The implementation of Six Sigma methodology offers measurable parameters of process

performance and valuable conclusions of the consequence of statistical analysis.
By implementing a permanent change for the transfer system and transport
inter-doors, disassembled for a vehicle, in order to make the installation of interior
components, it is removed completely the possibility of an error, a fact detected by a
100 % control of the machined parts.
From the statistical point of view, it can be seen that improvements have been
implemented effectively.
After implementing the corrective actions, it can be mentioned that improvements
applied to the analyzed process leaded to:
• significant reduction of losses by improvement of more than 85 % of the assembly
process for the monitored critical quality characteristics;
• improving the assembly flexibility by taking into account aspects of ergonomics in
the workplace;
• operational efficiencies growth;
• increase labor efficiency.
References
1. Chang, S.-I., Yen, D.C., Chou, C.C., Wu, H.C., Lee, H.P.: Applying six sigma to the
management and improvement of production planning procedure’s performance. Total Qual.
Manag. Bus. Excellence 23(3), 291–308 (2012)
2. Pyzdek, T.: The Six Sigma Handbook: A Complete Guide for Green Belts, Black Belts, and
Managers at all Levels. McGraw- Hill, New York (2003)
3. Virgil, P.: The new statistical approach to total quality management. Six Sigma (in Romanian:
Noul demers statistic în managementul calităţii totale. Six sigma). “Valahia” University of
Targoviste (2015). http://www.ecr-uvt.ro/informatii_noutati/pdf/Sesiune_de_comunicari_J.M.
JURAN’/6_sigma.pdf
4. Baral, L.M.: Integrating of knowledge management concepts with six sigma metodology (in
Romanian: Integrarea conceptelor managementului cunoașterii cu metodologia Six Sigma).
Doctoral thesis, “Lucian Blaga” University of Sibiu (2014)
5. Park, S.H.: Six sigma for quality and productivity promotion. Asian Productivity Organi-
zation, pp. 37–38 (2003). ISBN: 92-833-1722-X
6. ISO 13053-1: ISO guideline on quantitative methods in process improvement- six sigma- part
1: DMAIC methodology (2011)
7. ISO 13053-2: ISO guideline on quantitative methods in process improvement- six sigma- part
2: tools and techniques (2011)
Advanced Transport Systems
and Road Traffic
Analysis of the Influence of One-Way Streets
on the Urban Road Networks Connectivity
Daniela Florea(&), Dinu Covaciu, and Janos Timar

{d.florea,dinu.covaciu,jamcsika_timar}@unitbv.ro
Abstract. The paper presents a theoretic analysis of the indicators used to

evaluate the road network connectivity. The experimental research was per-
formed to compare the influence of the one-way streets in the network (Scenario
1) with a network containing only two-way streets (Scenario 2) using the
indicators of the full road network (network diameter, cyclomatic number) and
also indicators that assess the relative position of each network node in relation
to the network as a whole (i.e. the Alpha, Beta, Gamma, Eta indicators). Each of
the two cases analyzed starts from the primal graph composed by the most
important intersections as nodes and the connecting streets as arcs, and the
connectivity matrix and also for the dual graph, in which the links are nodes and
the intersections are arcs. The analysis involves the determination of changing of
the indicators when changing the network model. In this way can be identified
the nodes with highest and lowest connectivity, and the most accessible nodes.
Keywords: Connectivity Graph theory Road networks One-way streets
1 Introduction
An analysis of circulation evolution in urban road networks highlights some tenden-

cies. Thus, in United States of America, at the middle of XX Century, in the Pre-
Freeways era, when the cities started to face the congestion phenomenon, was adopted,
as a traffic fluidization solution, the conversion of two-ways streets into one-way streets
[1]. As a result, the number of lanes was doubled and, consequently, the traffic capacity
and the travel speeds have increased. In Romania, only after the year 2000 the big cities
were confronted with excessive agglomerations and thus, this kind of changes are more
and more frequent.
The economical development of areas surrounding the main roads, in the years
1980s, Post-Freeway Era, has lead to generation of increasingly high traffic volumes,
and as effect, the accessibility in the area from one-way arterials becomes lower,
leading also to the reducing of economic development and of the interest for businesses
and commercial activities in those zones [1].
As a reply to those tendencies, after 1990s, was arised a dilemma that has to be
solved: how much appropriate is the transformation of one-way streets in two-way
streets, especially in residential and commercial areas, in order to rehabilitate and
revitalize the central areas of the cities [2]. In this period, named Reemerging Era, the
urban communities rediscovered the downtown attraction [1].

654 D. Florea et al.
The return to urban networks with two-way streets involves conflictual views of
different interest groups with different perspectives: the traffic engineers, the road
users – vehicle drivers, providers of public transport services, pedestrians – as well as
the local community through political and business leaders.
It is considered that the return to two-way streets is beneficial for the local busi-
nesses that depend by visibility and by the capacity to cross the street. Also, this streets
category is safer than the one-way streets, that are associated with a high travel speed
and a reduced attention of drivers. The pedestrians prefer the two-way streets for
crossing because the drivers have the tendency to slow down and, therefore, the
conflicts with vehicles are more predictable and the severity of crashes is lower. Also,
operation of two-way streets allows installation of median separator areas that can be
used as pedestrians refuges [3].
Studies show that the residents and the travellers prefer two-way streets for eco-
nomical and livability reasons, but the traffic engineers and the transport planners
consider one-way streets as more efficient especially when the travels in the network
are shorter [4].
Long debates on the dilemma “one-way versus two-way streets” resulted in the
reconversion of one-way streets into two-way streets in many cities and agglomerated
metropolis from highly developed countries. These actions were correlated, also, with
the recent transport policies that encourage the transport modes environmental friendly:
public transportation, walking, biking [5].
2 One-Way Streets Particularities
Organizing of the road circulation in one-way streets networks is a relatively simple

and inexpensive tool, especially in central and residential urban areas [6, 7]. Certainly,
this solution offers a series of advantages, like:
• Capacity of one-way streets rise depending on local conditions (distribution of
directional flows, street widths, cornering in intersections).
• The supplementary lanes, not used before, may be used at their full capacity.
• Left turns in intersections are not delayed by the traffic from opposite direction.
• Signalized intersections coordination is easier.
• Travel durations on arterials are shorter in case of signals synchronisation and
higher travel speeds.
• The operation mode of the public transportation service is more efficient, because it
is possible to reserve dedicated lanes for public transportation.
• There are less conflict points with pedestrians that crosses a one-way street when the
left turn on the crossing direction is not permitted.
• The number of conflict points can be reduced from 24 (Fig. 1.a) to 11 (Fig. 1.b) or
even 6 (Fig. 1.c).
One-Way Streets on the Urban Road Networks Connectivity 655
Fig. 1. Influence of one-way streets on the number of conflict points in intersection
However, the disadvantages of this organisation mode of urban road networks must
not be neglected. From these, can be mentioned:
• Longer travel distances to reach the destination and, as a consequence, longer travel
times, higher fuel consumption and higher levels of chemical pollution (especially
in case of the residential areas).
• The public transport stations for the opposite direction must be relocated, leading to
longer walk distances to and from stations.
• Number and severity of accidents when the vehicles from lateral streets enter in the
main street are increased as an effect of higher speeds on the one-way street.
• In case of multiple lanes, the number of conflict points increases, caused by the
interweaving of traffic flows.
• The non-motorized traffic participants are disadvantaged by de reduction of number
of pedestrian crossings and the rise of injury risk.
• The traffic flows from canceled directions are moved to complementary streets from
residential areas, which leads to increasing of travel speed and of number of
accidents, as well as to losing the attractivity of these areas.
The traffic organization based on one-way streets is recommended especially in the
case of grid-type networks. The radial, linear or annular networks present even more
advantages. Beside the presented aspects, in the urban road network is important to be
possible to assess the traffic from the network connectivity perspective.
3 Connectivity in Road Networks
3.1 Road Networks Representation

Optimization of road traffic in urban areas may refers to decisions regarding the con-
nectivity in the road networks that involve major issues like: safety, travel speed,
distances and modal split. If the goal is increase of the travel speed it is obvious that it
is necessary to reduce the connectivity by increasing the distances between junctions
and reducing the number of pedestrian crossings. The choice of solutions depends by
the model of network organisation [8]. Connectivity in road networks refers to the
density of connections and directness of links. A well connected road network includes
many short links, many junctions and a reduced number of dead-ends. When the
connectivity increases, the travel distances decrease and the number of travel routes
increases. So, the routes between different origin-destination points are more direct,
which leads to the design of a more accessible and flexible road system.
Conectivitatea se poate analiza în cazul străzilor din interiorul unei zone (conec-
tivitate internă) sau referitor la conexiunea cu arterele majore de circulaţie şi cartierele
unui oraş (conectivitate externă). Acest tip de analiză necesită reprezentarea reţelei
rutiere astfel încât intersecţiile sunt nodurile, iar străzile sunt arcele unui graf denumit
primal graph, Fig. 2.a deci corespunde structurii spaţiale reale ale unui oraş.
Fig. 2. Primal graph (a) and dual graph (b) representation
The representation of road networks is known as dual graph when the streets are
nodes, and two nodes are connected when they meet in a junction; in other words, the
junctions are arcs that connect two intersecting streets, as per Fig. 2.b. This is a free
representation that does not involve attributes related to distances, durations or others,
being just an estimation of connectivity. In the dual graph representation, named also
“information city network” [9], are obtained relational (topological) measures [10], so
that in the connectivity graph each axial line is converted into a node, and the inter-
section of any pair of axial lines is converted into an edge. This way, the measure of
accessibility is calculated using the connectivity graph based on a topological distance,
or non-Euclidean distance (step-distance) [11].
3.2 Basics of Graph Theory and Connectivity

A graph G = (N, K) attached to a network is a mathematical object that consists in two
finite sets of data: the set of nodes or vertices, N (1 ≤ i ≤ N) and the set of arcs, or
edges, K, meaning the two unordered pairs of nodes k. A graph may be represented in
multiple ways. The best representation is the adjacency matrix, or the connectivity
matrix, A(G) = {aij}, a square matrix NxN whose inputs are the elements aij, where i,
j = 1, 2, …, n, defined below:

1; if i and j are connected
aij ¼ ð1Þ
0; if i and j are not connected
The matrix A(G) is symmetrical in case of unoriented graphs, and all the elements
on the diagonal are zero.
As terminology, it is recommended, in case of unoriented graphs (undirected) that
allow connections from i to j and back, to use the terms of vertex and edge, while for
the oriented graphs (directed graph or digraph) the terms of node and arc are preferred.
For description of the network characteristics, the recommended terms are those used
for oriented graphs.
3.3 Topological Measures of the Graph

First, from the most important topological measures that define a graph, should be
mentioned the node degree and the degree distribution.
Node degree, ki is the number of edge incidents to the node, or the number of
neighbours of node i with direct connection. Average degree is the mean of degrees of
all nodes in the network:
1X
k¼ ki : ð2Þ
n i¼G
The model of the streets network in which the arcs just pass through nodes is
represented as a planar graph. In a real network, obviously, not all the nodes have the
same number of edges (the same degree). In an important network as number of nodes
and connections it can be defined the distribution of node degrees, P(k), an essential
propriety for the calculus of the probability to find nodes with a certain number of
connections. So, for calculus is considered the absolute frequency N(k) of a certain
value k in relation with the total number of nodes n in the analyzed network.
NðkÞ
PðkÞ ¼ : ð3Þ
n
The previous studies, made on a significant number of networks represented by

random graphs with n nodes and k arcs (edges) randomly selected, had highlighted a
Poisson distribution of the parameter k. An arc represents actually a pair of nodes.
From the multiple indicators of connectivity proposed in the literature for different
network applications [12, 13], will be analyzed only a part, considered the most sig-
nificant for the study of road networks and for this particular study. For the evaluation
of the whole network the important parameters are: network diameter and cyclomatic
number.
Network diameter, D. Calculation of network diameter is based on the method of
transitive closing of a graph. It can be defined the transitive closing of the graph
G = (N, K) attached, the graph Gt = (N, Kt) attached determined so that for all the
pairs (i, j) 2 N2, the arc (edge) (i, j) belongs to Kt, if and only if there is a link (chain)
from i to j. Calculation of transitive closing of a graph can be done by adding the

successive powers to the adjacency matrix with the meaning that for each step of
multiplying with the initial matrix there are determined the possible links between node
i and node j, composed by two, three or more arcs. The calculation is done when in the
sum matrix all elements are not zero, and the network diameter is given by the rank of
the last matrix.
The cyclomatic number, u, represents the maximum number of independent
cycles of the graph attached to the network. A cycle is the complete coverage of a route
(from a node i to himself) having the arcs oriented in the same direction. The calcu-
lation formula takes into account the number of nodes n, and arcs k, of the network, as
well as the number of sub-graphs p, of the network:
u ¼ k n þ p: ð4Þ
In order to assess the individual elements of the network (nodes and arcs), was
given importance also to: the degree of connectivity gc, Alpha Index, α, Beta Index, β,
Gamma Index, γ and the Aggregate Transportation Score, ATS.
The degree of connectivity, gc, compares the maximum connectivity with the
observed one (the real one) of the network, namely the ratio between the maximum
number of connections kmax and the total number of graph connections:
kmax n ðn 1Þ=2
gc ¼ ¼ 100 ½%: ð5Þ
k k
The degree of connectivity increases when the number of nodes increases, having
values between 1 and n/2 but, at the same time, increases if the number of connections
decreases for a constant value of nodes number. The comparison of networks is pos-
sible only if the degree of connectivity is given as percents.
Alpha Index, α, as indicator of network connectivity is calculated, in case of planar
graphs, as the ratio of number of cycles and the maximum number of cycles:
k nþp
a¼ 2 ½0; 1 for n 3; ð6Þ
2n5
and in case of non-planar graphs:
k nþp
a¼ : ð7Þ
ðn 1Þ ðn 2Þ=2
Alpha Index measures the connection level, independent by the number of nodes,
being as high as the network is more connected. Value of 1 indicate the existence in the
network of the maximum number of arcs and thus the maximum degree of connec-
tivity. A value higher than 1 is specific to an over connected network.
Beta index, β, measures also the connectivity level of the graph, being calculated
as a ratio between the number of arcs and the number of nodes, of the graph.
k
b¼ ; 0 b 3: ð8Þ
n
Values of this index depend on the complexity of the network, as well as on the
type of the structure. As example, for a network with a tree structure or a simple
network, the index β has a value lower than 1. Values higher than 1 are obtained for
complex networks, hence with more travel routes between any two nodes. For the
planar graphs the maximum value is 3, and for the non-planar, ∞.
Gamma index, γ, is a measure of connectivity which express the correlation
between the number of existing connections and the number of possible connections.
For the planar graphs:
k
c¼ 2 ½0; 1 for n 3; ð9Þ
3ðn 2Þ
and for the non-planar graphs:
k
c¼ : ð10Þ
nðn 1Þ=2
The low values of the γ index are characteristic for non-connected networks or for
tree structures.
Based on the indicators described above, it was defined the Aggregate Trans-
portation Score, ATS, calculated using the relation [14]:
ATS ¼ a þ b þ c þ u: ð11Þ
The values of this indicator are arranged descending and are included in high,
medium and low level classes. The higher values indicate a higher connectivity and
greater efficiency of the network.
Eta Index, η, represents the sum of arcs k and nodes n divided by arcs number:
kþn
g¼ : ð12Þ
k
If the abstract form of the relation is replaced with a concrete representation where
attributes are assigned to arcs, like length units and it is considered the total length of
the network (sum of length of the arcs that define the graph), Eta Index represents the
measure of the median arc calculated with the relation:
L
g¼ ½km: ð13Þ
k
When there are available data about the traffic flows on arterials that compose the
network, and the evaluation is done using dedicated software, it can be also calculated
Theta Index, θ (which take into consideration the mean values of the traffic flows in the
network), and when the analyzed network is completely defined (by attributes like the
length of each arc) it can be calculated the Pi Index, π (as the ratio between the total
length of the network and the length of the route measured across the arcs that compose
the network diameter). Another useful indicator that characterize the network efficiency
is Detour Index, defined as a ration between the direct distance (straight line) between
two nodes and the real distance traveled between them.
3.4 The Influence of One-Way Streets on the Connectivity Indicators

in Intersections
First, it is considered the analysis of the simplest network attached to an intersection,
where the possible connections can be represented as a graph, Fig. 3, and, therefore, it
can be defined the connectivity matrix. In this case, the nodes are represented by the
intersection entries, and the arcs represent the movement or connection of them.
Node 1 2 3 4 Node 1 2 3 4 Node 1 2 3 4

1 0 1 1 1 1 0 0 0 0 1 0 0 0 0
2 1 0 1 1 2 1 0 0 1 2 0 0 0 0
3 1 1 0 1 3 1 1 0 1 3 1 1 0 0
4 1 1 1 0 4 1 1 0 0 4 1 1 0 0
a) b) c)
Fig. 3. Representation of an intersection as graph and the attached matrix
It can be noticed in case shown in Fig. 3.a that the degree of each node is 3, being
possible entries from each of the other nodes. As certain connection are suppressed
after introduction of one-way roads, the degree of nodes decreases (Fig. 3.b and c).
By analysis of this simple example can be observed the influence of diminishing the
number of connections on the evolution of certain indicators previously described.
Only for the graph in Fig. 3.a which allows all the possible movements in intersection,
it can be considered that a well connected network exists (the highest value for Alpha
Index and for Gamma Index), respectively that the network is complex (Beta Index).
As some movements are suppressed (Fig. 3.b and c.), these indicators evolves towards
inappropriate values (Table 1).
Table 1. Values calculated for conenctivity indicators for the three graphs in Fig. 3.
Network indicators (a) (b) (c) Recommended values
Number of nodes, n 4 4 4 –
Number of arcs, k 12 7 4 –
Number of sub-graphs, p 1 1 1 –
Degree of connectivity, gc (Eq. 5) 0.5 0.86 1.5 1.0 – n/2
Connectivity indicators
Cyclomatic number, u (Eq. 4) 9 4 1 –
Alpha Index (Eq. 6) 3 1.33 0.33 0–1.0
Beta Index (Eq. 8) 3 1.75 1 0–3.0
Gamma Index (Eq. 9) 2 1.17 0.67 0–1.0
Eta Index (Eq. 12) 1.3 1.57 2 –
4 Analysis of the Civic Center Area of the Brasov City

Network
The case study conducted on an area in the Civic Center of road network of Brasov
city, Fig. 4, consists in comparing the network connectivity indicators for the case of
one-way travel on some of the road segments (Scenario 1 - real situation), and intro-
ducing two-way travel on the all road segments (Scenario 2).
Fig. 4. Labels for the nodes in the map of Civic Center area in Brasov City [15]
The analysis is based on two different approaches of the graph representations

(primal and dual) attached to the road network, as it is described further, without using
specialized software. The objective of this analyses is to compare the connectivity
indicators for a small size network, considering the nodes on the network periphery as
dead-end streets. Obviously, renunciation of external links of the analyzed area has a
significant influence on connectivity and accessibility of these nodes.
4.1 Primal and Dual Graph for the Road Network with One-Way Streets
(Scenario S1 – Real Situation)
The first step consists in representation of the primal graph, oriented, where the 22
labeled nodes represent the main intersections in the analyzed area, and the arcs rep-
resent the connecting streets, taking into consideration the existing travel directions.
Fig. 5. Primal graph of City Center area for the network with one-way streets
Fig. 6. Dual graph of City Center area for the network with one-way streets
This way, it resulted the graph shown in Fig. 5 and a connection matrix (connectivity
matrix) filled according to Eq. 1.
In the next step it was made, following the rules previously mentioned, the directed
dual graph shown in Fig. 6, which was at the basis of the connectivity matrix.
4.2 Primal and Dual Graph for Two-Way Road Network (Scenario S2)
In this scenario, all the connections between nodes allow travels in both directions. The
primal non-oriented graph attached to the network in this case is shown in Fig. 7, and
the connection matrix was also filled following the known rules.
Fig. 7. Primal graph of City Center area for the network with two-way streets
Fig. 8. Dual graph of City Center area for the network with two-way streets
The analysis of the urban road networks using representations as primal and dual
graph allows the calculation of connectivity indicators and objective interpretation of
results. Thus, the analysis using the primal graph scenario of a network containing
one-way streets (real situation), one intersection (node) allows an average of 2.36
connections (arcs) with the neighbor intersections, while the scenario where all the
streets are two-way, a number of 2.91 connection.
In case of the analysis based on dual graph (Fig. 8) it can be observed that any
street (node) allows an average of 3.25 connections passing through intersections (arcs)
to the neighbor streets for the one-way scenario, and 4.44 connections for the scenario
based only on two-way streets. This prove a limitation of the travel possibilities and of
choice of alternative routes, in case of one-way streets. Consequently, the entire input
flow on these network parts will be forced to follow the same route and to pass through
multiple intersections.
The diagrams in Fig. 9 provide supplementary informations regarding the con-
nectivity, based on the statistical analysis of network nodes having the node degree as
Fig. 9. Frequency distribution, P(ki) and cumulative frequency, P(k < ki) of nodes degree for
the dual graph (one-way and two-way scenario)
variable. Thus was calculated relative frequency distribution of nodes having a certain
degree and cumulative frequency as a percentage. There can be defined the so called
percentiles like the median or 50th percentile, the 85th percentile and others. As
example, the median of a node degree is 2.6 for one-way network and 4 for two-way
network, meaning that 50 % of the network nodes have the node degree below or over
these values.
The synthesis of results for the calculus of network diameters using the transitive
closing method for the primal graph, for the two scenarios, is presented in Table 2.
Table 2. Results obtained for the network diameter calculation (primal graph).
One-way (S1) Two-way (S2)
Routes with 8 arcs 9–15, 9–21, 9–22, 21–1 –
and 7 intermediate
nodes
Routes with 7 arcs 1–15, 1–21, 1–22, 2–15, 2–21, 2–22, 8–15, 1–21, 1–22, 9–20, 9–
and 6 intermediate 8–21, 8–22, 9–14, 9–15, 9–19, 9–20, 9–21, 21 (8 pairs of nodes)
nodes 9–22, 14–3, 15–2, 15–3, 19–1, 20–7, 20–9,
21–1, 21–2, 21–7, 21–9, 22–1, 22–2 (27
pairs of nodes)
The most connected 19 (Zizinului Street – Traian Street) 5 (15 Noiembrie
node Blvd. – Iuliu Maniu
Street)
Nodul cel mai putin 9 (Mihail Kogalniceanu Blvd. – Oasului 9 (Mihail
accesibil Street) Kogalniceanu Blvd.–
Oasului Street)
The comparative results obtained from the analysis of the connectivity indicators
calculated using Eqs. 1–13 are presented in Table 3.
The connectivity indicators given in Table 3 are interpreted relatively to the optimal
values.
Table 3. Network characteristics and connectivity indicators for primal and dual graph (Civic
Center of Brasov City).
Conectivity indicators One-way (S1) Two-way (S2)
Primal Dual Primal Dual
Number of nodes, n 22 32 22 32
Number of arcs, k 52 104 64 142
Number of sub-graphs, p 1 1 1 1
Maximum number of arcs, kmax 231 496 231 496
Average node degree, k 2.36 3.25 2.91 4.44
(continued)
Table 3. (continued)
Conectivity indicators One-way (S1) Two-way (S2)
Primal Dual Primal Dual
Degree of connectivity, gc 4.44 4.77 3.61 3.49
Network diameter 8 8 7 7
Cyclomatic number, u 31 73 43 111
Alpha Index 0.795 1.24 1.10 1.88
Beta Index 2.36 3.25 2.91 4.44
Gamma Index 0.87 1.16 1.07 1.58
Eta Index 1.42 1.31 1.34 1.23
5 Conclusions
This paper was aimed to establish the influence of road network configuration and of
the mode of representation using the graph theory. By using the topological repre-
sentation and measures, cannot be analyzed important indicators as the characteristic
path length, network and global network efficiency based on Euclidian distance, along
the minimum path or the indicators of network centrality (Betweenness, Closeness,
Gravity Index, and Straightness) and clustering coefficient.
Analysis of the connectivity indicators showed that the cancelation of movements
along a road artery (suppression of a link in the primal graph or a node in the dual
graph) by introducing one-way streets influences both connectivity of a node and his
accessibility.
Thus, to connect any node with other nodes of the network, shortest path contains
fewer intersections in case of networks with two-way streets (network diameter).
Reducing the connections through one-way streets lowers the network connectivity
(Alpha index, Beta Index, Gamma Index).
Beyond the benefits of one-ways streets mentioned in the paper, should not be
forgotten that in urban road networks the accessibility function (characterized by the
relation volume/capacity tending to 1 and low speed) has priority over the mobility
function (characterized by high speed and reduced volume/capacity ratio).
References
1. Walker, W., Kulash, W., McHugh, B.: Are We Strangling Ourselves on One-Way
Networks? http://onlinepubs.trb.org/onlinepubs/circulars/ec019/Ec019_f2.pdf
2. The return to two way streets – Copies of research articles, prepared for “The Sydney
tourism improvement group”. Talbot Consultants International Inc., British Columbia,
Ontario, New York, Marked Research, Anlysis and Strategies (2013)
3. Chemmannur, J.: Convertion of one-way street pairs to two-way operations in downtown
Birminghan, a thesis submitted to the graduate faculty of the University of Alabama at
Birminghan, Master of science in Civil Engineering (2008)
4. O’Flaherty, C.A.: Transport Planning and Traffic Engineering. Elsevier Ltd., Oxford (1997).
ISBN 13: 978-0-340-66279-3
5. Dill, J: Measuring Network Connectivity for Bicycling and Walking, School of Urban
Studies and Planning, Portland State Univrsity
6. Jiang, B., Claramunt, C.: Topological analysis of urban street Network. Environ. Plann. B:
Plann. Des. 31, 151–162 (2004). Pion Ltd.
7. Latora, V., Marchiori, M.: Efficient behavior of small-world networks. Phys. Rev. Lett. 87
(19) (2001). Article number 198701
8. TDM Encyclopedia, Roadway Connectivity – Creating More Connected Roadway and
Pathway Networks, Victoria Transport Policy Institute, Decembrie 2015
9. Crucitti, P., Latora, V., Porta, S.: Centrality measures in spatial networks of urban streets
(2005). arXiv: physics/0504163 [v.soc-ph]
10. Masucci, A.P. et al.: Random planar graphs and the London street network. Eur. Phys. J. B
71(2), 259–271 (2009)
11. Porta, S., Crucitti, P., Latora, V.: The network analysis of urban streets: a primal approach.
Environ. Plann. B: plann. Des. 33(5), 705–725 (2006)
12. Kansky, K.J.: Structure of transport networks: relationships between network geometry and
regional characteristics. University of Chicago, Department of Geography, research paper
#84 (1963)
13. Rodrique, J.-P., Comtois, C., Slack, B.: The Geography of Transport Systems. Routledge
Taylor and Francis Group, London and New York (2006). ISBN 10: 0-203-00111-7, 13:
978-0-203-00111-0
14. Mukerji, A.B.: Road transportation network structure and levels of urbanization in
Rajasthan. Nat. Geogr. J. of India 20. Part I (1974)
15. GIS map. http://www.brasovcity.ro/index.php
Cooperative Smart Intersection as an Enabler
of Advanced Traffic Management Systems
Laura Coconea1(&), Gonzalo Alcaraz1,

and Renato Augusto Lira de Andrade2
1
Funded R&D and Strategic Programs Division, SWARCO Mizar, Turin, Italy
{laura.coconea,gonzalo.alcaraz}@swarco.com
2
Politecnico di Torino, Turin, Italy
renato0807@gmail.com
Abstract. Modern society is facing the reality of living in a world suffering the
consequences of past negligence regarding environmental issues: as the
awareness of natural resources limitations increases, governments are imple-
menting measures to contain the progress of the environmental damage. In
transportation (one of its main contributors), this leaded to the creation of reg-
ulations aimed at decreasing fuel consumptions and pollutant emissions. Intel-
ligent Transportation Systems are able to tackle this challenge, mainly because
they have a potential to enable more environment-friendly driving patterns. In
this work, a cooperative smart intersection concept, inducing a smoother driving
style, was deployed and tested under real-life driving conditions. Furthermore,
simulations were performed to investigate how this application would interact in
relation to the association with hybrid technology. The outcomes, besides pro-
viding indications of the application impact on the single vehicle, highlight that
a high penetration rate is mandatory for complete application success.
Keywords: Advanced traffic management systems Traffic light assistant

Hybrid vehicles
1 Introduction
As well known, an important part of the total amount of pollutant is due to the
transportation area [1]. In order to mitigate the environmental impact of transport the so
called Intelligent Transportation Systems (ITS) have already demonstrated their effi-
cacy and indicate to have a rather great potential yet to be explored. As these systems
may be designed to render traffic more efficient, they allow for the development of
more steady-state speed patterns, which result in fuel savings [2, 3].
As described in [4], there may be found 6 major areas in ITS; among these areas the
Advanced Traffic Management Systems (ATMS) allow the traffic managers to gather
information regarding the network (for instance, flow and speed), as well as to control
the on-field devices so as to modify the necessary parameters and render the traffic
conditions as efficient as possible. Roadway loop detectors, variable message signs and
ramp metering are examples of devices that are included in this category. Since these
technology-embedded systems are capable of bringing about improvements on quality,

Cooperative Smart Intersection as an Enabler 669
efficacy, safety and efficiency of the transport service, they result to be beneficial for
many stakeholders of the transport system as well as for the environment.
In addition to these technologies, Cooperative ITS systems (C-ITS) are a rapidly
growing sector in the field and they allow vehicles to communicate with each other and
with the infrastructure, in order to increase their awareness about the road environment,
through the means of innovative services related to specific situations: dangers on the
road, unexpected manoeuvres, traffic lights status, ecc. The C-ITS application presented
in this paper, is called the Traffic Light Assistance (TLA) and enables the provision of
custom assistance for the crossing of equipped intersections in order to reduce number
and duration of stop time, energy consumption, emissions and at the same time increase
both the comfort level of the drivers (either public or private) and the traffic flow.
Section 2 of the document is providing insights on the TLA application while
Sect. 3 following is describing in detail an impact assessment campaign carried on in
Turin within the context of EC-funded project TEAM. Finally, Conclusions and
Acknowledgements sections close the document.
2 Cooperative Smart Intersection
TLA is the main related service actually deployed in urban environments. The so called
cooperative smart intersection provides in-vehicle information about signal phase and
timing of traffic lights, and intersections topologies. [5] provides a clear example of the
benefits that ITS can procure for energy consumption reduction. In this case, the object
of study were the trams operating in Turin and the influence of traffic light priority on
their energy consumption. In relation to the urban traffic signal control (UTC) [6] – that
summarizes the influence of various ITS technologies – states that the UTC procures
fuel savings from 1.6 % up to 50 %, with the majority of them below 20 %. During peak
times, CO2 emissions are reduced by 4 % and by 2 % during periods of fluid traffic.
2.1 Application Insights

The TLA application is based on SPAT/MAP data structures and uses the communi-
cation of the signaling information from the traffic light controller (e.g. red, amber,
green) to all individuals and vehicles using an intersection (e.g. vehicles, bicycles,
pedestrians), in order to support the following use-cases:
• In-vehicle-information of SPAT for the preparation of the driver for increase of the
awareness
• Green light optimal speed advice for a more smooth and fluent traffic flow with less
stops and less emission
• Start-Stop suppression enables the vehicle control to avoid inefficient motor starts
on stops at red lights with only a view seconds.
670 L. Coconea et al.
2.2 Challenges
Having an universal predictor able to give reliable forecast to the phase duration
knowing only the states of the lights in the intersection and their historical data it would
be for sure the key component for large scale TLA deployment. In [7] the authors use a
Kalman filter to predict future probability distributions instead of future switching
points. This method can give probability distributions for different kinds of actuated
traffic lights, and can work also with high delay in data transmission, but is not able to
return an exact value for the switching time. In [8], the authors deal with the challenges
of an extensive prediction of complete urban areas’ traffic light networks, they intro-
duce a real-time prediction algorithm and back-end implementation that is able to
generate signalling predictions for adaptive traffic lights of an entire urban area.
A similar study is presented by AUDI [9], describing an implementation of TLA tested
in Ingolstadt. The system tries to predict the time to the next switch between phases
using probabilities associated to any possible state the traffic lights can provide.
The prediction algorithms that run within the TLA service presented in this paper
are, as described in [10] based in most cases on a “statistical forecasting” method. In
the operation phase, the algorithm choice is automatically executed by OMNIA [11],
by analysing in real-time the model with the most accurate prediction.
2.3 Benefits and Future Evolutions

TLA, as C-ITS systems in general, create a series of benefits for more than one
category of stakeholders [12]: the driver, the passenger, the operator, the freight
managers, the service suppliers. Thus, information generated by the connected vehicle
is used to bring transportation system to a higher plateau in which the actors involved
in it are able to access relevant data so as to make better decisions. Large scale
utilization of diverse technologies and applications allows for the delivery of amplified
benefits on traffic safety, efficiency and environmental impact.
ITS that focus on the operation of intersections and have a prevailing effect on the
urban traffic flow can contribute to the efficiency of urban traffic as showed in previous
sections. One of these ITS solutions, aiming at a reduction of the negative effects of
vehicular traffic and the optimization of traffic flows, is the implementation of adaptive
UTC. In light of arising C-ITS, information from become available and can be used as
an additional data source. In [13] the scope of this work is to explore the potentials that
rely in the high amount of data from connected vehicles in relation to adaptive traffic
control systems in urban areas, exploiting in such way also the feedback channel, from
the vehicle to the infrastructure. This work demonstrates that Adaptive UTC, that aims
at optimizing the traffic flows at signalized intersections, which now relies on road
infrastructure sensors that are costly and bring limited information will be able to
exploit in the near future the C-ITS by using the connected vehicle as an additional data
source.
3 Impact Assessment
The impact assessment of the TLA application is a challenge itself. One of the first works
addressing this issue is [14], where is proved that using an optimal driving strategy it
could be possible to save an average of 30 % of fuel consumption and also if just the first
vehicle is equipped with this system the average speed of the platoon is increased. In [15]
a TLA is tested in a simulation environment. Results suggest that the TLA application has
a positive effect on reduction in stop time and fuel consumption. Further performance
analysis can be found in [16, 17]. In [18], the authors show that the implementation of the
TLA can generally be recommended from the economical point of view. This result is
achieved by applying a simulation study combined with principles of a Cost-Benefit-
Analysis to a real-world demonstration site in Braunschweig. In [19], authors present the
results of a sensitivity analysis and identifies gear choice and the distance from the traffic
light at which vehicles are informed as key influencing factors. Results indicate that a
suboptimal gear choice can void the benefits of the speed adaptation.
The uncertainty of TLA effects in specific situations, and therefore the lack of the
evidence regarding its benefits represent one of the barriers to deployment. Hence the
exploration of these systems is proceeding with both field tests and traffic simulations
[20], like the impact assessment described in [10], based on a dedicated simulation
tool-kit.
3.1 Evaluation Context and Experiments Set-up

The impact assessment campaign carried in this work has been performed in con-
nection to operational activities concerning EC-funded project TEAM, that exploits the
use mobile devices such as smartphones to significantly improve transportation safety
and efficiency, implementing environmental aspects. The applications of the project
have been demonstrated in 7 pilots across Europe, one of this locations being the city of
Turin. The stretch of road, object of this work, in Turin, is 1.6 km long and has a speed
limit of 50 km/h.
One cooperative vehicle was repeatedly driven along this stretch and the trace
recording was planned so as to generate roughly the same amount of traces along the
two driving directions, at the different time periods during the day and involving the
use or less of the TLA application.
In order to diminish the effects of variabilities in driving style on the results, driving
standards were defined based on some eco-driving tips in relation to gear shift and
maintenance of a steady speed in as high a gear as possible.
3.2 Data Elaboration Process

The analysis consisted first in a verification of the capability of the collected data to
represent the real variations in traffic conditions experienced along the experiments. The
second step was then evaluating the impact of the TLA application on the performance
variables chosen. In the third step, simulation scenarios were performed to assess the
influence that the hybrid technology would have on the TLA impact. These scenarios are:
• Simulation scenario 1: simulates the implementation of a Start and Stop system,

present in micro- and mild-hybrids.
• Simulation scenarios 2, 3, 4 and 5: TLA impact on fuel consumption is calculated in
a scenario with a full hybrid vehicle with a parallel configuration.
Next, an assessment of the TLA impact as the speed threshold is modified was carried
out in order to indicate what should be taken into consideration from the point of view
of energy management when hybrid technology is associated with this C-ITS service.
Lastly, the alterations in the TLA impact was evaluated considering a situation in
which the vehicle is able to use the electric motor when the TLA predicts an imminent
stop. The simulation scenarios 4 and 5 were modified and used for the evaluation of the
alterations on the TLA impact.
3.3 Discussion on Results

The experimentation generated 110 trips, distributed along day time periods and both
driving directions. Despite the absence of direct information on the traffic flow, the
observation of the available variables provides as a result that the data collected reflects
what is expected regarding the traffic features along the day. It is rather straightforward
the identification of the peak hours, since more intense load traffic is characterized by
lower average speeds, longer average trip duration and larger number of stops, which
are also longer (Fig. 1).
Fig. 1. Analysis of TLA impact on single connected vehicle

Independently from the motion direction, the impact of TLA resulted to be mostly
positive for the two earlier time periods of the day, however, predominantly negative
for the other three ones (except number of stops, which showed reductions in the last
period and also in the 18–19 period towards South). During the peak hours, the driver
is more limited in one’s attempt to follow TLA instructions because the more loaded
road implies limited control on one’s own speed, which depends on the other drivers’
behaviour, besides the amplified impact of delays and dynamic issues on the traffic
flow. The results therefore highlight how crucial it is to provide the appropriate
incentive for the cooperative smart intersection to reach the desired impact. The lit-
erature already supported this and the results found along the experimentations
developed for this study render even clearer that, in order to take advantage properly of
the benefits the TLA is able to rise, a minimal penetration rate must be provided from
the beginning.
The analysis of the simulation scenarios demonstrated that, when traffic load is
largely high, the TLA impact becomes significantly less favourable as the speed
threshold for switching from ICE to EM is raised. However, this impact tends to be
predominantly negative in the cases of full electric hybrids, whereas some improve-
ments might generated with the use of micro- or mild-hybrids. The TLA produced a
reduction of 50 % of the time during which the vehicle would be using electric power
in the cases where the traffic load was most and remarkably loaded. Taking into
consideration all the time periods and travel directions in simulation scenarios 4 and 5
(which have 30 km/h and 40 km/h as speed thresholds) the variations ranged from a
reduction of 50 % to an increase of 60 %. If the TLA is used not only to modify the
speed patterns but also to work as an input variable for the vehicle control system in
order to allow it to apply electric propulsion instead of the internal combustion engine
when the vehicle is, according to the TLA, about to stop, the TLA impact might be
reduced by around 10 to 15 %.
4 Conclusions
This work aimed at performing an experimental analysis of the impact generated by the
operation of an ATMS based on the concept of cooperative smart intersection pro-
viding TLA in-vehicle service. Moreover, the study also considered how the TLA
would interact with hybrid vehicle technology. It could be developed to further
larger-scale experiments, or in test sites where the traffic lights are not managed by an
adaptive UTC; the fuel consumption analysis could be extended to an energy analysis,
with a simultaneous evaluation of the effect on the electric energy consumption.
Nevertheless, it can be stated that with the raise of the penetration rate C-ITS will
be able to drive the traffic demand, in order to enable reaching the global optimisation
of the transport network and therefore to increase its capacity. In the specific case of the
UTC this is realized by providing the vehicles with traffic light phases and timing
information while at the same time collecting feedbacks that support relevant tactical
control actions.
Acknowledgements. The authors would like to thank the European Comission for the financial
support provided for project TEAM (7th FP, GA no 318621).
References
1. Uherek, E., et al.: Transport impacts on atmosphere climate: land transport. Transport
impacts on atmosphere and climate: the ATTICA assessment report. Atmos. Environ. 44
(37), 4772–4816 (2010)
2. Ennio, C., et al.: La ricerca sui trasporti per la riduzione dei consume energetici e delle
emissioni inquinanti (Italian: The research on transport for reduction of energy consumption
and pollutant emissions). Conferenza Nazionale Energia e Ambiente, Rome, Italy (1998)
3. Tong, H.Y., Hung, W.T., Cheung, C.S.: On-road motor vehicle emissions and fuel
consumption in urban driving conditions. J. Air Waste Manage. Assoc. 50(4), 543–554
(2000)
4. Figueiredo, L., Jesus, I., Machado, J.A.T., Ferreira, J.R., de Carvalho, J.L.M.: Towards the
development of intelligent transportation systems. In: Proceedings IEEE Intelligent
Transportation System Conference, pp. 1206–1211 (2001)
5. Berger, M., et al.: Effetti della priorità semaforica e degli ITS sui consumi energeti-ci del
trasporto pubblico tramviario (Italian: Effect of traffic light priority and of ITS on tramway
public transportation energy consumption). Tesi di Laurea, Politecnico di Torino (2014)
6. Shaheen, S.A., Lipman, T.E.: Reducing greenhouse emissions and fuel consumption:
sustainable approaches for surface transportation. IATSS Res. 31(1), 6–20 (2007)
7. Protschky, V.: Adaptive traffic light prediction via Kalman filtering. In: Intelligent Vehicles
Symposium Proceedings, pp. 151–157. IEEE (2014)
8. Protschky, V., et al.: Extensive traffic light prediction under real-world conditions. In: 80th
Vehicular Technology Conference (VTC Fall), pp. 1–5. IEEE (2014)
9. Bodenheimer, R., et al.: Enabling GLOSA for adaptive traffic lights. In: Vehicular
Networking Conference (VNC), pp. 167–174. IEEE (2014)
10. Bottero, M., Coconea, L.: A. Malano: Impact assessment of an adaptive cooperative
decentralised traffic control system. In: Mobil.TUM – International Scientific Conference on
Mobility and Transport: Technologies, Solutions and Perspectives for Intelligent Transport
Systems Munich (2015)
11. Swarco: OMNIA. http://www.swarco.com/mizar/Prodotti/OMNIA. Accessed 08 July 1993
12. Chiara, B.D., et al.: ITS nei trasporti stradali. Ed: Egaf (2013)
13. Politi, S., Kaths, J., Pasquero, M., Coconea, L.: Integration of floating vehicle data for online
traffic signal control optimization. In: 11th ITS European Congress, Glasgow, Scotland
(2016)
14. Sanchez, M., et al.: Predicting traffic lights to improve urban traffic fuel consumption. In: 6th
International Conference on ITS Telecommunications Proceedings, pp. 331–336. IEEE
(2006)
15. Katsaros, K.: Performance study of a green light optimized speed advisory (GLOSA)
application using an integrated cooperative ITS simulation platform. In: 7th International
Wireless Communications and Mobile Computing Conference (IWCMC), pp. 918–923.
IEEE (2011)
16. Stevanovic, A., et al.: Impact of signal phasing information accuracy on green
light-optimized speed advisory systems. In: Transportation Research Board 92nd Annual
Meeting (2013)
17. Eckhof, D., et al.: Potentials and limitations of green light optimal speed advisory systems.
In: VNC, pp. 103–110 (2013)
18. Niebel, W.: Cost-benefit-based implementation strategy for green light optimised speed
advisory (GLOSA). In: Mikulski, J. (ed.) TST 2013. CCIS, vol. 395, pp. 313–320. Springer,
Heidelberg (2013). doi:10.1007/978-3-642-41647-7_38
19. Tielert, T., et al.: The impact of traffic-light-to-vehicle communication on fuel consumption
and emissions. In: Internet of Things (IOT), pp. 1–8. IEEE (2010)
20. Baur, M., et al.: Realizing an effective and flexible ITS evaluation strategy through modular
and multi-scaled traffic simulation. IEEE ITS Mag. 2(3), 34–42 (2010)
Bus Routing Safety for the Transportation
of Children to School
Eleni Chalkia1(&), Jose Maria Salanova Grau1, Evangelos Bekiaris1,

Georgia Ayfandopoulou1, Chiara Ferarini2, and Evangelos Mitsakis1
1
Center for Research and Technology Hellas – Hellenic Institute of Transport
(CERTH-HIT), Thessaloniki, Greece
hchalkia@certh.gr
2
University of Modena and Reggio Emilia (UNIMORE), Reggio Emilia, Italy
Abstract. The school transportation has a lot of hidden issues that make it
difficult to be organized and managed (i.e. safety, punctuality, time, etc.). This
paper analyses the school bus routing problem and proposes routing algorithms
that could be applied when planning transportation for students who go to school
using school buses. Furthermore, this paper proposes a routing methodology for
pedestrians (from home to the bus stop and school bus), in order to provide to
children a safer and more efficient way to school from door to door. Also, various
solving algorithms are used, presenting a good performance (the genetic algo-
rithm present up to 12 % overestimation of distance in relation to the best
solution, with an average overestimation of 5 %) and are being compared for a
real case, taking into account the safety framework presented in this paper. These
algorithms have been developed within the context of the European Project
SAFEWAY2SCHOOL.
Keywords: School bus routing problem Routing algorithm TSP VRP
1 Introduction
Going to and from school is a daily transportation necessity for millions of children of
age 6–16, within Europe. For example in Sweden, 250.000 children go to school daily,
in Poland approximately 700.000 and about 450.000 in Austria [2]. This transportation
is realized using various modes (car and bus primarily) and it involves various actors
(from the parents themselves to private companies). Accidentability is a significant
issue in this kind of transportation. According to accident statistics, school bus travel is
much safer than car travel to school [3]. This is due to various reasons, including the
strict policy applied to the professional transportation of pupils to school when buses
are used, from routing issues to behavioral issues within the bus, but this is not the
scope of the current paper.
In recent decades, there has been a growing concern about the safety of children
during their transportation to and from school. SAFEWAY2SCHOOL1 EU Project has
developed methods and tools enhancing the safety of children who travel to and from
1
http://safeway2school-eu.org/ - 2009–2012.

Bus Routing Safety for the Transportation of Children to School 677
school by the bus. These methods and tools focus on bus transport, taking into con-
sideration the door-to-door safety, and the methods and tools which have been
developed to provide support both in the planning and monitoring of school bus
services. This paper describes in particular the Safe Route Planning approach and the
‘Safe Map’.
The vehicle routing problem (VRP) is one of the most complicated optimization
mathematical models; the school bus routing problem (SBRP) is an important and
practical branch of this problem. Today, several methods are used to solve the VRP and
many of these methods can be applied to the school bus routing problem. However, due
to the heuristic nature of the methodologies used, the near optimal solutions may vary.
In addition to the numerous studies that addressed the vehicle routing problem, various
software methods have been developed that can be utilized to minimize the operating
cost. School bus routing problem is a complicated transportation problem. Three are the
major factors make school bus routing unique [4]:
• Efficiency (the total cost to run a school bus).
• Effectiveness (how well the demand for service is satisfied).
• Equity (fairness of the school bus for each student).
2 Methodology for Solving the SBRP

in SAFEWAY2SCHOOL
From the literature, we have seen that many researchers have tried to solve the SBRP
by introducing various constraints and objectives. Some of the most recent examples
used the branch-and-cut methodologies, while others used the
branch-and-cut-and-price methodologies, or the two-commodity network flow formu-
lation methodology.
Identifying and solving this problem, taking into account all the possible con-
straints, is of great complexity and the difficulty. Nevertheless, one major issue taken
into account that has not been identified by any of the researchers in the literature is the
part of the safety of the children while on board. In SAFEWAY2SCHOOL, we con-
sidered safety as one of the major issues that need to be taken into account. Thus, we
defined a concept that separates the SBRP in three sub-problems, which are the
following:
• The SAFE MAP.
• The pedestrian routing.
• The school bus routing.
Each one of the above topics is elaborated in detail at the following sections.
2.1 Safe Map

The ‘Safe Map’ is a cartographical representation of safety-related data connected to
school bus travel. It is a digital tool that can provide decision support to the organisations
678 E. Chalkia et al.
or persons responsible for planning school bus services (i.e. bus operators, local
authorities, school travel organisers, etc.).
The Safe Map concept involves constructing a representation of the journey to
school which reflects the safety (or risk) level of each link of the journey. The Safe Map
is then used as a basis, to select the safest bus route, location for the bus stops and also
to define the safest pedestrian routes to and from bus stops.
In order to build the map, it is necessary to carry out the following steps:
1. Define a set of safety criteria for the assessment of each part of the journey.
2. Identify the data requirements for each set of criteria; establish how the data can be
collected;
3. Establish how the resulting scores will be represented on the Safe Map.
2.2 Safety Criteria Used in the Presented Methodology

For a school bus operator or local authority responsible for school transport services, it
is very useful to have a systematic way of assessing the safety level of the road network
of the transportation service. In order to do this, it is necessary to define criteria which
can be used to produce a safety-related ‘score’ for each road section (link).
There are various ways of making such an assessment. For example, a systematic
examination can be made of the accident statistics to find out which parts of the
network have a record of accidents, in other words, are recognised ‘black spots’.
Another approach is to examine the characteristics of the road and traffic in order to
identify the parts of the road network which have the greatest potential risk. If the
necessary data is available, and also the time to undertake such a survey, the ideal
solution would be to combine these two types of approach.
Accident data can be represented on the Safe Map in the relevant geographical
location using symbols of different dimensions and colour to indicate the number, type
and gravity of accidents which have occurred in a given period, e.g. in the last five years:
• Fatal accident involving a pedestrian: child/adult.
• Accident with serious injuries: child/adult.
• Accident with minor injuries: child/adult.
• Different symbols could be used for accident involving a school bus and for
accident involving other types of vehicle.
Among others factors, the potential safety of a link is affected by its geometrical
characteristics, features such as crossings and intersections, the lighting and signage,
and the traffic behaviour, as well as various ‘dynamic’ factors relating to weather
conditions, e.g. areas prone to ice. Examples of features which can be responsible for
safety risks in rural and urban road environments are listed below:
• Rural areas:
– Sharp bends
– Lack of protective barriers on the road shoulder
– Narrow two way roads without a central partition
– Lack of lighting
• Urban areas:
– Lack of good visibility for making turns (due to buildings or on street parking)
– Fast roundabouts
– Unprotected (left) turns
– Complex intersections
With regard to traffic conditions, the criteria normally taken into consideration are:
• Traffic flow (volume of traffic/hour)
• Traffic speed (legal speed limit on a given section).
In order to determine the safety potential of a road, it will therefore be necessary to
select the most relevant factors according to the local context. This is because different
factors are relevant for different road types (as indicated above for rural and urban
areas) and also for different regions and countries. For example, in Sweden the risk of
snow and ice makes certain roads dangerous. Good road lighting is also important
because of short days in winter. On the other hand, on many country roads in Italy the
presence of roadside barriers is a critical safety factor where there are steep drops, as
well as central barrier when the roads are narrow.
2.3 Pedestrian Routing

In real life the pedestrian routing is a problem that really concerns the families of the
students. It really affects only the children of a certain age, since the very small ones are
picked up from their home door. On the contrary to the actual school bus routing
methodology, the literature for the pedestrian routing is quite poor. Additionally, there
is only some literature that refers to the allocation of the students at the bus stops, based
upon the distance criterion, but without referring at all to the route that is going to be
taken from the child.
A holistic methodology for the school routing was developed in 2006 that also
included the allocation of the children to the bus stops [5]. This methodology, deter-
mined a set of potential stops, so as each student has at least one stop that he/she can
walk to. Afterwards, it determined the school bus routes so that all students are picked
up at a stop they are allowed to use, while making sure that the capacity of the buses is
not exceeded. In this way it was possible that a stop is not used from the bus, as long as
the student is allowed to go to another stop.
In the process proposed in this paper, the allocation of the children to the bus stop is
going to be based upon the distance criterion, but the final allocation of the students to
the bus stops is going to be based only upon the safety criterion. The safety of the route
that the students will take is going to be calculated by using the pedestrian safety
criteria of the previous section. Thus, on the contrary to most vehicle routing formu-
lations, in which a set of stops is given and only the school bus routes need to be
determined, we will present the pedestrian routing problem as a standalone issue, which
is not affected by the vehicle routing, but it is only affected by the children’s safety. In
SAFEWAY2SCHOOL pedestrian routing problem we identified a set of potential stops
based upon specific bus stop safety criteria, find the set of stops each child can visit
according to distance criterion and finally determine for each student to which stop
should move to, so that the safety of the walking route is maximum.
The algorithm that is going to be used for the final allocation of the students to the
bus stops is going to be a variation of the Dijkstra’s algorithm that is used to solve the
shortest path problem. In our case we will use the Dijkstra’s algorithm in order to solve
the safest path problem. According to Dijkstra the node at which we are starting is
called initial node and the distance of node Y is the distance from the initial node to
Y. Dijkstra’s algorithm will assign some initial distance values and will try to improve
them step by step. On the contrary, SAFEWAY2SCHOOL pedestrian algorithm will
assign safety values to each link.
The steps go as following:
1. Assign to every node a safety value: set it to zero for our initial node and to infinity
for all other nodes.
2. Mark all nodes as unvisited. Set initial node as current.
3. For current node, consider all its unvisited neighbors and calculate their tentative
safety.
4. When we are done considering all neighbors of the current node, mark it as visited.
A visited node will not be checked ever again; its safety recorded now is final and
minimal.
5. The next current node will be the node with the highest safety in the unvisited set.
If all nodes have been visited, finish. Otherwise, set the unvisited node with the
maximum safety (from the initial node, considering all nodes in graph) as the next
“current node” and continue from step 3.
2.4 School Bus Routing

The “Safe Map” presented above, is now the basis for the routing algorithms to be
implemented. First the pedestrian algorithm is implemented, which follows the Dijk-
stra’s algorithm framework that is used to solve the shortest path problem, which is
used in order to solve the safest path problem. When the children are assigned to the
bus stops, the final branch of the problem follows, which is their route to the school, on
the bus.
As should be clear from the previous discussion, most papers on the routing of
school buses have focussed on capturing as many aspects of reality as possible,
building intricate multi-objective models with many complex constraints. On the other
hand, we focus on a basic version of this problem to develop efficient methods that take
the safety issue of the problem into account. Our problem formulation has only a single
school, one type of student and one type of bus, with fixed capacity and has been based
upon the multi-trip elementary shortest path problem with resource constraints [1].
We will try to optimize the following two vehicle routing criteria:
– The total time travelled by each student.
– The safest path taken from the school bus.
Assumptions:
– The route starts and ends at the school (depot).
– The cost of the route is proportional to the safety of the route and the time travelled
from each child.
– Travel times between each stop are known and accurate.
– Demands at each stop are known.
– Loading time per child is constant for every child (=1 min per child).
– The route will be executed only if there are more than 10 children on board.
– If the bus stops at a bus stop, the children will be boarded.
– Each stop is served only once.
– Each stop is served by one route.
Constraints:
– The bus capacity is not exceeded by the number of children on board.
– Each child should not be in the bus for more than one hour.
Constants:
– M = Set of nodes.
– B = Set of links.
– I = Set of customers.
– Δ = Set of routes.
– a = weight of the Time factor.
– b = weight of the Safety factor.
Variables:
– Sik: The safety mark of arc (i, k).
– xik,δ: binary variable (0–1) of arc (i, k) for route δ.
– yik,δ: Flow at the arc (i, k) in route δ.
– LV: Capacity of the school bus.
– LT: Time limit of the route.
– Tik: Travel time of the arc (i, k)
– Di: Demand of node i.
Mathematical Programming formulation:
X X
Min xik;d ðaTik þ bSik Þ ð1Þ
d2D ði;kÞ2B
XX
xik;d ¼ 1 8i 2 I ð2Þ
d2D k2M
X X
xik;d xki;d ¼ 0 8i 2 I; 8d 2 D ð3Þ
k2M k2M
X
xsk;d ¼ 1 8d 2 D ð4Þ
k2M
X
xks;d ¼ 1 8d 2 D ð5Þ
k2M
yik;d LV xik;d 0 8ði; kÞ 2 B; 8d 2 D ð6Þ

X X X
yik;d yki;d þ Di xik;d ¼ 0 8i 2 I; 8d 2 D ð7Þ
k2M k2M k2M
X
Tik; xik;d LT 8d 2 D ð8Þ
ði;kÞ2B
X X
Di xik;d 10 8d 2 D ð9Þ
i2I k2M
xik;d 2 f0; 1g 8ði; kÞ 2 B; 8d 2 D ð10Þ
yik;d 0 8ði; kÞ 2 B; 8d 2 D ð11Þ
Where xik;d ¼ 1, if arc (i, k) is used in route δ and 0 otherwise; and yik;d ¼ 1, if
equal to the amount of flow on arc (i, k) for route δ for all (i, k) that belong to B.
The objective function (1) indicates that the total cost of all routes, which is the
weighted sum of the arcs safety and travel time, should be minimised. Constraint
(2) specifies that the customer node i may be visited only once. Constraint (3) requires
that the vehicle should enter and leave a customer node an equal number of times in the
same route. Constraints (4) and (5) define that each route starts and ends at the school
(depot). Constraints (6) guarantee that the vehicle capacity is not exceeded in any route.
Constraints (7) require conservation of flow of each route at each customer node.
Constraint (8) limits the total time of a vehicle’s route to the time limit. Constraint
(9) puts a lower bound in the number of students of each route. Constraints (10) and
(11) are the integrality and non-negativity requirements on the variables.
3 Conclusions
The current paper is an extract of an extended research realized within

SAFEWAY2SCHOOL EU project, about the safety and the security of children
transportation to and from school, following a holistic approach that covers the
transportation from door (home) to door (school).
In SAFEWAY2SCHOOL we have followed an approach that first of all, splits the
SBRP in three separated ones and also, we focused at the safety perspective of it. Thus,
three individual problems emerged which needed separate development in order to get
the greater picture of the SBRP solved. These three problems are the “Safe Map”, the
“Pedestrian Routing” and the “School Bus Routing”. These three problems were solved
separately and at the end were combined in order to provide efficient (the total cost to
run a school bus), effective (how well the demand for service is satisfied) and equal to
all students (fairness of the school bus for each student), school bus transportation.
Acknowledgements. The research presented in this paper has been conducted within the con-
text of the SAFEWAY2SCHOOL (SAFEWAY2SCHOOL) project “Integrated system for safe
transportation of children to school”, funded by the Research Directorate General of the Euro-
pean Union under the Seventh Framework Programme for Research and Technological
Development.
References
1. Akca, Z., Ralphs, T.K., Berger, R.T.: Solution Methods for the Multi-trip Elementary Shortest
Path Problem with Resource Constraints. Optimization Online the Mathematical Program-
ming Society (2010)
2. Anund, A., Larsson, J., Dukic, T., Pauzié, A., Gadegbeku, B., Tardy, H., Chalkia, E.,
Ferrarini, C., Aigner-Breuss, E., Pripfl, J., Jankowska, D.: SAFEWAY2SCHOOL Deliverable
1.4 “Accident Analysis” (2010)
3. Kelsi, L.: School buses safer than cars for students (2013). http://www.fredericknewspost.
com/news/education/education_topics/policy/article_b1891a6f-2efb-5c2d-9b8c-eff56fef5bb3.
html
4. Nayati, M.A.K.: School bus routing and scheduling using GIS. Master thesis, May 2000
5. Schittekat, P., Sevaux, M., Sorensen, K.: A mathematical formulation for a school bus routing
problem. In: IEEE Proceedings of the 2006 International Conference on Service Systems and
Service Management (2006)
Improving the Road Traffic Regulation
in the Area of Roundabout Intersections
Function of the Traffic Streams Size
Elena Neagu(&), Andrei Alexandru Boroiu, Ionel Vieru,

and Alexandru Boroiu

{elena.neagu,ionel.vieru,alexandru.boroiu}@upit.ro,
boroiu_alexandru@yahoo.com
Abstract. When it is not possible to improve the level of service of a roundabout

by increasing its capacity, but only by diverting some of the traffic flows from the
respective area, it is required to identify each traffic flow from that intersection.
The determining of these traffic streams can be achieved by the proposed ana-
lytical relationships, depending on the traffic flows measured by fixed observers.
This method of analytical determining the respective traffic streams in roundabout
intersections was used for analyses of the traffic in areas with traffic problems in
the municipality of Pitesti, completed with proposals for a better regulation of
road traffic in the area of the Podul Viilor roundabout. The introduction of these
measures is expected to lead to the elimination of traffic congestion in the
roundabout analyzed and to improve road traffic in this type of intersection.
Keywords: Stream Flow Roundabout Congestion Regulation
1 Stating the Problem and the Objectives
In recent years, in the major cities of Romania a phenomenon of roundabout prolif-

eration has been remarked, being noted that this type of intersection brings forward a
number of advantages to the specifics of urban traffic. As a result, there is constant
concern to improve the service level of these intersections.
The main mathematical models that evaluate the ability of each arm of the inter-
section are based only on the volume of the conflict flow that may easily be recorded by
a traffic operator (an operator for each arm of the intersection). There are two alternative
mathematical models that can be used for a roundabout with a single lane on each
runway [4, 5]: one model has only one variable, i.e. the volume of the conflict flow, and
the second has two variables - the volume of the conflict flow and the volume of the
outgoing flow on the arm - dimensions that can be determined equally simply, by one or
two observers for each arm of the intersection. So, in order to assess the analytical
capacity of a roundabout, it is sufficient to measure all circular flows (and, for one of the
methods, all outgoing flows) - and this is absolutely feasible by using 1 or 2 traffic
monitors placed on each arm of the intersection - and the verisimilitude of the analytical
model can be verified by an effective measuring of the actual incoming or outgoing flow
(Fig. 1) when congestion occurs on the entry lanes of the intersection arms.

Improving the Road Traffic Regulation in the Area 685
Fig. 1. The flows that can be measured with fixed observers in a roundabout intersection
(Source: CNADNR 2009, HCM 2000)
As the setting up of a road junction (with the regulation of the route followed by
each access lane, with delimitation of the circular lanes by means of continuous
markings or lane dividers) is very important to increase its capacity [2], there emerges
the need to know the traffic streams from the intersection (what is the arm of entry in
the intersection and what is the arm of exit from it). But even more important is the fact
that if these traffic streams are known, one can reconsider regulating the traffic in the
area where the roundabout lies, by diverting some of the flows from the vicinity of the
intersection so as to reduce the load of the intersection in question and, therefore, to
eliminate the potential for traffic congestion.
It is very difficult to count these traffic streams because this would mean recording
the route followed by each vehicle that enters the respective intersection (the arm of
entry into the intersection and the arm of exit from the intersection for each vehicle),
which, in fact, would mean tracing the route of each vehicle. This is practically
impossible to be done by traffic monitors, the only feasible solution being that of
recording the traffic in the whole intersection and then identifying the route followed by
each individual vehicle.
This method requires the use of expensive equipment and, moreover, is very
laborious and consumes a lot of time for the processing of the recordings.
As a result, the following question arises: is it possible for an intersection with
roundabout-regulated traffic, based on measured values (with fixed observers) for
ingoing flows, outgoing flows and conflict flows (circular flows), to determine the
volume of the traffic streams - Fig. 2?
Such a question resulted in the following case in point: after 12/10/2015, when the
Calea Bucuresti - Calea Bascov roundabout was put into service in Pitesti, before the
upstream overpass was commissioned, there appeared road congestion at peak hours
and the question was asked by local authorities if this phenomenon would disappear
after the overpass road artery was functional, i.e. after completing and putting into
service the entire road facility.
In order to formulate a scientifically-founded answer, there emerged the need to
determine the U-turn flow on the Eastern arm and, reviewing the literature of specialty,
686 E. Neagu et al.
Fig. 2. The traffic streams that correspond to a roundabout arm (Source: HCM 2000)
it was found that there are no analytical relations to allow the determining of this flow
function of the streams that can be measured by fixed observers (incoming, conflict,
outgoing or even right-turn flows).
As a consequence, the objectives of this paper were formulated: is it possible to
identify traffic streams in a roundabout based on the flows measured by fixed observers
and, if so, what are the analytical relations between the volume of such traffic streams
and the volume of the flows measured in fixed points?
2 Mathematical Modeling of Traffic Flows in a Roundabout
To find the answer to this question, road circulation in the roundabout was formalized
by noting the volume of traffic flows (number of standard vehicles in a particular
timeframe) as similar as possible to the notations in specialized literature (written in
Romanian or in a foreign language), as follows:
• ingoing flows volume: Vi, m;
• outgoing flows volume: Ve, n;
• conflict (circular) flows volume: Vc, m;
• traffic streams volume, Vm, n.
In the above notations, the index m represents the sequential number of the ingoing
arm (origin), m = 1, 2, 3, 4, …, and n is the sequential number of the outgoing arm
(destination), n = 1, 2, 3, 4 ….
The first three categories of flows are the flows that will be measured, so they
represent the given data for the previously stated problem, and the last category of
flows represents the values to be determined, therefore they constitute the unknown
data of the problem. As a result, this issue will be analyzed to discover if it can be
solved, starting with the simplest intersection.
2.1 The 3-Arm Roundabout
• The case with the fewest variables is the one in which there are no U-turn flows
(return flows), a situation in which theoretical research [3] led to the following
conclusion: it is sufficient to measure the 3 circular flows (which will be needed to
calculate the level of service) and the 3 ingoing flows or three outgoing flows
(preferably, the 3 outgoing flows because they will be useful when determining the
intersection capacity through the alternative method specified above as well).
• But the most general case for the intersection under study is that in which possible
U-turn flows are taken account of on each of the three arms of the intersection, a
situation in which [3] if the three right-turn flows are measured, all the other traffic
streams can be determined analytically (3 left-turn flows and 3 U-turn flows).
The relations on the 3 arms are similar and, using a single notation for all arms of
the intersection, the analytical relations for U-turn flows and left-turn flows can be
formulated as follows:
leftarm: rightarm rightarm:
VUturn ¼ Vconflict ðVingoing Vright: Þ ð1Þ
and, from here:
Vleft ¼ Vingoing ðVright þ VUturn Þ ð2Þ
where:
• Vingoing, Vright, Vleft, VU-turn are the volumes of the flows related to the arm for which
the calculation is performed (arm under analysis);
leftarm
• Vconflict is the conflict volume for the arm to the left of the arm under study;
rightarm
• Vingoing is the entry volume for the arm to the right of the arm under study;
rightarm
• Vright is the volume of right-turn for the arm to the right of the arm under study.
In the first analytical relation, the flow that enters the circular path on the right arm,
rightarm rightarm
ðVentry Vright: Þ, was highlighted in between parentheses because, when there is a
lane reserved for the right turn (as happens on the Northern arm from the Podul Viilor
roundabout), the flow that enters frontally through the annular path can be measured
directly by a fixed observer (it shall not be calculated as difference of the two flows
measured by two fixed observers).
Based on these relations, using the data measured by fixed observers in the 3-arm
Calea Bucuresti – Calea Bascovului roundabout of Pitesti, an answer could be provided
to the former question: on the Eastern arm, the U-turn flow represents 70 % of the
incoming flow (i.e. about 1700 standard vehicles/hour of the 2400 standard
vehicles/hour representing the incoming flow during peak traffic) and it was estimated
that this flow, which increases the conflict flows for each of the other two arms, will be
reduced substantially after the overpass artery road is put into service, so the traffic
congestion reported will no longer take place. The new measurements, made after
688 E. Neagu et al.
24/03/2016 – when the Calea Bucuresti overpass was put into service – confirmed the
prediction: the incoming flow, as well as the U-turn flow, were reduced by about 700
standard vehicles/hour (to be found now in the North-South flow on the artery that
overpasses Calea Bucuresti), but the U-turn flow still has a high value that cannot be
ignored: approximately 1000 standard vehicles/hour. This result, obtained analytically
using the formulas above, revealed the need for further analysis in order to organize
traffic in the roundabout so as to obtain a further reduction of the vehicles that make a
U-turn on the Eastern arm (Fig. 3).
Fig. 3. Traffic streams generated by the incoming flow on the Eastern arm of the Podul Viilor
roundabout
2.2 The 4-Arm Roundabout

From the very beginning, the object of analysis was the simplest scenario possible,
when no U-turn flows are taken into account on any arm, and it was found [3] that in
the case of the 4-arm intersection, the 12 traffic streams cannot be determined ana-
lytically function of the 12 flows that can be measured by fixed observers. However, if
one considers the flows of vehicles that turn right as parameters (there being, however,
the possibility to measure them by fixed observers), the other 4 × 2 = 8 streams of
traffic can be determined analytically – the flows of vehicles going forward and the
flows vehicles turning left.
Thus, it was concluded that the 4 flows going forward and the 4 left-turn flows can
be determined analytically, only if the 4 conflict volumes, the 4 right-turn volumes and
four other streams - either the 4 incoming flows or the 4 outgoing flows - are measured.
The research that was carried on in [1] showed that, as in the case of the 3-arm
intersection, in this case relations for the 4 arms are similar and, by using a single
notation for all arms of the intersection, analytical relations for the forward and left-turn
flows can be formulated as follows:
forwardarm rightarm rightarm
Vleft ¼ Vconflict ðVincoming Vright Þ ð3Þ
and, from here:

Vforward ¼ Vincoming ðVright þ Vleft Þ ð4Þ
where:
• Vincoming, Vright, Vleft, Vforward are the volumes of the flows related to the arm for with
the calculation is performed (arm under analysis);
forwardarm
• Vconflict is the conflict volume for the arm in front of the arm under analysis;
rightarm
• Vincoming is the incoming volume for the arm to the left of the arm under analysis;
• Vdr:
bratdr:
is the right-turn volume for the arm to the right of the arm under analysis.
3 Synthesis of the Theoretical Results Obtained
Cumulating the results derived from the theoretical research performed for the papers
[1, 3] conducted to discover whether it is possible to analytically determine the traffic
streams in roundabouts based on traffic measurements performed with fixed observers,
the following results can be summarized:
• In the case of 2-arm roundabouts (actually, circular road facilities, for purposes of
traffic calming and to enable the vehicles to perform a U-turn), analytical solving is
not required.
• In the case of 3-arm roundabouts, the determining of the flows of conflict and the
incoming or outgoing arms to get the 6 streams of traffic is necessary. Instead, the
scenario with U-turn flows proved unsolvable based on the flows measured with
fixed observers. There is one single possibility in this case: measuring the flows of
vehicles that turn right with fixed observers (this is, however, possible), a case in
which the 6 traffic streams for the vehicles that turn left and perform a U-turn in the
intersection can be determined - rel. (1) and (2).
• In the case of 4-arm roundabouts, it is possible to determine the traffic streams only
for the scenario “without U-turn flows”, but only if the flows of vehicles that turn
right (which can be measured, however, by fixed observers) are considered as
parameters, there being a unique solution for the unknown eight traffic streams
(vehicles going forward or vehicles turning left) - rel. (3) and (4).
This means that for 3-arm intersections with U-turn flows and for 4-arm inter-
sections without U-turn flows it is possible to fully know the traffic streams based on the
flows measured by fixed observers (mandatorily including right-turn flows), which is an
absolutely exciting achievement, as knowing these traffic streams will allow the
reconsideration of the regulation of the traffic in the roundabout area, for purposes of
making the traffic smooth in the area and eliminating the possibility of road congestion
occurrence.
690 E. Neagu et al.
4 Research Based on Traffic Observations for the Purposes

of Submitting Proposals on Road Traffic Organization
in the Podul Viilor Area
The calculating relations for the 3-arm intersection with U-turn flows yielded results
that led to predictions about the way in which road traffic would be conducted in the
Calea Bucuresti – Calea Bascovului roundabout in Pitesti after 24 March 2016, when
the road artery that overpasses Calea Bucuresti was put into service - predictions that
were confirmed by the new measurements carried out. But, as previously stated, after
the Calea Bucuresti overpass was commissioned, the incoming flow on the Eastern arm
and the U-turn flow decreased by only 700 standard vehicles/hour (vehicles that are to
be found from now on in the North-South flow on the artery that overpasses Calea
Bucuresti), but the U-turn flow still has a high value that cannot be ignored: about
1000 standard vehicles/hour.
The hourly amount of the U-turn flow for the Eastern arm (Fig. 4) was calculated
each time using the relation (1) and with consideration of the temporal length of traffic
recordings (5 min):

Easternarm
VUturn ¼ 12 ½ðVconflict
Westernarm
Vingoing
Northernarm
Vright
Northernarm
ð5Þ
Although, initially, it was hoped that this substantial U-turn flow will decrease over
time as drivers that enter the city via Podul Viilor will understand that there appeared an
alternative left-turn solution (a new road that will be better signaled), things have not
moved forward in this respect, as seen from the traffic measurements re-executed 3
months after the commissioning of road passage. After the measurements performed by
fixed observers, it was found that, during peak hours, the length of the waiting line on
the Northern arm is maintained at high values, even 3 months after the commissioning of
the road passage (approximately 40 vehicles waiting in that line during the peak hours).
Without any doubt, the main solution to this problem still remains that of identifying
ways to reduce the amount of U-turn flows on the Eastern arm. In order to achieve this,
Fig. 4. The use of the passage that overpasses Calea Bucuresti by vehicles coming from the
Arges Bridge
Fig. 5. Degrees of difficulty for alternative routes on the way of return on the Eastern arm
the only answer is to find ways in which a large part of the vehicles going this way in
order to reach either the Market area from the Ceair neighborhood, the Popa Şapcă
neighborhood or the Târgul din Vale – Gh. Şincai roundabout (the University of Pitesti
Rectorate) can take the passage that overpasses Calea Bucuresti at Podul Viilor (Fig. 4).
Analyzing the current opportunities that vehicles crossing the Arges bridge have for
the purposes of reaching the destinations previously mentioned, one can notice the
following (Fig. 5):
1 – In the Market area from the Ceair neighborhood, the only way forward is
through the Calea Bucuresti – Calea Bascovului roundabout (easy access) because, if
they used the road passage, they could not enter the Crinului Street, under the railway
bridge, after getting off the overpass, the access being forbidden (there is a sign
ordering “Forward or to the left”).
2 – The access to the Popa Şapcă neighborhood is simpler when following, in a
similar manner, the return route on the Eastern arm of the roundabout, because then one
can enter either the first street to the right, the Crinului Street or the Sf. Vineri Street
(even in this case, because even if access is permitted after using the road passage on
Sf. Vineri Street, this access entails ceding priority to the flow of vehicles going on Gh.
Şincai Street - difficult access); the same is true for Râurilor Street, the route being even
longer.
3 – for those who want to reach the Rectorate roundabout (and, most of them, from
here to the Maior Şonţu intersection to take the Calea Craiovei), the fact that after
crossing the Calea Bucuresti – Calea Bascov roundabout (into which, basically, they
enter quite easily, because the flow of conflict they encounter is almost non-existent,
and the very wide annular path has 3 lanes materialized by tridimensional separators,
markings for lane separation being introduced only recently, but with a discontinuous
character) do not encounter “obstacles”, other than three pedestrian crossings that are
692 E. Neagu et al.
Fig. 6. The 3 alternative routes for the Podul Viilor - Maior Sontu Street - Targul din Vale Street
intersection
not very much used up to the Rectorate intersection, the following alternative being
preferable to the other two possible alternatives (Fig. 6): either use the overpass artery
and the Rectorate roundabout intersection, or follow the Calea Bucuresti – 1 Mai Street
– Maior Şonţu Street.
5 Proposals for Improving Traffic Regulation
The proposals for regulating road traffic resulting from this research are as follows:
1 – a reconfiguration of the central island from the intersection so that vehicle
access is temporized, the delineation of lanes on the roundabout annular path by
continuous and/or discontinuous longitudinal markings and pre-signaling of movement
directions on the annular path - so that the circulation of motor vehicles in the
roundabout could approach as much as possible the “equal opportunities for inter-
section access” desideratum.
2 – setting up a roundabout at the intersection of the Crinului and Gh. Şincai
Streets (Fig. 7), in order to reduce the U-turn traffic on the Eastern arm, with the
following new road regulations:
• removal of the ban to turn right, underneath the railway bridge, for the motor
vehicles that follow the overpass in the North-South direction, so that the motor
vehicles coming from the Depozitelor Street and those coming from the Arges
Bridge, with the intention to enter the Popa Sapca neighborhood, will be encour-
aged to follow this new path (the motor vehicles descending from the overpass
artery in a southward direction will be able to turn right to enter the Crinului Street);
Fig. 7. The timing and disciplining of traffic in a roundabout intersection via a suitable
configuration of the central island and by means of longitudinal markings
• the motor vehicles descending on Crinului Street will be able to enter Costache
Negri Street under the railway bridge;
• the motor vehicles coming from the South on Gheorghe Sincai Street will be able to
turn left on Crinului Street.
Given the grounding of the proposals on the theoretical and experimental research
submitted, the introduction of these measures is expected to lead to the elimination of
traffic congestion in the roundabout under study.
References
1. Boroiu, A.A., Boroiu, A.: Evaluarea curenţilor de trafic din intersecţiile în sens giratoriu pe
baza fluxurilor măsurate în puncte fixe (Evaluation of traffic streams in roundabouts based on
flows measured in fixed points), TRANSLU 2016, Bucharest (2016)
2. Kimber, R.M.: The traffic capacity of roundabouts, Crowthorne (1980)
3. Neagu, E., Boroiu A.A., Rizea, V., Boroiu, A.: Study of the opportunities of analytical
determination of the origin-destination flows in roundabout intersections based on data
provided by fixed observers. In: ACME 2016, Iaşi (2016)
4. Highway Capacity Manual (HCM): Transportation Research Board, National Academies of
Science, USA (2000)
5. CNADNR Normativ pentru amenajarea intersecțiilor la nivel pe drumuri publice (Standard for
the Development of Intersections at Public Roads), Search Corporation (2009)
Traffic Optimization in Urban Area -
Roundabout Versus Lights Case Studies
Nicolae Filip(&) and Claudiu Golgot
Technical University of Cluj Napoca, Cluj Napoca, Romania

nicolae.filip@auto.utcluj.ro
Abstract. More local administration choose roundabout as an optimum way to

organize the traffic in urban cross roads. If roundabout has some benefits as a
way to increase the traffic safety not in all cases it assure traffic flow in good
conditions. Taking into consideration the effect of roundabout through the
congestions produced in traffic we make some evaluations regarding the Service
Level of roundabout in accordance with the traffic flow. We analyze four
roundabouts from three cities (Cluj Napoca, Bistrita and Alba Iulia). The aim of
our work was to identify the roundabout effect on traffic flow. Some conclusions
were writing about the possibilities to increase urban traffic quality in accor-
dance with good practice regarding the traffic organization in cross roads case.
Keywords: Roundabout Lights Optimization Safety Congestion
1 Introduction
The roundabout represents the most way to organize the cross roads in all the cities of
country in the last decade. There is an explanation of that and it consists of [1–4]:
• reducing the cost comparing with signalized cross roads;
• low cost of maintenance;
• low vehicle speed;
• reducing the accidents effect (Accident Rate Cost).
As a result of mentioned benefits a lot of roundabouts were developed in all the
cities from country. But not in all cases its increase the traffic flow and cover the traffic
demands.
Traffic signals cause unnecessary delay for many reasons [5]:
– The need to provide a minimum green time to each movement in every cycle creates
time intervals in which no vehicles are entering the intersection.
– The need to provide for the most critical of two or more movements that proceed
simultaneously results in an ineffective use of green time by non-critical
movements.
– The need to provide a common red light for all directions simultaneous for assure a
safety exit for permitted displacements.
– The “lost time” associated with startup and termination of a green phase detracts
further from the amount of time that is available for moving traffic.

Traffic Optimization in Urban Area - Roundabout Versus Lights Case Studies 695
– Left turns that take place from shared lanes impede the other movements in the
shared lanes unnecessarily. This results in a very inefficient utilization of the
available roadway space.
– Heavy left turns, even from exclusive lanes, require dedicated phases that rob time
from the major movements and increase the total time lost due to startup and
termination of traffic movements.
Signals are mechanical devices that not only require maintenance but also peri-
odically malfunction. They are also dependent upon electrical power and do not,
therefore, provide any control during power failures. Many signal violations occur at
higher speeds so that the severity of accidents is often high. Permitted left turns and
right turns on red introduce additional conflicts.
A roundabout overcomes all of these disadvantages. There is no sequential
assignment of right-of-way and therefore no wasted time. Left turns are not subordi-
nated to through traffic. Vehicles enter under yield control instead of stop control and
therefore have lower headways and higher capacities. There are no electrical compo-
nents to malfunction.
Roundabouts, on the other hand, have their own limitations:
– Since roundabout operation is not periodic, it is not possible to coordinate the
operation of roundabouts on an arterial route to provide smooth progression for
arterial flows.
– Roundabouts offer the least positive form of control. Each vehicle entering the
intersection must yield to all traffic that has already entered, but they still experience
less accident than a comparable signal or stop control.
– Roundabouts impose a new form of traffic control that is not familiar to many
drivers, but experience has found that drivers learn quickly how to drive in a
roundabout.
Therefore, roundabouts are not the solution to all traffic problems at all locations.
Careful study is required to identify the most appropriate control method at any given
location. The studies required to justify the installation of traffic signal control and
all-way stop control are based on the warrants and requirements set [6, 7].
2 Roundabout and Signals Traffic Capacity Evaluation
The roundabout capacity is estimated using empirical mathematical models or some

programs as SIDRA and Synchro.
The limited capacity of roundabouts is the upper limit on the volume of traffic load
(amount of conflict volume of the ring in the right way and access input on the volume
of such access).
Traffic conflict on the median ring is calculated with the relation [8]:
Vci ¼ Vturn
i1
þ Vleft
i2
þ Vline
i3
ð1Þ
696 N. Filip and C. Golgot
where:
Vi is the volume related conflict on arm i, calculated arm;
Vi−1 turn is the volume of input return for the right arm located in the calculated
(counter clockwise direction);
Vi−2 left is the volume of left-turning entrance, afferent arm opposite calculated;
Vi−3 line is the volume of input going forward, for the arm to the left of the
calculated (clockwise).
The capacity of a roundabout with one lane path is determined with the equation:
Vc tc
Vc e 3600
Ca ¼ Vc tf ð2Þ
1 e 3600
where:
Ca representing capacity of the arm (veh./hour);
Vc is volume related conflict arm (veh./hour);
tc access critical time between 4.1–4.6 s;
tf is the follow time between 2.6–3.1 s.
The control delay is calculated with the relation:
2 vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3
u 2 3600 Vi
u
3600 6 Vi t Vi 7
di ¼ þ 900 T 4 1 þ 1 þ Ca Ca 5 þ 5 ½sec =veh ð3Þ
Ca Ca Ca 450 T
where Vi is the entry volume for the arm i.

The cumulated delay for vehicles is finally calculated with the relation:
di;1 Vi;1 + di;2 Vi;2 + di;3 Vi;3 + di;4 Vi;4

dg ¼ ð4Þ
Vi;1 + Vi;2 + Vi;3 + Vi4
The roundabout service level is assigned using the data from Table 1.
Table 1. The roundabout service level.

Level of service Delay [sec]
A <10
B 10–15
C 15–20
D 20–25
E 25–30
F > 30
Regarding the sign control evaluation we estimate will be relevant the results
offered by Synchro 7 simulation, because it offer as a results the same parameter:
delay/vehicle.
3 Case Study
Analysis of individual roundabouts usually shows acceptable levels of service. But the
theory provides little information on roundabouts cascade. If in the case of signs we
may apply successfully “green light”, to the roundabout there is not a similar alter-
native. In this context the intention was to determine the effect of the succession of
roundabouts analyzed using case studies. As a case study was chosen a sequence of
four roundabouts (Fig. 1).
Fig. 1. Four roundabouts placed on the same road.
The distance between two successive roundabouts is between 300 meters and 550
meters. Thus, between the first and the last roundabout was measured a distance of
1185 meters.
In this study we analyzed the effect of each individual roundabout and determined
their cumulative effect.
3.1 Traffic Data

Traffic data were collected over a period of 2 h by a team of three persons, each person
having one account traffic device (Fig. 2) [9, 10]. Every intersection was made a
recorder for a period of about 15 min.
Fig. 2. Traffic data counter (patent of Road Traffic and Safety laboratory from Technical
University of Cluj Napoca), and traffic data file
The data obtained from the survey of traffic were

• total volume of vehicles during observation;
• total volume of vehicles per direction of travel;
• breakdown by categories of road users of vehicles;
• time passing vehicle in front of the observer.
Traffic data through the interface software provides information on:
• the flow of vehicles;
• sequencing interval between vehicles;
• absolute and relative frequencies detailed;
• directions of travel;
• categories of road users.
We perform traffic data collection for each cross road. An example of the results of
the processing of data files is shown in Fig. 2 and Table 2.
3.2 Results and Discussions

With the relations (1)–(4) it was calculated the level of service of each roundabout. The
results are presented in Table 3.
We also consider a proposal to hold intersections with traffic lights. For this case
were determined in Synchro 7 program the following parameters: delays/vehicle and
ICU (Intersection Capacity Utilization) (Figs. 3 and 4).
From Fig. 4 intersections with traffic lights coordination results that lowers the
vehicle in traffic delays. Given by roundabout causes delays are:
• reduce speed when approaching a roundabout;
• crossing the road by pedestrians;
Table 2. Vehicles volume data and travel direction
Table 3. Level of service and traffic indicators.

Roundabout no. Rouandabout capacity Traffic signals
indicators
Dealy/veh Service level index Delay/vehicle ICU
1 12 B 8 65.1 %
2 32 F 12 72.4 %
3 28 E 11 67.7 %
4 16 C 11 69.4 %
5 39 F 28 79.2 %
Fig. 3. Delay/vehicle Synchro simulation results

Fig. 4. Complete Synchro report for an intersection.
• imbalance between the volumes of vehicles moving in conflicting directions;

• too little distance between two consecutive roundabouts.
4 Conclusion
The case study showed that the roundabout is not the best alternative routing inserted in
any situation. Of course, if we analyze this in terms of local administrations, the
roundabout is an alternative requiring low cost and does not involve continuous
monitoring of volumes of vehicles.
From another point of view to assist drivers by means of ensuring the safety on the
road in this case it is low. If we chose to detail cascade roundabouts because they are
found in many cities, what are their effects on traffic and environmental pollution:
• significantly longer trips; increased fuel consumption;
• urban pollution growth;
• reducing the safety of participants in traffic.
Of course, there are cases when the roundabout is the best solution. We demon-
strate this by: displacement of heavy traffic; junction between a main stream with a
second stream of traffic with low values; small towns where the arrangement of traffic
lights is not justified by financially point of view.
References
1. Brilon, W.: Roundabouts: a state of the art in Germany. Paper Presented at the National
Roundabout Conference, Vail, Colorado, 22–25 May 2005
2. Ambros, J., Slabý, P.: Comparison of roundabout accident prediction models: challenges of
data collection, analysis and interpretation. In: Votruba, Z., Jeřábek, M. (eds.) Conference
20th Anniversary of Faculty of Transportation Sciences, pp. 42–52. Czech Technical
University, Prague
3. Lambertus, F.G.H.: Turbo roundabouts: design principles and safety performance. The
National Academies of Sciences, Engineering, and Medicine. Vol. 2096. doi:10.3141/2096-03
4. Mahdalova, I., Krivda, V., Petru, J.: Disorders of structure elements on roundabouts. In:
SGEM2015 Conference Proceedings, 18–24 June, vol. 2, pp. 89–96 (2015). ISBN
978-619-7105-43-8/ISSN 1314-2704
5. Roundabouts vs. Signals and stop signs. http://www.roundaboutsusa.com/design/
roundabouts-vs-signals.html
6. Ren, L., Qu, X., Guan, H., Oh, E.: Evaluation of roundabout capacity models: an empirical
case study. J. Transp. Eng. (2016). http://dx.doi.org/10.1061/(ASCE)TE.1943-5436.000087
7. Highway Capacity Manual, Transportation Research Board, National Research Council,
Washington, DC, November 2004 Update
8. Roundabouts: Informational Guide, U.S. Department of Transportation, Federal Highway
Administration, Publication No. FHWA-RD-00-067
9. Filip, N., Chiriciuc, M.: Electronic device for traffic count, patent no. 019017/26.02.2010
10. Filip, N., Moldovan, M., Golgot, C.: Access optimization in an urban area: a case study. In:
(Conference Paper)/WIT Transactions on the Built Environment, vol. 128, pp 347–358
(2012). ISSN 1743-3509. doi:10.2495/UT120301
Mathematical Algorithm for Calculating
the Total Traffic Lights Cycle in Junctions
Lucian Matei(&), Ilie Dumitru, and Laurentiu Racila
University of Craiova, Craiova, Romania

mateiclucian@gmail.com, dumitru_ilie@yahoo.com,
racila_laurentiu@yahoo.com
Abstract. Because the current development of cities and increased motoriza-

tion, mobility in the city/metropolis suffers greatly. In these conditions the
development of new traffic control methods in junctions is important. Condi-
tioned by the current technical development, control methods are developed
more in the area of mathematical algorithms which can be used to control the
traffic equipment’s with a new efficiency. So the new intelligent traffic control
modes involve the use of even more efficient and faster mathematical
algorithms.
Keywords: Traffic Algorithm Saturation Flow Lane grouping
1 Introduction
The focus of interest on intelligent transport systems should be seen as a response to the
current context of increasing demand problem in the transport sector. On the one hand,
profound structural changes in the automotive industry and even disturbances in the
transport industry, creates new requirements of the structural framework of the trans-
port sector. On the other hand, congestion and environmental problems make it
impossible to satisfy, this increased demand, by simply extending the existing road
network and transport services. Thus the concept for an intelligent system for moni-
toring and controlling traffic, requires not the integration of an existing system but a
combination of several equipment specially grouped in the area and constantly inte-
grated with existing and prevailing needs to form an intelligent system.
This paper aims, is to take an overall concept (calculation of traffic light cycle) and
transpose it to a own mathematical model into an automatic calculation process by
using arrays to generate all combinations of movements existing in an intersection with
four arms [1]. The mathematical algorithm calculus will be based on these base
structures:
• Channelling movements in intersections
• Identify intersection right of access based on the conflict matrix
• Grouping movements and movements matrix creation
• Traffic composition – vehicles categories, percentage of heavy vehicles, etc.
• Calculation of saturation flow
• Calculation traffic light timings

Mathematical Algorithm for Calculating the Total Traffic Lights 703
2 Grouping Movements in Intersection
To calculate a traffic light junction we should initially define the ways in which
movements are made so that through those means we have to define in an mathematical
way how lanes are grouped [1, 2]. Such groups are combinations of lanes and
movements that can be performed in traffic light phases, thus identifying two kinds of
groups of movement generated based on the conflict between them:
• Protected movement – is the movement that runs on a green traffic light without
having any other movement conflicts (vehicles, pedestrians or cyclists) to take its
right of access.
• Allowed movement – is the movement that run on green traffic light and on its path
there are one or more conflicting movements (vehicles, pedestrians or cyclists) to
which must yield right of access.
The next step is to define in a mathematically representation what a group of lanes
looks like, how many maximum groups of lanes exist for an arm and the most
important factor: what a conflict between movements within an intersection means [3].
For this we used the software Mathcad that allowed us to use a mathematical
approach to the concept and for the creation of the algorithms than using a simple
programing software that can’t give us the depth and understanding of the complex
problem like determining a traffic light cycle.
2.1 Possible Combinations of Moves for the 4 Arms

According to the conflict matrix at an intersection, several initial steps for calculating
the total traffic light cycle were identified and the following, we think are most relevant
in the mathematical process.
Movement group generation. All possible moves without conflict for all arms were
identified and correlated with combinations based on the following assumptions and
considering that all movements are allowed (moves to the left, forward and right and
the turning movement is correlated with the left one). This part can be explained using
the signal plan phase system (the algorithm is used to identify in the end all possible
combination of movements that will give us for the entire junction all the possibilities
of group movement that can be combined into maximum of 4 traffic light phases) [3, 4].
The Fig. 1 shows the mathematical algorithm:
After this the next phase is to remove from the results all combinations that cannot
be achieved or are identical and to renumber them so that in the final step of this stage
we will renumber them in traffic light phases, following the following criteria below
and the algorithm from the Fig. 2:
• Right movement will always be grouped with the forward movement (line 2 and 4
of the matrix result from the mathematical algorithm in the Fig. 2)
• Movement to the left is separated (line 1 and 3 of the matrix result from the
704 L. Matei et al.
Fig. 1. Calculation and storage of possible combinations at a traffic lights coordinated junction
Fig. 2. Mathematical algorithm used to calculate all the possible movements groups for a
maximum of 3 phases.
mathematical algorithm in the Fig. 2)

• Each group of movement will be at least one time alone
• All opposite movements will be grouped at least once
The next step is to create all possible combinations of traffic light phases within a
junction starting from the premises that in a junction with 4 arms the maximum of
signal phases is 4 (this is because if we add more phases it will increase the delay time
in the intersection). The algorithm will calculate the possibilities based on the following
criteria:
• Combinations of 2 grouped movements in one phase (two movements are free, two
are grouped) – (this means that the result will be a maximum of 3 phase matrix)
• Combinations for 3 grouped movements in one phase (a movement is free the rest
are grouped) – (this means that the result will be a maximum of 2 phases matrix)
• Combinations for 4 movements grouped in one phase (all movements are grouped
together without conflict)
The Fig. 3 shows the mathematical algorithm and the result of one of the calcu-
lating criteria (F1 through F3 are the phases used in the signal plan of a traffic light
cycle).
Fig. 3. Mathematical algorithm for creating all combinations of 3 phase matrixes.
Final matrix generation. This is the last stage of the first section of calculation of the
traffic light cycle for an intersection with four arms. As a result of previous calculations,
multiple matrices of movements groups will result [5]. So the next step is to group
these matrices in a final matrix that will group all possible moves and also can group
these movements in traffic light phases (the mathematical model developed uses up to
four separate traffic light phase for the whole junction).
Fig. 4. Augmenting of the group matrix in the final traffic light phase matrix
The important aspect of the above figure is that the first four lines are West and East
arms and the next are South and Nord. This division is done because of a matrix
conflict between arms in case of a traffic light. The significance of the matrix lines is as
follows:
• 1 and 5 - left movement in West and respectively South as in the Fig. 4
• 2 to 6 - forward and right movement to the West and respectively South and as said
above the forward movement is combined with the right movement as in the Fig. 4
706 L. Matei et al.
• 3 and 7 - left movement in the East and respectively North as in the Fig. 4
• 4 and 8 - forward and right movement in the East and respectively North as in the
Fig. 4
Also traffic light phases were made, and for each group of arms 3 phases are used
so that the overall intersection phases will be 4 like in the Fig. 4.
3 Traffic Light Cycle Calculation
3.1 Input Data in the Mathematical Algorithm

Any calculation for calculation of traffic light timings assumes the existence of some
basic information about the characteristics of the intersection and the volume and
structure of transit vehicles (this can be seen in the Fig. 5). To use this information in
the mathematical algorithm we need to takes this data automatically from a file of
constants parameters and import it in the calculations.
Fig. 5. Input information’s that enter in the mathematical algorithm
3.2 Saturation Flow

The next step is to calculate the saturation flow (Eq. 1) for each group of lanes,
respectively for each traffic light phase extracted from the matrix finals [5, 6].
The saturation flow can be calculated with the following formula:
s ¼ s0 N fW fHV fg fp fbb fa fLU fLT fRT fLTp fRTp ð1Þ
Where:
• s = saturation flow rate for the lane group
• s0 = base saturation flow rate
• N = number of lanes in the lane group

• f = adjustment factor for prevailing condition (w = lane width; HV = heavy vehi-
cles; g = grade; p = parlung; bb = local bus blockage; a = area type; LU = lane
use; RT = right turn; LT = left turn; Rpb = pedestrian/bicycle interference with
right turns; and Lpb = pedestrian/bicycle interference with left turns)
For this calculation we use in addition to a normal saturation flow calculation
additional arrays that simplify the calculus for both saturation flow and traffic light
cycle (gradient lane factor for the vehicle group matrix, calculation of final number of
lane usage, etc.). In the Fig. 6 above you can see some of the mathematical formulas
relevant to the process of calculating the saturation flow like:
• The adjustment factor for parking (this is important because it calculates how the
flow is modified when parking maneuvers are used in the intersection Fig. 6a)
• Adjustment factor for lane usage (Lane usage is calculated when multiple move-
ments share the same lane, like the forward movement with the right movement and
the mathematical algorithm can be seen in the Fig. 6b)
• Adjustment factor for right turn (because the right movement is in general combined
with the forward movement we need to calculate the factor of usage when the right
turn lane is utilized, this can be seen in the figure Fig. 6c)
Fig. 6. Algorithm for: a. Adjustment factor for parking, b. Adjustment factor for lane usage, c.
Adjustment factor for right turn
3.3 Traffic Light Cycle

To calculate the traffic light cycles following steps are needed to be realized:
• Determining the interval for switch and protection time (Fig. 7) – a portion of each
yellow change plus red clearance period is not usable by vehicles. The sum of these
two periods comprises the switch and protection time for the phase.
708 L. Matei et al.
• Determine the minimum time required for pedestrian crossing (Fig. 8) – a minimum
time for pedestrian crossing is calculated based on the number of pedestrians
crossing, crossing distance, etc.
• Determine the length of traffic light cycle (Fig. 9) – this approach analyzes the
intersection with the heaviest traffic to determine a minimum cycle length and used
that to set the remaining intersections.
• Determine the length of traffic light phases (Fig. 9) – this algorithm calculates the
phase time based on the cycle length, traffic volumes, etc.
Fig. 7. Lost time calculation for all the group lane possibilities
Fig. 8. Green time calculation for pedestrian for all the group possibilities
Fig. 9. Total cycle and phase time calculation
It should be mentioned that the mathematical process, like in previous stages,

involves the use of the final group matrix, matrix that contains the necessary infor-
mation on traffic light phases and also on group movements, and information from the
input file described above [6].
The result of this mathematical process is an array with multiple columns, meaning
that the mathematical model calculates data for the same primary base input set for all
primary traffic light phases (signal plans) possibilities [7].
4 Conclusion
The upper limit of the mathematical algorithm is a junction with 4 arms and 3 lanes for
each direction of travel. The mathematical algorithm presented takes into account all
aspects of a signalised intersection as calculated traditionally but the novelty that it
brings is that it takes into account all traffic light phases possibilities that can be used in
an intersection (taken all modes of movements groups and signal groups up to four
phases). In this way, the mathematical algorithm can be used for an intelligent traffic
monitoring and control system, so that it could bring a benefit in the following
situations:
• The model takes into account all the possibilities of movement groups based on a
matrix of conflict so that you can use in real time to ensure the best possible
functioning of traffic when used in link with traffic controllers.
• Overall traffic is divided into two zones: peak zone and non-peak zone. Using the
created algorithm we could generate an optimal signal plans for each intersection
individually and for each set of values for vehicles volume. Such algorithm can be
used it to planning an adaptive signal or can be used to create plans for an automatic
signal traffic controller in fixed system based on multiple signal plans.
• Because the algorithm automatically calculates all the possibilities for traffic light
times it can be used in the calculation and maintenance of a green light coordination
system.
References
1. Treiber, M., Hennecke, A., Helbing, D.: Congested traffic states in empirical observation and
microscopic simulation. Phys. Rev. E 62(2), 1805–1824 (2000)
2. Florea, D.: Road Traffic Management (Ingineria Traficului Urban), 2nd edn. University
Transylvania of Brasov Publisher, Brasov (2000)
3. Webster, F.V.: Traffic Signal Settings. Road Research Laboratory, London (1958)
4. Miller, A.J.: A computer control system for traffic networks. In: Proceedings of the 2nd
International Symposium on Traffic Theory, pp. 200–220 (1963)
5. Mirchandani, P., Head, L.: Rhodes Traffic-Adaptive Control Systems, The Atlas Research
Centre, Systems and Industrial Engineering Department, The University of Arizona, Tucson,
Arizona (2011)
6. Ilie, D., Nicolae, D., Lucian, M., Laurentiu, R.: Saturation flow mathematical model based on
multiple combinations of lane groups. In: XII Conference on Transport Engineering -
Cit2016, Valencia (Spain) (2016)
7. Ilie, D., Lucian, M., Matei, V., Laurențiu, R., Theodor, O.: Aspects regarding priority settings
in unsignalized intersections and the influence on the level of service. In: Proceedings of the
European Automotive Congress EAEC-ESFA 2015, pp. 687–694. Springer (2016). ISBN:
978-3-319-27275-7 (Print) 978-3-319-27276-4 (Online)
Study on the Influence of Intersections
with Forest Roads upon the Traffic Flows
on Highways
Constantin Alexandru Bratu and Dinu Covaciu(&)

{alexandru.bratu,dinu.covaciu}@unitbv.ro
Abstract. The intersections between forest roads and highways are the con-
nection points where the forest road detaches from the route of the highway.
A heavy vehicle, like a log hauling truck, that enters to or exits from a forest road,
having low speed, will create queues on the highway. The present study is based
on records made on a truck with trailer, that travelled on a national road, passing
through different intersections with forest roads. Data were collected using GPS
receivers mounted on different points of the truck-trailer assembly, and processed
using traffic statistics methods, implemented in custom software applications.
The results of the tests are useful both for optimizing the geometrical parameters
of the intersections, and for increasing the safety and comfort of drivers during
the steering maneuvers in forest road – highway intersections.
Keywords: Road traffic Road safety GPS data acquisition Heavy vehicle
1 Introduction
The intersections forest road – highway are connection points where the forest road
detaches from the route of the main road (highway) [1]. They are civil-engineering
constructions (common platforms) where the road traffic shows a number of specific
characteristics. This kind of engineering amenities intersects two different types of road
traffic. Thus, slow forest vehicles, with increased mass and dimensions, travels along
with cars on the same road surface.
The intersections between the two types of roads require a series of studies to establish
how the traffic is carried out on the national road, influenced by the steering maneuvers in
the intersections. The results of the respective tests are necessary both for optimizing the
geometrical parameters of the intersections, and for increasing the safety and comfort of
the drivers during the steering maneuvers when entering/exit the forest road.
In the framework of this experimental study, an objective was to highlight how the
traffic flows on the highway (national road) are influenced by the steering maneuvers of
the heavy vehicles (trucks carrying logs) that passes through the intersections of forest
roads with the national road.
Study on the Influence of Intersections with Forest Roads 711
2.1 Research Method

The method used within the experimental tests was the Moving observer. This method
is used in the case of attaching to the test vehicle, which covers certain routes or
performs the steering maneuvers, the measuring equipment. The equipment is used to
acquire data related to the kinematics of the vehicle (position, time and instantaneous
speed of the vehicle) [2].
Preparation of the instrumented vehicle consisted of the following steps:
• checking the documents which certify the drivers’ skills (identity card, driving
license, etc.);
• verifying the technical condition of the vehicle;
• adding the trailer or the load of logs (when the tests are intended to be made with
the payload given by the carried logs);
• placing the GPS antennas in the following points: on the roof of driver’s cab, at the
rear of truck, on the logs loading crane, on the stakes at rear-left of truck and trailer.
The GPS receivers were placed in open areas, where least signal screening should
occur;
• connecting the data acquisition computers, or the logging devices, to electric energy
sources (external batteries or the vehicle’s cigarette lighter socket);
• connecting the GPS antennas to the acquisition computers/loggers;
• switching on the devices and starting the software applications.
The software tools used for logging data are those provided by the manufacturers of
the acquisition equipment (including the research teams from Transilvania University).
For data processing it was used AutoCAD as main software support, on which custom
Lisp applications were running [2].
2.2 Instrumented Vehicle

The vehicle used in the experiments was a straight truck equipped for forestry oper-
ation: Scania N3 164C with a trailer, type D Rema 04, shown in Fig. 1. The log hauling
trucks are equipped with powerful engines and transmissions that provide high torques
at low speeds. Installation of data acquisition equipment is influenced by the vehicle
superstructure, which consists mainly in bunks with stakes (made from steel in this
case) and a loader mounted at the rear of the straight truck.
The test vehicle was driven by two drivers duly certified for this type of operation.
For confidentiality reasons, the two drivers will be hereinafter referred to as Driver 1
(D1) and Driver 2 (D2).
2.3 Research Site

The tests were conducted in/ between the intersections of forest roads with the national
road DN 1A, in the sector between km 167+000–183+000. This road connects the
towns Sacele and Ploieşti, Romania, as shown in Fig. 2 and Table 1.
712 C.A. Bratu and D. Covaciu
Fig. 1. The instrumented vehicle during tests
Fig. 2. Site of the tests (Google Earth)
Table 1. Localization and identification of the studied intersections

No. Intersecting roads
National Road (Public Forest Roads
Road) DN1A
Name Position kilometer Detaching direction Name
1 DN1A km 179 + 900 left D177 (FE032)
2 DN1A km 175 + 600 right D135 (FE022)
3 DN1A km 169 + 100 left D152 (FE010)
2.4 Data Acquisition Equipment

The GPS based data acquisition devices used for performing the tests were made at
Transilvania University of Brasov, Department of Automotive and Transport Engi-
neering. Two types of devices were used, with the following technical characteristics:
• GPS 1. DS-5, a system based on the GPS 18x-5 Hz receiver, from Garmin: this is a
GPS receiver mostly used for operating the machinery in civil engineering and
agriculture; such a device is used due to its proven accuracy in measuring the
position, time, speed and course [3].
• GPS 2, 3 – DL-10, a logger that uses as motherboard an Arduino Leonardo pro-
totyping platform, with a microcontroller ATMega32U4; on this was installed a
GPS Logger Shield, made by Adafruit; the GPS receiver is a PA6H module [4].
2.5 Conducting the Experiments

The test vehicle performed a total number of 24 steering maneuvers in the intersections
forest road – national road, with both drivers. Thus, each driver performed individually
a number of 12 steering maneuvers. During these tests the vehicle had no payload.
A complete route includes two steering maneuvers (exit and entry from and to the
forest road), and the road sector between the two intersections (Figs. 3 and 4).
Fig. 3. Example of a test track represented in Google Earth
The tracks represented in Fig. 4 are drawn in AutoCAD using a custom developed
application. Each record from the log file is converted in a point (as geometric entity)
represented as a small cross, with custom data attached to the entity. These data include
speed, direction (course, as angle relative to North), time and position (longitude and
latitude) For a correct representation, the geographic coordinates (recorded in WGS84
datum) are transformed in cartesian coordinates according to the local projection
system (Stereo 70). The three different tracks represented in the figure are recorded
using the three GPS antennas installed on the truck/trailer assembly.
Fig. 4. Part of a track recoded with the test vehicle (GPS1 on the cab roof; GPS2 at the rear of
the straight truck; GPS3 at the rear of the trailer).
A close look to Fig. 4 will give information about how the truck was positioned
during the steering maneuver. Further in this paper, only the data from GPS1 (front of
the truck) and GPS3 (rear of the trailer) are used.
3 Data Processing and Results
Data processing is based on the microscopic traffic flow analysis [5–8]. Each data
record represents not only a position on the track, but also a speed value in the frame of
the traffic flow that characterize the road sector. Multiple travels of the same type of
vehicle, with different drivers, represent the basis for statistical processing of traffic
parameters. Key statistics used in the traffic theory to describe the speed distributions
include [7]: the average speed (time mean speed), standard deviation of speed values,
85th percentile speed V85, 50th percentile speed or the median speed V50, the pace
interval. The mean speed, the median speed and the pace are all measures of the central
tendency, describing the approximate center of the distribution. The standard deviation
is a measure of dispersion, describing the extent to which data spreads around the
center of the distribution.
Formula used for calculation of the mean speed, when multiple values are recorded
at fixed time interval, is:
P
fi vi
i
v ¼ ð1Þ
N
where v is the average speed, fi is the frequency of observations in the speed group i, vi
is the middle value of speed in the group i and N is the total number of observations.
Standard deviation is calculated using the following relation:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P
ðvi vÞ2
s¼ ð2Þ
N1
where s is the standard deviation, vi is the value of the speed i, v is the average speed,
and N is the number of observations.
Fig. 5. Variation of the travel speed, depending on space. GPS 1. Route D152 – DN1A – D135.
Driver 2.
Fig. 6. Variation of the travel speed, depending on space. GPS 3. Route D152 – DN1A – D135.
Driver 2.
Fig. 7. Variation of the travel speed, depending on space. GPS 1, GPS 3. Route D152 – DN1A
– D135. Driver 2.
Fig. 8. Zoom. Variation of the travel speed, depending on space. GPS 1, GPS 3. Route D152 –
DN1A – D135. Driver 2.
The median speed, as well as the 85th percentile and 15th percentile speeds (V85,
V50 and V15) are determined graphically, using the diagram of the cumulative fre-
quency distribution [8]. The time interval used for calculating the speed distribution is
the interval of logging, which is 0.1 or 0.2 s, depending by the logging device.
3.1 Speed Variation Depending on Distance

In order to determine which is the distance needed for the vehicle to accelerate or
decelerate in the vicinity of the intersection, there can be used the time and speed
values recorded by the GPS loggers. Samples of the recording speed between some of
the test intersections are given in the figures below.
In the figures above it can be seen the speed recorded for the test vehicle when
detaching from the forest road, travels approximately 200 m until reaching the speed of
the traffic flow on the national road. A similar distance seems to be necessary in the
deceleration phase, when the vehicle is preparing to steer for approaching a forest road.
The last two figures show the accuracy of the measurements. This is defined by the
precise overlapping of the curves which show the travel speed. The speed values were
recorded by two independent GPS devices, placed at the ends of the vehicle (GPS 1 –
driver’s cab; GPS 2 - trailer).
All diagrams of speed variation, as well as the diagrams of speed distribution
frequency, used in the study, are generated with a custom software application that runs
under AutoCAD [2].
3.2 Determination of the Speeds V15, V50 and V85

The 15th percentile speed is defined as the speed under which are 15% of the records.
A vehicle moving at or under speed 15th, is considered to produce an effect on the traffic
flow, being an obstacle that determine queues formation [5–7]. Thus, this consideration
is important in the present research, as there will be highlighted the space on which the
forest vehicle is equivalent to an obstacle on the national road. The 50th percentile speed
is the speed which is reached or exceeded in 50% of the records. The 85th percentile
speed is the speed at which the vehicles are traveling less than 85% of time.
The percentile speeds are determined graphically, based on the diagrams of
cumulative frequencies (see Fig. 9) of speed distribution.
Fig. 9. Determination of the speeds V15, V50 and V85. GPS 1, GPS 3. Route D152 – DN1A –
D135. Driver 2.
In Fig. 9, one can notice reaching the percentiles speeds 15th, 50th and 85th for the
coverage by the test vehicle of the route 4-3. In this way, the values obtained are:
• for the route covered by the driver’s cab (GPS 1), the following values were
obtained: 15th: 40.79 km/h; 50th: 47.90 km/h; 85th: 57.16 km/h;
• for the route covered by the trailer (GPS 3), the following values were obtained:
15th: 40.05 km/h; 50th: 47.68 km/h; 85th: 56.85 km/h.
Hence, these diagrams were plotted for each maneuver, obtaining the final results of
the percentile speeds 15th, 50th and 85th. For 15th, there were determined the distances
traveled under this speed value, with respect to the center of the intersections.
The complete results of the tests are presented in Table 2.
The values recorded in the table above stand for the following notifications:
– Intersecting Roads, National Road/Forest Road, Symbol (columns (1), (2), (3)) are
in accordance with Table 1;
– Maneuver (4) – represents the type of maneuvers performed in the intersections of
forest road with national road, according to Fig. 10;
– Driver (5) – represents one of the two drivers (D1 and D2, respectively);
– DS15 (6) [m] – is the distance traveled by the test vehicle in the vicinity of
intersections, at a travel speed lower than V15. These values were determined by
the graphical measurement of the distances covered by the GPS 1 and GPS 3
receivers, since the time of departure from the intersection, until the stop time in the
adjacent intersection (according to Figs. 5 and 6). The moments of start, respec-
tively of stop, were considered to be the moments in which the vehicle was com-
pletely outside the carriageway of the national road;
– DDS15 Max (7) [m] – is the maximum distance measured from the center of the
intersection to the farthest point where the travel speed of test vehicle is lower than
V15. These values were determined graphically;
– MSD [m] (8) – is the sight distance, measured on field;
– NSD [m] (9) – is the stopping distance necessary for a vehicle traveling at maximum
legal speed, so that the vehicle might stop under safety and comfort conditions, at
the emergence of a fixed obstacle on the carriageway. The value of this distance was
calculated as the maximum value between the following:
d ¼ 0:27 v þ 0:0125 v2 þ 5½9 ð3Þ
d ¼ 0:555 v þ 0:025 v2 þ 10 ½10 ð4Þ
d ¼ v=5 þ v2 =100 þ 10 ½11 ð5Þ
where d is the traveled distance and v is the initial speed.

Table 2. Distances traveled under V15, in the vicinity of intersections between forest roads and
national road
No. Intersecting Symbol Maneuver Driver Dist. Design Stopping distance [m]
roads DS15 Dist.
Nat. For. [m] DDS15 Meas. Ned.
Road Road Max [m] stop. dist. stop. dist.
MSD NSD
0 1 2 3 4 5 6 7 8 9
1 DN1A D177 1 (ItS) D1 27.35 64.42 119.28 182.35
D2 48.25
2 (IsS) D1 42.54
D2 54.36
3 (ItD) D1 40.85 66.86 134.85
D2 62.24
4 (IsD) D1 58.36
D2 62.97
5 DN1A D135 2 (ItS) D1 45.65 86.40 21.78 182.35
D2 43.63
6 (IsS) D1 52.97
D2 83.40
7 (ItD) D1 165.21 150.14 86.10
D2 146.84
8 (IsD) D1 149.67
D2 148.61
9 DN1A D152 3 (ItS) D1 53.38 175.22 112.07 182.35
D2 70.90
10 (IsS) D1 81.88
D2 182.36
11 (ItD) D1 91.58 142.01 104.06
D2 80.39
12 (IsD) D1 111.88
D2 136.45
Fig. 10. Type of steering maneuvers

4 Conclusions
Generalizing the aspects of the diagrams shown in Figs. 5, 6, 7 and 8, for all covered
routes, one could conclude that the travel speed of heavy vehicles (log trucks) is under
the travel speeds of the other vehicles, on the national roads. According to the men-
tioned figures, the maximum travel speed reached on the national road DN1A, by the
test vehicle, is 74.30 km/h (Fig. 5) and 73.47 km/h (Fig. 6). The speed limit on the
respective route is of 90 km/h.
From Table 2 results the following:
• the maximum distance covered by the test vehicle when entering to the national
road, accelerating up to 15th percentile speed (DS15 in the table) was recorded in
the intersection 3, DN1A – D152, maneuver IsS (see Fig. 10), driver D2. The
distance is 182.36 meters. Thus, along this distance from the intersection, the test
vehicle can be considered a moving obstacle;
• the maximum distance measured from the center of the intersection up to the
extreme point where the test vehicle reaches the 15th percentile speed (DDS15 Max
in the table) was recorded in the intersection 3, DN1A – D152, left side (relative to
the sketch in Fig. 10). The distance is 175.22 meters. Thus, along this distance, it is
compulsory to observe the sight distance necessary for the vehicles to decelerate
under safety and comfort conditions;
• in order to ensure the safety conditions for the traffic flow in the area of the studied
intersections, the following condition must be fulfilled:
DDS15 Max \ NSD ð6Þ
DDS15 Max \ MSD ð7Þ
One can see that condition (6) is fulfilled in all cases within Table 2, yet condition
(7) is fulfilled only in 2 from 6 cases.
5 Proposals
The following proposals are made for future researches in the field of this study:
– to extend the study over a higher number of intersections between forest roads and
highways;
– to highlight the duration in time, in the framework of the steering manoeuvres
performed by heavy (forest) vehicles, when the travel speed is below the 15th
percentile level;
– to emphasize the mathematical relation between the values of the distances on
which the heavy (forest) vehicle is moving slower than the 15th percentile speed,
and the value of the distance necessary for the vehicles which travel on the highway
to decelerate under safety and comfort conditions.
References
1. Ministry of Environment Water and Forests:, Normative for forest roads design PD – 003 –
11, pp. 42–43, Bucharest (2011)
2. Covaciu, D.: Solutions on the study of dynamic and in traffic behavior of automotive
vehicles through CAD/PLM and GPS integration. Ph.D. Thesis, Transilvania University of
Brasov (2010)
3. Covaciu, D., Preda, I., Ciolan, Gh., Câmpian, O.V.: Data acquisition system based on GPS
technology, for vehicle dynamics analysis. In: The XI-th International Congress on
Automotive and Transport Engineering CONAT 2010, vol. 5, pp. 31–36, Transilvania
University Press, Brasov (2010)
4. Adafruit, Ultimate GPS logger Shield. https://learn.adafruit.com/adafruit-ultimate-gps-
logger-shield/overview. Accessed June 2016
5. Florea, D.: Managementul traficului rutier, Ediţia a II-a (Road Traffic Management, Second
Edition), Transilvania University of Braşov (2000). ISBN 973 –9474-55-1
6. Hobbs, F.D.: Traffic Planning & Engineering –, 2nd edn. Pergamon Press, Oxford (1979)
7. Roess, R.P.: Traffic Engineering –, 4th edn. Person Higher Education, Upper Saddle River
(2011)
8. Florea, D., Covaciu, D., Preda, I., Timar, J.: Comparative analysis of the moving observer
kinematics in the urban road network. Acta Technica Corviniensis-Bulletin Eng. 6(3), 135
(2013)
9. Olteanu, N.: Drumuri Forestiere. Proiectarea Drumurilor Forestiere. Retele de Drumuri
Forestiere (Forest Roads Design. Forest Road Networks), Transilvania University of Brasov
(2008)
10. Cososchi, B.: Drumuri Trasee (Roads. Routes), Ed. Societatii Academice “Matei – Teiu
Botez”, Iasi (2005)
11. Hoel, L.A., Garber, J.N, Sadek, A.W.: Transportation Infrastructure Engineering – A
Multimodal Integration SI Edition, pp. 112–115. Cengage Learning, United States of
America (2011)
Smart Solar Electric Tricycle
Dahbi Mohamed1(&), Said Doubabi1, and Ahmed Rachid2

1
Department of Applied Physics, Cadi Ayyad University, Marrakesh, Morocco
dahbi.mohamed2@gmail.com, s.doubabi@uca.ma
2
Laboratory of Innovative Technologies,
University of Picardie Jules Verne, Amiens, France
rachid.greenway@gmail.com
Abstract. The transport sector in Marrakesh city has known significant

increasing in the number of tricycle. This latter has become used in different
areas such as transporting goods or for tourist uses. This fact however, has a
disadvantage of the bad impact on the environment, since, till now, all the
circulating tricycles in the city are fuel-based ones. In this paper we propose a
study to realize a solar electric tricycle with including the smart vehicles concept
in such way to increase the autonomy of this tricycle. Reaching this goal will be
by avoiding braking phases in the traffic lights. This control will depend on the
GPS position of the tricycle and that of the traffic lights, in addition to their state
when communicating (Green or red, still 30 s… etc.). The simulation results,
using a regular path of the tricycles in Marrakesh city will be discussed.
Keywords: Solar panels BLDC motor Battery autonomy Speed control

Start-up phases Power consumption
1 Introduction
Smart cities have become an important subject treated by both the industry and the
universities to solve different problems and to improve the lifestyle of the human been.
To designate this international subject many words may be used such as: digital city,
green city, connected city, Eco-sustainable city… the studies in term of a smart city
develop many new performed services in different sectors such as: environmental
sustainability and responsible and intelligent urban habitat…
The transportation sector is also one of the important sectors treated to change a
“normal” city to become a “smart” one, or to build a new one. Communication between
vehicles and infrastructures represents a big challenge in term of data treatment.
Depending on these data, many vehicle comportment may be considered to decrease
the traffic jam which go back to reduce greenhouse gases, to reduce waiting time in
traffic lights and also getting information on the most crowded streets to avoid them
when driving etc. …
In recent years, Marrakech city has known an increasing in the number of operating
vehicle in the city and this is due to many reason such as: (i) Marrakesh is a touristic
town known throughout the whole world and attracts more and more tourist,

722 D. Mohamed et al.
(ii) increasing in the numbers of citizens, (iii) it became easier to buy a mean of
transport, etc. …
Among these means of transport, we can mention the tricycle. They are used for
their benefits such as the easy access to various narrow area of Marrakesh city, their
low cost, and more advantages like using them in various domain including: trans-
portation of goods and transport of tourists.
Nevertheless, and so far, these tricycles are fuel-based ones. Thus represents a
constraint and a bad impact on the environment, especially as some Marrakesh
neighborhood are narrow and don’t have adequate ventilation.
Therefore the use of transportation based on renewable energy such as electricity,
solar or fuel cells etc. … may presents an alternate to replace the fuel-based tricycles. In
previous work [1], we introduced an analysis of a range electric tricycle with an
adaptation to become equipped with solar panels to increase its autonomy. Other
studies [2], presents a photovoltaic charging system to increase the battery autonomy of
an electrically assisted tricycle for touristic purposes. The work presented in [3],
suggests an electric system design for electric reverse tricycle for the Elderly. The
vehicle to vehicle communications (V2X) impact is shortly discussed on helping
electric bicycles saving energy in [4]. The batteries of a solar charging station are used
for charging electric tricycles; detailed schematics are presented in [5].
In this paper, we propose a design of an “intelligent” tricycle able to minimize stops
in traffic lights. These tricycle stops cause an important call of energy to restart
operation of the vehicle. The impact of this control on the autonomy of the tricycle will
be presented by simulation using the example of a conventional tricycle path in the city
of Marrakesh. Communication between the tricycle and infrastructures will treat the
state and GPS position of the nearest traffic light. The right command will be calculated
and displayed to the driver before 50 m of the traffic light position.
The constitution of this paper is arranged as follows, Sect. 2 will present a global
electric schematic of the tricycle. In Sect. 3, the tricycle performance depending on
different resistive forces will be treated to present after the traffic light simulation for a
usual tricycle path in Marrakesh city in Sect. 4. Section 5 will treat the simulation
results with interpretation of results followed by a conclusion in Sect. 6.
2 Electric Schematic of the Solar Tricycle
The global schematic (Fig. 1) of the tricycle is composed by a specific number of solar
panels, a maximum power point tracker to extract the maximum power from the solar
panels that will charge the battery which supplies energy the a brushless DC motor.
This latter is controlled by an inverter responsible also of controlling the tricycle speed
and gives the right command to the motor.
The size of the tricycle has limited the number of the implemented solar panels:
fours solar panels will be placed on the top of the tricycle. Two solar panels are
connected in series connected in parallel with the left two solar panels. The charged
battery supplies the BLDC motor via the inverter, and depending on the driver com-
mand, the motor inverter adjusts the right command to control the MOSFETs inverter.
Smart Solar Electric Tricycle 723
Fig. 1. Global schematic of the designed tricycle
Fig. 2. Tricycle schematic used for communication vehicle-traffic light
Using the communication between the tricycle and different infrastructure needs the
implementation of an antenna to gather data and display the right command to the
driver. The new schematic taking in consideration this communication is presented in
Fig. 2:
The driver will take the decision of deceleration or acceleration to avoid stopping in
Red traffic lights instead of regulating the motor speed. This configuration and speed
controlling depends on the traveled distance and the speed of the tricycle. Those
parameters depend on the performance of the tricycle.
3 Resistive Forces and Tricycle Performance
When rolling the tricycle is subject to different resistive forces to its progress, we can
cite the aerodynamic force, rolling force, hill climbing force and the force of the
acceleration. The tricycle BLDC motor will need to deliver a superior force to the sum
of these forces to move forward, which refers to say more energy consumption and
higher power.
The total resistive force is presented in (1) [6, 7]:
Ftot ¼ Facc þ Faer þ Frol þ Fclim

dv 1 ð1Þ
¼ m þ :Cr :A:q:Vr2 þ m:g:Cr þ m:g: sin a
dt 2
Where:
– m mass of the vehicle (Kg)
– V vehicle speed (m/s)
– Cd drag coefficient
– A Frontal area (m2)
– q air density (Kg/m3 )
– Vr relative speed (m/s)
– Cr the rolling coefficient
– g gravitational force (m/s2)
– a road gradient (rad)
To present the tricycle performance, two scenarios were used, each one having two
test cases. The first scenario represents the performance of the tricycle when empty
(300 kg including the driver weight) and the second scenario is when the tricycle is
rolling with its maximum weight (600 kg including the driver weight). For each sce-
nario, the tricycle performance is presented for two cases: when rolling on a flat road
and when rolling on a 5 % hill climb.
Simulation results show the speed limits of the tricycle depending on the used
power, and show the torque that must deliver the motor to overcome different slope of
the road. Using Matlab/Simulink environment, Fig. 3 presents the power and the torque
versus the speed respectively.
Fig. 3. Electric power (a) and electric torque (b) versus vehicle speed
These results show that using 1000 W as motor power, the maximum speed
reached 35 km/h for 0 % as a hill climb, which represent for us an acceptable speed.
The needed torque in this case reaches 22 N.m. For a 5 % and with the maximum load
(600 kg), the speed reaches 15 km/h and the motor needs to deliver almost 60 N.m.
In that simulation, only the aerodynamic, the rolling and the hill climbing forces are
taking in consideration. The acceleration force was not added because it’s less
important in the permanent regimes.
Those results are calculated using these parameters: Cd = 1.2, A = 1.2, Cr = 0.007,
g = 9.81.
4 Traffic Light Simulation and Data Processing
Different and important information must be known on the traffic light such as, the state
of the lights (RED, YELLOW or GREEN), and the remaining time before changing the
type of light to the next Another important information is the remaining distance
needed by the tricycle to reach the traffic lights position. In this study, and for a usual
path of the tricycles in Marrakesh city, the calculated speed will begin until the distance
left to reach the traffic light is 50 m. Figure 4 presents an image of this path with the
traffic light position, and Fig. 5 shows the method of calculating the elevation data of
the same path.
Fig. 4. Simulated path and traffic light position with hill climb values
The elevation data will be converted to hill climb degree. The obtained data will be
added in the simulation as presented in previous paragraph. To save the data of the
slope, we have taken the step of 7 m to recalculate the degree of the slope (Fig. 5).
Fig. 5. Calculated degre of the hill climb

Fig. 6. Algorithm of control speed strategy
The state of the traffic light is presented as follow: 30 s for the red color followed
by the Green light for 25 s and then the yellow light for 5 s. The calculated speed
method, taking in consideration that the maximum speed is 30 km/h, is done using the
algorithm presented in Fig. 6.
Using the state flow environment in Matlab software, the traffic light information
are programmed to get the situation of the on and off lights (RED, YELLOW or
GREEN) and the remaining time for each on light before becoming off.
Figure 7 presents the state of one of the different traffic lights selected by random
way. The needed information such as the remaining time, the state of lights can be
recovered using these proposed results.
The simulation was done for a regular path in Marrakech city for some tricycles.
Simulation results took in consideration the command given by the driver without any
driving strategy and when driving with speed control strategy. For this latter, and in
simulation, the optimal calculated speed to avoid stopping in RED lights, is imposed.
Unlike the practice, for more security, this calculated speed will be displayed to the
driver to give him the priority to control his vehicle. Simulation results are presented in
tricycle speed, and absorbed current from the battery and the travelled distance.
Fig. 7. RED, YELLOW and GREEN lights states for the first traffic light
Fig. 8. Speed of the tricycle (a) without any strategy, (b) when speed control strategy
From the speed figure (Fig. 8), the tricycle has stopped every time he met a RED
light, more energy consumption in restart-up phase’s accurses and may decrease the
autonomy of the vehicle. The difference, between this mode of driving and the speed
control strategy, is noticed when the speed does not reach zero and neither the absorbed
current from the battery (Fig. 9), the difference of energy consumption is consumed to
maintain the speed of the tricycle to still in the one calculated (Fig. 10).
This difference in driving style, has shown that using speed control strategy con-
sume less energy and can avoid the motor energy consumption spikes.
Fig. 9. Absorbed current from battery (a) without any strategy, (b) when speed control strategy.
Fig. 10. Travelled distance (a) without any strategy, (b) when speed control strategy.
6 Conclusion
In this paper, changing the driving style from stopping in RED lights to controlling
speed strategy is proposed to increase an electric vehicle autonomy. The simulation
was done for a solar electric tricycle, and has revealed that avoiding stops in RED lights
may increase the battery autonomy. This is achieved by processing data before 50 m
from the nearest traffic light, then calculating speed, depending on the saved data, will
be imposed on the motor. This speed will allow the tricycle when it arrived at the traffic
light to find it in the GREEN state. For more security, in practice, the calculated speed
will be displayed to the driver to give him the priority to choose between his speed
command and the calculated one.
References
1. Dahbi, M., Doubabi, S., Rachid, A.: Autonomy analysis of a solar electric tricycle. In: 2015
3rd International Renewable and Sustainable Energy Conference (IRSEC), pp. 1–6 (2015)
2. Caruso, M.M., Cecconi, V., Cipriani, G., Di Dio, V., Di Tommaso, A.O., Genduso, F.,
LaCascia, D., Lauricella, R., Liga, R., Miceli, R., Prestigiacomo, G., Galluzzo, G.R.,
Trapanese, M.: A photovoltaic charging system of an electrically assisted tricycle for touristic
purposes (2013)
3. Choi, Y., Son, Y., Joe, U.: Electric system design for electric reverse tricycle for the elderly
steering structure design driving power performance analysis for selecting motor (2013)
4. Tal, I., Zhu, T., Muntean, G.M.: Short paper: On the potential of V2X communications in
helping electric bicycles saving energy. In: IEEE Vehicular Networking Conference VNC,
pp. 218–221 (2013)
5. Faraz, T., Azad, A.: Solar battery charging station and torque sensor based electrically
assisted tricycle. In: Proceedings of the 2012 IEEE Global Humanitarian Technology
Conference GHTC 2012, pp. 18–22 (2012)
6. Adnan, A., Ishak, D.: Finite element modeling and analysis of external rotor brushless DC
motor for electric bicycle. In: SCOReD, pp. 16–18 (2009)
7. Prochazka, P., Pazdera, I., Vorel, P., Cervinka, D.: Design of Small Electric Car, pp. 359–364
(2012)
Programming an Autonomous Mini-vehicle
in Narrow Environments
Emanuel Giurgiu, Teodora Girbacia(&), and Gheorghe Leonte Mogan
Transilvania University of Braşov, 29 Eroilor Blvd, 500036 Braşov, Romania

emanu.giurgiu@gmail.com,
{Teodora.Girbacia,Mogan}@unitbv.ro
Abstract. In order for a mini-vehicle to accurately follow a trajectory with a

constant velocity even while avoiding an obstacle, the transition between driving
straight ahead and making a turn is realized by driving along a clothoid arc. In
narrow environments the generated trajectory requires a higher number of
maneuvers in order to avoid the obstacles placed in the environment. In this
paper is presented the hardware structure of a mini-vehicle with two electric
servomotors and an algorithm developed for achieving an autonomous
mini-vehicle that allows to accurately establish the speed with which the car is
going to travel, calculate the distances between specific points of the trajectory
and the necessary time required to reach transitional or final targets.
Keywords: Autonomous vehicle Trajectory following Clothoid Narrow

spaces
1 Introduction
Nowadays the development of autonomous vehicles has become an important research

area especially because of its priority in solving the problem regarding trajectory
following. General path planning methods have been described in great detail by many
authors, to such extent that they have become mainstream knowledge in the fields of
computer science, artificial intelligence and autonomous systems [1, 2]. The classical
problem formulation in path planning is finding the optimal path between two points.
The optimal trajectories could be paths that maximize the obstacle avoidance, minimize
the number of vehicle turns.
A problem frequently analyzed is the generation of a safe path for parking
autonomous vehicles that has common characteristics with generating optimal paths in
narrow environments due to the necessity of performing numerous maneuvers in order
to reach the target [3]. In [4] is presented a method of generating safe parking paths for
an Ackermann steering vehicle that is smooth due to the usage of Bezier curves. The
path planning algorithm for an Ackermann steering vehicle can be composed from two
parts: planning the method for performing the turns and distance calculation according
to [5]. Besides the path planning for parking autonomous vehicles, another problem
that aroused and requires the application of this algorithm is obtaining an optimal path
in intersections [6]. Also, planning a trajectory for a vehicle that requires changing the
lane while steering in approached in [7]. In [8] is generated a smooth path applied to

730 E. Giurgiu et al.
changing lanes based on the use of clothoids. A method of planning a path off-road is to
use predefined landmarks aiming to generate a set of alternative paths from which the
optimal one is chosen [9].
The purpose of this paper is to present the necessary changes made to the hardware
structure of Traxxas Rustler VXL model in order to follow a predefined trajectory for
narrow environments (indoor) using dedicated Arduino command functions.
2.1 The Hardware and Software Architecture of the Autonomous

Mini-Vehicle
The hardware structure of a mini-vehicle comprises two electric servomotors, one for
propulsion and the other for steering, based on the Traxxas Rustler VXL radio con-
trolled auto model. The propulsion engine of the vehicle was modified to allow the car
to travel at a slower speed in order to verify the accuracy with which it follows the
trajectory. The control of the mini-vehicle in order for it to follow an imposed trajectory
is realized through the use of an Arduino microcontroller due to the easiness of
hardware connection and the ability to develop a new set of functions in the C ++
programming language which are subsequently transmitted to the compiler and
transformed into executable code for the processor of the vehicle.
The main parts of the developed autonomous mini-vehicle (Fig. 1) are: an Traxxas
Rustler VXL RC car that has a steering based on servomotor and the propulsion based
on a regular DC motor, an Arduino mega2560 Microcontroller board for controlling the
entire circuit, a SD card reader, a SD card, a L298 M motor driver, a 12 V rechargeable
battery and a couple of jumper wires. The trajectory that the car must follow up is
stored on a SD card, in a.txt file named “PATHDATTA.txt” as string commands.
In Fig. 2 is represented the schematic of the vehicle hardware. The source of the
current is from a 12 V LiPo Battery that provides continuous current for all the devices
including the motors and microcontroller. The servo motor has only three pins, VCC,
GND and PULSE. The ground pin from servo is connected to the common ground of
Fig. 1. The developed mini-vehicle

Programming an Autonomous Mini-vehicle in Narrow Environments 731
Fig. 2. The hardware parts connections
the circuit, the VCC pin is connected to the Arduino 5 V and the pulse signal is
connected to the pin number 2 from Arduino board.
To control the propulsion motor a driver was needed. This hardware driver is
connected first directly to the power source (12 V battery) and it provides PWM signal
to the DC motor connected to the output3 from it. To control this driver, 3 connections
had to be made: the chip enable pin from the L298 N driver which is connected to the
pin number 9 from Arduino and the two directional pins (en3 and en4) to pin number 8
and respectively pin number 7 from microcontroller board. When the en3 is enabled
and en4 not, the motor rotation direction is clock-wise. When the en4 is enabled and
en3 not, the motor rotation direction is anti-clock-wise. When both of enable pins are
not enabled or enabled, the motor will stop immediately. The SD card reader module
implementation, as it can be seen in the picture, is the most complicated from the
circuit. It works on a master-slave technology which is a model of communication
where one device or process has unidirectional control over one or more other devices.
The connections med for the SD card are as follows: VCC to Arduino +5v, GND to
common ground of the circuit, MISO to Arduino 50, MOSI to Arduino 51, SCK to
Arduino pin number 52 and CS (Chip Select) to Arduino pin number 53. This con-
nection allows the serial communication between the Atmega2560 microcontroller and
the SD car reader module and all the features that a communication between a com-
puter and a SD card offers, such as read files, write files, create or delete files, read card
information and some format options.
2.2 The Trajectory Planning Algorithm and Programing

In order to obtain a smooth path the turns made for driving around obstacles are
composed of: a clothoid arch of entering the turn, a circular arc needed to avoid the
obstacle and a clothoid arch of exiting the turn (Fig. 3). The path is defined by the
starting and finishing points and the orientation of the car when starting and arriving at
the destination. In order to avoid the obstacles placed in narrow environment the angle
Fig. 3. Path planning geometry
(φc) and length (Lc) of the clothoid used for entering and exiting the turn has to be the
maximum provided by the characteristics of the mini-vehicle and the values of
parameters Lc and φc are obtained from experiments.
The variables that define the characteristics of the path are: P is the starting point;
Lp is the length of the first straight line; uc is the angle between the tangent line of the
clothoid and the elongated Lc ; h is the angle of the circular arc; Rmin is the radius of the
circular arc; S is the finishing point; Ls is the length of the last straight line; a is length
of tangent to stirring curve.
In order to determine the angle of the circular arch, knowing when parameters the
angle of the clothoid used for entering and exiting the turn is the maximum provided by
the vehicle, in correlation with Fig. 3, the following relations:
BE ¼ Lc sin uc ; EO ¼ a Lc cos uc ð1Þ

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
OB ¼ ðLc sin uc Þ2 þ ða Lc cos uc Þ2 ð2Þ
Lc sin uc
a1 ¼ arctg ; ð3Þ
a Lc cos uc
pa pa Lc sin uc

a2 ¼ a1 ¼ arctg ; ð4Þ
2 2 a Lc cos uc
h
BF ¼ Rmin sin ; BF ¼ OB sina2 ; ð5Þ
2
h OB sina2
sin ¼ ; ð6Þ
2 Rmin
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðLc sin uc Þ2 þ ða Lc cos uc Þ2 sinðpa Lc sin uc
2 arctg aL cos u Þ
h ¼ 2arcsin c c
: ð7Þ
Rmin
After establishing the value for the angle of the circular arch θ, it si possible to find
length of the arch La = θ Rmin and all the variables that define the characteristics of the
trajectory can be obtained and then the steering motor and the propulsion motor are
correlated in order for the vehicle to follow the predefined trajectory [10].Thus in the
same moment in which the propulsion motor changes speed of operation, the steering
motor must also begin to operate, these are the crucial conditions that must be fulfilled
in order to maintain the trajectory. In Fig. 4 it is presented characteristics of propulsion
and stirring velocities (v and ω) that will be used in this case for testing and showing
that two motors controlled with PWM values on a time scale can reproduce a previous
calculated and well determined trajectory from Fig. 3.
Fig. 4. The characteristics of velocities in correlation
In order to programing in Arduino language these characteristics we define dedi-

cated command function Forword (L, v,PWMf), Back(L, v, PWMb), Right (La, ω,
PWMr) and Left (La, ω, PWMl) for go ahead, go back, turn right and turn left
respectively. The functions for turn left and right contain the clothoid curve initially
predefined.
3 Test Cases
The measurement for the steering calibration started by steering to a maximum in one
way, and then steering to the maximum again but in the opposite direction in order to
set the command signal corresponding to the absolute value for the steering angle. In
Fig. 5 is shown the correlation between the command signal and actual angle of the
front wheels.
Fig. 5. The correlation between PWM signals and the steering angle
In order to propulsion calibration by testing we measure the displacement of vehicle

in straight line using different PWM frequencies. Knowing the times of displacements
was possible to find the velocities. In Fig. 6 is presented the correlation between the
PWM signal and the vehicle speed.
Fig. 6. The correlation between PWM signal and the vehicle velocity
Using the calibration characteristics (Fig. 5, 6) it is possible to find the PMM

frequencies parameters PWMf, PWMb, PWMr and PWMl from command functions
Forword (L, v,PWMf), Back(L, v, PWMb), Right (La, ω, PWMr) and Left (La, ω, PWMl).
The conducted experiment consists in defining a complex trajectory in a narrow
environment and then commanding using generated functions the steering motor and
the propulsion motor so that the vehicle will follow it.
As shown in Fig. 7, the vehicle starts from the position P and drives until it reaches
an intermediate arriving point (I1), reaching a checkpoint placed along the circular arch
component of the path, then it drives to an intermediate arriving point (I2) before
driving to the parking spot in point S, as can be observed in Fig. 8.
Fig. 7. The complex trajectory to be tracked
Fig. 8. Backward parallel parking in final part of the trajectory (Fig. 7)
In order to follow the predefined path, the minivehicle performs the following
functions: forward acceleration followed by continuous driving, turning to the right
110, further continuous driving for 5 m, stop, backward acceleration followed by
continuous driving, turning right 110, driving ahead until it reaches position I2, stop,
accelerate forward followed by continuous driving for 3 m slowing down, turning left
following a clothoid with a 20 degree turning angle, maintaining the steering and then
turning left following a clothoid with a 20 degree turning angle, stop, backward parallel
parking, stop. For all these stages was established the PWM frequencies values.
4 Conclusions
This paper presented the hardware and software architecture of a mini-vehicle based on
the Traxxas Rustler VXL radio controlled auto model that has been modified in order to
autonomously follow smooth trajectories generated for narrow environments. The
proposed trajectory planning with dedicated functions assure a very flexible and
modular structure of a programs to command mini-vehicle contains two electric ser-
vomotors controlled using PWM signals. Tests were made in order to provide the
necessary PWM signals to ensure that the minivehicle drives according to the prede-
fined trajectory. Future work will be focused on evaluation the accuracies of tracked
trajectories and establish prediction related to using the results for autonomous
vehicles.
Acknowledgment. This paper is supported by the Romanian Government, specifically MEN –

UEFISCDI authority under the program PNII “Partnerships in priority areas”, under the project
number 240⁄2014 - NAVIEYES, supporting the collaboration between the company Route 66
and University Transilvania of Brașov.
References
1. Souissi, O., Benatitallah, R., Duvivier, D., Artiba, A., Belanger, N., Feyzeau, P.: Path
planning: a 2013 survey. In: Proceedings of 2013 International Conference on Industrial
Engineering and Systems Management, pp. 1–8 (2013)
2. Pérez, J., Godoy, J., Villagra, J., Onieva, E.: Trajectory generator for autonomous vehicles in
urban environments. In: IEEE International Conference on Robotics and Automation,
pp. 409–414 (2013)
3. Muñoz, P., Barrero, D.F., R-Moreno, M.D.: A statistically rigorous analysis of 2D
path-planning algorithms. Comput. J. 58(11), 2876–2891 (2015)
4. Liang, Z., Zheng, G., Li, J.: Automatic parking path optimization based on bezier curve
fitting. In: IEEE International Conference on Automation and Logistics, pp. 583–587 (2012)
5. Gupta, A., Divekar, R., Agrawal, M.: Autonomous parallel parking system for Ackerman
steering four wheelers. In: IEEE International Conference on Computational Intelligence and
Computing Research, pp. 1–6 (2010)
6. Au, T.C., Fok, C.L., Vishwanath, S., Julien, C., Stone, P.: Evasion planning for autonomous
vehicles at intersections. In: IEEE/RSJ International Conference on Intelligent Robots and
Systems, pp. 1541–1546 (2012)
7. Guo, L., Ge, P.S., Yue, M., Zhao, Y.B.: Lane changing trajectory planning and tracking
controller design for intelligent vehicle running on curved road. In: Mathematical Problems
in Engineering, pp. 1–9 (2014)
8. Funke, J., Gerdes, J.C.: Simple clothoid lane change trajectories for automated vehicles
incorporating friction constraints. J. Dyn. Syst. Meas. Control 138(2), 1–9 (2016)
9. Chu, K., Lee, M., Sunwoo, M.: Local path planning for off-road autonomous driving with
avoidance of static obstacles. IEEE Trans. Intell. Transp. Syst. 13(4), 1599–1616 (2012)
10. Gîrbacia, T., Margarit, A., Mădălin, S., Mogan, G.: Path and trajectory planning for vehicles
with navigation assistants. In: Olaru, A. (ed.) Applied Mechanics and Materials, vol. 811,
pp. 300–304. Trans Tech Publications, Zurich (2015)
About Uncontrollable Reactions of the Driver
Due to Skin Mechanical Stimuli
Alexandru Bogdan and Silviu Butnariu(&)
Transilvania University of Braşov, Braşov, Romania

bogdanalex91@gmail.com, butnariu@unitbv.ro
Abstract. The goal of this paper is to study the reaction of a car driver to skin
mechanical stimuli. This study meets the various applications that aimed
supervising the driver and involving the attaching of sensors or equipment on
his clothes or skin (pulse, pressure, temperature sensors, etc.) that may disturb
activity and influence on driving. The research consists of mounting vibrating
buzzers in tiny pockets on a clothing item (a very elastic shirt, sticking to the
human body), triggering them by considering a certain algorithm and then
tracking the reaction of the car driver by using optical tracking devices (e.g.
Kinect). Twelve vibrating buzzers will be attached on the t-shirt, seeking the
more sensitive areas of the human body. The mechanical stimuli system will be
build using Arduino development platform.
Keywords: Mechanical stimuli Reaction Optical tracking Vibrating

buzzers Haptics Driver posture
1 Introduction
With the rapid development of the automotive industry systems, visual and auditory
systems integrated in the vehicles, it comes as a normal step to analyze the integration
of piezoelectric buzzers in the driver warning systems.
The purpose of this study is to analyze the human reaction to external mechanical
stimuli, to track and observe this reaction as it comes more and more clear that there is
a need to develop new methods of accidents prevention than the existing ones.
This study marks the beginning of such systems that soon may be integrated in our
vehicles to provide a much better communication between the car monitoring systems
and the driver. At this point, this communication is done at a visual and auditory level,
but these systems come with a distraction of the human attention from the road.
What we propose is a system that will provide all this information, more at an
intuitive level, then at a conscious level. By activating vibrating buzzers, and by
knowing the humans reaction to each stimuli, we can further develop an algorithm
which can encompass all systems, and which can provide guidance and safety for all
traffic participants.

738 A. Bogdan and S. Butnariu
2 State of the Art

2.1 About Vibro-Tactile Stimuli
Considering the complexity of this study, several topics must be taken into consider-
ation, such as neuro – biological implications of human stress reaction, the develop-
ment of the mechanical stimuli generation platform and the data acquisition system.
Several researches have been conducted on similar topic such as driver reaction to
road stimuli – referring to different road stimuli such as slope inclination, curvature and
speed of driving; other topics considered were listening to music, actively engaging in a
conversation and how these are influencing brake time reaction and how can these
factors conduct to accidents.
First studies in the drivers behavior in traffic scenes began in the 1950’s when this
topic was intensively studied in order to provide answers on the best way to build cars
and roads, by describing traffic flows using mathematical and statistical ideas [1].
The idea of using vibro-tactile feedback in cars is new and until now has not
reached its full potential. There are, however, already some systems that have been
implemented in mass produced car models. So far these systems are designed for lane
departure warning, while in the future many other applications are possible.
The Lane Departure Warning Citroën (LDWS) system uses 6 infrared sensors to
detect when the car moves across road markings. The system is designed to work at
velocities above 80 km/h and can manually be activated or deactivated. In case of a
lane departure with no directional indicator given the system will presume it’s
unwanted. This will conduct a signal to the driver seat and trigger a vibration motor on
either the left or right side. The vibro-tactile signal thus alerts the driver about the
upcoming situation and will hopefully prevent a potential accident [2].
In [3], several students were tested, simulating the driving experience and were
distracted by several means and their reaction time and pulse were measured. In this
study, the subjects were sat before a monitor and when a cat would appears on the
display they had to press the space key. While they were sitting loud music was played,
they were engaged in conversations and also they were asked to write text messages on
the mobile phone. These topics were chosen based on the following considerations:
music is related to passive thinking, conversation is related with active thinking and
texting is related to physical and cognitive distractions. The result of this study showed
that the highest influence in reaction time was given by the text message writing,
followed by the conversation and then the music. The difference between the men-
tioned study and the topic of the present paper is that in this paper the subjects will be
stimulated by external mechanical stimuli, with the purpose of analyzing the movement
they have when exposed to such stimuli.
In the research literature, the time required for the driver to perceive and respond to
the danger is usually called time-response or a perception of the response time. The
response time is defined as the amount of time that begins when an obstacle at first
becomes visible to the driver or the first actions in a way that could be perceived as
dangerous for the driver and ends when the driver has initiated a response, such as
touching the brake pedal [4].
About Uncontrollable Reactions of the Driver Due to Skin Mechanical Stimuli 739
Theoretically, response time is to be perceived as the series of steps such as:

detection, identification, decision, and response [5]. Basically, the response time is most
commonly divided into two stages: perception time (which begins when a stimulus is
presented and ends at the onset of foot motion) and the movement time (which starts at
the start of foot motion and ends when the foot had touched the brake pedal).
In [6], the authors have studied the effect of a multi distraction event, such as visual,
motor and cognitive distraction with respect to the braking time. However, that
experiment was conducted using a driving simulator, so direct comparison between the
two studies is a relative difficult and not so precise task. In the study hereinbefore
presented, the participants were asked to press the brake pedal as soon as possible, while
in the driving simulator they had to perceive a real life situation, and this may have led to
other results as the urgency perceived there may had also affected other senses.
In [7], the purpose of the study is to mount vibration devices on the driver seat and
triggering them when possible threats occurred on the road. In [8] the authors have used
a vibrating device mounted in the waist seatbelt, which was triggered anytime a car was
approaching with great speed, and the participants were asked to look in front and in
the rear view mirror. The experiment took place in a driving simulator, in order to
ensure the safety of all participants. In [9], the study goal was to analyze the drivers
reaction when performing a secondary task, besides driving, talking on a cell phone.
Several experiments were done in order to assess the effeteness of vibro-tactile
warning in increasing driver safety [10]. The participants were asked to follow a green
dot on the monitor and initiate the brake pedal when they saw the object approaching.
2.2 About Neuro Biological Implications of Mechanical Stimuli Reaction

The human body is a complex system that is able to receive all kind of information
from the surrounding environment, process them in accordance with their type
(physical, visual, auditory, energy) and react to them with respect to each stimuli type
and each person experience, physical and mental state. The perception of such stimuli
is done via the five senses each person has (except for some cases when a health issue
arises). The senses are the human sensors, which help us to gather data from the
environment and to perceive our surroundings. The senses were of great concern in
many studies done through time in fields such as neuroscience, cognitive science and
philosophy of perception. Each of our five senses is broadly connected to our body and
the nervous system. The five senses traditionally recognized are: hearing, sight, smell,
taste and touch. Touch is a perception resulted by the activation of neural receptors in
the skin, tongue, throat and mucosa. This sense is triggered by a variety of receptor,
which respond to the variation in pressure. There are also some itch specific receptors,
which are triggered usually by insects, allergies. The loss or impairment of the ability to
feel anything touched is called tactile anesthesia. (www.wikipedia.com).
For the mechanical stimuli reaction, the human body uses the tactile sense, in
collaboration with the peripheral nerves. Basically, the receptors in the skin detect a
change in the pressure and transmit this information to the brain, where it is processed.
Here, the peripheral nervous system kicks in, as it is responsible for the automatic
reaction and unconsciously processes in the human body.
The developed device system consist of one Arduino board, one test board, two 25 pin
connection (and F-M and one M-F), twelve vibrating buzzers and one shirt on which
these buzzers are attached (Fig. 1).
Fig. 1. Developed test device [11]
The program created triggers each of the twelve buzzers at a time, in a random
order. The signal is send to the test board using jumper wires. From the test board, the
signal is transmitted via two 25 pin connection to the buzzers by using SCART cable
wires. The buzzers were mounted on the upper body of the driver, in order to generate a
physical signal for passengers or arousal by wearing the seatbelt in order to perform the
simulation as close as possible to real life conditions.
Fig. 2. Experimental stand

Experimental stand is shown in the figure below (Fig. 2). Moreover, the position
of the driver can be analyzed in a professional car simulator, located in our research
department. Experiments were conducted on a group of people with driving licenses,
respecting the rules of ethics available in the institution.
It has prepared a scenario, which consists of a number of start/ stops of different
buzzers, previously established. First, we used the scenario, for the same subject in two
different positions (sitting down on a chair or standing). In the second phase, we used
more subjects. The results was gathered and stored during the scenario application, they
were analyzed and presented as charts. The charts have comments and conclusion
considering the reaction the driver had during the observation time.
After the experiments, it can be concluded on areas where the various sensors can
be installed without affecting the quality of driving. The processing of the buzzers
triggered- reaction was done using a Kinect two device. The influence of the buzzers
represent the body reactions of the car driver and were tracked by using Kinect SDK 2.
The device tracks the main articulation of the body, while the arms, feet and the rest of
the body are seen merely as connections between articulations (Fig. 3). The shirt had
twelve buzzers mounted in the position shown in the picture (Fig. 4). The sensors
positions were chosen subjectively by the authors, considering the sensitive areas of the
skin. They were triggered in group of two, or three in order to reproduce as realistic as
possible all types of interactions. The buzzers were split in several categories, as
follows: 1-2-3-4 were left hand and shoulder(further named LHS); 5 and 6 were left
body (LB); On the other side, 8 and 9 were right body (RB) and 10-11-12-7 were
considered as being right hand and shoulder (named RHS). In total, there were four big
categories that were triggered, and which result were tracked and further analyzed in
this paper.
Fig. 3. Kinect body tracking Fig. 4. T-shirt with buzzers
4 Results
Next we present the results of four tests done with the Kinect sensor. The visual
observations were done on four subjects, with ages between 23 and 25, three male and
one female. The tests were performed once. In these tests, the car was the only driver
seat, without the other components of the cockpit. These details will be examined in
future research. The representation of their reaction is described below.
Fig. 5. Left hand and left shoulder
Case one – Left hand and left shoulder (Fig. 5). In this experiment, we proposed to
trigger the vibrating motors corresponding to the left hand and left shoulder. The
buzzers triggered are buzzers 1, 2, 3, 4. Movement inducted by the triggering on these
buzzers was tracked and observed and further analyzed. As seen in the pictures below,
there is a clear difference between the initial position, and the position after the buzzers
were triggered. It is clear that the spinal cord and the left hand (shown in the right side
in the picture) has tilted toward the right side (left in the picture). Also, the left hand is
closer to the body in the test position than the initial position, which confirms once
again the above stated fact.
Another type of observation was also conducted within this study. This refers to
visual observation as stated hereinbefore. The vibration buzzer trigger was, same as for
the first type of observation, of two seconds, with a five second pause.
The observation was conducted on four participants, which have were dressed with
the shirt with vibrating buzzers, and their reaction was tracked by visual observation. For
this first case, two types of reaction have been observer, considering the magnitude of
the reaction. In two of the four study participants, there was only a negligible movement
(for human observer) of the left hand towards the body. A variety of reaction can be
triggered, each person tested having different reactions from others (Fig. 6).
In the second case (cases 2, 3, 4 are shown in Fig. 7), a big amplitude reaction was
triggered on the other two participants, observing a movement starting at shoulder and
Fig. 6. Reaction LHS 1 and 2 [11]

Fig. 7. Case studies
finishing to wrist. Also, a small movement of the shoulder toward up and the back was
observed.
After the four experiences, we can observe a difference between the three results
shown below. One, from the Kinect sensor stated also a small movement of the spinal
cord. This could not be observed clearly by visual observer. This second statement
speaks of small movement whatsoever and the third one of a greater movement,
involving all the left side of the participants. The conclusion is that there is at least a
slight effect of the buzzers triggered, and the main direction of the human body reaction
is toward the right side.
In the same manner, the research was conduct in the other three cases. In these three
cases, measurements and observations were made exactly as in the first case presented.
The reactions of the participants are presented in Table 1. For the first case we pro-
posed an arousal of the left arms and shoulder. In the case is clear that both the Kinect
method and visual observation have come with the same results, both showing that
there is a human reaction in this case, and that is by moving toward the right side and a
short movement of the shoulder. It is clear what the reaction is when using such a
system and now it remains only a question of power level can be used and in which
situation this should be enabled.
Table 1. Reactions of the participants

Participants P1 P2 P3 P4
LHS reaction √ √ √ √
LB reaction √ 0 0 √
RHRS reaction √ √ √ √
RB reaction 0 √ 0 √
In the second scenario we proposed to observe and analyze the reaction the arousal
came from the left side of the body, by triggering the corresponding vibrating buzzer.
Here a more interesting reaction appeared, as differences appeared between the par-
ticipants in the study. In this case, the information gathered from the Kinect sensor has
showed a reaction of the tracked subject, and when manual observation was done,
differences were found between the participants. There were two participants having
the same reaction as the one identified by the Kinect sensor, and two participants which
had no reaction to these vibrating buzzers.
The third scenario is similar to the first scenario, but looked in the mirror,
watching the reactions of the right hand and right shoulder. The reaction identified here
were similar in all cases, both with Kinect sensor and human observation. The reaction
is always in the same direction, with small differences between the subjects. This side
of the body can easily be used to generate stimuli that can help the driver during the
driving task, to provide a feedback on the road ahead. All the observations indicated a
movement to the left side.
The last scenario discussed was with vibrating buzzer mounted on the right side of
the body. Here, same as on the other cases both computer and human tracking was used
and the result show again slightly different result. In four out of five cases a discernable
movement was tracked, concluding with a slight movement to the left side. This result
shown, similar to the second case study, we can have reaction only in certain cases, not
all human reacting in the same way to the same stimuli.
5 Conclusion
After registering the human reactions to different stimuli we observe that these vibro
tactile stimuli can be used for the future prevention on accidents. Further studies need
to be conducted to provide clear statements on the exact reactions in real driving
conditions, in order to quantifying the reaction by measuring the movement of the
human body after stimulation. As stated before, the purpose of this study was to
observe the human reaction to external stimuli in order to analyze the feasibility to
integrate vibro tactile system in vehicles to prevent and reduce the accidents caused by
lack of attention and to ensure a much better communication between the active safety
systems integrated in vehicles and the human driving the vehicle. Four analyses were
made by causing these stimuli on different parts of the body, they were tracked and
analyzed in order to determine the human reaction and the best areas in which to act.
Registrations were made by using a Kinect sensor and by visual observation.
Acknowledgements. This article contains the results of research conducted in preparation of

master’s thesis of the first author and was realized within the Partnership Programme - PN-II,
which runs with the financial support of MEN-UEFISCDI Romania, Project no. 227/2014,
System for Diagnosis and Therapy of Spine Diseases (SPINE).
References
1. Wardrop, J.G.: Some theoretical aspects of road traffic research, paper presented at Road
Engineering Division Meeting, 24 January 1952. http://www.its.uci.edu/*yangi/papers/
Wardrop.1952.pdf
2. The Lane Departure Warning Citroën. http://www.citroen.co.uk/about-citroen/technology/
lane-departure-warning-system. Accessed July 2016
3. Anderson, E., Bierman, C., Franko, E., Zelko, A.: The effects of audio and visual distractions
on reaction time, University of Wisconsin – Madison, laboratory notices (2012). http://jass.
neuro.wisc.edu/2012/01/Lab%20603%20Group%2014%20FINAL.pdf
4. Olson, P.L., Sivak, M.: Perception response time to unexpected roadway hazards. Hum
Factors J. Hum. Factors Ergon. Soc. 28(1), 91–96 (1986)
5. Olson, P.L.: Driver perception response time, SAE Technical Paper 890731 (1989). doi: 10.
4271/890731
6. Lee, J.D., Hoffman, J.D., Hayes, E.: Collision warning design to mitigate driver distraction.
CHI 6(1), 2004 (2004)
7. Rosengren, P., Wennerholm, K.: Design of a vibro-tactile warning system in an automobile
applications, Mst thesis LiU-IEI-TEK-A-08/00334-SE, Linkopings Universitet (2008)
8. Ho, C., Tan, H.Z., Spence, C.: Using spatial vibrotactile cues to direct visual attention in
driving scenes. Part F Traffic Psychol. Behav. 8(6), 397–412 (2005)
9. Mohebbi, R., Gray, R., Tan, H.Z.: Driver reaction time to tactile and auditory rear-end
collision warnings while talking on a cell phone. Hum. Factors 51(1), 102–110 (2009)
10. Meng, F., Gray, R., Ho, C.: Dynamic vibrotactile signals for forward collision avoidance
warning systems. J. Hum. Factors Ergon. Soc. 57, 329–346 (2014)
11. Bogdan, A.: Study concerning the driver reaction under surface mechanical stimuli, Master
thesis, Transilvania University of Brasov (2016)
Development of an Advanced Driver
Assistance System Using RGB-D Camera
Alin Pantea, Florin Girbacia(&), and Teodora Girbacia
Transilvania University of Braşov, 29 Eroilor Blvd., 500036 Braşov, Romania

alingpantea@gmail.com,
{garbacia,teodora.girbacia}@unitbv.ro
Abstract. In recent years, the automotive industry has shown increased interest
in Advanced Driver Assistance Systems (ADAS), especially those based on
bio-signals. Recent advances in RGB-D technologies have provided effective
solutions for tracking human activity based on depth data. In this paper is
presented an ADAS system based on Kinect RGB-D camera for the identifi-
cation of the driver’s distraction. Using depth and colour information the pro-
posed ADAS system is be able to identify the driver’s head orientation and eye
position. Based on this data, the driver’s inattention is detected and the driver is
warned by audio signals. The proposed ADAS system was evaluated using a
Virtual Reality driver simulator for manual and visual distraction. The results
show accurate recognition of driver’s distraction.
Keywords: Driver distraction RGB-D sensor ADAS
1 Introduction
Distracted driving is a state of inattention that can damage the vehicle occupants as well
as pedestrians and bicyclists. Diver distraction is encountered when drivers divert their
attention from the driving task to some other activity. Many studies regarding this
situation were focused on using mobile devices, but distracted driving also includes other
activities such as adjusting vehicle controls, discussing with the passengers or eating.
According to the US National Highway Traffic Safety Administration, distracted
driving played a significant role in 12 % of the teenager’s fatal car crashes accidents
[10]. Based on [8] in 2014 alone, 3,129 people were killed in distracted driving crashes.
Today’s mobile phones are so appealing to subject’s visual attention that even crossing
the street while using a smart phone can prove to be risky [1]. ADAS are developed to
enhance the safety inside the vehicle and in its vicinity and to provide better driving
skills. This kind of systems can be used to detect and prevent districted driving which
can lead to fatal accidents.
Over the last years, several studies describe solutions for ADAS systems [2–7, 9].
In [3] is presented a multimodal solution for detecting driver’s fatigue and distraction
using a driving simulator equipped with several sensors. The following data is pro-
cessed: audio, color video, depth map, heart rate, steering wheel positions and then the
signals are processed and analyzed. The authors show that by combining different types
of data is possible to estimate driver’s fatigue and distraction.

Development of an Advanced Driver Assistance System Using RGB-D Camera 747
Another study conducted on this topic is presented in [7]. It is described a novel

approach for studying the movements of the driver inside the vehicle. Four users were
asked to drive on one route along the freeway and arterial segment. The authors noted
that during the merging manoeuvre, the movement of the driver, although reduced, is
still perceived and can offer useful information of the driver posture while performing
the merging manoeuvre. In the study [5], is presented an algorithm for tracking of facial
features and face orientation estimation based on a single depth camera. The authors
show that the setup can be used for monitoring driver behavior.
The purpose of this paper is to present an ADAS system based on RGB-D
Microsoft Kinect camera and to evaluate the capability of the RGB-D camera to
become a useful part of the ADAS in identifying the situations in which the driver is
distracted using a driving simulator. The study is part of the NAVIEYES Project [14]
which aims to develop an intelligent assistant for mobile devices while driving. In a
few words, NAVIEYES aims to build a new navigation software product that will help
people to drive safer and easier.
2 Development of the ADAS System
The ADAS system was developed in Transilvania University laboratory, using the
Microsoft Kinect V2 RGB-D camera, Logitech G29 steering wheel and audio system
connected to a PC running a car driving 3D virtual environment. The whole layout of
the system is presented in Fig. 1. The advantage of using a RGB-D camera compared
to a 2D video device is the possibility of using the same device to obtain both depth and
color information regarding dynamic environments.
Fig. 1. The developed RGB-D based ADAS system

748 A. Pantea et al.
Microsoft Kinect V2 RGB-D camera was used to collect depth information with the
resolution of the time-of-flight camera of 512 by 424 pixels from the real environment
and sends it to a PC. The image generation rate was around 30 frames per second
(fps) [13]. The Kinect sensor was positioned above the visual display, and at 130 cm
from the user. The positioning of the Kinect sensor depends on the driving workspace.
The Logitech G29 steering wheel is an electronic steering wheel designed mainly for
driving video games with realistic force feedback that also includes a set of three pedals
which complete the immersive perception of the user in the virtual environment [11].
The system allows the identification of the driver’s eye and signals the situation in
which the driver’s gaze is off the road when driving in the 3D simulator. The eye
position was calculated using face orientation angles (pitch, roll and yaw) obtained
based on the face model captured using Kinect SDK [12]. We can determine the
distraction of the driver using the face rotation angles.
The event signalising that the driver’s eyes are off the road occurs in the following
situations:
• The user is looking down towards the floor, pitch angle > −10 °;
• The user is looking up towards the ceiling, pitch angle >10 °;
• The user turned towards the right shoulder, yaw angle > −30 °;
• The user turned towards the left shoulder, yaw angle >30 °;
When the driver eyes off the road event are detected, an audio beep signal is played
until the user is not distracted anymore and his gaze is focused on the road in front of
him while driving.
3 Evaluation of the Developed ADAS System
The conducted evaluation study tried to identify the effects of the proposed ADAS on
the user’s behaviour when driving in the VR simulator in three scenarios:
• Normal driving, with no distraction and no usage of ADAS referred to as driving
scenario 1.
• Normal driving with no ADAS enabled, then the driver engages in two distracting
actions– SMS writing and selfie photo making – followed by normal driving,
referred to as driving scenario 2.
• Normal driving with ADAS enabled, then the driver engages in two distracting
actions– SMS writing and selfie photo making – followed by normal driving,
referred to as driving scenario 3 (Fig. 2).
For the evaluation of the developed ADAS, we selected four students in engi-
neering (male, mean age 24.75, SD = 0.82) from Transilvania University of Brasov
without previous experience with using ADAS systems. Each subject filled in a
questionnaire, providing information about age, driving license, previous experience in
using driving simulators and ADAS systems. All participants had a valid driving
license and were in a good health, without vision problems. They did not wear glasses
throughout the simulations in order to not influence in any manner the reaction of the
ADAS system to the encountered events.
Fig. 2. User performing the experiment on the developed RGB-D based ADAS System
There were three driving sequences in a country road virtual environment scenario
for every subject, each sequence lasting for five minutes. The subjects were asked to
comply with all traffic rules while driving in the virtual environment (90 km/h legal
driving speed limit, giving priority, traffic signs, and lane departure and collision
avoidance). Each subject had up to 10 min to become familiar with the driving sim-
ulator and ADAS system.
In order to determine the effects of the proposed ADAS on the user’s behaviour, the
following parameters were investigated in this experiment: the number of improper
lane positions, speed infractions, turn signals missed. It was counted for each parameter
the number of times it occurred.
In the Fig. 3 below are presented the results from the conducted study. After
evaluating the results it was established that there is a significant difference between the
number of improper lane positions and speed infractions in the case of simulating
scenario 2 (with the user engaging in distracting activities and ADAS disabled)
compared to simulating scenario 1 when the user is driving safely. Also, user 3 had a
crash during the simulation of scenario 2, which happened due to improper lane
positioning. The results from testing scenario 2 clearly showed that smart phone dis-
tractions are very dangerous activities while driving a car.
When ADAS was enabled in the simulation of scenario 3, quantitative analysis of
the obtained data suggested improved overall driver performances related to the
number of improper lane positions compared to the results achieved from simulating
scenario 2. Also, the number of speed infractions and missed turn signals was lower
than in the second scenario. The number of the identified infractions is higher than in
the case of the scenario 1. This suggests that the using sound signals to alert the
distracted user is not an efficient method. In the future work other alerting methods, like
vibrotactile sensors, will be investigated.
750 A. Pantea et al.
Fig. 3. The parameters investigated: number of improper lane positions, speed infractions, turn
signals missed and total number of infractions for every user
4 Conclusions
This paper presents the development of an ADAS system based on a RGB-D camera.
The advantage of RGB-D sensors is that they provide both depth and color information
using the same device. It was shown from a test case that the proposed ADAS system
enables with higher accuracy the detection of the situations in which the driver’s eyes
are off the road. The developed system has the potential of a promising tool for the
detection and warning of distracted drivers. Further work is focused on the identifi-
cation and analysis of the user’s posture based on skeletal data received from the
Kinect sensor. Also, we plan to implement other output modalities (like vibrotactile
haptic feedback) for the perception of the warnings in case of eyes off the road event.
Acknowledgment. This paper is supported by the Romanian Government, specifically MEN –

UEFISCDI authority under the program PNII “Partnerships in priority areas”, under the project
number 240⁄2014 - NAVIEYES, supporting the collaboration between the company Route 66
and University Transilvania of Brasov.
References
1. Byington, K.W., Schwebel, D.C.: Effects of mobile Internet use on college student
pedestrian injury risk. Accid. Anal. Prev. 51, 78–83 (2013)
2. Duguleana, M., Dumitru, A., Postelnicu, C., Mogan, G.: Video-based evaluation of driver’s
visual attention using smartphones. In: IEEE 6th International Conference on Information,
Intelligence, Systems and Applications (IISA), pp. 1–5 (2015)
3. Craye, C., Rashwan, A., Kamel, M.S., Karray, F.: A multi-modal driver fatigue and
distraction assessment system. Int. J. Intell. Transp. Syst. Res. 14(3), pp. 1–22 (2015)
4. Gallahan, S.L., Golzar, G.F., Jain, A.P., Samay, A.E., Trerotola, T.J., Weisskopf, J.G., Lau,
N.: Detecting and mitigating driver distraction with motion capture technology: distracted
driving warning system. In: Systems and Information Engineering Design Symposium
(SIEDS), pp. 76–81 (2013)
5. Hu, Z., Uchida, N., Wang, Y., Dong, Y.: Face orientation estimation for driver monitoring
with a single depth camera. In: 2015 IEEE Intelligent Vehicles Symposium (IV), pp. 958–
963 (2015)
6. Trivedi, M.M., Gandhi, T., McCall, J.: Looking-in and looking-out of a vehicle. Selected
investigations in computer vision based enhanced vehicle safety. IEEE Trans. Intell.
Transp. Syst. 8(1), 108–120 (2007)
7. Kondyli, A., Sisiopiku, V.P., Zhao, L., Barmpoutis, A.: Computer assisted analysis of
drivers’ body activity using a range camera. IEEE Intell. Transp. Syst. Mag. 7(3), 18–28
(2015)
8. NHTSA: Distraction.gov, official US government website for distracted driving (2016).
http://www.distraction.gov
9. Postelnicu C.C., Machidon, O., Girbacia, M., Voinea, D., Duguleana, M.: Effects of playing
mobile games while driving. In: International Conference on Distributed, Ambient, and
Pervasive Interactions, pp. 291–301 (2016)
10. Selfies while driving. http://edition.cnn.com/2013/11/06/tech/mobile/selfies-while-driving/.
Accessed July 2016
11. Logitech G29 driving force. http://gaming.logitech.com/en-roeu/product/g29-driving-force.
Accesed July 2016
12. Kinect for Windows SDK. https://developer.microsoft.com/en-us/windows/kinect
13. Microsoft Kinect V2. https://developer.microsoft.com/en-us/windows/kinect/hardware
14. NaviEyes research project site. http://navieyes.unitbv.ro
A Method for Transforming Electric Vehicles
to Become Autonomous Vehicles
Mihai Kocsis1(&), Andreas Schultz2, Raoul Zöllner2,

and Gheorghe Leonte Mogan1
1
Auto Vehicle and Transportation Department Brasov, Transilvania University,
Brașov, Romania
mihai.kocsis@hs-heilbronn.de, mogan@unitbv.ro
2
Automotive Systems Engineering, Heilbronn University, Heilbronn, Germany
{andreas.schultz,raoul.zoellner}@hs-heilbronn.de
Abstract. In order to develop autonomous navigation tasks for transportation

systems or for fulfilling certain agriculture or industrial tasks, vehicle prototypes
are needed. These vehicles cannot be reinvented every time. Existing vehicles
are transformed by using a data processor that replaces the driver. Furthermore,
depending on the vehicle type, actuators (e.g. steering or braking), sensors that
observe the environment and interfaces to the information provided by the
vehicle must be integrated. This process differs from vehicle to vehicle. In this
paper we present a (further) solution for such a reconstruction, which is, indeed
not completely general, but due to its modular architecture, the open software
implementation and the use of open source components, easy to apply on new
systems. The paper presents the hardware and software implementation on two
vehicles. A third and a fourth one are planned. The electric power unit is not
mandatory but helpful for the transformation, because an intervention in an
already electric vehicle is easier to carry out.
Keywords: Autonomous vehicles Electric vehicles
1 Introduction
To provide a car with the ability to navigate autonomously, the driver must be replaced
by technical components. Therefore, actuators are needed to realize the control and
sensors to capture the environment and certain relevant driving states. Finally, the
reactive and cognitive ability of the driver must be fulfilled by such a system. Through
the autonomous functions, the vehicle must be able –similar to a person– to respond
fast and safe to all danger and changes that occur in the environment.
The first requirement regarding the reconstruction of the vehicle is the control of the
accelerator pedal, the brake pedal and the steering wheel. Although, it’s a general
requirement the reconstruction strategy differs from vehicle to vehicle e.g. brake type
(electric, hydraulic) or steering type (mechanical, electrical, steer by wire). Further-
more, the choosing and positioning of sensors is also depended on the task the vehicle
has to fulfil.

A Method for Transforming Electric Vehicles 753
An important aspect of the system is its modularity and scalability. The extension
the system with other components (e.g. unit to perform an agricultural task) should be
possible without changing the existing hardware and software components.
Nevertheless, the system should be able to detect if one of the components fail to
work properly and react fast and precise. Regarding on the importance of the com-
ponent, the whole system has to be able to change its behaviour avoiding damage to
itself and to other obstacles around it.
The electric vehicle has the advantage that many components are electrical and the
most important one, the traction, is realized by an electric motor. The access to these
units in order to control them is usually easier fulfilled. Mechanical components like a
mechanical steering have to be extended by electrical motors combined with an
electronic control unit (ECU). From the architectural point of view these ECUs together
with further introduced internal sensors (potentiometer etc.) or actors (e.g. motors)
build the operational execution hardware system. The interface to this system is gen-
erally realized by a Communication-Bus. A further component contains the external
sensors mounted to the car used for localization, navigation and object detection.
A main controller (MCU) coordinates the execution and the sensing, and is usually
realized by one (or if needed more) master ECU.
2 State of the Art
2.1 Electric Vehicles Architectures

Usual electric cars consist of a battery, an electric motor and a controller that dictates
how much of the motor’s energy is transmitted to the wheels. When the driver presses
the accelerator pedal, the controller signals the motor to increase the speed. As soon as
the driver removes his foot from the pedal, the motor power will be set to zero.
A potentiometer and a closing contact is attached to pedal and is connected to the
controller. Every time the driver presses down the pedal, the contact will close and
depending on the potentiometer value, the voltage will be adapted through the con-
troller increasing the motor power. The brake and the steering differ from vehicle to
vehicle. The brake pedal can work in the same way as the accelerator pedal, the
controller gets the pedal position with the help of a potentiometer and controls for
example the motor (recuperation) or a hydraulic brake. Steering wheel systems are
often mechanical, but especially in nowadays cars equipped with a servo steering
mechanism, an extra electric motor is included. For special application external steering
systems exist, which are guided by GPS, e.g. EZ-Pilot System [2]. These systems
consist of a motor that is mounted on the steering wheel and a controller which gets the
actual position from a GPS System and has a defined path to follow. These systems are
mainly used in agricultural vehicles, where the driver should concentrate on the agri-
cultural task, without the need of steering the vehicle. In order to make the vehicle
autonomously, these three components need to be controlled.
754 M. Kocsis et al.
2.2 Sensors
In order to get information about the vehicle and to interact with the environment,
sensors and sensor data fusion are required. A bright overview of sensors available in
the car including positioning sensors, acceleration sensors, angular rate sensors and
obstacle detection sensors can be found in [1]. Regarding the obstacle detection sen-
sors, the positioning in the car is very important and has to be made individual
according to the type of the car and the purpose of the task. Furthermore, the com-
munication and data fusion between the sensors is relevant for obtaining more accurate
and precise information. A state of the art survey regarding sensor data fusion is
described in [1].
2.3 Mapping and Path Planning Algorithms

Based on the control of the actuators, the information of the vehicle and of the envi-
ronment, a map and a path planning algorithm is required to fulfill the autonomous
navigation to a certain position. The fundament to perform a proper navigation is an
efficient and accurate simultaneous localization and mapping algorithm (SLAM).
Hereby, with the data from the sensors, the vehicle system builds a map while local-
izing itself relatively to this map. Meanwhile several SLAM algorithms can be found in
the literature [9, 10]. Having the map, a global path planning algorithm (e.g. grid path,
simple potential calculate, A*) calculates a path for the navigation from the actual map
point to the given destination point. A possibility to do this is presented by the authors
in [13]. With the map information and the information provided by sensors during the
drive, a local path planner recalculates the route in case that obstacles are detected on
the initial path. Local planer algorithms are the TEB Local Planer (described in [6, 7])
or the SBL Lattice Planner whose functionality is described in [5, 8]. Many SLAM and
Path Planning algorithms exist as a pre-built function in the Robotic Operating System
(ROS) framework [3].
3 Hardware Architecture
In order to transform an electric vehicle into an autonomous one, from the hardware
point of view, three main steps have to be taken into account: the control of the steering
system, the control of the pedals and the integration of sensors in order to observe the
environment and perform the autonomous navigation.
The systems have to be integrated in an overall design consisting of different levels
of control and safety loops (Fig. 1).
The reactive layer has the task to control the actuators and to react properly in case
of any dysfunction. The sensors used for controlling the vehicle are usually placed on
the actuator or integrated in the power electronics. The data processing on this layer
occurs at a high frequency (up to 1 kHz).
The system component extends the control loop with further sensors connected to
the system. The computation of this loop has a slower computation rate (50–100 Hz)
and requires more system knowledge, e.g. the kinematics of a component.
The last loop integrates all partial systems of the vehicle to a complete system.
Here, on the one hand, information about the environment are gathered and on the other
hand, kinetic and dynamic models of the complete system are considered. Here the
computation frequency is in the range of 1–10 Hz.
Fig. 1. Hardware architecture with closed loops.
Figure 2 shows the overall hardware architecture of one of the vehicles and the
locations of parts in the car. The components of the three systems are marked with a
distinct colour. A component consists usually of an actuator with power electronics, a
microcontroller and sensors. It has to be attached to a defined vehicle part. The actu-
ators and the sensors are placed beside the vehicle components that have to be con-
trolled. The power electronics is placed near the actuators due to electromagnetic
compatibility. The controlling MCUs are placed together in a control box in order to
minimize communication distance. This communication is done via CAN Bus, Eth-
ernet (LAN), USB or in analogic mode. The ECUs use a separate power supply of
12 V. Therefore, an additional battery is integrated in the battery tray. For reliability
reasons, a further battery is integrated into the system, powering the actuators.
Fig. 2. Hardware components: wheel steering components (pink), pedals control (orange) and
sensors (green). (Color figure online)
3.1 Steering System

Most of the electrical vehicles available on the market have a mechanical steering. In
order to control the direction, an electric motor need to be integrated. The motor is
controlled through a PWM signal. The logic board (ECU) gets from the CAN Bus the
desired steering value, calculates the steering angle and sends the appropriate PWM
signal to the H-Bridge connected to the motor. A steer angle sensor with a CAN
interface mounted on the steering shaft sends the actual steering angle to the logic board.
Figure 3 shows the used components for the steering and the interaction between them.
Fig. 3. Driving direction steering components
The vehicle velocity, which is available on the CAN Bus, is used in the control
algorithm that generates the PWM Signal. In case the velocity is high, the wheel
steering speed and the maximum steer angle are reduced accordingly.
3.2 Brake and Acceleration System

For electric vehicles the control of the driving motor is done by a potentiometer
attached to the gas and a brake pedal. The voltage is then passed to the vehicle engine
controller which controls the electric motor. Alternatively, the brake can be hydraulic
and in order to control it, another electric motor must be added. The brake pipe of the
car is connected to this motor (Fig. 4 right bottom). The pipe is closed by the motor and
pressure is generated in the braking system.
For the control of the pedals a board (ECU) with a microcontroller was connected
between the pedals and the vehicle controller (Pedals Board in Fig. 4 left). On two
analog inputs the voltage value (in range of 0 to 5 V) from the potentiometer attached to
the pedals is received. The analog value is transformed to a digital one for further
processing. The processed value on the controller is retransformed to an analog signal
and send via an analog output to the inbuilt controller of the vehicle. In order to control
the pedal switch, a relay for each pedal is needed. The inbuilt end switch of the brake
pedal opens when the pedal is pressed. The electric circuit to the magnetic clutch is
opened and the driving wheels are blocked. The same principle is used when controlling
the pedal from the board. An opener is preventing an undesired brake maneuver in the
case of loss of voltage. The relay is connected in serial to the inbuilt switch. When
pressing the gas pedal the inbuilt switch closes the circuit and the magnetic clutch is
activated. The relay controlled from the board is connected in parallel to the inbuilt
Fig. 4. Pedals steering components (left); Brake pedal control electric (right top) and hidraulic
(right bottom)
switch. The driver can oversteer the vehicle at any time. The controller is connected to
the CAN Bus. In the autonomous driving mode another board calculates the values for
the pedals (PID Board). Here, a proportional integral derivate controller is running. It
gets the required speed from the main controller (Car PC) and the actual position of the
pedals from the first controller and calculates the required position of the pedals.
3.3 System Level Sensors

For steering, localization and path planning several sensors for getting information
about the vehicle and the environment must be used. In order to measure the speed of
the vehicle two Hall sensors are mounted on the wheels and connected to the ECU. The
actual velocity is send by the board over the CAN-Bus to other components which use
this information. An inertial measurement unit and a GPS system can be used to get the
global location of the vehicle. The information from these sensors is then merged to
obtain a more accurate position. In order to avoid collision with other objects or
participants, laser scanners combined with cameras are used. These are connected via
Ethernet to the Car-PC (Main ECU).
Fig. 5. Used sensors and communication protocols (left); inside control box (right)
4 Software Architecture
The software architecture is also subdivided in the reactive, the system and the vehicle
layer.
The reactive layer is implemented on the microcontrollers that get control inputs
from the vehicle layer. It consists of the controlling of the hardware components by
using fast communication channels. The µCs react to current, voltage, temperature and
pressure.
The system layer is implemented also on microcontrollers and deals with high level
controlling e.g. position or momentum. The µCs get the information from sensors by
using a bus communication system (CAN-Bus). On this layer the calculation of
kinematic models occurs. This includes calculation of position dependent maximum
values. For example, the maximal angular speed or maximum acceleration.
The vehicle layer consists of several sensors and a main controller (Car-PC) which
is connected to the sensors by several communication protocols (Fig. 5 left). The laser
scanners, the cameras, an inertial measurement unit (IMU) and a Real Time Kinematic
Global Positioning System (RTK GPS) are used in the autonomous driving process.
Furthermore, cameras are currently used to detect markers near a charging station. The
Robotic Operating System (ROS) [11] runs on the Car PC, which is the main com-
putation component of the system. ROS is a software framework that offers a large
number of libraries to develop robotic application. It provides hardware abstraction,
low level device control, message passing between processes and package manage-
ment. A ROS system consists of independent nodes which communicates with each
other through a publish/subscribe messaging model. The pcan node uses the CAN
socket drivers to access the CAN-Bus. Two other nodes (can_to_ros and ros_to_can)
are used to provide the information from the CAN bus to the main system and from the
main system to the CAN bus. Other nodes provide information from the sensors
(inertial measurement unit, GPS, Hokuyo laser scanner, Ibeo laser scanner). After the
information are gathered, three main steps have to be taken: map generation, calcu-
lation of an accurate odometry and the path planning.
4.1 SLAM, Odometry and Path Planning

In order to create the map, several laser scanners are used. A ROS node (laser_tool) is
used for the fusion of the data from all laser scanners. The result of this fusion is used
as an input for the simultaneous localisation and mapping (SLAM) node. ROS provides
several nodes with different SLAM algorithms (e.g. G mapping, Karto slam or Hector
slam). Best results were achieved with the G mapping algorithm. Figure 6 shows the
map with a part of the parking place at the Heilbronn University.
From the car model (steering angle and speed), based on an Ackerman model [12] a
first odometry is calculated. The RTK GPS provides a second odometry. Based on this
two odometries combined with the data from the inertial measurement unit (IMU), a
Kalman filter as described in [4] calculates a more accurate odometry.
Based on the generated map, a path planning algorithm calculates the path that has
to be followed by the vehicle. ROS provides for global path planning several algo-
rithms: Grid Path, Simple Potential Calculate or A*.
Fig. 6. Map created with the G mapping algorithm used within the ROS Node
With the information from the map and the information provided by the
laser-scanners, a local planner calculates the new route in case objects are detected on
the initial calculated path. Local planner algorithms provided by ROS are: base_local,
teb_local_planner, DWA or carrat_planner. The best results were achieved using the
teb local planner.
5 Safety Concept
Another important aspect of the system is the safety strategy. What happens if one of
the control board fails? For this case a watch dog solution was implemented. A separate
board checks the cyclic messages from the boards. Figure 7 shows the connection of
the watchdog board with the other boards. If a cyclic message from one of the boards
Fig. 7. Watchdog, safety solution

fails for five times consecutive, the watchdog board sends to each board adequate
messages. In critical situation the system is safety turned off.
Although the safety layer provides a reliability in controlling the vehicle autono-
mously, we have to point out that for the driverless steering of the vehicle a complete
redundancy of all the systems is required. This can be achieved through the multi-
plication of the µCs, the use of the sensors with different principle of measurement and
up to the duplication of the communication busses.
6 Test and Validation
The above presented architectures were successfully implemented by the authors on

two vehicles: a Golf cart EZGO Freedom and a Piaggio Porter. The two vehicles were
able to drive autonomously on a given path, or plan a path with the help of a given
self-acquired map.
Beside the functionality tests, safety tests were made, by inducing the failure of
single sensors, actuators etc. on the real system. As previously described, a safety
driver has to be always on board since we don’t implement a fully redundant system.
The presented transformation is suitable particularly through the ROS integration as
a testing environment for new sensors, sensor clusters or functions. In this way several
different GPS sensors (D-GPS or RTK-GPS) were tested regarding their reliability on
path tracing. Another advantage in using ROS is offered by the simulation environment
Gazebo, which is integrated in ROS. Sensor values recorded during driving scenarios
can be analysed offline and visualized in the simulation environment (Fig. 8 left).
Figure 8 (right) shows the actual available vehicles (EZGO Golfcart, Piaggio Porter
and Audi e-tron) at the Automotive Systems Engineering department at the Heilbronn
University and the Smart at the DATR department at the Transilvania University in
Brasov.
Future experiments will investigate the right choice for sensors for different driving
tasks e.g. agricultural tasks.
Fig. 8. The Piaggio Porter in the simulation environment Gazebo (left); EZGO Golfcart, Piaggio
Porter, Audi A3 e-tron, Smart (right)
7 Conclusion
The paper describes a method to convert vehicles in order to make them useful to
execute autonomous tasks. The hierarchical Software–Hardware concept is modular
and enables the integration of diverse sensors and µCs. Although the execution
depends of the type of the vehicle, the consistent pursuit of the concept brings a clear
benefit, due to the fact that adaption can be made fast, easy and safe. The ROS inte-
gration in the vehicle layer supports the functionality development through the pre-built
functions, the integrated simulation environment (Gazebo) and the possibility to use
and analyse real data in offline mode.
References
1. Winner, H., Hakuli, S., Lotz, F., Singer, C. (eds.): Handbook of Driver Assistance Systems.
Springer, Cham (2016)
2. http://www.trimble.com/Agriculture/ez-pilot.aspx
3. Quigley, M., Gerkey, B., Conley, K., Faust, J., Foote, T., Leibs, J., Berger, E., Wheeler, R.,
Ng, A.: ROS: an open source Robot Operating System. In: IEEE International Conference on
Robotics and Automation (ICRA), Workshop on Open Source Software (2009). https://
www.willowgarage.com/sites/default/files/icraoss09-ROS.pdf
4. Negenborn, R.: Robot Localization and Kalman Filters, Copenhagen, September 2003 http://
www.negenborn.net/kal_loc/thesis.pdf
5. http://wiki.ros.org/sbpl_lattice_planner
6. http://wiki.ros.org/teb_local_planner
7. Rösmann, C., Feiten, W., Wösch, T., Hoffmann, F., Bertram. T.: Trajectory modification
considering dynamic constraints of autonomous robots. In: Proceedings of 7th German
Conference on Robotics, Munich, pp. 74?79 (2012)
8. Rösmann, C., Feiten, W., Wösch, T., Hoffmann, F., Bertram, T.: Efficient trajectory
optimization using a sparse model. In: Proceedings of IEEE European Conference on Mobile
Robots, Barcelona, pp. 138?143 (2013)
9. Durrant-Whyte, H., Bailey, T.: Simultaneous Localisation and Mapping (SLAM): Part I the
essential algorithms. IEEE Robot. Autom. Mag. 13, 99?110 (2006). doi:10.1109/MRA.2006.
1638022. http://people.eecs.berkeley.edu/*pabbeel/cs287-fa09/readings/Durrant-Whyte_
Bailey_SLAM-tutorial-I.pdf
10. Chen, Y.: Algorithms for simultaneous localization and mapping, 3 February 2013
11. https://www.willowgarage.com/pages/software/ros-platform
12. Weinstein, A.J., Moore, K.L.: Pose estimation of Ackerman steering vehicles for outdoors
autonomous navigation, Industrial Technology (ICIT). IEEE (2010)
13. Gall, R., Tröster, F., Mogan, G., Luca, R.: Trajectory planning for mobile car-like robot
navigation in an extra-road environment. Solid State Phenomena. Trans Tech Publications,
Switzerland (2010)
Accident Research and Analysis
Cars Crashes with Cars of Same or Different
Generation – Occupant’s Loads, Timings
and Acceleration’s Effects
Nicolae Ispas(&) and Mircea Nastasoiu
Transilvania University of Brașov, Brașov, Romania

{inicu,m.nastasoiu}@unitbv.ro
Abstract. Improving occupant safety is a main target of today cars designers.

Known as active or passive safety, many technological solutions were devel-
oping over the time for an actual better car’s occupant safety. In the real world,
in traffic accidents are often involved cars from different generations with var-
ious safety historical solutions. The main aim of these papers are to quantify the
influences over the car and car’s occupants loads and vehicle accelerations in
cases of same or different generation of cars involved in side car crashes. Both
same and different cars generations were used for the study. Other goal of the
paper was the study of in time loads conformity for occupants from both cars
involved in crash. The paper’s experimental results were obtained by support of
DSD, Dr. Steffan Datentechnik GmbH - Linz, Austria. The described tests were
performed in full test facility of DSD Linz, in “Easter 2015 PC-Crash Seminar”.
In all crashes we obtaining results from both dummy placed in impacted and
hits car. The novelty of the paper are the comparisons of data set from each of
occupant (dummy) of two cars involved in each of seven experimental crashes.
Another novelty of this paper consists in possibilities to analyze the influences
of structural historical cars solutions over deformation and loads in each cases of
traffic accidents involved. Paper’s conclusions can be future used for car
designers and for passive safety improvement.
Keywords: Crash test Traffic accident Acceleration level Occupant safety
1 Introduction
RTD (Research Technological Development) in the field of occupant safety often

involves crash tests that are strong tools for improve both occupants and road vehicle
safety. Compatibility between cars has for a long time been reduced to the simple
image of heavy against light cars. Over the last years, vehicle structure has been
improved thanks to materials, new improved restraint systems and new structural
solution for crash energy management, that provide a better understanding of the front,
end and side design energy absorption.
Front end and side design is at the cross road of numerous contradictory con-
straints: self-protection of occupants, protection of vulnerable users such as pedestrians,
reparability, styling, aerodynamics, engine cooling and so on.

766 N. Ispas and M. Nastasoiu
Therefore, each manufacturer has developed its own solution to solve this difficult
equation which resulted in a wide variety of front end and side designs, structure and
stiffness regardless of the overall mass of the vehicle. Solutions however have been
optimized against a rigid wall or soft obstacle but not in car to car configuration [1, 4, 5].
Side-impact collision occurs when one car hits the side of another car. This kind of
collision, commonly referred to as a “T-bone” or “broadside collision” is often caused
by merging into another vehicle’s lane, speeding through a red light, or by t-boning
another car at an intersection [1]
Side-impact collisions most likely result in car damage, though side impact bars
and side airbags should serve as buffers to impact. Side-impact bars are steel bars that
are fitted into the doors of the vehicle, forming part of a protective cage around the
occupant. This kind of technical solution are very different for different generations of
cars.
For the goals of this paper, in order to obtain valuable experimental data, we
obtained the support of DSD, Dr. Steffan Datentechnik GmbH - Linz, Austria. The
follow described tests were performed in full test facility of DSD Linz, in “Easter 2015
PC-Crash Seminar” [2, 3].
Seven crash experiments were made for estimate the accelerations over the car
driver head and in crash cars accelerations, using cars and crashes design showed in the
Figs. 1, 2, 3, 4, 5, 6 and 7.
The main technical data of the cars used in the crashes experiments together with
the cars shapes are presented in the same Figs. 1, 2, 3, 4, 5, 6 and 7.
Fig. 1. First crash, cars involved details and cars pre-crash positions.
Cars Crashes with Cars of Same or Different Generation 767
Fig. 2. Second crash, cars involved details and cars pre-crash positions.
Fig. 3. Third crash, cars involved details and cars pre-crash positions.
Fig. 4. Fourth crash, cars involved details and cars pre-crash positions
Fig. 5. Fifth crash, details of cars involved and cars pre-crash positions
Fig. 6. The sixth crash, details of cars involved and cars pre-crash positions
Fig. 7. The seventh crash, details of cars involved and cars pre-crash positions
2 Experimentals Tests
Seven crash tests were performed according with EN1317 and EN12767. In each of
them, except of last one, was used dummy as car driver in both in crash cars. Complex
accelerometers were mounted for dummy’s head, chest, neck and pelvis (Fig. 8).
Accelerations, velocities and rotations of the cars were also measured.
In the Figs. 9, 10, 11, 12, 13, 14 and 15 on see heads acceleration data together
with acceleration of each car involved in respect of the crash timing.
For quantify the influences over the car and car’s occupants of loads and vehicle
accelerations in cases of same or different generation of cars, we used valuable data
obtained from real world experimental crashes.
All tests were performed under the conditions that the driver was not protected by
an air bag and seat belt pretensions.
This kind of test data enable the possibilities to analyze the influences of structural
historical cars solutions over deformation and loads in each cases of traffic accidents
involved. The analyze can be future used for car front, end and side design
improvement.
For driver’s heads accelerations, sensors were placed as can be seen in Fig. 8 [6].
Sensors for cars acceleration were mounted in the COG of each car.
Fig. 8. Dummy – placement of sensors [6]

Fig. 9. Car’s accelerations and driver’s heads accelerations in the time of first crash test
Fig. 10. Car’s accelerations and driver’s heads accelerations in the time of second crash test
Fig. 11. Car’s accelerations and driver’s heads accelerations in the time of third crash test
Fig. 12. Car’s accelerations and driver’s heads accelerations in the time of fourth crash test
Fig. 13. Car’s accelerations and driver’s heads accelerations in the time of fifth crash test
Fig. 14. Car’s accelerations and driver’s heads accelerations in the time of the sixth crash test
Fig. 15. Car’s accelerations and driver’s head accelerations in the time of the seventh crash test
3 Conclusion
Using for each crash case cars with or without technical/age compatibility, we can write
the followers main conclusions:
A. First crash test:
1. Cars crash was at low impact speed (19,4 km/h) between vehicles from different
generation (Fiat Punto 55 – from year 1996 versus Chrysler PT Cruiser from year
2006);
2. The cars are dissimilar weights (Fiat Punto 55-895 kg, Chrysler PT Cruiser-
1470 kg);
3. For 2006-year car generation the car side from neighborhood of “B” pillar is more
stiffness compared with stiffness of the same zone from a cars made in 1996;
4. The Fiat Punto car is made with a low absorber of crash energy in the car front
structure;
5. The maximum absolute amounts of accelerations of Fiat Punto car driver head were
7.6 g on “Z” direction and 8.7 g on “X” direction;
6. The maximum absolute amounts of accelerations of P01 car driver head were 4 g on
“Z” direction and 9.8 g on “Y” direction.
7. Side crash occurs on the right side of the Chrysler PT Cruise car;
8. Maximum lateral acceleration of the impacted car (Chrysler PT Cruise) – 61 [m/s2]
occurs earlier with 0,02 s than maximum of longitudinal acceleration of car hits –
52 [m/s2].
B. Second crash test:
1. Cars crash was at low impact speed (19,3 km/h) between vehicles from different
generation (Fiat Punto 55 – from year 1994 versus Chrysler PT Cruiser from year
2006);
2. The cars are dissimilar weights (Fiat Punto 55-850 kg, Chrysler PT Cruiser-
1470 kg);
3. For 2006-year car generation the car side from neighborhood of “C” pillar and rear
axle is more stiffness compared with stiffness of the same zone from a cars made in
1994;
4. The Fiat Punto car is made with a low absorber of crash energy in the car front
structure;
5.5 g on “Z” direction and 6.82 g on “X” direction;
6. The maximum absolute amounts of accelerations of P01 car driver head were 5 g on
“Z” direction and 7.85 g on “Y” direction;
7. Side crash occurs on the left side of the Chrysler PT Cruise car;
8. Maximum lateral acceleration of the Chrysler PT Cruise– 27 [m/s2] occurs earlier
with 0,015 s than maximum of longitudinal acceleration of Fiat Punto – 53 [m/s2].
C. Third crash test:

1. Cars crash was at relatively low impact speed (40,7 km/h) between vehicles from
the same generation (Fiat Punto 55 – from year 1996 versus Fiat Punto 55 – from
year 1997;
2. The cars are similar weights (Fiat Punto 55, 1997-830 kg, Fiat Punto 55,
1996-895 kg);
3. For 1996-year car generation the stiffness of front door is the same compared with
stiffness of the front door of a cars made in 1997;
4. Both Fiat Punto cars are made with a low absorber of crash energy in the car front
structure;
5. The maximum absolute amounts of accelerations of Fiat Punto 1996-year car driver
head were 13 g on “Z” direction and 13.9 g on “X” direction;
6. The maximum absolute amounts of accelerations of P02 car driver head were 12 g
on “Z” direction and 42 g on “Y” direction.
7. Side crash occurs on the right side of the impacted car;
8. Maximum vertical acceleration of the impacted car – 170 [m/s2] occurs practically
in the same time with vertical acceleration of the car hits – 100 [m/s2] and later with
0,01 s than maximum of longitudinal acceleration of Fiat Punto 55 from 1996 – 100
[m/s2].
D. Fourth crash test:
1. Cars crash was at relatively low impact speed (40,7 km/h) between vehicles from
the same generation (Fiat Punto 55 – from year 1994 versus Fiat Punto 55 – from
year 1996;
2. The cars are similar weights (Fiat Punto 55, 1994-890 kg, Fiat Punto 55, 1996–
895 kg);
3. For 1994-year car generation the neighborhood of “C” pillar and rear axle have the
same stiffness compared with stiffness of the same zone from a cars made in 1996;
structure;
head were 10.6 g on “Z” direction and 8.8 g on “X” direction;
7. Side crash occurs on the left side of the Fiat Punto-1996 car;
8. Maximum lateral acceleration of the impacted car – 78 [m/s2] occurs earlier with
0,018 s than maximum of longitudinal acceleration of Fiat Punto from 1994 –
122 [m/s2].
E. Fifth crash test:
1. Cars crash was at medium impact speed (59,4 km/h) between vehicles from very
different generation Opel Astra from year 2014 versus Fiat Punto 55 – from year
1994);
2. The cars are dissimilar weights (Opel Astra from year 2014 – 1440 kg, Fiat Punto
55 from year 1994-850 kg);
3. For 1994-year car generation the front side door is les stiffness compared with
stiffness of the same zone from a cars made in 2014;
4. The Opel Astra car is made with a high absorber of crash energy in the car front
structure;
18,5 g on “Z” direction and 33 g on “X” direction;
on “Y” direction and 73 g on “X” direction.
7. Side crash occurs on the right side of the Fiat Punto 55 car;
8. Maximum lateral acceleration of the impacted car – 338 [m/s2] is recorded at the
same time with maximum of longitudinal acceleration of Opel Astra – 153 [m/s2];
vertical acceleration maximum of Opel – 212 [m/s2] occurs later than maximum of
vertical acceleration of Fiat – 220 [m/s2].
F. The sixth crash test:
1. Cars crash was at relatively medium impact speed (60.7 km/h) between vehicles
from the same generation (Fiat Punto 55 – from year 1997 versus Fiat Punto 55 –
from year 1994;
2. The cars are similar weights (Fiat Punto 55, 1997-890 kg, Fiat Punto 55,
1994-830 kg);
3. For 1997-year car generation the stiffness of the side in place of rear axle is the same
compared with stiffness of the side rear axle from a car made in 1994;
structure;
head were 14.8 g on “Z” direction and 14 g on “X” direction;
7. Side crash occurs on the left side of the Fiat Punto-1997 car;
8. Maximum lateral acceleration of the impacted car – 482 [m/s2] occurs in the same
time with maximum of longitudinal acceleration of Fiat Punto from 1994 –226
[m/s2].
G. The seventh crash test:
1. Cars crash was at relatively high impact speed (80 km/h) between vehicles from
almost the same generation (VW Golf 1.4 – from year 1998 versus Honda Accord
4D – from year 1992);
2. The cars are similar weights (VW Golf 1.4, 1998 - 1137 kg, Honda Accord 4D,
2002 - 1272 kg);
3. For 2002-year car generation the stiffness of the side in B pillar area is comparable
with stiffness of the side from a car made in 1998;
4. The maximum absolute amounts of accelerations of VW Golf (P09) year car driver
head were 35 g on “X” direction and 7 g on “Y” direction.
5. Side crash occurs on the right side of the VW Golf-1998 car;
6. Maximum lateral acceleration of the impacted car – 290 [m/s2] occurs later with
0,0124 s than maximum of longitudinal acceleration of Honda from 1992 – 172
[m/s2].
Accelerations above all values are indicates as absolute values. For almost all cases of
collision can be seen as a driver’s head acceleration levels are higher than levels
accelerations cars involved.
REFERENCES
1. Du Bois, P., et al.: Vehicle Crashworthiness and Occupant Protection, vol. 4875. American
Iron and Steel Institute, Southfield (2004)
2. Ispas, N., Năstăsoiu, M.: Crash Tests and the Loads over Driver Head in Different Side. In:
Impact Cases Current Solutions in Mechanical Engineering, pp 181–186 (2016)
3. Ispas, N., Năstăsoiu, M.: A study concerning the loads over drivers heads in the cases of cars
crashes with involved cars of the same or different generation. In: Andreescu, C., Clenci, A.
(eds.) Proceedings of the European Automotive Congress EAEC-ESFA 2015, pp. 325–336.
Springer, New York (2015)
4. 54th GRSP 2013 Research on Chest Injury Criteria
5. Thorsten, A., Eggers, A., Thomson, R., Koji, M.: Comparison of the dummy response in two
different restraint system crash tests. In: IRCOBI Conference 2014, pp 545–561 (2014)
6. Zellmer, H.: Dummy design and issues. In: International Course on Transportation Planning
and Safety, New Delhi, December, 2010
Crash Tests Data Acquisition and Processing
Dinu Covaciu(&) and Dragoş Sorin Dima

{dinu.covaciu,d.dima}@unitbv.ro
Abstract. Crash tests are helping the automotive manufacturers to improve the
security of the vehicles, and are also useful for validating computer simulations
of road traffic accidents. In crash tests are used real vehicles instrumented with
sensors that measures speed, forces and accelerations. The human body,
pedestrian or vehicle occupant, is replaced with a dummy, also instrumented
with sensors. The sensors must be connected to data acquisition systems that are
able to record information at high rates and being at the same time shockproof.
For different types of tests (vehicle-vehicle, vehicle-pedestrian, lateral or frontal
impact) are required different sensor’s characteristics. Some recommendations
regarding vehicle instrumentation for impact tests, including data processing, are
included in standards like SAE J211. In this paper are presented data acquisition
devices used in real tests, and the methods for processing data, including
sampling rates and filters. There are used sensors integrated in data acquisition
systems from well recognized manufacturers and also sensors mounted in new,
original developed systems.
Keywords: Crash test Data acquisition Acceleration Digital filter

Accident reconstruction
1 Introduction
Crash tests are destructive tests performed in order to ensure safe design standards in
crashworthiness and crash compatibility for vehicles, or automotive systems and
components. Nowadays, there are many well known car safety performance assesment
programmes, but there are also crash testing programmes performed by research
institutes (like DSD - Doctor Steffan Datentechnik) or academic institutions like
Transilvania University of Brasov [1]. The various types of crash tests include: frontal
impact tests, overlap tests, side impact tests, roll-over tests, vehicle-pedestrian tests or
vehicle-two wheeler tests, and roadside hardware crash tests. The vehicle(s) must be
instrumented according to the type of test, and data processing should also be done
accordingly. Recommendations regarding the instrumentation for impact tests are given
in SAE J211 standard [2]. This standard is focused on the use of accelerometers as
impact sensors.
In the report of the Data Processing Vehicle Safety Workgroup [3] are described
some criteria for crash analysis, for different parts of the dummy (head, neck, chest,
extremities), and criteria for measuring and filtering the acceleration values. The fil-
tering procedure recommended by SAE J211 is to use the channel frequency classes
(CFC) of low-pass filters. Filtering is the most critical phase in the processing of impact
Crash Tests Data Acquisition and Processing 783
signals [4]. The filter’s primary function is to eliminate the high-frequency noise.
Filtering reduces the peaks in the signal.
SAE J211 designate four CFC filters for crash data processing: CFC 60, 180, 600,
and 1000. The CFC filters are derived from analog Butterworth filters that have the
corner frequency (the frequency at which the signal loses one half of its power) equal to
the CFC designation divided by 0.6. The filter types and their parameters are listed in
Table 1.
Table 1. Filter types, SAE J211

Filter type Filter parameters
CFC 60 3 dB limit frequency 100 Hz
Stop damping −30 dB
Sampling frequency at least 600 Hz
Sampling frequency at least 1800 Hz
Sampling frequency at least 6 kHz
Sampling frequency at least 10 kHz
The expertize of the testing engineer is important when selecting the frequency
response class appropriate for the application. The selected frequency class will
determine the necessary sampling frequency (according to Table 1).
2 Data Acquisition
2.1 Acceleration Measurement
Sensors. The sensors used for acceleration measurement, the accelerometers, are
MEMS devices [1]. MEMS defines miniaturized mechanical and electro-mechanical
elements made using the techniques of microfabrication. The accelerometer is basically
a capacitive or piezoresistive device consisting of a suspended pendulum proof
mass/plate assembly. Acceleration acts on the proof mass and the micromachined
plates sense a change in acceleration from deflection of the plates [5]. Depending by the
type of output signal, the accelerometers can be digital or analogue. In the crash tests
presented further, in this paper, are used analogue accelerometers. The most important
parameter that define the appropriate accelerometer to be used in the crash test is the
input range. This can vary from ±2 g to ±200 g or even more. The accelerometer must
be chosen depending by the expected values to be measured. The impact duration is
784 D. Covaciu and D.S. Dima
very short, so the bandwidth is also important. The sensor bandwidth, also referred to
as frequency response, is an indication of the sensor’s ability to respond to changes in
the measured displacement. Higher bandwidth sensors can measure higher frequency
motion. This means that the sensor bandwidth influences the sampling frequency of the
acquisition device. The MEMS accelerometers usually can have a bandwidth between
0.5 Hz and 1600 Hz, and can be set using external capacitors. In some tests this can be
a limitation. Another parameter is the sensitivity, and depends by the input range and
maximum output voltage. For example, if the input range of the accelerometer is ±5 g,
and the output voltage is between 0 and 3.3 V, this means the output resolution is
0.33 V for 1 g. The sensitivity is influenced by the temperature.
Data loggers. The sensors output is a voltage. This is an input for the data acquisition
board, and is converted in values to be stored on memory cards. Two types of devices
were used in the tests that are at the basis of this work: PicDAQ, made by DSD (Doctor
Steffan Datentechnik) [6] and Loka, an original device constructed by authors. Each
acquisition device have two components: the phisycal device, or hardware, and the
logical component, or the software.
PicDAQ is a complete acquisition system made for crash tests and also for braking
tests. It includes two sensor units: for braking tests, an accelerometer on three axes with
a ±5 g input range, and a gyroscope that measures the pitch-, roll- and gyration rate;
for crash tests, an accelerometer on three axes with an input range of ±50 g. Data are
stored on SD card, in a custom binary format that is converted to a easy to use text
format with the help of the dedicated software. A dedicated software application is used
for processing the data (PockedDaqAnalyzer, with the user interface in Fig. 1).
Loka is a small acquisition device comprising a data logger board and two analogue
accelerometers. The input ranges of accelerometers are ±3 g and ±200 g. The logger
is based on the LPC2148 ARM7 microcontroller and records data on a micro-SD card.
Fig. 1. Screen capture of PocketDaqAnalyzer

Fig. 2. Screen capture of Loka software
For data processing was designed a custom software application, with the user interface
An important phase in preparing the test equipment is the accelerometers calibra-
tion. This is done using the gravity as reference. The procedure is described in the
application notes of the sensor manufacturers [9], as well as the methods for measuring
tilt with accelerometers [10].
2.2 Speed Measurement

Vehicle speed is important in crash tests, as it is important also in accident recon-
struction. It can be calculated by integrating the acceleration, but the result is affected
by accelerometer’s bias. Better results are obtained with GPS receivers, when the test is
done outside, with clear view to the sky. The devices used in the tests conducted by
authors are GPS loggers with a sampling rate of 10 Hz. Each sample is a $GPRMC
sequence according to the NMEA 0183 standard. The sequence contains information
like: time, position (longitude and latitude), speed (in knots) and course (degrees
relative to North). The most important in a crash test are time and speed.
The acceleration can be calculated as variation of speed in the given time, but the
sampling rate is too small for a crash test; it cannot be captured precisely the moment of
the impact, and it is not possible to determine the crash pulse [11].
3 Data Processing - Filtering
One of the goals of a crash test is to identify the deceleration curve that describe the
impact. Depending by the obstacle with which the vehicle collides, there is a crushing
that reduces the speed of the vehicle, and the crushing duration will determine the
amplitude of the deceleration pulses. The series of decelerations (with a total duration
of hundreds of milliseconds), may be called a crash pulse or deceleration curve [11].
The crash pulse first affect the vehicle, then the interior components, and then the
occupants. The severity of injuries depends by the amplitude of deceleration.
One of the objectives of data processing is to identify the crash pulse, in order to
find better solution for vehicle protective structure. The measured signal (acceleration
amplitudes) is affected by noise that causes peaks which have to be removed. It is
important to remove only the peaks caused by noise, not the useful signal peaks.
The analog-digital converters featured by the controllers that are the cores of the
data acquisition devices have usually a 10 bits input, which limits the resolution. This is
sufficient for most of crash test applications, but a higher resolution can improve also
the signal/noise ratio. A technique commonly used is oversampling and averaging,
which is described in the application note AVR121 from Atmel [7] and also in other
documents [8]. The data acquisition device Loka has the oversampling feature included
in the firmware. The same technique can be applied also in the stage of data processing.
Another important operation that may be performed before data are analyzed using
the recommended criteria [3], is filtering. Filter’s primary function is to eliminate
undesired noise [4]. The digital filter that should be used, according to SAE J211, is a
4-channel Butterworth low-pass filter with linear phase. The filter sequence is described
by the difference equation:
YðtÞ ¼ a0 XðtÞ þ a1 Xðt 1Þ þ a2 Xðt 2Þ þ b1 Yðt 1Þ þ b2 Yðt 2Þ ð1Þ
where the coefficients a0, a1, a2, b1, b2 are calculated as follows, and they depend by the
filter class and sampling frequency [2].
x2
a0 ¼ pffiffiffi a ð2Þ
1 þ 2 xa þ x2a
a1 ¼ 2 a0 ð3Þ
a2 ¼ a 0 ð4Þ
2 ðx2a 1Þ
b1 ¼ pffiffiffi ð5Þ
1 þ 2 xa þ x2a
pffiffiffi
1 þ 2 xa x2a
b2 ¼ pffiffiffi ð6Þ
1 þ 2 xa þ x2a
where
sinðxd T=2Þ
xa ¼ ð7Þ
cosðxd T=2Þ
xd ¼ 2 p CFC 2:0775 ð8Þ

Fig. 3. Sample raw data from a calibration procedure
where CFC is the filter class (e.g. 60) and T is the sampling period, in seconds.
A sample diagram based on the unfiltered acceleration data, measured during a
calibration procedure, is shown in Fig. 3.
By applying a CFC filter on the signal shown in Fig. 3, using formulas (1–8), it is
obtained a smoothed curve, like the examples shown in Fig. 4 for 60, 30 and 15 class,
respectively. It is obvious that for lower classes (15 in the example) the signal is
smoother than for higher classes (60), but the amplitude is more influenced.
Fig. 4. Effect of different filter classes (CFC 60, 30 and 15)
4 Crash Tests - Results
One of the crash tests carried out involved a vehicle and a pillar (Fig. 5). The speed of
the vehicle was 50 km/h and the duration of the impact was about 130 ms. There were
measured accelerations, on the vehicle body, up to 25 g on x axis (longitudinal
acceleration) and 10–15 g on the other axes (Fig. 6). In the diagram it can be seen the
crash pulse, with high amplitude and short duration. The filter used is CFC 60, for a
sampling frequency of 1 kHz. A higher level of filtration (CFC 15) reduces the
amplitude with about 5 g. The sampling rate is also very important: at 1000 samples
per second, in an interval of 100 ms there are only 10 samples. A lower sampling rate is
likely to lose critical values for test. The bandwidth of the MEMS sensors available on
the market must be also taken into consideration.
Fig. 5. Crash test: vehicle - pillar
Fig. 6. Accelerations measured during the vehicle-pillar impact (unfiltered and CFC 60)
Fig. 7. Vehicle-pedestrian impact, data acquired with GPS device and Loka
Another type of tests are those that involve vehicle and pedestrians or bicyclists,
namely objects with much lower mass and speed. An example is the vehicle-pedestrian
impact illustrated in Fig. 7, conducted at a vehicle speed of 37 km/h with the pedes-
trian dummy being stationary (diagrams from devices mounted on vehicle body). In
this case, the deceleration of vehicle is given mainly by braking, and the measured
value is close to the value that can be calculated from GPS speed and time. A short
pulse before braking is also around the value of 1 g. The measured values lead to the
conclusion that in vehicle-pedestrian impacts the input range of the sensors should be
much lower than in the previous example.
5 Conclusions
The objective of this paper was to highlight the importance of the data acquisition
devices used in crash tests, as well as of the parameters involved in data processing.
The different types of tests involve the use of sensors with different characteristics.
The input range of accelerometers used in impacts of vehicles with stationary obstacles
must be large (±50 g, ±100 g), in comparison with those used in vehicle-pedestrian
impacts, with a lower input range (±5 g) and higher sensitivity. The accelerometers
installed inside the dummies must have higher input ranges (±200 g).
An important step in data processing is filtering, in order to eliminate the high
frequency noise. The digital filters CFC are recommended by SAE J211, and the filter
class must be chosen according to the sampling frequency, which is chosen depending
by the expected duration of the impact.
References
1. Dima, D.S., Covaciu, D.: Use of MEMS sensors for data acquisition in crash tests, In:
Proceedings of EVU 2016 Congress, Bratislava (2016)
2. SAE: J211 Standard, Instrumentation for impact tests, Rev., March 1995
3. Cichos, D., de Vogel, D., Otto, M., Schaar, O., Zoelsch, S.: Data Processing Vehicle Safety
Workgroup - Crash Analysis Criteria Version 1.6.1 (2004)
4. Alem, N., Perry, M.: Design of digital low-pass filters for time-domain recursive filtering of
impact acceleration signals. USAARL Report No. pp. 95–13 (1995)
5. PRIME Faraday Partnership: An Introduction to MEMS. Loughborough University (2002).
ISBN: 1-84402-020-7
6. DSD: PicDAQ Technical Data. http://www.dsd.at. Accessed May 2016
7. Atmel: AVR121: Enhancing ADC resolution by oversampling, Application note (2005)
8. Kester, W.: Oversampling Interpolating DACs, MT-017 Tutorial, Analog Devices (2009)
9. STMicroelectronics: Tilt measurement using a low-g 3-axis accelerometer, AN3182
Application note (2010)
10. Fisher, C.J.: Using an Accelerometer for Inclination Sensing, AN-1057 Application Note,
Analog Devices (2010)
11. Peters, G.A., Peters, B.J.: Automotive Vehicle Safety. Taylor & Francis, London (2003).
ISBN 0-203-26092-9
Research Regarding Occupant’s Movement
in the Case of Frontal Collision Using
High-Speed Video Recording
Alexandru Ionut Radu1(&), Corneliu Cofaru1, Bogdan Tolea1,

Daniel Dragos Trusca1, and Horia Beles2
1
alexandru.radu@unitbv.ro
2
Abstract. The paper is proposed to study the analysis of a frontal vehicle

collision using high-speed video recording and dedicated software to study the
occupant’s movement. From the high speed footage, data can be acquired
regarding the occupant’s head acceleration, speed and distance, but also data
regarding the vehicle’s velocity, and acceleration. These data will be used to
assess the injury criteria of the occupant.
Keywords: Occupant Collision Acceleration Velocity Video Analysis

Recording
1 Introduction
Frontal crashes have proven to be the most frequent type of vehicle accidents in which
front seat occupants get injured. To further increase the vehicles protection for its
occupants vehicle manufacturers conduct crash tests in similar condition to the real
world conditions. However, there is no real evidence to prove the validity of these tests
in real world situations, the crashworthiness of the vehicles [1]. The main reason is that
the fatality rate is influenced by the type of people who are driving the vehicles [2]. In a
study, Wenzel and Ross, 2000, reported that “Drivers of different vehicles differ in their
distributions of driver’s age and sex but also differ in innate driving style and
propensity for riskier driving behaviors” [3].
Since its introduction on motor vehicles, the seat belt has been the most effective
system to protect the vehicles occupants in case of an accident. In Europe the standard
setup of the seat belt is the 3 point configuration, covering the lap and shoulder of the
passengers. It has been proven that the seatbelt will save the passenger’s life if it used,
and not using it in the same circumstances will lead to death [4].
Seat belts are set on preventing the most serious of injuries, for example protecting
the head and thorax from an impact to the dashboard, or steering wheel, as well as
keeping the occupant in its place [5]. The seat belt is also influenced by the position of
the seat, the legs are exposed to contact with the dashboard and steering wheel, and the
upper body, to hit the instrument panel. The rear seat occupants have a greater chance

Research Regarding Occupant’s Movement 791
to injury, about 60 % more than front occupants [6]. Seat belts may prove more
efficient in protecting rear seat occupants.
2 Objectives
The main goal is to use a high-speed recording of a car-to-car frontal collision to study
the driver occupant’s kinematics.
From the video analysis the velocity and acceleration of the vehicles can be
determined using the derivation and integration methods.
The secondary objective is to determine the occupant’s head acceleration and the
assessment of the head injury criteria.
3 Methods and Equipment Used
The method used includes a controlled collision between 2 vehicles: a impacted vehicle
and an impactor one. In the impacted vehicle, a male test dummy with equipment
mounted to measure the acceleration will be positioned. The impact was recorded with
a high speed recording camera, to analyze every frame in a dedicated software. Also,
the velocity of the vehicle was measured by using a GPS receiver. Video cameras were
mounted inside the vehicle at different angles to capture the motion of the dummy. The
used crash test vehicle was equipped with an airbag positioned in the steering wheel.
Also, the dummy was provided with an acquisition system mounted inside to acquire
the acceleration values during the impact.
The equipment was as follows: accelerometers, mounted in the vehicles and in the
test dummy, high speed camera, capable of recording at 1200 FPS, advanced male
crash test dummy, 2 vehicles (mid-size sedan), GPS equipment, video cameras and
DSLR cameras.
To create the collision in a safe environment, a test was conducted in a zone clear of
any other vehicles or pedestrians, where a vehicle with the crash test dummy was
positioned at a marked area, where the high speed camera was pointed to, and the other
vehicle at a distance, far enough to reach the speed of 35 km/h. The position of the
vehicles is presented in Fig. 1.
Fig. 1. Controlled collision procedure

792 A.I. Radu et al.
To accelerate the second vehicle, a system of cable, pulley and a third traction
vehicle were used. The traction vehicle had a 140 HP engine so it could pull and get to
impact speed the second vehicle.
In Fig. 2, the collision phases are presented, the pre-impact phase where the
vehicles first come into contact with each other, the impact phase, where the energies
are transferred, and the post-impact phase where the two vehicles detach and come to a
standstill. In this test, the impacted vehicle with the crash test dummy has an airbag
mounted in the steering wheel that deploys in the impact. Also, the seatbelt has a
pretension system that is also activated at the moment of impact.
Fig. 2. Collision phases
After the test, the high-speed footage was analyzed in two dedicated software: one
is Tracker: Video Analysis and Modeling Tool, capable of tracking an objects posi-
tions, velocity and acceleration and generate data, and the other is Origin Pro, a
software used for data analysis and graphing.
4 Results
The first results are generated from the equipment installed in the vehicle and dummy,
in Fig. 3 the velocity of the impactor vehicle is obtained from the GPS system installed
on it. Using the velocity, the acceleration was calculated.
The velocity at which the vehicles collide was 35 km/h and the deceleration at the
moment of impact was 19.5 m/s2.
Fig. 3. Vehicle velocity and acceleration
And for the dummy, the head’s acceleration was obtained from the tri-axial
accelerometer mounted in the head. Acceleration is presented in Fig. 4.
Fig. 4. Dummy’s head acceleration
The head acceleration of the occupant-dummy was pretty low, about 80 m/s2,
which means that it would not suffer any major injuries or breakings. The reason the
acceleration was low is because the airbag had great influence in reducing the risk of
the head’s hitting the steering wheel.
From the equipment, the acceleration magnitude can be calculated and concluded in
Fig. 5, the maximum head acceleration is 90 m/s2.
Using the analysis software and the high speed video recording, the head accel-
Fig. 5. Acceleration of the occupant’s head
Fig. 6. Acceleration of the occupant’s head obtained from video recording
eration of the occupant can also be obtained, presented in Fig. 6. It can be observed that
the data obtained from the high speed video, matches the data obtained from the
equipment, with little difference, in the case of head acceleration, the difference is about
5 m/s2.
Also, from the video recording, the velocity for the impacted vehicle can be
obtained, as well as the acceleration of this vehicle at the moment of impact. These are
From the software the velocity of the impacted vehicle, during the impact, went
from 0 at the moment of impact, up to 18 km/h, in a time of 0.3 s, generating an
acceleration of 54 m/s2. Compared to the second vehicle, the acceleration value was
much higher because of the energy transfer. The graph shows only the time period of
the impact, and the vehicle was later braked, but that period is not irrelevant.
By means of the results provided by the dedicated software, the speed of the
occupant’s head can be obtained by integrating the acceleration values. Also the
Fig. 7. Impacted vehicle’s velocity and acceleration during the impact
distance of the dummy’s head can also be obtained by integrating the velocity. In
Fig. 8, the dummy’s head velocity is presented.
The maximum velocity of the head was 7 m/s during the forward movement of the
head, and reduced at 3 m/s during the flexing of the neck backwards (Fig. 9).
Fig. 8. Dummy’s head velocity during the impact
The occupant’s head distance, relative to the vehicle’s interior, is about 0.6 m; this
is the maximum distance the head traveled during the flexion phase and 0.1 m during
Fig. 9. Distance of the occupant’s head

the extension phase. The flexion was at an angle of 42°, and the extension 24°. The
flexion and extension phases are presented in Fig. 10.
These angles are good, because there are elements that limit them. In the front there
was the airbag that stopped the head from hitting the steering wheel and in the back
Fig. 10. Flexion and extension of the head during impact
there was the seat headrest that also blocked the head and neck from possible injury.
Also, because of the pretension on the seatbelt, the upper torso was secured in the seat
and had little movement during the course of impact.
To assess the injury criteria, the HIC (head injury criteria) formula can be applied to
the head acceleration of the occupant. The formula is presented below.
Zt2
1
HIC361 ¼ max½ðt2 t1 Þð aðtÞdtÞ2:5 ¼ 6:3 ð1Þ
t1;t2 t2 t2
t1
Where:
t2 – final value of the time (s)
t1 – initial value of the time (s)
a(t) – the resulting head acceleration (g) [7].
Because of low acceleration values obtained during the collision, the HIC value is
very low as well. This means that the occupant would not suffer any major injury
during this type of crash and these conditions. Values that are above 700 are considered
dangerous for occupants, with a 5 % of severe injury [8].
5 Conclusions
From the video recording, data were acquired regarding the impacted vehicle and its
occupant. One observation is that the acceleration of the impacted vehicle was higher
than the impactor vehicle, because of the energy transfer at the moment of collision. It
can be concluded that, by using high speed footage and special software, accurate data
can be obtained, similarly to using equipment on the vehicles. Also, from the test
conducted, the restrain systems are very effective, especially the airbag. It reduced the
possible impact of the occupant’s head with the steering wheel. Also, the head
acceleration measured was low, about 9 g’s, which results in a low injury risk, and
during the impact the flexion and extension of the occupant’s head and neck are limited
between the airbag and the headrest, resulting in a low injury risk of the neck.
References
1. National Highway Traffic Safety Administration. New Car Assessment Program Public
Meeting. US Department of Transportation, February 28, 1995
2. Kahane, C.J.: Correlation of NCAP performance with fatality risk in actual head-on collisions.
NHTSA Technical report. National Highway Traffic Safety Administration, Department of
Transportation, Washington, DC (1994)
3. Wenzel, T.P., Ross, M.: The effects of vehicle model and driver behavior on risk. Accid.
Anal. Prev. 37, 479–494 (2000)
4. Høye, A.: How would increasing seat belt use affect the number of killed or seriously injured
light vehicle occupants? Accid. Anal. Prev. 88, 175–186 (2016)
5. Elvik, R., Høye, A., Vaa, T., Sørensen, M.: The Handbook of Road Safety Measures.
Emerald, Bingley (2009)
6. Smith, K.M., Cummings, P.: Passenger seating position and the risk of passenger death or
injury in traffic crashes. Accid. Anal. Prev. 36(2), 257–260 (2004)
7. Trusca, D., Tolea, B., Radu, I.: Research Regarding the Severity of the Injury of the
Pedestrian’s Head with the Vehicle’s Bonnet
8. Guidelines for Rating Injury Measures, IIHS (2009)
Research Regarding the Influence of Vehicle’s
Safety Restraint Systems on Its Occupants
in Case of Rear-End Collision
Alexandru Ionut Radu(&), Corneliu Cofaru, Bogdan Tolea,

Daniel Dragos Trusca, and George Toganel

alexandru.radu@unitbv.ro
Abstract. This paper studies the effect of a failed automotive front seat in the
case of a rear-end collision and its effects on the occupants. Also, a comparison
is made between a normal seat and a failed one. Two collision tests are con-
ducted on the same conditions to analyze the similarities of the two. The
occupants used are biofidelic crash test dummies. The occupant kinematics is
analyzed and also the head acceleration is measured. The data is obtained from
equipment mounted in the vehicles and occupants but also obtained from a
high-speed video recording.
Keywords: Occupant Collision Whiplash Seat Restrain Rear-end
1 Introduction
Rear impact collisions are mostly low severity, but carry a very high societal cost due
to reported symptoms of whiplash and related soft tissue injuries [1]. Rear-end collision
are most common in urban areas where vehicles emergency stop at intersections and
pedestrian crossways, and accidents occur due to a high velocity and driver’s lack of
attention. In these types of accidents, the occupants’ head and neck tend to slip
backwards and then forward, also called the whiplash effect. Whiplashes injuries on the
human body are caused by sudden deceleration events, most often caused by accidents
the rear-end motor vehicle is subject to, or by other types of collisions [2].
Statistics show that about 28 % of motor vehicle accident emergency department
visits had neck sprain/strains the predominant injury [3]. The best way to reduce the
whiplash effect is with headrests. In motor vehicles, they protect the head and the upper
spine in case of crashes. Another advantage is that they provide comfort that people
experience by the support of their heads by the headrest [4].
2 Objectives
The target of this paper is to assess the occupant injury in the case of the seat failing in
a rear end collision. Also conduct two tests to compare the injury assessment when the
seat fails and when it is normal. From the comparison the occupant kinematics can also
be analyzed using high speed footage.
Research Regarding the Influence of Vehicle’s Safety 799
3 Methods and Equipment Used
In order to obtain accurate data, two rear end crash tests were performed using four
different vehicles. In one of the impacted vehicles, the driver’s seat was deteriorated
before the test started such that during the collision it will fail such that our study could
be conducted. In the other vehicles the seats were working properly. The crash tests
were performed at the same collision velocity using the same crash configuration: one
of the vehicles was standing still while the other was accelerated to the desired velocity.
In all of the tests, on the driver’s seat of the impacted vehicles, a crash test male
dummy that was equipped with a 3-axial accelerometer mounted in the head was used.
All the vehicles were equipped with GPS devices such that it could gather accurate data
regarding the collision velocity, accelerations/decelerations during the collision,
respectively the traveling distances to the rest positions.
As a safety measure, a remote control braking device was used on all of the vehicles
during the collision such that they can be emergency braked in case it was needed.
All the crash tests were recorded with a high speed recording camera, capable of
recording at 1200 FPS. This recording can be used to determine the occupant kine-
matics during the collision.
In the first test, the collision was between 2 vehicles of the same class, compact
class, and the seat was set to fail during the impact. As it is shown in Fig. 1, in the
phase of the impact, the occupant moved out of position when the seat fell backwards.
Fig. 1. Impact phases of the first test
In the second test, two different vehicles were used, of a different class, a mid-size
class, where the impact was done on similar conditions with the first test, but the
occupant’s seat behaved normal and the headrest also protected the occupant from
possible whiplash effect. This test is presented in Fig. 2.
Fig. 2. Impact phases of the second test
After these tests had been conducted, the data obtained were processed using
special software to determine the velocity of the vehicles. The acceleration of the
impacted vehicles and occupants was determined.
4 Results
The first results are the vehicles’ velocities, there are the two impactor vehicles that
were accelerated at a certain velocity, and the two impacted vehicles that were stopped.
The results will be presented in the succession of the tests.
Test 1
In Fig. 3, the velocity and acceleration of the impactor vehicle is presented. The
velocity at the moment of impact was about 36 km/h and the acceleration generated
was 35 m/s2 or 3.5 g’s.
Fig. 3. Impactor vehicle’s velocity and acceleration

In Fig. 4, the velocity and acceleration of the impacted vehicle is presented. During
the impact the velocity of this vehicle went from 0 up to 19 km/h in the time of 0.2 s,
generating an acceleration of almost 80 m/s2 or 8 g’s. This is higher than the impactor
vehicle because the energy transferred from one vehicle to the other during the
collision.
Fig. 4. Impacted vehicle’s velocity and acceleration
From this test the occupant’s head acceleration was obtained and presented in
Fig. 5.
Fig. 5. Occupant’s head acceleration
In this test, the seat failed and the occupant fell backwards, and the acceleration on
the head was about 250 m/s2, or 25 g’s. This is a high value that can cause injuries to
the head and neck. Also, from the high speed recording the occupant kinematics can be
analyzed. In Fig. 6, the occupant’s movement is presented.
The dotted lines represent the initial position of the occupant before the impact, and
the solid lines represent the maximum displacement of him/her during the impact
phase. The upper torso moved at an angle of 38° relative to the initial position, and the
head and neck moved at an angle of 29° relative to the torso. This movement can cause
major injuries to the torso, but more important, the massive whiplash can injure the
occupant’s neck and head. Also, there is the possibility of the head’s hitting the rear
seat and cause breaks.
Fig. 6. Occupant’s movement
Test 2
In the second test, similar graphs can be obtained, in Fig. 7, the acceleration and
velocity of the impactor vehicle are presented.
Fig. 7. Impactor vehicle’s velocity and acceleration
The maximum velocity of the vehicle at the moment of the impact was about
31 km/h, and had an acceleration of 55 m/s2 or 5.5 g’s. The velocity of the impact was
5 km/h lower than that obtained in the previous test, but it compensates as there were
used different types and different weights of the vehicles during the test, compared to
the first test when the vehicles were lighter and smaller.
In Fig. 8, the impacted vehicles’ velocity and acceleration are presented.
One can observe that even though the velocity of the impact was lower than the
previous test, the velocity and acceleration of the impacted vehicle, with the occupant
inside, was almost the same during the impact, in the second test the velocity went up
from 0 to 18 km/h and the vehicle had an acceleration of 80 m/s2 during the impact
phase. Also, the occupant’s head acceleration is presented in Fig. 9.
In the second test, the occupant’s seat was behaving normally and in this way, a
lower acceleration value was measured on the occupant’s head, 165 m/s2, or 16 g’s.
This is 54 % lower than on the previous test, which represents a lower risk of injury in
the case of a rear impact.
Fig. 8. Impacted vehicle’s velocity and acceleration
Fig. 9. Occupant’s head acceleration
Also, from the test, the occupant’s movement can be analyzed and compared to the
previous test. The result is presented in Fig. 10.
In this test, the position of the torso moves at a maximum angle of only 18° from
the initial position (upright position before the impact) during the impact phase, and the
head and neck at 25° relative to the torso. The seat’s headrest prevented the whiplash
effect, limiting the displacement of the neck.
Fig. 10. Occupant’s movement

Compared to the previous test, the torso had a displacement angle twice as low with
the seat behaving normally compared to the failed one. The neck and head had only a
13 % less angle than in the previous test.
5 Conclusions
The condition of the occupant’s seats is an important factor for the occupant’s safety in
case of a rear-end collision. It is well known that the most common whiplash cases
have come from this type of collisions and the seat’s condition is also influencing this
phenomenon.
The occupant in the vehicle where the seat failed had a 54 % higher acceleration
value due to the whiplash effect and the leaning backwards toward the back seat of the
vehicle. Also, the displacement angle of the occupant’s torso leaning backwards is
twice as high, and 13 % of the head and neck.
In can be concluded that a failed seat in the case of a rear-end collision may double
the possibility of the occupant’s hitting objects in the back seat and more than 50 % an
increase of head injury due to the increased acceleration also adding the possibility of
neck injury due to the effect of whiplash.
References
1. Nishimura, N., Simms, C.K., Wood, D.P.: Impact characteristics of a vehicle population in
low speed front to rear collisions. Accid. Anal. Prev. 79, 1–12 (2015)
2. Seroussi, R., Singh, V., Fry, A.: Chronic whiplash pain. Phys. Med. Rehab. Clin. N. Am.
26(2), 359–373 (2015)
3. Department of Rehabilitation Medicine, University of Washington, 1959 Northeast Pacific
Street, Box 356490, Seattle, WA 98195, USA; b Seattle Spine & Sports Medicine, 3213
Eastlake Avenue East, Suite A, Seattle, WA 98102, USA
4. Franz, M., Durt, A., Zenk, R., Desmet, P.M.A.: Comfort effects of a new car headrest with
neck support. Appl. Ergon. 43, 336–343 (2012)
The Assessment of the Head Injury
of a Pedestrian in Comparison with a Cyclist
Bogdan Tolea(&), Stelian Tarulescu, Daniel Dragos Trusca,

George Toganel, and Alexandru Ionut Radu

bogdan.tzolea@gmail.com
Abstract. The main goal of this paper is to determine the differences regarding
the head injury risk of a pedestrian in comparison with a cyclist which is not
wearing a helmet. The secondary objective is to study the kinematics behavior
for both pedestrian respectively the cyclist during the collision. In order to
obtain accurate results, two crash-tests were performed using a dummy which
had mounted a tri-axial accelerometer in the head. Using head acceleration
values we were able to determine the head injury risk.
Keywords: Pedestrian Cyclist Crash test HIC Head accelerations
1 Introduction
Together with the pedestrians, the bicyclists represent the vulnerable road users [1].
Cycling represents one of the most important ways to reduce air pollution and traffic
congestion [2].
Usually, the cyclists use the same infrastructure as the other vehicles, but their vul-
nerability is higher than other motorized vehicles [2]. A study conducted by Daniels et al.
in 2009, shown that the by replacing the signal-controlled intersection with roundabouts
jeopardize bicyclists safety and increases the number of bicycle accidents [6].
Yan et al. reported in 2011 that more than 6 million people are riding bicycles daily,
being exposed to severe injuries in case of a collision with a motorized vehicle [1, 3].
In case of a car-to-bicycle accident the most common injury part is the rider’s head
[4, 5]. Some studies have shown that by wearing a helmet, the head injury severity may
be reduced [7, 8].These conclusions were drawn based on comparative studies about
the injury characteristics reported in the hospitals, either when the use of helmet was
due to a public information campaign or when the helmet use was mandatory in some
of the countries [9].
The study of Nie et al. from 2014 concluded that the most important parameters that
influence the cyclist kinematics during a collision are the vehicle frontal shape struc-
ture, collision velocity, initial posture of the rider [1].

806 B. Tolea et al.
2 Methodology
In order to determine the pedestrian’s injury severity risk, was necessary to perform
two different crash-tests which can give us good results. For the first crash test a vehicle
equipped with a GPS, necessary to determine the collision velocity, and a dummy
equipped with two tri-axial accelerometers mounted in the head and thorax were used.
The configuration of the crash-tests is presented in the Fig. 1.
Fig. 1. Crash test configuration for the car-to-pedestrian respectively car-to-bicycle crash test
In the case of car-to-pedestrian crash-test, the dummy experienced the same phases
of impact as a real pedestrian, and this phases are described below:
• Contact phase – this is the phase where the bumper and the bonnet leading edge
enter in contact with the pedestrian
• Carrying on the bonnet phase – in this phase the pedestrian is carried out on the
bonnet moving with the same velocity as the vehicle
• Sliding phase – this is the phase where the velocities of the pedestrian and the
vehicle are no longer the same, and usually the vehicle is braked and the pedestrian
has the tendency to continue with the same velocity as in the previous phase
These phases are summarized in the Fig. 2.
For the second crash-test, where the collision was between the car and a bicycle, the
collision phases have an additional phase, called the “car-to-bicycle contact phase”.
These phases are the following:
• Car – to –bicycle contact phase – this phase starts when the vehicle’s frontal profile
enters in contact with the bicycle
• Car-to-cyclist contact phase- this phase is described by the contact between the
pedestrian with the vehicle
The carrying on the bonnet respectively sliding phase are identical with the phases
described above, in case of the pedestrian. All these 4 phases are described in the
Fig. 3.
The Assessment of the Head Injury of a Pedestrian in Comparison 807
Fig. 2. Car-to-pedestrian impact phases
Fig. 3. Car-to-bicycle collision phases
3 Results
After the crash-tests were performed, data stored in the acquisition devices was
interpreted and filtered. Our main interest was to determine the head injury severity,
and in order to do that, head acceleration data was necessary.
One of the ways to assess the head injury severity risk is by determining the HIC
(Head Injury Criteria) values. HIC is a scale which by definition is the integration of the
head resultant acceleration over a period of time (36 ms or 15 ms) [10].
The resultant head acceleration over the period of collision, in the case of the
car-to-pedestrian crash tests is presented in the following diagram (Fig. 4).
It can be observed that the head acceleration peak value is to 932 m/s2 (approxi-
mately 93 g, where g is the gravitational acceleration), obtained after 200 miliseconds.
This is the moment where the pedestrian’s head collides with the vehicle’s base of the
windshield (wipers area), generating high accelerations due to the stiff surface.
For the second crash-test (car-to-bicycle), the obtained resultant head acceleration
diagram is presented in Fig. 5.
In this case, it was determined a resultant head acceleration peak value of
1465 m/s2 (approximately 146 g)m after 240 miliseconds. The delay of 40 ms between
the obtained head acceleration of the two cases comes due to the additional impact
phase in the case of the car-to-bicycle impact, described above.
808 B. Tolea et al.
Fig. 4. The pedestrian’s resultant head acceleration for the first crash-test
Fig. 5. The cyclist’s resultant head acceleration for the second crash-test
The resultant head acceleration of the cyclists presents an increasing with 63 % in

comparison with the pedestrian’s head accelerations. The main reasons of this
increasing is the contact point between the dummy’s head in the two analyzed cases. If
in the first case the dummy’s head enters in contact with the wipers area, which have a
higher elasticity, in the second crash-test, the cyclist head enters in contact with the
separation of the ceiling with the windshield, creating a significant deformation that
generates high accelerations. During a collision with a pedestrian, the bonnet leading
edge height plays an important role in the collision kinematics, by stopping contacting
the pedestrian in the femur or hip, and decreasing the wrap around distance (WAD).
The wrap around distance is defined as the distance from the ground to the head impact
location point on the vehicle. In the case of the cyclist, the dummy slides above the
Fig. 6. Wrap around distances in the two crash-tests
bonnet leading edge entering in contact directly with the vehicle’s bonnet, and its head
with the separation of the ceiling with the windshield. The two WAD are presented in
the Fig. 6.
It can be observed that the WAD for the cyclist is much higher than the pedestrian,
therefore the head impact point height is higher in the case of the cyclists leading to a
greater potential energy in the moment of collision. This is the second reason why the
pedestrian experience much lower head accelerations than the cyclist.
4 Assessment of the Head Injury Risk
An accurate way to assess the head injury risk is by using HIC value. The dummy’s
HIC values in the two crash-tests are the following:
Zt2
1
HIC151 ¼ max½ðt2 t1 Þð aðtÞdtÞ2:5 ¼ 1307 ð1Þ
t1;t2 t2 t2
t1
Zt2
1
HIC152 ¼ max½ðt2 t1 Þð aðtÞdtÞ2:5 ¼ 3510 ð2Þ
t1;t2 t2 t2
t1
where:
HIC15 1- HIC value for the car-to-pedestrian crash test
HIC152 – represents the HIC values for the car-to-bicycle crash-test
t2 – the end value of the time where the peak acceleration started to decrease
t1 – the initial value of the time where the acceleration started to increase
a(t) – the resulting head acceleration (g)
Δt = t2-t1 ≤ 0,015 (s)
As was expected, the head injury criteria for the car-to-bicycle impact is much
higher than the one for the pedestrian, and this is due to the higher resultant acceler-
ation. In the Fig. 7, the risk of brain injury for the two crash-tests can be assess.
By analyzing this diagram in function of the obtained HIC values from our
crash-tests, we can argue that the pedestrian risk of a brain injury is almost 25 % while
for the cyclist case, the risk of brain injury is 100 %.
810 B. Tolea et al.
Fig. 7. Risk of brain injury in function of HIC 15 [11]
5 Conclusions
After analyzing all the results, we can conclude that the risk of a severe head injury of a
pedestrian differs from the cyclist which is not wearing a helmet. The chance of a
cyclist of getting a severe brain injury at a collision velocity of 30 km/h is 100 % while
for the pedestrian the chance is reduced with 75 %.
Therefore, the most vulnerable road users can be considered the cyclists, which
may suffer serious injuries at even lower collision velocities than the pedestrian, and
their only way to protect themselves in case of a collision is the helmet, which may be
the only safety accessory that makes the difference between life and death.
References
1. Nie, J., Yang, J.: A study of bicyclist kinematics and injuries based on reconstruction of
passenger car–bicycle accident in China. Accid. Anal. Prev. 71, 50–59 (2014)
2. Vanparijs, J., et al.: Exposure measurement in bicycle safety analysis: a review of the
literature. Accid. Anal. Prev. 84, 9–19 (2015)
3. Yan, X.P., Ma, M., Huang, H., Abdel-Aty, M., Wu, C.: Motor vehicle-bicycle crashes in
Beijing: irregular maneuvers, crash patterns, and injury severity. Accid. Anal. Prev. 43(5),
1751–1758 (2011)
4. Mizuno, K., Kajzer, J.: Head Injuries in Vehicle-Pedestrian Impact. SAE Technical Paper
2000-01-0157 (2000)
5. Mizuno, Y., Ishikawa, H.: Summary of IHRA pedestrian safety WG activities – proposed
test methods to evaluate pedestrian protection afforded by passenger cars. In: Proceedings of
the 17th ESV Conference, Amsterdam, the Netherlands, Paper NO. 280 (2001)
6. Daniels, S., Brijs, T., Nuyts, E., Wets, G.: Injury crashes with bicyclists atroundabouts:
influence of some location characteristics and the design of cyclefacilities. J. Saf. Res. 40(2),
141–148 (2009)
7. Kopjar, B.: Population preventable fraction of bicycle related head injuries. Inj. Prev. 6(3),
235–238 (2000)
8. Thompson, D.C., Rivara, F.P., Thompson, R.: Helmets for preventing head and facial
injuries in bicyclists. Cochrane Database of Systematic Reviews 2, CD001855 (2000)
9. Richard, J.-B., Thélot, B., Beck, F.: Evolution of bicycle helmet use and its determinants in
France: 2000–2010. Accid. Anal. Prev. 60, 113–120 (2013)
10. Handbook of Accident Reconstruction – Heinz Burg, Andreas Moser (2013)
11. Mertz, H.J.: Injury risk assessments based on dummy responses. Accidental Injury, pp. 89–
102. Springer, New York (2002)
Determination of Kinematic and Dynamic
Behavior in the Driver’s Skull upon the Impact
with the Steering Wheel
Oana Victoria Oţăt(&), Nicolae Dumitru, Victor Oţăt,

and Ilie Dumitru
Faculty of Mechanical Engineering, University of Craiova, Craiova, Romania

otatoana@yahoo.com, nicolae_dtru@yahoo.com,
otatvictor@yahoo.com, dumitru_ilie@yahoo.com
Abstract. At European level, one of the main causes of death rates is generated
by severe injuries following traffic accidents. The most serious consequences of
traffic accidents result from lesions in the head.
In this study, we aim to determine the kinematic and dynamic behavior in the
driver’s skull after the occurrence of a frontal collision by computer modelling.
The frontal collision between the vehicle and the rigid wall was taken into
consideration, while observing the driver’s behavior, by means of the LS-Dyna
software. To increase the accuracy of the results, in the last phase of the study,
the dynamic modeling was carried out on a skull model, which has been scanned
previously. The skull motion law depends on the speed variation computed after
determining the kinematic and dynamic behavior of the Hybrid III 5th percentile
dummy. To simulate the contact between the head (skull) - wheel, it was
assumed that the passive safety system, namely, the airbag system is not
activated.
Keywords: Frontal collision Modeling Ls-Dyna Skull model Dynamic
1 Introduction
The European Union Member States report yearly traffic accidents that cause more than
37,000 casualties and about 1.2 million injured people, as well as property damage
estimated at around 145 billion Euros per year.
The main negative effect caused by road congestion is undoubtedly the significant
increase in the number of accidents, particularly those resulting in serious conse-
quences. The identification and the analysis of the joint factors leading to the occur-
rence of traffic accidents has become nowadays an extremely complex issue, given the
large number of disturbing factors and the disparity of traffic situations, but also due to
the behaviour and the different reactions manifested by each driver individually.
A theoretical overview of the most relevant research studies, as put forward in
[1–5], highlight current approaches to the investigation of head and cervical spine
lesions due to traffic accidents. Thus, the research studies in [2–4] aim at determining
the head area injury degree of the vehicle occupant. In [3] the setting out of the

Determination of Kinematic and Dynamic Behavior 813
investigation underpins the design of a 3D model of the skull - brain assembly. Mazdak
Ghajar et al. take over this research dimension and seek in [2] to determine the injury
degree in the head area by using a helmet as the safety measure during the impact.
Meijer et al. develop in [4] a human model with an active-controlled behavior con-
trolled involving the design of the neck muscles. The results obtained conclude that this
model facilitates a better anticipation of the human subjects’ kinematics in comparison
to the situations that do not involve the contraction of the neck muscles, hence a
completely passive behavior.
According to data provided by the Traffic Office of Dolj County Police Inspectorate
within the period of 01.01.2012-31.12.201, an average over 800 accidents were
recorded in Dolj county per year, of which about 260 accidents had serious
consequences.
Regarding the type of road users involved in the traffic accidents throughout the
county of Dolj, the statistical analysis we have carried out indicates that a considerable
frequency of fatalities were recorded among vehicle drivers. Thus, throughout the
period considered, namely the years 2012–2014 over 1,100 deaths annually were in the
category of vehicle drivers, see Fig. 1.
Fig. 1. Type of road traffic users between 2012–2014
2 The Virtual Prototyping of the Frontal Impact

for a Complex Dummy-Vehicle-Safety Systems Structure
2.1 Vehicle Law of Motion During the Impact
The complex dynamics analysis of the frontal collision involved the virtual prototyping
of a dynamic complex encompassing the dummy – the vehicle – and the safety systems.
Through this dynamic system virtual prototyping we aimed at analyzing the dynamic
response of the dummy, taking into account the normal position of the dummy.
For the virtual prototyping we have developed CAD models, kinematic and dynamic
models while using as input data the vehicle motion, geometrical-mass-and-inertia
characteristics, as well as the active loads and material properties.
The analysis of the driver’s kinematic and dynamic behavior set out by identifying
the collision phases, i.e.: the pre-crash, the crash (see Fig. 2) and the post-crash phase.
814 O.V. Oţăt et al.
We have modeled a frontal collision between a middle class sedan-type vehicle [6]
and a rigid wall with the direction of the main deformation force (PDoF) oriented after
the 12 o’clock direction and a 100 % degree of overlapping.
The vehicle law of motion during the pre-crash phase was determined via the
LS-DYNA software, as we have developed a simulation of a frontal collision using the
LS-DYNA software. The initial velocity of the vehicle at the pre-crash was set at
40 km/h.
Based on the analysis of the vehicle speed variation diagram (Fig. 3) we can
identify the phases of the accident. In pre-crash phase the vehicle is moving at the
constant velocity of 11.2 m/s. This phase occurs over the time interval of 0 s up to
approximately 0.02 s. At the moment of first contact between the vehicle and the rigid
wall we can observe a sharp decrease in vehicle velocity until the 0.09 s moment, when
the velocity reaches about –2 m/s. The sign indicates the travel direction of the vehicle;
after impact, in the recoil phase, the vehicle bounced, hence changing the travel
direction backwards.
The collision between the head of the dummy and the steering wheel occurs during
the crash phase, as shown in Fig. 2.
Fig. 2. The complex model: vehicle – dummy – crash phase
Within the context of a frontal collision, at a speed of 40 km/h, while considering

the passive safety systems, i.e. the fastening system of the dummy (the seat belt) active
and the airbag system deactivated, we could observe the occurrence of the collision
between the driver’s head and the vehicle steering wheel.
2.2 The Dummy-Vehicle Model

The analysis of the kinematic and dynamic behavior following a frontal collision upon
the vehicle driver triggered the virtual prototyping of the frontal collision for a complex
system encompassing vehicle - dummy - passive safety systems.
As for the virtual modeling to reproduce the influence of the impact phenomenon
upon the driver, as faithfully as possible, we have included in the virtual model all
elements from the interior of the vehicle, i.e. the dash board, the steering wheel, the
steering column and the pedals, as shown in Fig. 4.
Fig. 3. Vehicle velocity variation diagram in relation to time
seat belt
steering
wheel
daashboard
seat
vehicle
dummyy
Fig. 4. Dummy-vehicle structure
The positioning of the dummy in the driving position was done in accordance with
the real driver’s positioning mode, since a series of injuries occur due to the interaction
with the vehicle’s interior elements.
2.3 The Kinematic and the Dynamic Behavior in the Dummy Head Area
In order analyze the influence of frontal collision upon the driver’s head we have
obtained from the undertaken numerical modeling the variation diagrams in relation to
time for the linear velocity and linear acceleration, relative to the global system of axes
(Fig. 5).
To assess the injury degree in the head area we have used as reference element the
HIC factor - Head Injury Criteria (Fig. 6).
According to [7], Head Injury Criteria is defined as follows:
8 " #2:5 9
< 1 t2
Z
=
HIC ¼ ðt2 t1 Þ aðtÞdt max ð1Þ
: t2 t1 t1 ;
where:
– t1 and t2 are the initial and respectively the final time moments (expressed in
seconds) of the time interval when HIC reaches a maximum value
Fig. 5. The acceleration variation in the Fig. 6 Head injury criteria HIC 36
head of the dummy
– a(t) is the resulting acceleration (expressed in g) measured at the center of mass of

the head.
The HIC maximum values, compatible with the survival, in compliance with the
regulations imposed by the Federal Motor Vehicle Safety Standard (FMVSS) No. 208,
Occupant Crash Protection and CMVSS 208 (Canada Motor Vehicle Safety Regula-
tions), in case of a frontal collision, are as follows:
• HIC 15 (for a time interval of 15 ms) < 700;
• HIC 36 (for a time interval of 36 ms) < 1000.
As indicated by the recorded results – Fig. 6, we can notice that the head area
injury, indicated by the HIC factor value, exceeds the limits imposed by regulations in
force. In conclusion, following the impact between the head of the dummy and the
vehicle steering wheel, the lesions suffered in the head area are not compatible with the
driver’s survival.
3 Skull-Steering Wheel Impact Modeling
To model the frontal impact in the driver’s head area, in accordance with the typical
conditions of traffic accidents, we have investigated the impact phenomenon for a
virtual model consisting of a board and a skull, see Fig. 7.
In order to develop the CAD modeling of the skull we have applied the Marching
Cubes method, which reproduces with a high accuracy its anatomical structure. Fol-
lowing the mesh design, the skull-board model is composed of 8559 elements and 8146
nodes.
Aiming at obtaining as highly accurate results as possible, we have used in the skull
modeling material properties similar to the human skeleton [8], i.e.: mass density:
ρ = 2100 kg/m3; Young’s modulus: E = 6000 MPa; Poisson’s ratio: PR = 0.3
The properties of the material used to model the board are identical to those used
for the steering wheel in the previously investigated complex model, as shown in
Fig. 7.
The kinematic parameters defining the skull displacement were obtained via a
numerical simulation of the previously described model. Hence, we aimed to identify
Fig. 7 The skull model – the board and the material properties for the board
Fig. 8 The acceleration variation of the skull
the dynamic response of the driver in a frontal collision under critical situations that are
typical for road accidents incidence: airbag malfunction.
Based on the virtual model processing we sought to determine the variation of the
linear acceleration in the skull (Fig. 8) as well as the variation of the impact force in the
skull-board interaction, and the graphical representation of stresses and strains.
In compliance with the Directive No. 208 regarding the Federal Motor Vehicle
Safety Standard (FMVSS), Occupant Crash Protection, the maximum value of the
acceleration in the head area is of 80 [g]. Thus, we notice that due to an airbag
malfunction, the head acceleration value exceeds the limits in force.
Figure 9 illustrates the variation of the collision force registered in the head skull area
following the interaction with the board. The maximum value registered is of 1.4 kN,
at a 0.02 s time, a moment that coincides with the crush phase identified in the complete
model encompassing the dummy and the vehicle.
The Von Mises stresses distribution has been graphically represented in Fig. 10.
The maximum stress reaches the value of 50 N/mm2, being placed in the nasal bone
and maxilla area, though the impact occurred in the frontal bone area.
As indicated by the distribution diagram in Fig. 11 we can notice that the maximum
deformation occurs in the nasal bone and maxilla area as well.
Fig. 9 Impact force variation of the skull
Fig. 10 Effective stress Von Mises [N/mm2]
Fig. 11 Effective strain
4 Conclusion
Aiming at developing a model for the study of a complex frontal collision, we have
also considered the complex system vehicle-driver-seat belt. Based on our recorded
results, we could conclude that the accelerations display higher values in case of a
skull-board impact. Such phenomenon occurs mainly due to the lack of kinematic
strains and of damping elements imposed by the dummy modeling.
Under the circumstance, we could record that the evolution of the strains and the
deformations for our investigated case, i.e. a velocity of 40 km/h, does not register
critical values.
The analysis of our results highlights that the maximal values recorded in the
evolution of the kinematic and dynamic parameters are mainly due to a malfunction of
the passive safety system, i.e. the airbag system.
Thus, we consider that the processed virtual models put forward in the present
research study can be successfully applied to investigate particular issues regarding the
occurrence of severe injuries in the driver’s head area, in accordance with the typology
of road accidents, hence providing valuable outcomes in skull-implant testing.
References
1. Cronin, D.S.: Finite element modeling of potential cervical spine pain sources in neutral
position low speed rear impact. J. Mech. Behav. Biomed. Mater. 33, 55–66 (2014). Elsevier
2. Ghajari, M., Galvanetto, U., Iannucci, L., Willinger, R., Influence of the body on the response
of the helmeted head during impact. Int. J. Crashworthiness 16(3), 285–295 (2011). doi:10.
1080/13588265.2011.559798. Taylor & Francis
3. Hollowell, W.T., Gabler, H.C., Stucki, S.L., Summers, S., Hackney, J.R.: Updated review of
potential test procedures for FMVSS No. 208, Office of Vehicle Safety Research (1999)
4. Meijer, R., Hassel, E., Broos, J., Elrofai, H., Rooij, L., Hooijdonk, P.: Development of a
Multi‐Body Human Model that Predicts Active and Passive Human Behaviour, IRC-12-70,
IRCOBI Conference (2012)
5. Naif, A.-S., Neal-Sturgess, C.: An investigation into neck injuries in simulated frontal
impacts. In: Proceedings of the Institution of Mechanical Engineers, Part K: Journal of
Multi-body Dynamics, Sage Publications (2012)
6. http://www.ncac.gwu.edu/vml/models.html
7. Hans-Wolfgang, H.: Crash tests and the head injury criterion. Teach. Math. Appl. 17(4),
162–170 (1998)
8. Motherway, J.A., Verschueren, P., Van der Perre, G., et al.: The mechanical properties of
cranial bone: the effect of loading rate and cranial sampling position. J. Biomech. 42(13),
2129–2135 (2009). Elsevier
Pedestrian Dynamics During the Contact
Phase with the Vehicle
Adrian Șoica(&)
Transilvania University of Brașov, Brasov, Romania

a.soica@unitbv.ro
Abstract. The paper presents a simplified mathematical model for the contact
phase between the pedestrian and the vehicle during collision. After the primary
impact with the lower limbs, the movement of the human body is influenced by
factors such as vehicle speed during the impact, position / posture of the
pedestrian and vehicle type. The secondary impact often causes pedestrian
injury to the upper body as well as head trauma that is most often severe, even
fatal. The study on the dynamics of the pedestrian hit by the vehicle is useful in
order to understand the phenomena that occur; analysing the influence of dif-
ferent factors that determine the throw distance of the pedestrian is essential in
conducting road accidents investigations.
Keywords: Accident reconstruction Accident analysis Pedestrian impact

Contact phase
1 Introduction
The dynamics of traffic accidents still arousing interest because of the multitude of
situations in which they occur. Researches in the field of accident reconstruction were
made in [3–5, 11]. Researches regarding vehicle speed on impact with pedestrians were
made in [6–9]. Experimental tests, theoretical models and empirical determination of
the throw distance of pedestrians have been made by [2, 10, 14–16, 18, 20, 23].
Specialized software applications have been developed and validated by the numerous
research in [1, 12, 13, 17, 19]. Limpert in [11] divides the vehicle-pedestrian collision
into three phases, namely the impact phase, the flight phase and the sliding and/or
rolling phase. Aspects of the first phase have been closely analysed by Soica in [21],
and it has been divided into two sub-phases.
Five categories of vehicle - pedestrian collisions are described for instance in
[24, 25] and others. These are wrap, forward projection, fender vault, roof vault and
somersault collisions. A wrap collision is one where a significant area of the pedes-
trian’s body extends above the forward or front end of the vehicle and the pedestrian’s
body “wraps” up onto another area of the vehicle such as the bonnet, windscreen and/or
roof; The wrap contact is accompanied by a rotation and bending of the pedestrian’s
body and is often followed by a secondary impact of the head and/or upper torso. In a
fender vault, the pedestrian falls or rolls off to the side of the moving vehicle [2]. For
the collision types described above, the movement of the pedestrian depends on a

Pedestrian Dynamics During the Contact Phase with the Vehicle 821
number of factors such as the speed of the vehicle during the impact, the position /
posture of the pedestrian and vehicle type, as also analysed in [22].
2 General Simplified Model Illustrating the Movement

of a Hit Body
Figure 1 illustrates the process by which a solid body is hit in point O1 = O2 eccentric
to the axis O2y2. The Oxyz is a fixed system of axes, linked to the ground, O1x1y1z1
system is mobile, in translation movement towards the fixed system, and the system
O2x2y2z2 is connected to the body whose center of mass is located in the point Cg.
Point O1 = O2 is thus the instantaneous center of rotation and the body rotates around it
with angles ψ, θ and φ.
The Oxyz system is linked to the mobile system O1x1y1z1 through the position
vector r0 and to the O2x2y2z2 system through the position vector of the center of mass
rc. O1 is the instantaneous center of rotation of the pedestrian during the primary impact
with the vehicle. Through rotation around the axes of the mobile system the versors of
the axes of the mobile system O2x2y2z2 linked to the body are determined. For an 2D
movement on the Oxy plane, the unit versor of y2-axis is j2, and it has the following
position towards the Oxyz system:
8 9 8 9
< ux = < sinðWÞ sinðuÞ cosðWÞ sinðhÞ cosðuÞ =
! !
j2 ¼ ucg ¼ uy ¼ cosðhÞ cosðuÞ ð1Þ
: ; : ;
uz cosðWÞ sinðuÞ þ sinðWÞ sinðhÞ cosðuÞ
Fig. 1. General diagram of the pedestrian movement

822 A. Șoica
For the position vector of the impact point which is also the instantaneous center of
rotation (r0) in the begining, we may choose a law of variation provided the vehicle is
in motion during the impact or the position vector may be null if the vehicle is
stopped/braked during the impact.
3 Analysis of the Pedestrian Dynamics During the Contact

Phase with the Vehicle
During the contact between the vehicle and the pedestrian, especially at low speed
collision, the pedestrian is carried on the bonnet of the vehicle after the secondary
impact for a short time period, before the flying/slip phase. This is what Han called the
“carry” phase.
An example of this phase is where a pedestrian rises up onto the bonnet, perhaps
suffers a secondary collision, and remains on the bonnet without being thrown forward.
In cases such as these, the pedestrian typically slides from the bonnet either to the side
or forward, depending on the braking and steering of the vehicle. In this case there is a
“carry” phase between impact and flight [2].
The work mentioned above does not include in-depth studies, the distance travelled
by the pedestrian between the primary and the secondary impact is approximated
through a bilinear law, without making any other reference to the distance travelled in
the “carry” phase. Batista maintains in [1] that from the primary impact to the flying
phase the vehicle travels a distance of about 2 meters. In [21] was conducted further
studies and showed that the approximations mentioned in the previous paragraphs do
not bear enough resemblance to the reality. Overall, on calculating the pedestrian throw
distance the errors obtained were between 1 and 13 %, but the distance traveled by the
vehicle in contact with the pedestrian was greater than the distances mentioned by Han
and Batista in [2] and [1].
Like in [21], we consider the following:
• Vehicle speed before the first contact with the pedestrian “vo”.
• The duration of sub-phase 1.1 of the collision between the vehicle and the pedes-
trian “t1”, until the secondary impact is known.
• The average braking deceleration “a” is known. This is constant throughout the
whole sub-phase 1.1 and the sub-phase 1.2. of the impact.
• The speed of the pedestrian is zero and the impact occurs in OXY plan.
3.1 Sub-phase Between Primary Impact and Secondary Impact

According to Soica [21] the distance travelled by the vehicle during the sub-phase 1.1
is represented as follows:
0
2 0 2
vo v
m
1 þ mpv
þ a t1 1 þomp
mv
x1 ¼ ð2Þ
2a
Fig. 2. Vehicle - pedestrian primary and secondary impact
Fig. 3. Schematic of the primary impact
Where “a” stands for the vehicle deceleration in braking movement, which may be
determined for eg. from the braking traces, “v” stands for the vehicle speed at the
moment of secondary impact, and “t1” stands for the time until the secondary impact,
“mp” is the pedestrian mass, and “mv” is the vehicle mass. Throughout the sub-phase
1.1 the pedestrian is hit, accelerated and bent on vehicle bonnet [21]. Impact dynamics
is schematically illustrated in Figs. 2 and 3 and the equations of forces and moments
acting on the pedestrian are described below.
(

xc ¼ a ðc1 hc Þ cosðaÞ
ð3Þ
yc ¼ a ðc1 hc Þ sinðaÞ
8
<
xc ¼ aðc1 hc Þ cosðaÞ þ a2 ðc1 hc Þ sinðaÞ
ð4Þ
:
yc ¼ a ðc1 hc Þ sinðaÞ a2 ðc1 hc Þ cosðaÞ
824 A. Șoica
In order to simplify the calculations an equation system is built:

8 9 2 3 2 3

< xc >
> = ðc1 hc Þ cosðaÞ no ðc1 hc Þ sinðaÞ
4 5 4 5
yc ¼ ðc1 hc Þ sinðaÞ a þ ðc1 hc Þ cosðaÞ a ð5Þ
2
: >
> ;
a 1 0
which can be more simply written like:

no
fag ¼ ½ A a þ ½B a2 ð6Þ
where [A] is the matrix of the coefficients of angular acceleration of the pedestrian;
[B] is the matrix of the coefficients of the square of the angular velocity of the
pedestrian;
{a} is the vector of translation and rotation accelerations of the pedestrian’s body.
According to Fig. 4, following the primary impact between the vehicle and the
pedestrian, the equilibrium equations of forces on axes x, y and the moments equation
will be written in relation to point A.
The first contact is with the lower limbs, at the height of the bumper “h” with force “F”.
In a very short time, considered negligible, the pedestrian impacts the front edge of the
bonnet, which is situated at the height “hc”, with the lower limbs. This point noted with A
becomes instantaneous center of rotation of the pedestrian’s body from the primary
contact until the secondary contact when the head impacts the bonnet-windshield area.
2 3 8 9 8 9
m 0 0 > xc1 >
< = < F =
40
m 0 5 yc1 ¼ G ð7Þ
: >
> ; : F ðhc hÞ cosðaÞ þ G ðc1 hc Þ sinðaÞ ;
0 0 JpA a
Fig. 4. Schematic of forces acting upon the pedestrian during the primary impact
which may be simply written as:
½M fag ¼ fQg ð8Þ
where: [M] is the matrix of the mass and inertia moment of the pedestrian.
[Q] is the matrix of forces acting on the pedestrian;
{a} is the vector of the rotational and translational acceleration vector of the body.
In order to find the unknowns in the Eqs. (6) and (8) by multiplying the left with
[A]T we will obtain:
no
T T
½ A ½M ½ A a þ ½ A ½M ½B a2 ¼ fQext g ð9Þ
where:
fQext g ¼ ½ AT fQg ð10Þ
Relation (9) may be written as:

no
Q1
½A1 a þ ½B1 a2 ¼ ð11Þ
Q2
Relation (11) represents the simplified form of the differential equation in unknown
α = α(t). For a 2D motion in the XOY plane we can find the coordinates of the center of
mass of the pedestrian body, see Fig. 1.

xc ¼ xO O2 C sinðhÞ
ð12Þ
yc ¼ yO þ O2 C cosðhÞ
3.2 Phase 1.2 After Secondary Impact Until the Flying/Falling Phase
After the secondary impact the vehicle-pedestrian assembly is still in braking motion
and, therefore, the pedestrian falls off the vehicle at time “t1”, after he has travelled the
space “x1”, when his speed becomes greater than or equal to the vehicle speed.
According to [21] we have:
0
2
vo
m
1 þ mpv
þ a t 1 ð g2 1Þ
0
x1 ¼ ð13Þ
2a
where “η” is the impact coefficient.
During the phase of contact between the pedestrian with the vehicle travels the
space is:
826 A. Șoica
Fig. 5. Schematic of forces acting upon the pedestrian during the secondary impact
0
2 0 2 0
2
vo v vo
m
1 þ mpv
þ a t1 1 þomp m
1 þ mpv
þ a t 1 ð g2 1Þ
0 mv
Saut ¼ x1 þ x1 ¼ þ ð14Þ
2a 2a
After the secondary impact between the head of the pedestrian and the vehicle, we
write the equations of equilibrium of forces on axes x and y and the equations of
moments in relation to point B. After impact the pedestrian impacts the
bonnet-windscreen area with the head, in point B. This point becomes the instanta-
neous center of rotation for the body of the pedestrian from the secondary impact until
pedestrian falls off the vehicle, when the flying phase begins (Fig. 5).
The equations of movement may be written as follows:
2 3 8 9 8 9
m 0 0 < xc1 >
> = < 0 =

40 m 0 5 yc1 ¼ G ð15Þ
>
0 0 JpB : >; : G hp c1 cosð90 bÞ ;

b
Like in precedent case, relation (15) represents the form of the differential equation
in unknown β = β (t), we can find the coordinates of the center of mass of the pedestrian
body, to establish his trajectory.
4 Conclusions
The complex movement of the human body after a collision with a vehicle may be
analyzed in a simplified way through two-dimensional models. These models do not
indicate a high degree of precision, which calls for more accurate analyses, and,
therefore, specific applications such as multibody or FEA are required. Highlighting the
transport phase after secondary impact and determining the carry distance on vehicle
constitutes a step forward in analyzing and reconstructing traffic accidents with

pedestrians. A large number of experimental data can help to develop a law of
regression to establish the total carry distance of the pedestrian in the phase of contact
with the vehicle.
References
1. Batista, M.: A simple throw model for frontal vehicle pedestrian collisions. Promet Traffic
Transp. 20(6), 357–368 (2008)
2. Han, I., Brach, R.M.: Throw Model For Frontal Pedestrian Collisions. SAE Technical Paper
Series 2001-01-0898 (2001)
3. Collins, J.C., Moris, J.L.: Highway Collision Analysis. Thomas Publishing, New York
(1979)
4. Dettinger, J.: Methods of Improving the reconstruction of pedestrian accidents: development
differential, impact factor, longitudinal forward trajectory, position of glass splinters
(in German), Verkehrsunfall und Fahrzeugtechnik, December 1996, pp. 324–330 (1996)
5. Dettinger, J.: Methods of Improving the reconstruction of pedestrian accidents: development
differential, impact factor, longitudinal forward trajectory, position of glass splinters (in
German), Verkehrsunfall und Fahrzeugtechnik, January 1997, pp. 25–30 (two parts) (1997)
6. Eubanks, J.J., Haight, W.R.: Pedestrian Involved Traffic Collision Reconstruction Method-
ology. SAE Technical Paper Series, No. 921591
7. Evans, A.K., Smith, R.: Vehicle speed calculation from pedestrian throw distance. Proc. Inst.
Mech. Engrs. 213, 441–447 (1999). Part D
8. Hill, G.S.: Calculations of vehicle speed from pedestrian throw, Impact, pp 18–20 (1984)
9. Jun, Xu, Li, Yibing, Guangquan, Lu, Zhou, Wei: reconstruction model of vehicle impact
speed in pedestrian-vehicle accident. Inte. J. Impact Eng. 36, 783–788 (2009)
10. Kuhnel, A.: Der Fahrzeug Fussganger Unfall und seine Rekonstruktion, Dissertation,
TU-Berlin (1980)
11. Limpert, R.: Motor Vehicle Accident Reconstruction and Cause Analysis, 5th edn, pp. 539–
554. Lexis Publishing, Charlottesville (1999)
12. Moser, A., Hoschopf, H., Steffan, H., Kasanicky, G.: Validation of the PC-Crash Pedestrian
Model. SAE Technical Paper Series, 2000-01-0847 (2000)
13. Moser, A., Steffan, H., Kasanicky, G.: The Pedestrian Model in PC-Crash – The Introduction
of a Multi Body System and its Validation, SAE 1999-01-0445, Accident Reconstruction:
Technology and Animation IX (1999)
14. Wood, D., Simms, C., Walsh, D.G.: Vehicle–pedestrian collisions: validated models for
pedestrian impact and projection. Proc. IMechE Part D J. Automobile Eng. 219, 183–195
(2005)
15. Searle, J.A.: The physics of throw distance in accident reconstruction, SAE Technical Paper
Series, No. 930659 (1993)
16. Searle, J.A., Searle, A.: The trajectories of pedestrians, motorcycles, motorcyclists, etc.,
following a road accident, SAE Technical Paper Series, No. 831622 (1983)
17. PC-CRASH. A Simulation program for Vehicle Accidents, Technical Manual, Version 8.1
(2008)
18. Rau, H., Otte, D.: Car to pedestrian collisions with high speed impact. J. Verkehrsunfall
Fahrzeugtechnik 11, 294–300 (2001)
828 A. Șoica
19. Stevenson, T.J.: Simulation Of Vehicle-Pedestrian Interaction, thesis submitted in partial

fulfilment of the requirements for the Degree of Doctor of Philosophy in Engineering in the
University of Canterbury, University of Canterbury (2006)
20. Toor, A., Araszewski, M.: Theoretical Vs. Empirical Solutions For Vehicle/Pedestrian
Collisions. 2003-01-0883, Accident Reconstruction 2003, Society of Automotive Engineers,
Inc., Warrendale, PA, 2003, pp. 117–126 (2003)
21. Soica, A., Tarulescu, S.: Impact phase in frontal vehicle pedestrian collision. Int. J. Automot.
Technol. 17(3), 387–397 (2016). 10.1007/s12239−015−0040−y
22. Soica, A., Ispas, N.: The influence of the vehicle’s front-end profile upon adult pedestrian
kinematics and dynamics, multibody approach. Int. J. Veh. Saf. 6(2), 134–148 (2012).
doi:10.1504/IJVS.2012.049019
23. Togănel, G.: Cercetări privind influenta designului caroseriei asupra siguranței pasive a
automobilelor, doctoral thesis (2008)
24. Ravani, B., Brougham, D., Masson, R.T.: Pedestrian post impact kinematics and injury
pattern, SAE Paper, no. 811024 (1981)
25. Brooks, D., et al.: A comprehensive review of pedestrian impact reconstruction, SAE Paper,
no. 890859 (1989)
Example of a High-Speed, Side-Impact,
Car Crash Reconstruction Using a Planar
Multibody Software
Petru A. Simionescu1(&) and Ion Preda2

1
Texas A&M University, Corpus Christi, USA
Petru-Aurelian.Simionescu@tamucc.edu
2
pion@unitbv.ro
Abstract. The paper presents an example of Working Model 2D multibody

software use in the reconstruction of a high-speed side-impact accident of two
automobiles. The simulation (part of a contested judicial litigation) allowed to
clarify whether at the time of the accident, the impacting vehicle was imper-
missibly passing to the left of the continuous center line.
Keywords: Accident reconstruction Side impact Multibody simulation
1 Introduction
The purpose of traffic accident reconstruction is to estimate the velocity of the vehicle
(s) at the moment of collision. Other objective of the reconstruction is the determination
of the location of the vehicle(s) in the cases when evidence is insufficient to estimate
their positions and orientation at the moment of impact.
The velocity of a vehicle in an accident can be determined using possible tire marks
left on the road surface, or by measuring the amount of deformation caused during
impact [1, 2]. In mapping the accident site, debris scatter analysis and driver and
eye-witness statements are also utilized. Nowadays, substantial use is made of com-
puters in traffic accident analysis. Crash analysis computer programs serve investigators
in determining the relative velocity between the two vehicles (Δv) prior to impact - an
essential piece of information in most accidents. There are instances however when a
post-impact analysis is also necessary to clarify the circumstances of the accident.
The present paper considers such a scenario, i.e. of a court-contested case where the
impacting vehicle was allegedly passing to the right of the continuous center line. The
reconstruction is further complicated because part of the information about the accident
has been lost, in particular the measurements of the deformation of the two vehicles.
A two-car side-impact accident will be analyzed using Working Model 2D multibody
simulation software [3], for the purpose of establishing the location of the two vehicles
prior to impact. Additionally, internet-based digital maps have been utilized to accu-
rately reconstruct the accident scene based on the police-report measurements.

830 P.A. Simionescu and I. Preda
Fig. 1. Post-accident photograph showing the two vehicles involved i.e. a SKODA Octavia (the
bullet vehicle), and a BMW Touring (the target vehicle)
2 Accident Description
The vehicles involved in the accident were a SKODA Octavia (the impacting vehicle or
bullet vehicle), and a BMW Touring (the impacted vehicle or target vehicle) – Fig. 1.
The accident occurred at the intersection of a four-lane road with a secondary road
(see Fig. 2). The BMW was entering the main road from South-East and was struck in
the middle of its left side, severely injuring the driver. The accident happened during
daylight, in heavy traffic. The road surface was in favorable conditions with no pre-
cipitation. There was only one short, inconclusive tire mark left on the road (Fig. 2).
Fig. 2. Accident scene showing in blue the intended direction of travel of the BMW, and in red
(the arrow) the alleged travel direction of the Skoda. The only tire mark identified at the scene has
been marked on the figure with a red plus sign
Example of a High-Speed, Side-Impact, Car Crash Reconstruction 831
Fig. 3. Intersection where the accident occurred, viewed in the intended direction of travel of the
BMW
The debris produced during impact yielded little information to the police crew that
mapped the accident scene.
What is contested about this accident were the pre-impact speed of the SKODA,
and whether it was passing to the left of the median at the moment of collision.
According to the statements of its driver, the SKODA was traveling on the second lane
at about 90 km/h, he was caught by surprise by the appearance of the BMW, and had
no time to apply the brakes to avoid a full impact. By contrast, the driver of the BMW
claims that at the moment of the collision the SKODA was passing left of the median
visible in Figs. 2 and 3.
3 Computer Simulation
There is a number of computer programs available for car-crash reconstruction that can
simulate either 2D or 3D road accident events. Examples are CRASH/EDCRASH,
SMAC/EDSMAC, WinSMASH, PAM-CRASH, PC-CRASH, LS-DYNA etc. [4–11],
or in-house software as described in [12–15]. Working Model 2D (WM 2D) considered
here is a general purpose planar multibody simulation software, able to perform
dynamic analysis of constrained bodies with contact and friction. Its capabilities can be
enhanced by the use of a native formula language, and by interfacing it with Microsoft
Office Excel and MATLAB.
Part of reconstructing the accident under consideration, a scale drawing of the scene
has been generated within AutoCAD, based on the police report data, and using an
aerial view of the intersection as background image [16]. A layout of the scene has then
been exported to WM 2D via the DXF format. Secondly, top views of two vehicles
(see Appendices 1 and 2) were drawn at scale in AutoCAD, and were then exported to
WM 2D as polygonal bodies. The mass of the SKODA was assumed to be 1495 kg,
and the mass of the BMW was assumed to be 1615 kg. Lacking precise data, the
centroidal moment of inertia of the two vehicles was considered to be the same and
equal to 2750 kgm2. No adjustment of the centers of mass of the two vehicles has been
performed beyond the default locations assigned by the software to the bodies repre-
senting the two vehicles. The common restitution coefficient was set equal to 0.22,
which according to references [17, 18] and [19] is a probable number in case of
high-speed collisions of automobiles. Also, as defined in WM 2D, it was adopted a
friction coefficient between the two vehicles equal to 0.28, which is the ratio between
the tangential and normal impact forces assigned by the software [20].
In addition, the effect of the drag between the wheel tires and the road surface has
been modeled using four conditional forces, programed using WM 2D formula lan-
guage to act only after the impact between the two vehicles has occurred (Fig. 4).
These forces (acting on the wheels) were set each equal to a quarter of the weight of the
respective vehicle, times the friction coefficient between the road and the tire (also
known as adhesion coefficient) and assumed to be equal to 0.7. According to his own
statements, the driver of the SKODA did not press the brake pedal. However, an
equivalent transmission drag force has been assumed in case of the impacting vehicle
and was added to the simulation (see Fig. 4-b).
Fig. 4. Working Model 2D simulation showing (a) the position of the two vehicles right before
impact, and (b) accumulated frames of the simulation that matches closely the resting position of
the bullet vehicle.
4 Results
The WM 2D model generated as explained above, fitted with control text boxes for a
faster editing of the simulation parameters, has been run multiple times. The simulation
for which the final position of the impacting vehicle matches closely that in the accident
scene is provided in Fig. 4. Due to the limitations of the software, the effect of the road
ditch which stopped the side slipping of the BMW has not been included in this
simulation.
None of the simulations with the SKODA moving on the second lane in its
direction of travel resulted in the final resting position as documented in the accident
report. It is not excluded however that the driver of the SKODA acted instinctively
upon the steering wheel to avoid collision, something that he no longer remembers.
5 Conclusions
A road accident between two cars - a SKODA and BMW - that resulted in the side
impact and the severely injuring of the BMW driver has been described. The litigation
required the dismissal of the claim that the SKODA, the impacting vehicle, was
traveling along the second lane, while in actuality it was passing outside the median.
The accident has been reconstructed using Working Model 2D motion simulation
software, with additional information about the accident scene being extracted from
Internet-based digital maps.
APPENDIX 1: Geometry of the 2004 SKODA Octavia

APPENDIX 2: Geometry of the 2004 BMW Series 3 Touring
References
1. Brach, R.M., Brach, R.M.: Vehicle Accident Analysis and Reconstruction Methods. SAE
International, Warrendale (2011)
2. Struble, D.E.: Automotive Accident Reconstruction: Practices and Principles. CRC Press,
Boca Raton (2013)
3. Working Model 2D, Design Simulation Inc., Canton, MI (2016). www.design-simulation.
com
4. McHenry, R.R., McHenry, G.: Effects of Restitution in the Application of Crash
Coefficients. SAE Technical Paper, 970960 (1997)
5. McHenry, R., McHenry, B.: Crash damage analysis. In: 2014 NAPARS Conference, Sept.
8-12, 2014, Portland, Maine, USA
6. Day, T.D.: An overview of the EDSMAC4 collision simulation model. SAE Technical Paper
1999-01-0102 (1999)
7. Solanki, K., Oglesby, D.L., Burton, C.L., Fang, H., Horstemeyer, M.F.: Crashworthiness
simulations comparing PAM-CRASH and LS-DYNA. SAE Technical Paper, 2004-01-1174
(2004)
8. Schran, R.: Accident Analysis and Evaluation of PC-Crash. Technical Report, Chalmers
University of Technology, Department Machine and Vehicle Systems, Sweden (2005)
9. Niehoff, P., Gabler, H.C.: The accuracy of WinSmash Delta-V estimates: the influence of
vehicle type, stiffness, and impact mode. Ann. Proc. Assoc. Adv. Automot. Med. 50, 73–89
(2006)
10. Sharma, D., Stern, S., Brophy, J., Choi, E.: An Overview of NHTSA’s crash reconstruction
software winsmash. In: Proceedings of the 20th International Technical Conference on
Enhanced Safety of Vehicles, Paper No. 07-0211 (2007)
11. Johnson, N., Gabler, H.C.: Evaluation of WinSmash accuracy in NHTSA side crash test
reconstructions. Traffic Inj. Prev. 13(1), 72–80 (2012)
12. Taoka, G.T., Reizes, H., Reizes, A., Kovel, D.: Use of a microcomputer in traffic accident
reconstruction. Inst. Transp. Eng. J. 59(11), 27–31 (1989)
13. Lozano, J.A., Vera, C., Félez, J.: A computational dynamical model for traffic accident
reconstruction. Int. J. Veh. Des. 19(2), 213–227 (1998)
14. Moura de Carvalho, F.V., Speranza Neto, M.: Flexible vehicle model for planar collisions
analysis. In: Proceedings of 18th COBEM International Congress of Mechanical Engineer-
ing, Nov. 6-11, 2005, Ouro Preto, Minas Gerais, Brasil
15. Preda, I., Ciolan, G., Covaciu, D., Seitz, N., Dima, D.S.: Contributions to the study of
vehicle impact and post-impact dynamics. In: Proceedings of the 9th CAR Conference, Nov.
2–4, 2005, Pitesti, Romania, Paper 1082
16. Simionescu, P.A.: Computer Aided Graphing and Simulation Tools for AutoCAD Users.
CRC Press, Boca Raton (2014)
17. Brach, R., Brach, M.: A review of Impact Models for Vehicle Collision. SAE Technical
Paper, 870048 (1987)
18. Ishikawa, H.: Impact Center and Restitution Coefficients for Accident Reconstruction. SAE
Technical Paper, 940564 (1994)
19. Rose, N.A., Fenton, S.J., Beauchamp, G.: Restitution Modeling for Crash Analysis: Theory
and Validation. SAE Technical Paper, 2006-01-0908 (2006)
20. Marine, M.C.: On the Concept of Inter-Vehicle Friction and its Application in Automobile
Accident Reconstruction. SAE Technical Paper 2007-01-0744 (2007)
Is the Virtual Homologation for Pedestrian
Protection Viable?
Silvia Santos(&), Carolina Álvarez-Caldas, and José L. San Román
Mechanical Engineering Department,

Carlos III University of Madrid, Madrid, Spain
{ssantos,calvarez,jlsanro}@ing.uc3m.es
Abstract. One out of five deceased in traffic accidents is a pedestrian. In

addition, pedestrians represent the 20 % of the hospitalized injured people. The
deadliness rate of a pedestrian crash is significantly greater than for the rest of
accidents. Thus, pedestrian crash is one of the more lethal traffic accidents and,
consequently, pedestrians are the most vulnerable road users. Vehicle’s design
can influence immensely in the risk of seriousness of the accident. Regulations
are the legal instruments in order to establish if a vehicle achieves the minimum
safety requirements. Nevertheless, homologation implies costly and destructive
tests. This problem could be solved by simulation techniques. Analyzing the
viability of a virtual homologation is the main goal of this article. After studying
pedestrian crash biomechanics, virtual tests will be performed using Finite
Element software (Ls-Dyna) to assess the influence of the design of vehicle and
the effect of a safety system (active bonnet). Comparison between virtual tests
results and real tests allows deducing if the virtual homologation for pedestrian
protection is viable.
Keywords: Pedestrian crash Homologation Virtual test Road safety
1 Pedestrian Safety Systems
The 22 % of global deaths from traffic accidents are pedestrians, so-called vulnerable
road users together with cyclists and motorcyclists. The pedestrian is the weakest
element of the road because above 55 km/h, 95 % of pedestrian crashes are fatal.
Among pedestrian passive safety systems, airbags and active bonnet highlight. With the
airbags displayed from the top of the hood on the windshield it is intended to prevent
the pedestrian’s head striking the windshield glass, and by means of the airbags
deployed on the front of vehicle to reduce the force of impact of the bumper against
lower limbs. Also there are actuators that lift the hood at its rear edge to increase the
distance between hood and engine components in order to the passerby’s head does not
hit extremely hard elements and decrease the severity of brain injury and thus the risk
of death. For its part, among the ADAS (Advanced Driver Assistance Systems), which
keep a thin line with active safety, image recognition highlights allowing some systems
recognize autonomously a pedestrian between the different obstacles that precede the
vehicle and calculate by the electronic control unit if there is danger of collision and the

Is the Virtual Homologation for Pedestrian Protection Viable? 837
distance and time to prevent it. These systems can automatically brake the vehicle and
even act on the power steering to avoid pedestrians. Thermal or infrared cameras
improve these systems allowing night vision.
2 Pedestrian Crash Dynamics
The main purpose of this article is to deduce if virtual homologation for pedestrian
protection is viable. To this end, it is essential, before performing the virtual test, to
understand the biomechanics of a pedestrian crash in order to interpret the results
properly. With this purpose, dynamics and injury patterns of a pedestrian crash are
explained in this and the following chapters. The sequence of events that happens in a
pedestrian crash, in most cases, follows a pattern that only in unusual circumstances
will be modified significantly. Depending on the point of impact between the vehicle
and the pedestrian and, specifically, the position of the center of gravity of the passerby
respect to the bumper and the prior edge of hood, we can distinguish five basic
typologies of pedestrian crashes [1, 2]:
• Wrap projection. The torso and head bend over the vehicle by contacting the hood
and sliding over it. When vehicle decelerates, the body is separated from it and is
thrown down to the floor. Usually it occurs at about 30 km/h, while the contact
head-windshield is commonly seen over 40 km/h.
• Forward projection. It is typical when it comes to a short pedestrian or child who is
hit by a passenger car, or when an adult pedestrian collides with a vehicle type SUV
or industrial, that is, when the center of gravity impacts at the height or below the
top of the car front. Habitually it occurs at about 20 km/h.
• Fender vault. It normally takes place when pedestrian is hit near the front corner of
vehicle, in such a way that passerby is moved outwardly by the sideways of the
hood. Usually it occurs around 40 km/h.
• Roof vault. It is characteristic when the pedestrian’s center of gravity remains high
relative to the hood, so that he is turned in the air as a result of the impact speed
and/or vehicle design. It often occurs at high speed, lower than 60 km/h, and not
usually to less than 30 km/h.
• Somersault. It requires high-speed impact (50–60 km/h) so that energy is such that
it causes the pedestrian spins around him in the air before falling to the ground,
usually in front of the vehicle (Fig. 1).
Fig. 1. Pedestrian crash typologies: wrap projection, forward projection, fender vault, roof vault
and somersault [3].
838 S. Santos et al.
3 Pedestrian Injury Biomechanics
Among influential factors in the distribution of injuries spotlight impact velocity,

vehicle type, stiffness and frontal geometry, pedestrian’s age and size and his position
with respect to the front of the vehicle at the impact moment. The most frequent
injuries due to pedestrian crashes are located in the head and lower extremities. For this
reason, homologation requirements for pedestrian protection are focused on these body
areas. Those placed in head are the fatal injuries more frequent. Not only are important
cranial fractures, but also the acceleration experienced by the brain mass. The Head
Injury Criteria (HIC) measures the risk of suffering brain injury. This criterion does not
take into account the influence of rotational acceleration of the head nor any effect on
the location of impact on the head. But even so, it is used universally in collisions
research. Regulation assigns a maximum value of 1000 corresponding to a 16 % risk of
brain injury with AIS (Abbreviated Injury Scale) equal or higher than 4, equivalent to a
serious injury. There are other criteria for assessing pedestrian injuries such as Neck
Injury Criteria (NIC), or Thoracic Trauma Index (TTI), but these areas are not eval-
uated in homologation. Other of most affected areas is pelvis, in which the most
common injuries are caused by external rotation, lateral compression and shear. The
impact with the vehicle front and subsequent acceleration transmitted to the lower
extremities result in injury mechanisms, emphasizing the lateral shear and lateral
bending of lower limbs. The most common lower extremities injuries are long bone
fractures (femur, tibia and fibula), knee injuries and sprains or fractures of ankle or foot.
Injury thresholds in these body areas are evaluated in homologation [4].
4 Pedestrian Protection Homologation Tests
Regulation imposes and harmonizes the minimum requirements that must meet vehi-
cles to ensure the safety of both occupants and pedestrians. Investigations conducted by
the European Enhanced Vehicle-Safety Committee (EEVC) led to the drafting of
directives and regulations concerning pedestrian protection systems homologation.
Unlike legislation, which only uses impactors simulating the legs and head in physical
testing, the European New Car Assessment Programme (Euro NCAP) performs tests
with anthropomorphic dummies. The disadvantage of the latter is the need for more
infrastructure, greater complexity in data collection, and not knowing the exact point
where the head hits the vehicle. The European Regulation (EC) 78/2009 collects the
homologation requirements for vehicles in protection to pedestrians and other vul-
nerable road users [5]. Specifically in its annex I are the maximum allowable limits
(HIC factor, displacement and bending angle of the knee, tibia acceleration, impact
forces, etc.) that can be achieved in the trials. The implementing rules of tests are
described in Annex I of European Regulation (EC) 631/2009 [6]. These approval tests
are perfectly summarized in Fig. 2, and they consist of the throw of a lower legform
impactor to bumper and upper legform impactor against the prior edge hood, both trials
at 40 km/h, a child/small adult headform towards the top of the hood and adult
headform impactor against the windshield at 35 km/h. In this article, impactors of
Fig. 2. Pedestrian protection homologation tests.
homologation are replaced by a dummy. This is because real tests with which virtual
tests results will be compared are based on dummies.
5 Virtual Test of a Pedestrian Crash
Homologation tests are, in most cases, destructive, besides too expensive both in time
and economically. It should be noted then the power of computers tools in the industry.
Nowadays, it is unthinkable to design a vehicle without the presence of virtual testing.
These virtual labs are becoming increasingly important both from lawmakers and car
manufacturers. This type of tests can be a future alternative to actual tests in the
approval of vehicles. The IMVITER European project (Implementation of virtual
testing in safety regulations), which intended to settle virtual testing in the field of
safety legislation, concluded that although the simulation technique is not able to
replace the physical tests to date, virtual testing provide clear benefits in cases where
repetitive tests are carried out.
In order to establish a first hypothesis about the reliability of virtual testing, it is
decided to perform the simulation of pedestrian crashes by the finite element software
LS-DYNA using a Crash Test Dummy with percentile 50. The virtual tests results will
be compared with actual results of trials conducted by Euro NCAP, for the purpose of
checking whether what is observed in the simulations corresponds to what happens in
reality. For this, two car models completely opposite concerning pedestrian protection
will be used; the sedan Citroën C6, equipped with one of the more notable pedestrian
safety systems, the active bonnet and was also a pioneer in obtaining four stars in
pedestrian protection in Euro NCAP, and, on the other hand, the SUV Mercedes
Benz ML with only one star. The particularity of this second model is a defense located
on the frontal area of car which in case of collision with another vehicle is beneficial
but if it does against a person injuries increase significantly [7]. On modeling of the
vehicle front it has been used Solid Edge CAD software. The rear of the car and wheels
have been ignored in the model since they are not significant in the study of a
pedestrian crash. This simplification reduces the complexity and time simulation.
Furthermore, the visualization of car deformation caused by the dummy is omitted,
not being useful information in this study. The inner structure engine can be critical in
the injury result of pedestrian. Thus, in order to supply the lack of internal structure
definition, the mechanical strength of the car front was increased.
Virtual tests are performed to study the influence of the safety of these vehicles.
They consist of a frontal pedestrian crash caused by the Citroën C6 with and without
active bonnet at 40 km/h and, on the other side, a frontal pedestrian crash caused by
Mercedes Benz ML with and without its front defense at 40 km/h. The reason why this
value of speed is because the percentage of pedestrian crashes increases considerably in
urban roads, where 48 % of deceased in traffic accidents are due to a pedestrian crash.
In these roads the limit is usually 50 km/h. A close and lower value is selected in order
to take into account the possible previous braking conducted by the driver (Fig. 3).
6 Results
In the simulation of pedestrian crash caused by the Citroën C6 without active bonnet, it
is observed a roof vault post-impact trajectory, as a result of the high impact speed and
the high position of the pedestrian’s center of gravity with regard to the impact point
with the vehicle front. Some studies affirm that this typology is not observed below
32 km/h or above 60 km/h, so apparently this simulation may reflect what would
happen in reality. The HIC criterion shows the severity head injury by measuring the
acceleration in the nodes of the virtual dummy’s head. In this case, it reaches a
maximum value of 4828. However, when in the test the active bonnet is applied, the
HIC decreased approximately 45 % up to a value of 2136. It is proved in this way what
would be expected in reality happens, i.e. the active bonnet reduces significantly the
severity of brain injury, preventing the pedestrian’s head impacts with extremely hard
areas which stay under the hood, such as engine, increasing the separation distance
between these hard elements and the hood veneer. That is why Euro NCAP awards
such a high score to this car model. Results from neck, chest and hip severity injury
pedestrian have been also obtained. They are reflected in Table 1 by the NIC (Neck
Injury Criteria), CSI (Chest Severity Index) and LEC (Lower Ext. Criteria) factors
respectively (Fig. 4).
On the other hand, in the simulation of pedestrian crash by Mercedes Benz ML
without front defense, it is observed a wrap projection post-impact trajectory. In order
to study the severity of this accident it is used again the HIC criterion, reaching a value
of 2836. When the front defense is adding in the test, it is surprising that the HIC value
is almost constant, so it is decided to analyze the global injury in the pedestrian by
means of the Injury Severity Score (ISS) which takes into account the three highest
Fig. 3. Real and virtual pedestrian and vehicle models for simulation.
Fig. 4. Simulation sequence of frontal pedestrian crash by Citroën C6 at 40 km/h with active
bonnet and HIC (resultant acceleration) results.
Fig. 5. Simulation sequence of frontal pedestrian crash by Mercedes Benz ML at 40 km/h

without front defense and HIC (resultant acceleration) results.
values achieved among the various factors that evaluate the different parts of the body.
These three factors considered by the ISS, in this simulation, are the HIC (head), the
CSI (chest) and the LEC (hip). In Table 1, it can be seen that by including the front
defense the chest injury is 4 times greater and the hip injury is 5 times greater. Thus,
this defense increases significantly the pedestrian injury severity and can be cause his
death due to internal bleeding and other injuries. For this reason, Euro NCAP penalizes
this car model in its pedestrian protection score (Fig. 5).
Table 1 shows the numerical results of the factors that assess the severity of the
injuries suffered by the virtual dummy in the head, neck, chest and hip. It displays a
comparison of the values obtained without safety systems in both models (the active
Table 1. Head (HIC), neck (NIC), chest (CSI) and hip injury simulation results with and
without safety systems.
Without safety systems HIC Neck Chest Hip (kN)
Citroën C6 4828 3,925 525,5 −7
Mercedes Benz ML 2836 5,55 2058 −21
With safety systems HIC Neck Chest Hip (kN)
Citroën C6 2136 – 362 −6,5
Mercedes Benz ML 2864 – 8231 −110
Reference values 700–1000 1,20 700–1000 −10,9
bonnet in Citroën C6 and a front defense in Mercedes Benz ML) versus those obtained
when these systems are included. Reference values based on data extracted from
cadavers, actual tests and traffic accident data, have been also include in order to reflect
the maximum tolerance injury for these body parts.
7 Conclusion
It exists lately a tendency in safety systems focused on pedestrians. Some systems, very
effective in the survival of the pedestrian, such as the active bonnet, begin to be
incorporated as standard on new vehicles. Nevertheless, homologation tests for these
systems are destructive and so greatly costly. Thus, finding a viable alternative to
physical tests would minimize enormously costs. Virtual labs could become in the best
option to achieve this aim. After analyzing simulations performed and comparing their
results with the score awarded by Euro NCAP, it is verified that finite element software
(specifically in this case, LS-DYNA) behaves as a tool able to reflect the actual
behavior of pedestrian protection systems. Undoubtedly, virtual representations leave
many factors without intervening together, while they do in real life. Nonetheless, there
is no doubt that virtual tests represent a trend or reliable quick overview of immediate
consequences of factors such as speed, front height, position or pedestrian safety
measures in car design without having to resort to costly and time-consuming full-scale
tests. Thus, they can reflect what factors are the most influential in pedestrian crash, and
lead the efforts in the development of safety measures or in the vehicle design to protect
increasingly in an effective way to pedestrians. In simulations of pedestrian crash
caused by passenger car Citroën C6, it has been quantified the importance of a sig-
nificant pedestrian safety system, the active bonnet. It has to be noted the reduction of
45 % in the HIC value corresponding to the risk of suffering a brain injury. These
values are consistent with real tests performed by Euro NCAP based on physical tests,
assigning a high score in pedestrian protection with four stars. There are also reductions
in injury severity of chest and hip but to a lesser extent by using the active bonnet.
Conversely, in simulations of pedestrian crash by SUV model Mercedes Benz, it is
finding out that risk of serious injury brain does not change when the front defense is
added to the vehicle. However, when making an overall assessment of the injuries
taking into account other areas as chest and hip, the severity increases significantly.
When front defense is added, chest injury severity is 4 times greater and hip values are
5 times greater. Hence, virtual tests have behaved as expected according to physical
tests performed by Euro NCAP, which emphasize that this model is extremely lacking
in pedestrian protection. This front defense is a protection system in a collision with
other vehicle, but if it impacts with a pedestrian increases greatly the risk of dying.
Although virtual tests conducted in this study have reached the same conclusion as
actual tests performed by Euro NCAP, it is considered that there is still a long way to
go in research of virtual laboratories and set up which simplifications and limitations
are acceptable to consider the results drawn from these virtual models as true repre-
sentative of reality. One of the most decisive factors to accept a virtual test is as reliable
validation and justification of the virtual model. However, in this research field there is
a need to collect enough test and actual accidents data in order to compare and verify
the accuracy and reliability of these virtual tests. This information is not generally
available or its access is extremely limited. Therefore, in this study we have not been
able to simulate the homologation tests by using legform and headform impactors,
since no actual data were available with which to compare the results of virtual tests.
Hence the decision to consult the physical tests provided publicly by Euro NCAP to
check the veracity of virtual labs and establish some first assumptions on the viability
of substituting the real approval tests with regard to pedestrian protection or at least to
reduce them in number by virtual tests. Virtual labs could reduce enormously both
temporal and economic cost of tests. Moreover, accepting as valid and reliable virtual
tests would allow vehicle manufacturers to reduce time to market of their models and
so increase their competitiveness.
As a final conclusion, to date, it is considered that the physical tests cannot be
replaced completely by virtual tests, but virtual labs can provide great benefits in
repetitive tests, such as in the case of homologation tests of pedestrian protection
systems in which the impactor must be thrown several times in the same conditions.
Although virtual tests are able to simulate the reality considering a large number of
variables and, therefore, each time with minor differences between real and virtual
model, it will be always necessary to have real data in order to validate the models. In
addition to this, it is necessary to perform a minimum number of virtual tests in such a
way that the level of associated uncertainty is acceptable. Results also conclude that
homologation tests for pedestrian protection require major revision and must add body
areas as chest or neck, since injury results obtained in this research exceed significantly
the injury thresholds assigned to these areas.
References
1. San Román, J.L., Campón, J.A., Díaz, V., Sanz, S.: Accident reconstruction: pedestrian crash.
J. Sociedad de Técnicos de Automoción (STA) 182, 9–19 (2007)
2. Condes, J., San Román, J.L.: Pedestrian crash tests. Description and simulation. Mechanical
Engineering Department, Carlos III University of Madrid, Spain (2009)
3. Mingueza, R.: Pedestrian crashes in urban roads. Valencia Local Police, Spain (2012)
4. Campón, J.A.: Pedestrian crash reconstruction. Doctoral Thesis, Mechanical Engineering
Department, Carlos III University of Madrid, Spain (2015)
5. European Regulation (EC) 78/2009
6. European Regulation (EC) 631/2009
7. Santos, S.: Vehicle safety systems and pedestrian crash analysis. Mechanical Engineering
Department, Carlos III University of Madrid, Spain (2014)
Comparative Analysis of Kinematics
and Dynamics Parameters that Characterize
Vehicle-Pedestrian Collision
Alexandru Petrescu1,2(&) and Liviu Petrescu1

1
Asig Tehnic Expert SRL, Bucharest, Romania
expertpetrescu@yahoo.com
2
Politehnica University of Bucharest, Bucharest, Romania
Abstract. Reconstruction of kinematics and dynamics parameters that char-

acterize vehicle-pedestrian collision represents a matter of great interest, which
preoccupies more and more specialists from automotive and related industries.
The paper refers to the analytical methods, including the ones with numerical
solving that allow computer-based reconstitution of the vehicle-pedestrian
impact, being used specialized software PC-Crash.
It presents a comparative analysis of kinematics and dynamics parameters that
characterize car-to-pedestrian collisions caught on camera (filmed in real time)
and simulate using computerized collision reconstruction program PC-Crash,
that is the most commonly simulation software used in Europe.
The aim of this paper is to evaluate and compare analytical methods that
allow the reconstruction of real accidents involving pedestrians.
Keywords: Automotive Pedestrian Collision Accident reconstruction

Accident simulation
1 Introduction
Determining the impact speed of the vehicle involved in a vehicle-pedestrian

traffic accident is one of the main requirements of traffic accidents analysis and
reconstruction [11].
Since the ‘60 s, there were many experimental studies and research on
vehicle-pedestrian impact, conducted by Wood, Searle, Apple, Stcherbatcheff, Kuhnel,
Rau, Brach, Han, Schmidt, Batista, Fugger, Toor etc.
This paper examines the assessment of the impact speed of the vehicle in relation to
the pedestrian throw distance, using analytical methods and the computer-based
reconstruction of the vehicle-pedestrian collision with the help of the PC-Crash soft-
ware. For this purpose, data were used from a traffic accident which was filmed, in real
time, by a surveillance camera.

Comparative Analysis of Kinematics and Dynamics Parameters 845
2 Circumstances of the Traffic Accident
The accident occurred on Regina Elisabeta Ave., in Bucharest, and was filmed by a video
surveillance camera that records at a 25 fps frequency, with a frame size of 704 × 576
pixels. The camera was located on the opposite side of the street to the side walk where
the pedestrian had started to cross the road. The accident involved a BMW 525 vehicle
and a pedestrian, male, aged 23, with a height of 1,73 m and a weight of 73 kg.
Figure 1 shows the area where the accident occurred, the direction of the vehicle
and that of the pedestrian, respectively.
Fig. 1. View of the area where the accident occurred
The pedestrian crossed the street running, behind the bus that had stopped at the bus
station, and he entered the lane of the BMW vehicle from the left side, as shown in
Fig. 2. The pedestrian was hit by the frontal left side of the vehicle and was thrown into
the air, so that he reached above the roof of the BMW vehicle, as shown in Fig. 3.
Fig. 2. The pedestrian enters the lane of the BMW vehicle from the left side.
846 A. Petrescu and L. Petrescu
Fig. 3. The pedestrian was hit by the frontal left side of the vehicle and was thrown into the air,
so that he reached above the roof of the BMW vehicle.
The damage caused to the BMW vehicle are shown in Fig. 4, which reads: A -
traces of deformation of the engine bonnet, on the left; B - frontal grids between
headlights, broken and detached from their supports; C - damage to the lower left side
of the windscreen; D - damaged left front headlight; E - broken support of the license
plate, on the left side.
The real-time recording of the traffic accident provided the opportunity to establish
the way it occurred, determining the time and place of the impact, as well as the
momentums and the final positions when both vehicle and pedestrian stopped on the
roadway, also allowing to ascertain the post-impact trajectory of the pedestrian and the
relative position between the pedestrian and the vehicle on initial impact. The total
duration of the accident was accurately determined at 2,9 s from the initial moment of
the impact until the pedestrian and the BMW vehicle stopped on the roadway.
Fig. 4. The damage caused to the BMW vehicle after the impact with pedestrian
3 Reconstruction of the Specific Parameters

of the Vehicle-Pedestrian Traffic Accident
3.1 Reconstruction of the Specific Parameters of the Vehicle-Pedestrian
Traffic Accident Using Analytical Methods with Numerical Solutions
In order to reconstruct the kinematic and dynamic parameters specific to this traffic
accident, it was used the comput-based collision reconstruction method, with the help
of the PC-Crash software. With PC-Crash can be reconstruct complex accidents, using
multibody dynamics formulations. This model allows simulating complex pedestrian
accidents [3].
The dimensional parameters, as a whole, of the road section where the accident
occurred and the positions where both the vehicle and the pedestrian stopped after the
impact were identified by the measurements taken by the police at the accident site.
A sketch of the accident site is shown in Fig. 5. It contains the measurements taken
by the police at the accident site, as well as the impact position and the final positions,
on the roadway, of the vehicle and of the pedestrian.
Fig. 5. The sketch of the accident site
The distance between the impact position and the final position of the vehicle
(relative to the centre of gravity of the vehicle) is of 24,20 m, and the distance between
the impact position and the final position of the pedestrian (relative to the centre of
gravity of the pedestrian) is of 26,26 m.
In the simulation generated using the computer-based reconstruction method, the
impact speed of the vehicle was of 69,40 km/h and the pedestrian speed was of 9 km/h,
given de fact that the vehicle and the pedestrian stopped, in the simulation, in the
positions recorded by the police in their on-site investigation; the initial positions and
trajectories of the vehicle and of the pedestrian are similar to those shown in the video
recording.
The dynamics of the pedestrian - vehicle impact, achieved using the
computer-generated reconstruction of the collision, provided by the PC-Crash software,
are shown in Fig. 6.
Fig. 6. The dynamics of the pedestrian - vehicle impact
Figure 7 shows the speed of the pedestrian, and Figs. 8 and 9 show the distances
relative to the axes, travelled by the pedestrian (by his centre of gravity).
Figure 10 shows the speed of the vehicle, and Fig. 11 shows the distance on the Ox
axis, travelled by the BMW vehicle.
Figure 12(a) shows the components of the vehicle’s accelerations/decelerations on
the axes of the Oxyz coordinate system, while Fig. 12(b) shows the axes components
of the external forces acting upon it.
Fig. 7. The speed of the pedestrian

Fig. 8. The distance on the Ox axis, travelled by the pedestrian (by his centre of gravity)
Fig. 9. The distances on the Oy and Oz axes, travelled by the pedestrian (by his centre of
gravity)
Fig. 10. The speed of the BMW vehicle
Fig. 11. The distance on the Ox axis, travelled by the BMW vehicle (by its centre of gravity)
The values of the external forces, shown in Fig. 12(b), are obtained by applying
Euler’s equations for a rigid body with a fixed point, i.e. the system of 2nd order
differential equations system for translations (with €x; €y; €z accelerations on the axes):
8 X 8 X
>
> m€x ¼ Fx >
> max ¼ Fx
< X < X
m€y ¼ Fy ) may ¼ Fy ð1Þ
>
> X >
> X
: :
m€z ¼ Fz maz ¼ Fz
a) b)
Fig. 12. The components of the vehicle’s accelerations/decelerations on the axes of the Oxyz
coordinate system (a) and the axes components of the external forces acting upon it (b)
3.2 Reconstruction of the Specific Parameters of the Vehicle-Pedestrian

Traffic Accident Using Analytical Methods
For the determination of certain specific parameters of traffic accidents involving
pedestrians, the literature includes a number of analytical methods with varying levels
of complexity.
Based on the theoretical approaches and on the experimental tests conducted using
crash test dummies that replicate the size and composition of the human body, various
models and equations (Kuhnel, Searle, Wood, Stcherbatcheff etc.) were suggested.
Most of these equations relate the speed of the vehicle at the time of the impact, v
[m/s] or V[km/h], to the pedestrian throw distance, Sp[m], for the situations where the
pedestrian is hit by the front of the vehicle, as it happened in the traffic accident that is
being analysed.
The paper shall use equations determined both analytically (Searle, Wood, Rich)
and based on the crash testing (Kuhnel, Rau, Toor, Stcherbatcheff, Fugger, Schmidt).
(1) Kuhnel suggests the following equation to calculate impact speed of the
vehicle, V, depending on the pedestrian throw distance [9]:
V2
Sp ¼ 0; 0178aV þ 0; 0271 ð2Þ
a
where: - a [m/s2] is the vehicle’s decelerations, a = 7,85 m/s2
- Sp [m] is the pedestrian throw distance, Sp = 26,26 m (Fig. 5)
(2) Rau suggests the following equation to calculate impact speed of the vehicle, V,
depending on the pedestrian throw distance [4]:
Sp ¼ 0; 0052V 2 þ 0; 0783V ð3Þ
(3) Toor suggests the following equation to calculate impact speed of the vehicle,
V, depending on the pedestrian throw distance [10]:
V ¼ 8; 25S0;61
p ð4Þ
(4) Stcherbatcheff suggests the following equation to calculate impact speed of the
v2
Sp ¼ þ 0; 03av ð5Þ
2a
(5) Fugger suggests the following equation to calculate impact speed of the
V ¼ 8; 3604S0;6046
p ð6Þ
(6) Searle suggests the following equation to calculate impact speed of the vehicle,
2lgSp
V ¼ 3; 6 ð7Þ
cos a þ l sin a
where: - μ is the friction coefficient between pedestrian and road surface, μ = 0,7
- α is the pedestrian throw angle.
As the value of this angle is usually unknown, by minimising and maximising the
throw angle in Eq. (7), it can obtain the minimum and maximum impact speed values:
sffiffiffiffiffiffiffiffiffiffiffiffiffi
2lgSp
Vmin ¼ 3; 6 ð8Þ
1 þ l2
Vmax ¼ 3; 6 2lgSp ð9Þ
(7) Schmidt suggests the following equation to calculate impact speed of the
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
V ¼ 3; 6 l2 hcgp þ 2lgSp lhcgp ð10Þ
(8) Wood suggests the following equation to calculate impact speed of the vehicle,
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2ffi
2lg Sp lhcgp ma þ mp
V ¼ 3; 6 ð11Þ
m2a
where: - ma [kg] is the total mass of the vehicle (including the mass of the vehicle,
occupants and luggage), ma = 1630 kg
- mp [kg] is the mass of the pedestrian, mp = 73 kg.
(9) Rich suggests the following equation to calculate impact speed of the vehicle,
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

da2
V ¼ 3; 6 g þ 2ldg ð12Þ
hcgp þ da
where: - da [m] is the pedestrian throw distance (the flight phase), da = 20,25 m
- dg is the pedestrian sliding distance (the phase of pedestrian motion along the
ground), dg = 6,01 m (Fig. 8).
It is worth noting that, on the basis of the available data, it is possible to use a
particular mathematical model of the vehicle-pedestrian impact to determine other
specific parameters.
Figure 13 shows the Han&Brach model diagram, where V is the impact speed of
the vehicle (for the analysed accident, V = 69,4 km/h), and Vp0 is the pedestrian throw
speed. The diagram also contains the positions of the vehicle and of the pedestrian, as
well as the throw angle of the pedestrian, θ. The distance travelled by the vehicle on the
Ox axis, Sv = dxa = 24,2 m (Fig. 11), and the distance travelled by the pedestrian on
the Ox axis, Sp = dxp = 26,26 m (Fig. 8), are also known. The pedestrian throw
Fig. 13. Sketch of wrap collision sequence (the Han&Brach model)

distance (the flight phase), da = 20,25 m (Fig. 8), are also known, as well as the
pedestrian sliding distance (the phase of pedestrian motion along the ground), dg =
6,01 m (Sp = da + dg).
Based on the mathematical model (the Han&Brach model [2]) and on the known
data, it can be determine:
– the impact speed of the vehicle, V [km/h]:
ma þ mp pffiffiffiffiffiffiffiffiffiffiffiffi
V ¼ 3; 6 2lgSv ð13Þ
ma
– the pedestrian throw speed, Vp0 [km/h]:

ma
Vp0 ¼ V ð14Þ
ma þ mp
1630
Vp0 ¼ 68; 57 ¼ 65; 63 km/h ð15Þ
1630 þ 73
with a difference of 0,6 % compared to the value of the pedestrian throw speed of
65,2 km/h, determined using accident reconstruction program PC-Crash.
The results obtained using the above-mentioned analytical methods for the analysed
traffic accident are summarised in Table 1.
As it appears from the foregoing, the Kuhnel equation generates, for this accident,
the lowest difference, i.e. 0,1 %, and as such it gives the best prediction of the value of
the vehicle’s impact speed. The values that are closer to the vehicle’s impact speed are
Table 1. The summary showing, for the analysed traffic accident, the value of the impact speed
of the vehicle (V) determined using the above computing models
Method name Throw Impact Difference Obs.
distance speed of (%)
(m) the
vehicle
(km/h)
PC-Crash 26,26 69,40 – The difference is determined
Kuhnel 26,26 69,30 −0,1 from the result obtained by
Rau 26,26 63,93 −7,8 comput-based collision
Toor 26,26 60,56 −12,7 reconstruction (PC-Crash)
Stcherbatcheff 26,26 66,74 −3,8
Fugger 26,26 60,30 −13,1
Searle 26,26 [56,01– [−19,3– −1,4]
68,37]
Schmidt 26,26 68,34 −1,5
Wood 26,26 70,44 1,5
Rich 26,26 59,12 −14,8
Han&Brach 26,26 68,57 1,2
generated by Wood’s relation (a difference of 1,5 %), Schmidt’s relation (a difference

of 1,5 %) and by Searle’s relation, the maximum speed case (a difference of 1,4 %).
The mathematical model of the vehicle-pedestrian impact (the Han&Brach model),
which takes into account a number of parameters that influence the pedestrian thrown
trajectory, allows for a good prediction of the value of the vehicle’s impact speed. The
difference is of 1,2 % compared to the 69,4 km/h value which resulted from the use of
the PC-Crash software, which validates an impact speed of 69,4 km/h, as determined
for the vehicle.
4 Conclusions
The analytical relations take into account a small number of parameters influencing the
pedestrian’s throw trajectory. It can be concluded that neither of these relations may be
used as a general rule. Considering of these limitations requires using several analytical
relations, in order to be able to assess an impact speed of the vehicle within a range of
values. For a greater accuracy these analytical relations are used with other computing
models, such as the analytical methods with numerical solutions, of which this paper
has presented the computer-based method, using the PC-Crash software.
Therefore, compared to the analytical methods used to reconstruct the collision, the
computer-based methods, which take into account a significantly higher number of
input parameters, including for the multibody structure related to the pedestrian, are an
appropriate means of analysing traffic accidents involving pedestrians.
References
1. Fugger Jr., T.F., Randles, B.C., Wobrock, J.L., Eubanks, J.L.: Pedestrian Throw Kinematics
in Forward Projection Collisions. SAE Technical Paper (2002-01-0019)
2. Han, I., Brach, R.: Throw model for frontal pedestrian collisions. SAE Technical Paper
(2001-01-0898)
3. Moser, A., Steffan, H., Hoschopf, H., Kasanicky, G.: Validation of the PC-Crash Pedestrian
Model. SAE Technical Paper (2000-01-0847)
4. Rau, H., Otte, D., Schulz, B.: Pkw-Fußgangerkollisionen im hohen Geschwindigkeitsbereich
Ergebnisse von Dummyversuchen mit Kollisionsgeschwindigkeiten zwischen 70 and
90 km/h. Das Fachmagazin fur Verkehrsunfall und Fahrzeugtechnik 12, 341–350 (2000)
5. Schmidt, D.N., Nagel, D.A.: Pedestrian impact case study. In: Proceedings of 15th
Conference American Association for Automotive Medicine (1971)
6. Searle, J.A., Searle, A.: The trajectories of pedestrians, motorcycles, motorcyclists, etc.,
following a road accident. SAE Technical Paper, 831622 (1983)
7. Stcherbatcheff, G., Tarriere, C., Duclos, P., Fayon, A., Got, C., Patel, A.: Simulations of
Collisions between Pedestrian and Vehicles using Adult and Child Dummies. SAE Tehnical
Paper, 751167 (1975)
8. Stevenson, T.J.: Simulation of vehicle-pedestrian interaction. A thesis submitted for the
degree of Doctor of Philosophy in Engineering in the University of Canterbury, Canterbury
(2006)
9. Şoica, A., Lache, S.: Case study regarding the motor vehicle – pedestrian collision. WSEAS
Trans. Appl. Theor. Mech. 2(10) (2007). ISSN: 1991-8747
10. Toor, A., Araszewski, M.: Theoretical vs. empirical solutions for vehicle/pedestrian
collisions, ASME Paper (2003-01-0883)
11. Wood, D., Simms, C., Walsh, D.G.: Vehicle–pedestrian collisions: validated models for
pedestrian impact and projection. In: Proceedings of IMechE, vol. 219 Part D: Journal of
Automobile Engineering (2005)
The Importance of Consumed Energy
for Deformation in Study of Collision
Mihai Clinciu(&), Anghel Chiru, and Stefanita Ciunel

mihaiclinciu@yahoo.com, achiru@unitbv.ro
Abstract. This paper intends to be a study on the importance of crush energy

which determines the mechanism who produces the lateral collision. Using
Newton’s laws of mechanics, we can discuss about the conservation of energy
and linear momentum law which determine the impact force and crush.
Therefore, the framework and methodology for achievement and conducting
experimental tests and measuring chains data is established. In this context, we
have defined a series of testing structures to design a technical system for
conducting experimental tests which contain a various lists of measurements and
IT systems. Analytical techniques who determine the occupant and vehicle
behavior from this stage allow to determine the mechanism who produces the
damages and biomedical parameters, which are representative and correspond
real applications from the practice investigation of the traffic incidents too. The
results obtained from theoretical and experimental processing data are presented
as graphical crush energy - deformation. The information are extremely useful to
expert investigators who activates in the field of the traffic events investigation
and reconstruction.
Keywords: Side impact Crush energy Kinetic energy Work done

Deformation
1 Introduction
Vehicles in general and cars in particular have experienced a spectacular development

in a relatively short time being improved continuously, so leading to ingenious solu-
tions that provide a high reliability and good road safety.
Over time the number of traffic events increased, because of the increasing number
of cars and their performance on speed, road infrastructure development and because
the traffic intensification.
Traffic events continue to be the main cause of traffic deaths and injuries both at
national and international level. Annually, in the European countries, over 28.000
traffic participants are killed and 1.4 million are injured, of which 250.000 seriously
injured this representing loss of about 120 billion euros, without quantifying medical
treatment of trauma victims. To counteract the lack of awareness of traffic participants
is necessary to optimize the structures of the medical intervention and traffic monitoring
by implementing all the prevention strategies of road events [9].

The Importance of Consumed Energy for Deformation in Study of Collision 857
Modeling and simulation of traffic events require laborious technical procedures. In

practice the engineers and experts investigators estimate the gravity of the traffic events
and initial condition of motion using data such as braking distance, the coefficient of
adhesion and energy consumptions. Sometimes traces of braking are not preserved or
collected because the vehicle is moving on surfaces that do not retain traces of rubber
(wet, snowy or areas where traffic is intense) [1].
Passive safety has begun to be studied gradually being investigated in the beginning
some car to car collision problems, of course as their simple form, these actually being
resumed to classic relations between body shape and space offered to the car occupants.
Has been developed several models that allow investigators to correlate the energy
consumed in the collision with the damage geometry. These analytical models put in
evidence the relationship between crush amplitude and initial speed of the vehicle he
had before the collision, while allowing to simulate the occupant behavior and the car
during the traffic event. The objective of all these models is to simulate the behavior of
the victim and the car during the road event, and the degree of accuracy being is
validated by comparing interest data obtained from the simulation with those obtained
experimentally in laboratory condition or real condition. Has been developed a variety
of programs to help engineers to estimate with accurately the vehicle speeds and their
behavior before the collision [4].
This study provide to the investigators information who leading to a better
understanding of the loads application who acts during the side impact and how they
are converted into deformation energy of the vehicle structure. The paper provides a
valuable reference material to the investigators experts for traffic accident investigation
and reconstruction [3].
The methodical investigation of collisions includes measurement and comparison
of data collected on the one hand by the place of the road event to those obtained from
experimental tests. As a criterion for comparing the results of various tests is the size of
the deformation energy absorbed by elements of the vehicle.
2 Side Impact Energy
The kinematic, dynamics and modeling analysis process of deformation are interde-
pendent because the parameters of pre-crash phase conditioning the body deformation,
which also influences the kinematic parameters of the post-crash phase. In the side impact
collision the driver and occupants are subjected to high solicitation loads in comparison
with other types of collision with the same level of crush energy absorbed [2].
The procedures for investigating road events suppose that to determine the crush
energy with the measured deformations of the vehicle it help in determining the vehicle
speed at that moment. So the amount of deformation energy for two vehicles can be
checked by providing energy losses with the help of dynamic timing analysis (Virtual
Crash software). Using analytical calculus, we can obtain reasonable estimations;
however, they are extensive, laborious and require deep knowledge in the field [9]. The
kinetic energy of a colliding vehicle with a massive obstacle and rigid structure (pole,
tree, bridgehead, roadside safety barrier) has the following form:
858 M. Clinciu et al.
G 2
Ec ¼ 0:5 mv2 ¼ 0:5 v ð1Þ
g
Where, m – mass, G – weight of the vehicle in motion, g – acceleration of gravity

and v – initial speed of the vehicle.
The analysis will consider true the assumption according to which before the
impact the car has only translational kinetic energy, and at the moments of the collision
the conservation of linear momentum is applied respectively the law of conservation of
energy.
The movement of the car in the collision phases is governed by Newton’s two and
third laws. Newton’s third law says that in case of a collision between two bodies, one
force is equal and opposite to the other:
F1 Dt ¼ F2 Dt ð2Þ
m1 v1 þ m2 v2 ¼ m1 v01 þ m2 v02 ð3Þ
Where, m1 – mass of the first body, m2 – mass of the second body, v1 – initial speed
of the first body, v01 – final speed of the first body, v2 – initial speed of the second body
and v02 – final speed of the second body.
From physics we know that any quantity of energy expressing at the same time a
mass, the reverse is true, that any mass corresponds a quantity of energy, this meaning
that the conservation of energy is equivalent to the principle of conservation of mass:
p21 p2 P02 P02

þ 2 þ Ei ¼ 1 þ 2 þ Ei0 ð4Þ
2m1 2m2 2m1 2m2
Since conservation of energy is equivalent to the principle of conservation of mass,

work done can be written as follows:

L ¼ Ec0 Ec ¼ Ei0 Ei ¼ DEi ð5Þ
Where, ~ p – total linear momentum of the system before the collision, ~ p0 – total
linear momentum of the system after the collision, Ei – internal energy of the system
before the collision, E0c – internal energy of the system after the collision, Ec – initial
kinetic energy of the system (before the collision), E0c – final kinetic energy of the
system (after the collision).
If the resultant force who acting on the system is variable, then:
mv02 mv
L¼ ð6Þ
2 2
Ideally it would be that the vehicle should be able to take over and dissipate own
kinetic energy, without allowing potentially deformation with a potentially high risk for
their own occupants. For a better compatibility, high-mass vehicles must to absorb an
additional relative kinetic energy compared to the lower class cars.
At the moment when the collision it produce the impact is divided into two periods,
compression and relaxation, which are developed differently in time, but their sum is
equal to the time of the collision. Into a real collision, deformation occurs such that not
all kinetic energy is converted into deformation energy; a part of it is converted into
other forms of energy.
The period of compression (Pd ) is limited in time of the initiated contact between
vehicles or between the vehicle – obstacle (pole, tree, bridgehead, roadside safety
barrier), and the moment when deformations of the car bodies achieve maximum
amplitudes and speeds are equal, that being characterized by converting a part of the
kinetic energy (Ec) in deformation energy (Ed).
m1 v1 þ m2 v2 ¼ ðm1 þ m2 Þ vd ¼ ðm1 þ m2 Þ vs ð7Þ
Pd ¼ m1 ðvd v1 Þ ¼ m2 ðvd v2 Þ ð8Þ
The period of separation (Ps ) begins when the maximum amplitude of deformation
of car body elements it is achieved and lasts until the separation of the vehicles begin
(out of contact) or separation vehicle - obstacle begin and the relative speed of them
will rise again. Characteristic of the elastic-plastic structures deformation it is that a part
of the accumulated energy during compression is returned and the car body elements
are returned to the original shape, the magnitude of deformation varies from the
maximum value to the residual deformations.

Ps ¼ m1 v01 vs ¼ m2 v02 vs ð9Þ

m1 m2 v02 v01
Ps ¼ ð10Þ
m1 þ m2
In which vd and vs represents the common vehicle speeds, at the end of deformation,
respectively common vehicle speeds at the beginning of separation (restitution).
If at the end of the impact speeds are known, then we can determine the kinetic
energy consumed during the collision in the form of deformation energy:

Ed ¼ DEc ¼ 0:5 m1 v21 þ m2 v22 m1 v02 02
1 m2 v 1 ð11Þ
m1 m2
Ed ¼ 1 c2R ðv1 v2 Þ2 ð12Þ
2ð m 1 þ m 2 Þ
When the relative speed is zero, the maximum deflection occurs. The relative speed
in the return phase will increase negative to the final separation speed, while the vehicle
rebound of obstacles. The impact period between the two masses include contact times
phases of deformation and restitution. When the relative acceleration becomes zero and
the relative speed of separation achieve the maximum recoverable value we have the
separation of the two masses.
Part of the vehicle’s kinetic energy before the collision is converted to crush energy
and another part in energy consumed through the movement of translation and/or
rotation. From energy considerations based on energy conservation law the energy
balance will be established:
Ect ¼ Ed þ Edc þ Er ð13Þ
Where, Ect – total energy, Ed – crush energy, Edc – translational kinetic energy after
collision, Er – rotational kinetic energy.
During vehicle collision, they suffer both elastic and plastic deformations. Struc-
tural behavior of the vehicle during the collision can be modeled in different ways. An
energy measurement uses the concept of compensation to identify the elastic properties
of the collision by defining rejection speed (recoil) as a percentage of the vehicles speed
before impact.
This method is presented below. The behavior of the side of a vehicle during the
collision is generally defined as a function of the deformation form [2]. The simplest
mathematical model considers the deformation forces as constant throughout the col-
lision, Ci. Under these circumstances the medium force during impact, Fi is constant
where i represent the number of collisions with impact in the same place, and the
average constant force is calculated using the total energy of the collisions structure, the
strain energy is given by:
1 Ci
Edi ¼ F ð14Þ
X
Edt ¼ E
i di
ð15Þ
Presentation of the simplified model of deformation is expressed as a linear func-

tion of the impact F = kx, is:
Z
Ed ¼ Fdx ¼ 0:5kc2 ð16Þ
c
These conditions are applicable at the stage preceding the appearance of the
deformation in ideal conditions of transmission of force.
To correlate the impact with a fixed vertical cylindrical obstacle (pole), in this
study, simple linear models were used.
3 Yaw Movement
During lateral collision with a vertical cylindrical obstacle of unique radius, the central
forces induce to the vehicle a particular rotating angular velocity. This induced angular
velocity (ω) will continue to grow, with the deformation of the vehicle body elements,
until it reaches the maximum level. For calculating the energy absorbed by deformation
of the vehicle structure resulting rotational energy must be subtracted from the total
energy of the impact. The rotation induced as a result of vehicle impact is composed of:
• Rotating phase of the vehicle around the obstacle collided;
• Rotation phase of the vehicle after separation from the obstacle collided.
At this stage, the rotational speed decreases thanks to interaction between tire and
road surface area. In general the angular velocity reaches the maximum value at the end
of the collision phase.
The moment of inertia during the rotation of the vehicle takes into account the two
components of inertia:
I0 ¼ mkx2 ð17Þ
G 2
Ip ¼ m kx2 þ n2 ¼ k x þ n2 ð18Þ
g
Where, m – mass of the vehicle, G – vehicle weight, kω – radius of gyration, n –

distance from the point of impact to the center of mass (Cg), I0 – moment of gyration
and Ip – moment of inertia of vehicle that rotates around the point of impact, Fig. 1.
Fig. 1. (a) trajectory of collision, through center of mass; (b) real test
Rotation, after separation, is supposed to be around the center of mass (Cg), since
contact with vertical cylindrical obstacle is to the final position.
Energy angular velocity is calculated using the inertia moment I0 from the center of
mass of the vehicle. Center of mass was determined by weighing the individual vehicle
measurements without load. Where not taken into account changes in the position of
the center of mass or I0 yaw moment during the collision. When collision occurs in a
direction perpendicular to the side of the vehicle and passing through the center of
mass, distance from the point of impact at the center of mass is quite common who
connecting the mass center with side of the vehicle, Fig. 1(a). Since the angular
velocity is small, rotational energy of the collision can be neglected.
So, to determine the vehicle deformation energy consumption, it reduces the

rotational energy of the kinetic energy of translation resulting in the equation:
Ed ¼ Ect Edc Er ð19Þ
Mass center velocity does not vanish, but change its position by gyration move-
ment. Using the mathematical model presented and experimental data obtained, it was
able to achieve an evaluation of addicts deformation based on the impact energy,
Fig. 2.
Fig. 2. Energy/Crush distance correlations
Where curve 1 corresponds to the center of mass (Cg) and curve 2 corresponds to
the B pillar.
4 Experimental Analysis and Electronics
The vehicle submitted to experimental test collisions was provided with a network of
sensors, Fig. 3(a), and tri-axial devices, Fig. 3(b), for measurement variables of interest
such as acceleration, gyration in the center of mass or in its immediate vicinity, further
recourse to eliminate errors by applying a relationship correction of measured data. It
was measured the initial speed of the vehicle before the collision. The layout of the
measuring equipment required it to be mounted on the opposite side of the place of
impact at the level A respectively B pillars, directly on the structural elements of the
body, and in the case of measurements at the level of the front axle the sensor was
located in the internal wing, fixation at the wheel suspension system elements.
Fig. 3. (a) MC-32 S acquisition system; (b) Dummy with accelerometers installed
5 Conclusions
This study helps to evaluate and compare the stiffness of the vehicle side structure at
various points.
Crush energy determined using data from tests reveal extensive information on the
variation of change in stiffness to the vehicle side.
The amplitude of residual deformations concurs significantly to calculate defor-
mation energy. The energy consumed for vehicle body structure deformation involved
in the same traffic event, characterizes the severity of the impact, performance and
structure resistance of kinetic energy dissipate by deformation.
The paper presented a test procedure where is performed a reversibly collision,
namely the obstacle collides the vehicle which otherwise is in rest. Although this
working mode is not always met in practice, the test procedure may be considered valid
to study the energy absorbed during road event.
The procedures for investigating traffic events suppose that the crush energy
determined using the measured deformations to assist in determining the vehicle speed.
During collision of fixed vertical cylindrical obstacles the energy turns in crush
energy of the side of the vehicle, and the information regarding the deformation can be
used in reconstructing traffic events. Only data obtained from tests vehicle collisions
are compared with similar technical characteristics.
References
1. Brach, R.M.: Mechanical Impact Dynamics. Wiley, New York (1990)
2. Fildes, B., Digges, K., Carr, D., Dyte, D., Vulcan, P.: Side impact regulation benefits.
Australian Federal Office of Road Safety, Report No. CR154 (1995)
3. Jennings, P.W.: Methods for assessing vehicle speed from impact damage. Accident
Reconstruction Seminar (1999)
4. Jones, I.S.: The effect of impact type and vehicle velocity on vehicle crush. In: 27th Annual
Proceedings American Association for Automotive Medicine (1983)
5. MacMillan, R.H.: Dynamics of Vehicle Collisions. Inderscience Enterprises Ltd., U.K. (1983)
6. Huang, M.: Vehicle Crash Mechanics. CRC Press, Boca Raton (2002)
7. Brach, R.M.: Friction, restitution, and energy loss in planar collisions trans. ASME. J. Appl.
Mech. 51, 164–170 (1984)
8. Rose, N.A., Fenton, S.J., Beauchamp, G.: Restitution Modeling for Crush Analysis: Theory
and Validation. 2006-01-0908 (2006)
9. Wood, D.P.: Determination of Speed From Crush, Forensic Accident Investigation: Motor
Vehicles Damask. Michie Butterworth, Charlottesville (1995)
Aspects Regarding the Reconstruction
of Traffic Events
Mihai Clinciu(&), Anghel Chiru, and Stefanita Ciunel

mihaiclinciu@yahoo.com, achiru@unitbv.ro
Abstract. The complexity of this study conducted by the authors in this paper
is intended to be a systemic approach to road side impact event in generally and
influence of the deformation amplitude on the conditions of survival of the
victims in special. Therefore, in this paper there are presented models for
determining the pulse collision and the speed variation of the vehicle during the
collision. Also, the principal direction of force (resultant force) respectively the
collision occupant behavior in this stage are determined. Impact initiation and
post-crash phase are controlled by a network of sensors commonly used in
passive vehicle safety system. The execution elements provide information on
driving conditions and the position of occupants in the vehicle. The information
is extremely useful to expert investigators who activates in the field of the traffic
events investigation and reconstruction.
Keywords: Collision Linear momentum Speed Occupant Injury Traffic

events
1 Introduction
The traffic events are one of the main causes of death in the modern society. Side
impact collisions is the second major cause of death and injury after the front impact
collisions in analysis of the traffic incidents. This mode of collision subdue the occu-
pants to violent forces and accelerations because the space between the victim and
vehicle components (non-lethal space) is limited. Serious injury result from the side
impact collisions because the side structure is less rigid than the front part [2].
The vehicle structural stiffnes of the door is designed to reduce the speed and depth
of intrusion on occupants during the traffic events. Evaluation of the effectiveness door
is significant to reduce serious injuries and even death. In the side impact collision the
driver and the other occupants are subjected to high solicitation loads in comparison
with other types of collision with the same level of crush energy absorbed. The doors of
a car tend to have a reduced depth, what it means depth should take account, however, to
provide resistance to intrusion. The correct seating position of the occupants must allow
the thorax, abdomen and pelvis to be at the level at which the structure provides a high
impact resistance [1]. Constructive solutions influencing compatibility and effectiveness
of side structures and systems who limit the movement (seat belt, airbag, head restraint)
because to the differences of mass, stiffness and geometry of the different vehicles
types [5].

The objectives of the passive safety research should be considered as an essential

part of any strategy to reduce injuries produce in traffic events. The final scope is to
avoid or eliminate the risk elements and the new technologies to concur for achieving
these goals. Improving passive safety can be achieved through both driver assistance
systems in order to avoid traffic events, and by changing the vehicle structural stiffnes
for better protection at the time of the collision.
It is understood that in short time or medium the preventing or reducing the severity
of injuries is a major challenge in improving vehicle safety.
Therefore, the research of traffic incidents is a topical field, which must constantly
adapt to technological developments in the field of road transport.
The study of passive safety began to be approached gradually, where first are
investigated some problems collision, of course in their simplest form, they actually
comes down to classical relations between vehicle body and the space it offers to the
occupants.
The procedures for investigating traffic events suppose that the determined crush
energy with the measured deformations of the vehicle to contribute in determining the
vehicles speed at that moment. Therefore the total energy consumed in the collision can
be calculated by summing the energy consumption for each stage of the traffic event.
Using the hand calculation methods we can obtain significantly close to reality estima-
tions, instead they are extensive, laborious and requires deep knowledge in the field [6].
2 Linear Momentum and Principal Direction of Force
Collision is a process of mutual conditioning between two or more vehicles or between

a vehicle and a rigid fixed obstacle (pole, tree, bridgehead, roadside safety barrier)
which takes place in a finite time, so this interaction is missing before and after the
collision. The contact time is finite and is called time of collision. It is much smaller
than the time observation of other motion conditions who varies in very wide limits.
During the collision occur big internal forces, but does not change the total linear
momentum of the system. Since the time of collision is very small, the interaction of
the impulse force and momentum is practically zero, this means that can not change the
total linear momentum of the system. Therefore, during the collision the total linear
momentum of the system is conserved, the sum of the force (impulse) vectors before
the collision is equal to the sum of force (impulse) vectors after the collision [8].
Considering a material point on which acts a constant resultant force ~F according to
the principle II of mechanics we have:
~ D~ v0 ~
v mð~ v0 m~
vÞ m~ v ~ p0 ~
p D~ p
F ¼ m~
a¼m ¼ ¼ ¼ ¼ ð1Þ
Dt Dt Dt Dt Dt
And it results:
~
FDt ¼ D~
p ð2Þ
Where, ~
FDt is the resultant force applied.
Aspects Regarding the Reconstruction of Traffic Events 867
Equation (1) represents the impulse-momentum theorem of the material point. This
theorem show that any resultant force ~ F applied to a material point, a certain time
interval Dt causes a variation of linear momentum Dp to the material point.
The law of conservation of linear momentum - if the material point is completely
isolated or if the resultant force applied to the material point is zero, ~ F ¼ 0, then
~
FDt ¼ 0, result D~ p ¼ 0, namely ~ p0 ¼ ~
p ¼ constant (a material point keeps constant
linear momentum if on the material point does not act any forces or if the the resultant
of forces is zero). If the force ~
F is not constant then the impulse-momentum theorem of
the material point is written like this:
t2
Z
FðtÞdt ¼ Dp ¼ p0 p ð3Þ
t1
Where, p – initial momentum and p0 – final momentum.

According to Newton’s third law, in case of a collision between two bodies, one
force is equal and opposite to the other:
m1 v1 þ m2 v2 ¼ m1 v01 þ m2 v02 ð4Þ
Where, m1 – mass of the first body, m2 – mass of the second body, v1 – initial speed
of the first body, v01 – final speed of the first body, v2 – initial speed of the second body
and v02 – final speed of the second body.
When the maximum deformation occurs (collision is complet), the vehicle speeds
are equal to each other:
v1 ¼ v2 ¼ vd ¼ vs ð5Þ
m1 v1 þ m2 v2 ¼ ðm1 þ m2 Þ vd ¼ ðm1 þ m2 Þ vs ð6Þ
In which vd and vs represents the common vehicle speeds, at the end of defor-
mation, respectively common vehicle speeds at the beginning of separation
(restitution).
2.1 Principal Direction of Force

Side impact collisions, under a certain angle, it is a configuration in which speed
vectors are not collinear and the principal direction of force (resultant force) of the
collision at least one of the vehicle does not pass through the center of mass. The
collision being inelastic, there is the possibility that the vehicle may have a induce yaw
motion, with the collision trajectory force non-central. Addittionaly, there is no
physical condition that the post-collision speed for each vehicle to be equal or have the
same direction [9].
Speed variation Δv can be calculated using the geometric vector form the cosines
and sines law for both vehicles involved in the collision under an arbitrary angle,
Fig. 1.
Fig. 1. (a) Real test; (b) A vector diagram solution to a collision using the conservation of linear
momentum
The law of cosines is applied in a triangle where two sides and the angle between
them are known.
While, v1 , v01 and angle h1 are known, then:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
c2 ¼ a2 þ b2 2abcosA ) c ¼ a2 þ b2 2abcosA ð7Þ
Where, a ¼ v1 ; b ¼ v01 ; c ¼ Dv1 and cosA ¼ cosh1

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Dv1 ¼ v21 þ v02 1 2v1 v1 cosh1
0 ð8Þ
Where, v01 – the magnitude of the post impact speed, v1 – the magnitude of the
impact speed, Dv1 – the magnitude of the velocity change durinmg impact, h1 – the
vehicle departure angle or the direction of v01 and a1 – principal direction of force angle
(direction of Dv1 ).
The law of the sines applies when two angles and one side are known, and the third
side can be calculated as:
0
v1 sinh1
a1 ¼ arcsin ð9Þ
Dv1
For the second vehicle, the direction of vectors Δv2 can be calculated using the
sines rule, resulting the principal direction of the collision force (resultant force), as
sketched in Fig. 1(b).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Dv2 ¼ v22 þ v02 2 2v2 v2 cosh2
0 ð10Þ
0
v2 sinh2
a2 ¼ arcsin ð11Þ
Dv2
For the side impact collisions, under a certain angle, vectors of the post-impact
linear momentum of the two vehicles can be written as:
p01 ¼ m1 v01 ð12Þ
And
m2 v02 cos/ m2 v2 cosw

v1 ¼ v01 cosh1 þ
m1 m1
p02 ¼ m2 v02 ð13Þ
And
m1 v01 sinh1 v02 sin/

v2 ¼ þ
m2 sinw sinw
pT ¼ m1 v01 þ m2 v02 ð14Þ
Where pT represents the total momentum of the both system (principal direction of
force), before and after impact.
During collision the speed varies continuously, permanently changing its sense and
direction (velocity becomes negative), period during which the vehicle changes its
geometric configuration by deforming. After measuring deceleration, we can express
the vehicle displacement. Because the impact is not centric but it’s eccentric we can
analyze recorded impulse on transverse direction (axis - y), but also yaw moments. By
integrating vehicle deceleration, we get a graphic that presents the relationship between
acceleration, speed and deformation, Fig. 2.
In Fig. 2, the curve 1 represents the speed for the central impact, curve 2 is the
speed for the non-central impact, curve 3 is the acceleration, curve 4 is the displace-
ment, curve 5 is the experimentally measured speed, t0–t01 is the reaction time of the
Fig. 2. Vehicle speed/Time diagram

occupant limiting motion systems (time in which occupant is in free flight to the impact
point) and t01–t02 is the ideal restraint.
When the collision speed (post-crash) is zero, one can make the claim that the
deformation is at maximum. The relative velocity in the rebound phase increases
negatively to the final separation velocity, during which the vehicle bounce out of
obstacle. The length of contact between the two masses includes times during defor-
mation and restitution phases. When the relative acceleration is zero and the relative
speed of separation achieves maximum recoverable value, we have separation of the
two masses.
3 Occupant Dynamics
Reduction of the vehicle speed in case of vehicle-vehicle collision or vehicle-rigid fixed

obstacle (pole, tree, bridgehead, roadside safety barrier) or reduction occupant speed
during impact by elements from vehicle interior body structure in the same traffic event
may occur as a result of applying the high-value force in a short time, or as a result of
application a low intensity force in a relatively large time interval. High value forces
occur during occupant impact with the vehicle interior body structure (pillars, doors,
dash board, steering wheel column, floor, ceiling) when the occupant limiting motion
systems is not in use (restraint system-seatbelt, airbag, head restraint) and lower loads
acting between the occupant contact surface and limiting motion systems [2].
It should be noted that the occupant injuries have been encountered even when they
are protected by the limiting motion systems and there no exist direct contact between
them and vehicle interior. Under these conditions, injury may be caused by:
• Aggressive contact between the occupant and the limiting motion systems (if the
systems is in use and/or occupant is out of position);
• The relative movement of the occupant’s body segments;
• Deformation of vehicle body structure;
• Aggressive contact with the unsecured objects from the vehicle interior;
• Intrusion of own vehicle body structure elements or fragments of glass or plastic.
Common initial kinetic energy of the occupant-vehicle consists in vehicle kinetic
energy and occupant kinetic energy. When the impact occur some of the vehicle initial
kinetic energy is transformed into deformation energy and the other part is consumed
on vehicle motion in the post-collision phase. The occupant kinetic energy will be
equal to the commun kinetic energy occupant-vehicle and when we refer to the vehicle
system this is null because there is no relative motion between the two
(occupant-vehicle) [3].
When the relative motion between occupants-vehicle reaches a maximum, the
kinetic energy associated with this motion is equal to one-half the product of mass and
velocity squared. Likewise, the relative kinetic energy of the occupant will be partially
converted to:
• Work done consumed in contact with occupant limiting motion systems or contact
between occupants;
• Work done consumed for deformation vehicle interior elements that change shape
or original dimensions (seat, anti-submarining device);
• Work done consumed for slipping occupant under the seat and seat belt;
• Work done consumed for muscular system reaction.
Therefore, the remaining relative occupant kinetic energy during the collision will
be consumed by loads and occupant injuries. The specific loads that act on different
parts of the human body represent the determining factor in producing occupant
injuries when their size exceeds the limits of human tolerance.
The size of the relative kinetic energy of the occupant can be used as a measure
injury risk prediction, and represent an important research objective in automotive
passive safety. In primary impact research of the vehicle-obstacle collision is intended
to reduce vehicle speed and implicit the kinetic energy of vehicle by turning it as much
as possible into controlled deformation energy (elements deformation only to the
occupant safety cell-the cockpit). By researching the secondary impact occupant-
vehicle is intended to reduce the relative velocity between them by dissipating the
relative kinetic energy of the occupant in the vehicle interior elements, elasticity of the
dashboard material, elasticity of the seats material, deformable steering wheel column
and antisubmarining devices and to limiting motion systems (restraint systems).
4 Experimental Analysis and Electronics
For evaluation of the experimental tests collisions in real conditions the vehicle who was
collided was equipped with a system type of networked sensors, Fig. 3(a), and triaxial
transducers, Fig. 3(b), for quantifying the sizes of the monitored and controlled variables.
In the paper speed is measured before, during and after the collision, the acceler-
ation and the gyration in the center of mass or in its immediate vicinity, which imposed
to finally eliminate errors by applying a correction relationship resulting data. Mea-
suring chains were placed as much as possible on opposite side to the impact resultant
force of the collision to the A and B pillars, meaning directly on vehicle body structure
elements to avoid parasitic noise.
Fig. 3. (a) Xsens sensors; (b) Accelerometers

5 Conclusions
Modern vehicles must meet consumer requirements regarding performance and safety
features. One should remember that the results of experimental tests could be erro-
neously interpreted and conclusion could be erroneously placed that the lateral limits of
the vehicles body structure are not capable to provide non-lethal space to the occupants.
The results of this study allow generalization of the characteristics of energy dis-
sipation in the lateral side of the vehicle.
Test procedures used in the present are expensive and requires time in relation to
the proposed solution.
The proposed solutions come to support traffic events research and implicitly
determine the causes of their occurrence.
In the future this experimental test procedure is intended to be extended to the rear
side behavior and rollover during road events research.
References
1. Watts, A.J., Atkinson, D.R., Hennessy, C.J.: Low Speed Automobile Accidents. Lawyers &
Judges Publishing Company, Tuscon (1999)
2. Fildes, B., Lane, J., Lenard, J., Vulcan, P.: Passenger cars and occupant injury: side impact
crashes, Australian Federal Ofice of Road Safety, Report No.CR134 (1994)
3. Brach, R.M.: Comments on energy loss in vehicle to vehicle impact by Dario Vangi.Int.
J. Impact Eng. 37(1), 114–115 (2010)
4. Mackay, G.M., Hill, J., Parkin, S., Munns, J.A.R.: Restrained occupants on the non-struck
side in lateral collisions. In: Proceedings of 13th ESV Conference, Paris France (1991)
5. Kahane, C.J.: Fatality Reduction by Safety Belts for Front-Seat Occupants of Cars and Light
Trucks (DOT HS 809 199). National Highway Traffic Safety Administration, Washington DC
(2000)
6. Lund, A.K., O’Neill, B., Nolan, J.M., Chapline, J.F.: Incompatibility in vehicle-to-vehicle
side impact crashes in the United States: real-world and experimental results. In: Vehicle
Safety 2000, Insurance Institute for Highway Safety, London UK. and Arlingto, VA (2000)
7. McHenry, R.R.: Mathematical Reconstruction of Highway Accidents. PB-240 867, NTIS
(1975)
8. Smith, G.C., James, M.B., Perl, T.R., Struble, D.E.: Frontal crush energy and impulse analysis
of narrow object impacts. In: 87-WA/SAF-5, ASME Winter Annual Meeting, Boston MA,
13–18 December 1987
9. Smith, R.: Reconstructions of the collisions staged at the field day at Norwich. In: ITAI 2nd
International Conference, Plymouth (1995)
Vehicle Driver Drowsiness Monitoring
and Warning System
Tiberiu Vesselenyi(&), Alexandru Rus, Tudor Mitran, Bogdan Tataru,

and Ovidiu Moldovan

tvesselenyi@yahoo.co.uk, alexandru.rus.61@gmail.com,
tudor_mitran@yahoo.com, mirceabogdant@gmail.com,
ovidiu30@gmail.com
Abstract. The paper presents a new method of monitoring and warning drowsy
drivers in order to prevent accidents. Analyzing current approaches we decided
to study algorithms which can be used for drowsiness/awareness detection when
applying techniques based on EEG and EMG signals measurement. There are
also a large number of studies which uses EEG signals for brain – computer
interface applications. This method has the advantage that is independent from
the equipment, vehicle or situ where it is used, and can be applied both for
drivers and equipment operators too.
Keywords: Drowsiness monitoring EEG EOG Signal processing

Automobile driver safety
1 Introduction
Major studies reported that 54 % of adult drivers admitted there were occasions when
they were drowsy during driving and the percentage of those who had fallen asleep
during driving was 28 % [1]. Based on these statistics it would be important to develop
methods and equipment to monitor and warn drivers in case they are driving in a
drowsy state and so preventing accidents.
Sleep cycle has the following phases: Non-Rapid Eye Movement (NREM) sleep and
Rapid Eye Movement (REM) sleep. The NREM stage also has four sub stages.
Drowsiness is the first sub stage of the NREM stage [5, 6]. So it is important to detect the
difference between the first sub stage of the NREM stage of sleep and the awaken state.
In order to achieve this goal many methods were tried out based on the use of: heart
rate, detecting of the eye movement and facial expression, electroencephalography
(EEG), electrooculography (EOG) and electrocardiography (ECG) signals.
The eye movement detection approach analyses eye images to detect changes of
image, such as eyelid movement and eye direction. A few methods used cameras and
computer processing methods to find out the percentage of eyelid closure. In this case
the drowsiness detection accuracy is affected by the lighting and driving conditions.
The signal detection method is to measure biological signals, such as the EEG or
others. These signals can give more accurate drowsiness monitoring than eye motion
detection. The problem of this technique is that the electrode and wiring means a
874 T. Vesselenyi et al.
discomfort for the driver and it is using a conductive gel. These problems can be solved
by using dry-contact, low-noise EEG sensors, as shown in [7]. The dry-contact EEG
electrodes were created by the authors [7] with micro-electrical-mechanical system
(MEMS) technology. Each channel of the analog signal processing device comes on a
custom-built, small-sized circuit board which contains an amplifier, filters, and
analog-to-digital converter. As the authors describe daisy-chain configuration between
boards with bit-serial output reduces the wiring needed and the system is capable to
detect alpha-band rhythms and eye-blink signals.
Also for driver awareness detection a number of methods were proposed which are
based on the technique of embedding biosensors into steering wheel or in the seat of the
vehicle in order to measure heart beat pulse signals [8]. Time series of heart beat pulse
signal can be used to calculate the heart rate variability. As it is shown in [5], the
frequency domain spectral analysis of heart rate variability shows that typically it has
three main frequency bands: high frequency band, low frequency band and very low
frequency. A number of psycho-physiological researchers have found that the low and
high frequency power spectral density ratio decreases when a person changes from
waking into drowsiness/sleep stage, while the high frequency power increases asso-
ciated with this status change. The authors in [8] show that this variability can be an
effective method for the detection of driver drowsiness. The problem with this method
is that the drowsiness detection equipment is not portable and it must be embedded in
the vehicle parts which will limit its applicability.
2 Preliminary Results
Analyzing the above mentioned approaches we decided to study algorithms which can
be used for drowsiness/awareness detection when applying techniques based on EEG
(Electroencephalography), EOG (Electrooculography) signals measurement and also
eye image processing. The EOG and eye image measurements are redundant giving a
better accuracy for drowsiness detection. There are also a large number of studies
which uses EEG signals for brain – computer interface applications [9–12]. If the EEG
device is built in the manner described in [7] this method has the advantage that is
independent from the vehicle where it is used. The preliminary results obtained during
researches conducted at the University of Oradea are presented in [2–4]. The useful
potentials measured on the skull are considerably low. In case of EOG (when mea-
suring eye movement or blinking) is about from hundreds to a few thousands of micro
volts. EEG potentials are much lower from a few to a hundred micro volts. The signal
conditioning methods and equipment (amplification and filtering) has to be of a very
good quality at these low levels of potential. Even so due to the electrical complexity of
the human body, noises from different sources can alter the useful signals (Fig. 1). One
of the major noises which are present in these signals is the 50 Hz component which is
due to the AC source of the EEG device (Fig. 2).
These noises are observable in the obtained results. The noise is much larger
compared to the useful signal, because the eye movement signals are much lower than
the eye blink signals. Another components which can be considered as noises are the
heart beat (or pulse) signals which cannot be always avoided. Power Spectral Density
Vehicle Driver Drowsiness Monitoring and Warning System 875
Fig. 1. Eye movement and heart beat time signal, filtered with a low pass FFT filter in the
0–20 Hz domain.
Fig. 2. Eye movement time signal, filtered with a low pass FFT filter in the 0–6 Hz domain.
of signals had been also analysed. Studying the possibilities of filtering these signals we
could assume that filtering eye blink and eye movement can be made using frequency
domain filtering using the same filtering parameters both for eye blinking and eye
movement signals (Fig. 4).
Regarding the EEG signals these were recorded both in aware and drowsy states.
Signals were recorded for a group of healthy subjects. From these signals we obtained
the PSD using a simple FFT algorithm. The PSD Amplitude versus frequency diagrams
had been studied. Analyzing these diagrams we could observe (as it is also shown by
different authors) that the peak representing the alpha rhythm (in the frequency domain
of 8–12 Hz) clearly increases (almost doubles) in drowsy state compared to aware state.
Fig. 3. EEG signals in aware state.
Fig. 4. PSD amplitude versus frequency of the EEG signal presented in Fig. 3, with the alpha
rhythm peak in the 10–12 Hz frequency domain.
Our approach is based on the increasing PSD amplitude in the EEG signal in the
frequency band of 8–12 Hz combined with eye images to differentiate awareness from
drowsiness. We also studied the possibilities to apply a fuzzy decision system to detect
drowsiness based on the measurement results [2, 3]. The results of the preliminary
studies showed that the system proposed in this project can be feasible (Fig. 6).
Fig. 5. EEG signals recorded in drowsy state.
Fig. 6. PSD amplitude versus frequency of the EEG signal presented in Fig. 5, with the alpha
rhythm peak in the 10–12 Hz frequency domain.
3 Presentation of the Concept
The novelty of the solution presented in this paper (drowsiness monitoring system),
consists in the simultaneous measurement and analysis of EEG, EOG signals and
automobile driver eye movement, eyelid position, eye pupils measurement, based on
which a unique decision making algorithm can give accurate information on the
drowsiness state of the automobile driver.
The feasibility of the described solution is based on knowledge of automobile
safety, knowledge of automobile driver behaviour, data acquisition, signal and image
processing, decision making algorithms and also in the use of high quality electronic
products.
Starting with the formulated concept which is based on research results on EEG and
EOG signals measurement and analysis and also on image processing algorithm
development results used for automobile driver drowsiness detection, going through
the development of a system which experimentally proves the initial concept (EPC in
Fig. 7) and finally the laboratory validated system is designed and developed (LVP in
Fig. 7).
Fig. 7. Concept of the drowsiness monitoring system.
The objective is to develop at first the EPC system. The EPC system will be used to
test EEG, EOG and eye image parameters corresponding to drowsy and aware states of
a statistically significant group of subjects and extract meaningful characteristics for
each state. Self-calibration methods in order to adjust system parameters for each user
will also be developed. In the second stage of the project, the LVP system will be
developed. The algorithms and methods developed in the first stage will be imple-
mented and tested on the LVP system. The steps to achieve the product are described in
the followings.
In order to achieve this goal the selection of devices and equipment needed to
develop the EPC system, has to be made. Acquisition of video cameras and pro-
gramming environment with image processing and recognition capabilities will be also
made.
The next step is to develop the EPC monitoring system using the acquired
equipment and devices ensuring the seamless interoperability of different devices,
mainly at electrical level in respect to the input output signals but also regarding the
protocols used for data transmission between digital devices in the system.
In the experimental phase realization of the laboratory tests for determining the
parameters and conditions of the eye monitoring procedure will be accomplished. The
measurements will be done in laboratory conditions using the CCD camera.
Based on the results of the experimental measurements to determine the system
parameters for eye monitoring and the resulted report the software application to
determine drowsiness based on eye monitoring will be realized. The application will
have to automatically process the acquired image and based on a series of parameters
determined in the previous activity will establish the drowsiness state of the subject.
Also in the experimental phase we will have to find the parameters and charac-
teristics of the EEG equipment and of the sensors and in accordance with the previous
results, selecting the optimal programming environment with advanced signal acqui-
sition and processing capabilities. For this step we will acquisition an EEG equipment
and sensors and programming environment with signal acquisition and processing
capabilities.
Integrating the acquired equipment and devices in order to realize an operational
system for monitoring of brain signals using EEG and EOG will be performed in the
next step, also ensuring the interoperability of all the devices and the possibility of data
acquisition from the system with minimal error and signal disturbance and solving
issues related to signal noise by applying hardware and software filtering techniques.
Realizing of measurements and collecting data in order to determine the actual
parameters, signal level and feasibility of the EEG and EOG signals has to be
accomplished in the final step and based on the test results obtained previously and
organized in databases, a software for signal analysis and processing will be realized.
Using the information collected previously and the systems developed the authors
will develop the integrated system that will include the image progressing subsystems
and the EEG – EOG signal processing modules.
Having a working LVP system for the drowsiness monitoring system the authors
will realize the system in an ergonomic shape.
References
1. Sahayadhas, A., Sundaraj, K., Murugappan, M.: Detecting driver drowsiness based on
sensors: a review. Sensors 12, 16937–16953 (2012). doi:10.3390/s121216937
2. Dzitac, I., Vesselenyi, T., Tarca, R.C.: Identification of ERD using fuzzy inference systems
for brain computer interface. Int. J. Comput. Commun. Contr. 6(3), 403–417 (2011). ISSN
1841-9836
3. Dzitac, S., Vesselenyi, T., Popper, L., Moga, I., Secui, C.D.: Fuzzy algorithm for human
drowsiness detection devices. Stud. Inform. Contr. 19(4), 419–426 (2010)
4. Vesselenyi, T., Dzitac, I., Dzitac, S., Hora, C., Porumb, C.: Preliminary issues on
brain-machine contextual communication structure development. In: IEEE Conference, 3rd
International Workshop on Soft Computing Applications, Szeged, Hungary, pp. 35–40
(2009)
5. Carlson, N.R.: Sleep and biological rhythms. In: Carlson, N.R. (ed.) Physiology of Behavior,
7th edn. Allyn & Bacon Publishers, Boston (2001)
6. Sleep Research Society, Basics of Sleep Behavior Syllabus, Web Sciences International and
Sleep Research Society (1997)
7. Sullivan, T.J., Deiss, S.R., Jung, T.P., Cauwenberghs, B.: A brain-machine interface using
dry-contact, low-noise EEG sensors. In: 2008 IEEE International Symposium on Circuits
and Systems Washington, US (2008)
8. Yu, X.: Real-time non-intrusive detection of driver drowsiness. Intelligent Transportation
Systems Institute Center for Transportation Studies, University of Minnesota (2009)
9. Daly, J.J., Wolpaw, J.R.: Brain-computer interfaces in neurological rehabilitation. The
Lancet Neurol. 7(11), 1032–1043 (2008)
10. Ting, J.A., D’Souza, A., Yamamoto, K., Yoshioka, T., Hoffman, D., Kakeif, S., Sergio, L.,
Kalaska, J., Kawato, M., Strick, P., Schaal, S.: Variational bayesian least squares: an
application to brain–machine interface data. Neural Netw. 21(8), 1112–1131 (2008)
11. Cvetkovic, D., Übeyli, E.D., Cosic, I.: Wavelet transform feature extraction from human
PPG, ECG, and EEG signal responses to ELF PEMF exposures: a pilot study. Digital Signal
Process. 18(5), 861–874 (2008)
12. Ting, W., Guo-zheng, Y., Bang-hua, Y., Hong, S.: EEG feature extraction based on wavelet
packet decomposition for brain computer interface. Measurement 41(6), 618–622 (2008)
Research Regarding Pedestrian Visibility
During Night-Time Using Photo Processing
Bogdan Tolea(&), Daniel Dragos Trusca, Csaba Antonya,

Alexandru Ionut Radu, and Dragos Sorin DIMA
Transilvania University of Brasov, 50068 Brasov, Romania

bogdan.tzolea@gmail.com
Abstract. The main goal of this paper is to determine the pedestrian visibility
using photo processing on the lower parts of the pedestrian, at different distances
and to determine how much is the pedestrian visibility influenced by the
upcoming traffic. The second objective of the paper is to determine the safe
speed in order to avoid an accident during night-time. For this experiment 5
vehicles were used, one from which the data was collected, called source
vehicle, and the remaining ones were positioned on the opposite lane in order to
simulate oncoming traffic. Using different test scenarios, data was collected and
processed in order to obtain relevant conclusions.
Keywords: Pedestrian Visibility Luminance Glare effect Garment color
1 Introduction
Pedestrians are considered the most “underprivileged” road users due to their lack of
protection during a collision with a car. A proper visibility of the pedestrians can
reduce the consequences of an accident. The pedestrian visibility is mainly influenced
by the way how the human visual system adapts from daylight to darkness. During
night-time driving, the perception and detection of objects positioned on the vehicle’s
path are critical [1].
The main parameters that influence the visibility are the luminance and the contrast
between the object and its surroundings. The most common car-to-pedestrian accidents
occur during night-time, and in 20 % of these cases the drivers reported that they
couldn’t percept the presence of a pedestrian [2, 3].
Glare effect represents one of the biggest problems during night-time driving, and
surprisingly few studies analyzed the effects of the headlamps glare upon the drivers,
and the driver’s perception on the pedestrian right after the glare effect [4]. One of this
researches concluded that the main influence in the pedestrian perception is the garment
color [5].
In order to reduce the glare effect, the manufacturers developed a new lighting
system, based on the LED technology, which is capable to modify the lighting cone in
a way that the oncoming traffic will not be affected by any light. This system is capable
to detect the pedestrians, and sends luminous signals to inform the driver of the
presence of the pedestrian [6]. But this technology is not available on all the vehicles,
and the visibility during night-time is still a big issue.

882 B. Tolea et al.
The main objective of our study was to determine the pedestrian visibility at
different distances in function of trousers color. The second objective was to determine
the limit velocity necessary for the drivers in order to stop to the pedestrian, depending
on the determined distances.
2 Methodology
In order to conduct this experiment, measurements on the testing track were performed.
Therefore, a road with a length of 200 m was chosen, without street lighting. On this
test track, marking on intervals of 10 m were realized, in such a way that the testing
track was split in 20 intervals.
The vehicle from which all the measurements were performed was positioned at
distance 0. We will call this vehicle as the “source vehicle”. On the opposite lane, 4
vehicles were positioned in such a way to simulate the oncoming traffic. The first
vehicle was positioned at a distance of 5 m from the source vehicle, being the optimum
distance to simulate the glare effect, while the rest of the vehicles were positioned at a
distance of 20 m from each other. The sketch of the experiment is presented in Fig. 1.
Fig. 1. Experiment configuration (Color figure online)
The pedestrian was using trousers of 4 different colors (black, blue, red and yel-
low), while the measurements were performed from the red vehicle. All the mea-
surements were performed for the given situations: the source vehicle had the low
beams turned on; the source vehicle had the high beams on; the oncoming traffic had
the low beams turned on while the source vehicle was using the low beams also. In the
source vehicle, 2 cameras and a lux meter necessary to measure the luminance, were
positioned. The next step was to position the pedestrian at a distance of 0.5 m of the
sidewalk, on the right lane.
The pedestrian moved from 10 to 10 m until the distance of 100 m was reached. At
each 10 m interval a measurement was performed on each trousers color. In the same
way, pictures were taken at each distance interval with all the clothing colors. Using
these pictures, we were able to photo-process them in order to obtain the visibility of
the pedestrian.
Research Regarding Pedestrian Visibility During Night-Time Using Photo Processing 883
Fig. 2. Area analyzed on the pedestrian
The visibility of the pedestrian was analyzed on the lower leg, and the main reason
is the fact that the lighting cone coming from the vehicle beams is inclined to a certain
angle, and we can talk about pedestrian visibility in the moment when the lighting cone
illuminates the pedestrian lower leg (Fig. 2).
3 Results
All the photo-processing was realized in the photogrammetry software PC-Rect version
4.2. This software is able to calculate the luminance on different areas of the picture and
can calculate the contrast between two analyzed areas. Therefore, the visibility of the
pedestrian is given by the software by comparing two luminance values and deter-
mining the contrast.
An example of photo-processing is presented in Fig. 3, where the two rectangles
represent the analyzed areas. In all the cases, the analysis was made on the pedestrian
lower leg and its surroundings.
Fig. 3. Example of photo-processed picture
The software is capable to extract all the data in a table that shows the visibility of
the object, based on the most important parameter ΔLu, which is the luminance dif-
ference between the two areas selected with the rectangles (contrast). The calculations
made were based on Adrian Werner method, which is a method to establish the
visibility of the object in different light conditions. By gathering all the ΔLu values
from all the pictures, 3 diagrams were extracted for all the test scenarios.
884 B. Tolea et al.
3.1 First Scenario – Source Vehicle Having Low Beams Turned on

In order to observer better the critical areas where the pedestrian was no longer visible,
the maximum values on the Y axis were excluded being obvious that in the case where
the pedestrian was visible at longer distances, then its visibility can’t be questioned at
shorter distances. Analyzing this diagram, it can be observed that the most visible color
garment is the yellow one, which gives a high visibility at high distances (70–80 m).
This color was visible on all distances (Fig. 4).
Fig. 4. Visibility diagram in the first scenario (Color figure online)
The next color with a high visibility in the first 50 m is the red color. The distance
where this color is losing visibility is after a distance of 70 m. The blue color presents a
lower visibility level in comparison with the previous discussed colors, having a greater
visibility in the first 30 m. In the interval 30–50 m its visibility is reduced and requires
greater attention from the driver in order to notice the pedestrian. After the distance of
50 m the visibility of the blue trousers is lost. Black is the color with the smallest
visibility, presenting an acceptable visibility in the first 20 m, while after that the
visibility is lost. This phenomena occurs due to the contrast given by this color with the
background. In this case we can talk about a 0 contrast.
3.2 Second Test Scenario – Source Vehicle Having the High-Beams On

In this scenario the pedestrian visibility was greater than in the first case, as was
expected. It can be observed that the colors yellow respectively red have a great
visibility on all distances, while the black and blue colors lose visibility after certain
distances.
For the red color was observed an increasing of visibility of 30 % while for the blue
color the increasing was of 20 %. The blue color loses its visibility after the distance of
70 m.
For the black color, were not observed visibility changes in comparison with the
first scenario. The only difference between the two tests is represented by the contrast
value, which is significantly higher at the distance of 10 respectively 20 m, leading to
the conclusion that the visibility of the pedestrian is higher in the first 20 m (Fig. 5).
Fig. 5. Visibility diagram for the second test scenario (Color figure online)
3.3 Third Test Scenario – the Source Vehicle and the Oncoming Vehicles
Have the Low Beams Turned on
This case represents a problem during night-time driving because the glare effect may
appear, and the pedestrian visibility is reduced.
As in the previous discussed cases, the yellow respectively red color were the most
visible ones, and their visibility was the same as in the first test scenario. The only
difference observed for the red color is the ΔLu value, where the value until the
distance of 60 m is higher than in the first test scenario (Fig. 6).
Modifications regarding the black color were observed. In the case of the black
color, the visibility presents an increasing of 10 % in comparison with the two dis-
cussed cases, and this is due to the light fascicles sent by the oncoming traffic beams.
886 B. Tolea et al.
Fig. 6. Visibility diagram for the third test (Color figure online)
These light fascicles lights the background around the pedestrian increasing the con-
trast value, and leading to an increased pedestrian visibility. For the black color, the
visibility is the same as in the first case.
4 Calculation of the Initial Car Velocity in Order to Avoid

the Collision with the Pedestrian
The braking distance is divided in 3 different distances:

Sr – Distance traveled by the vehicle in the reaction time of the driver
Sl – Distance traveled by the vehicle during the brake lag – where brake lag means
the time until the braked is fully applied
Sb – Distance traveled by the vehicle with the brake fully applied
ST – Total braking distance
ST ¼ Sr þ Sl þ Sb ð1Þ
The parameters known for each case are the following:

V = 0 – Velocity at 0.5 m of the pedestrian
tl = 0.2 – lag time – time until the brake is fully applied
tr = 1 – reaction time – time when the drive reacts
ab = µg = 7.8 m/s2 – braking deceleration

µ = 0.8 – dry wet asphalt [7]
For each color we determined the visibility distance using Kinematics Calculation
available in PC-Crash 10.2. The calculations were made on the first and third scenarios,
where the visibility was reduced for some of the colors.
In order to understand better the 3 phases of braking and the parameters that
influence the braking, a sketch was created and it is presented in Fig. 7.
Fig. 7. Sketch of the 3 phases of braking process
The parameters are:

Vav – avoidance velocity
Vr – velocity of the vehicle after reaction time (usually is equal to avoidance
velocity due to the constant traveling of the vehicle in the reaction time phase)
Vr = Vav
Vl – velocity at the end of the braking lag process – in this phase the braking
deceleration is considered half of the maximum obtainable acceleration
The calculation will be made backwards, starting from the end of the braking until
the beginning. By performing all the calculations we obtained the following table
(Table 1).
Table 1. Avoidance velocity depending on the garment color and the scenario
Velocity limit
Black Blue Red Yellow
Low beam 41 km/h 78 km/h 96 km/h 119 km/h
Oncoming traffic 53 km/h 78 km/h 96 km/h 119 km/h
888 B. Tolea et al.
5 Conclusions
After photo-processing all the pictures and analyzing all the data presented in this paper
was observed that visibility level during night-time driving is influenced by parameters
such as garment colors and oncoming traffic lights. By correlating the visibility dis-
tances with the avoidance velocities determined in this paper, we can argue that the
speed limit imposed in most of the urban areas of 50 km/h, where the street lighting is
poor, may not be small enough in order to avoid an accident with a pedestrian dressed
in black.
In case of the roadways where the speed limit is 100 km/h outside urban areas, the
safe color for the pedestrian is yellow in order to be seen in comparison with the other
colors, which may put them in danger. The main problem is when there is oncoming
traffic and all the drivers have to switch to low beams leading to a decreased visibility
outside urban areas.
References
1. Hankey, J.M., Blanco, M., Gibbons, R.B., McLaughlin, S.B., Dingus, T.A.: Enhanced Night
Visibility Series, vol. I: Executive Summary. No. FHWA-HRT-04-132 (2005)
2. Owens, D.A., Sivak, M.: Differentiation of visibility and alcohol as contributors to twilight
road fatalities. Hum. Factors 38(4), 680–689 (1996)
3. Hazlett, R.D., Allen, M.J.: The ability to see a pedestrian at night: the effects of clothing,
reflectorization and driver intoxication. Am. J. Optom. Arch. Am. Acad. Optom. 45(4), 246–
258 (1968)
4. Wood, J.M., Tyrrell, R.A., Carberry, T.P.: Limitations in drivers’ ability to recognize,
pedestrians at night. Hum. Factors 47, 644–653 (2005)
5. Wood, J.M., Tyrrell, R.A., Chaparro, A., Marszalek, R.P., Carberry, T.P., Chu, B.S.: Even
moderate visual impairments degrade drivers’ ability to see pedestrians at night. Invest.
Ophthalmol. Vis. Sci. 53(6), 2586–2592 (2012)
6. http://www.audi.com/com/brand/en/vorsprung_durch_technik/content/2013/10/audi-a8-in-a-
new-radiant-light.html. Accessed 10 June 2016
7. PC Crash user Manual
Estimating the Costs Caused by Road Traffic
Accidents in Romania
Alin Drosu(&) and Corneliu Cofaru

alin_drosu@yahoo.com, ccornel@unitbv.ro
Abstract. At European level, the costs caused by crashes are estimated at about
100 billions of Euros. This paper aims at estimating the costs of the road traffic
accidents occurred in Romania and the Human Capital method is used to esti-
mate that costs. Depending on the accidents’ severity, several costs have been
identified and the victims’ lost output has been evaluated afterwards to give
present-day values. Other costs, as the administrative ones regarding property’s
damage, funeral costs, and technical and legal expertise costs have been taken
into account. Also, a significant value has been added in order to reflect the
“pain, grief and suffering” of the crashes’ victims and to all those who care for
them as families, relatives or communities.
It is the first time in Romania when such a study is conducted and shows that
road accidents costs are approximately 1 % from the national GDP.
Keywords: Crash costs Road safety Accidents
1 Introduction
Yearly, over 1.2 million people are killed worldwide. The road accidents are among the
leading causes of mortality for the young people aged 15–29 years and cost the gov-
ernments about 3 % of Gross Domestic Product (GDP). Therefore, the road accidents
are becoming a real problem in the national health systems [1].
Globally, road crash injuries are estimated to be the ninth leading cause of death
among the population and it is predicted that by 2030, it will become the seventh
leading cause of death [2].
In addition to deaths in the road accidents, up to 50 million people are injured each
year in road events, those having consequences on the global health [3].
In 2013, more than 26.000 people have died in road accidents in the European
Union and an average of 52 people killed per million inhabitants has been reported.
The highest number of deaths in road accidents was recorded in Romania, which
reported 92 deaths per million inhabitants [4].
In order to reduce the number of death caused by crashes, national and local
simultaneous actions regarding the safety of vehicles, road user’s behavior and a new
design of roads, are required. Lessons learned from the developed countries about the
impact it has on the road safety infrastructure, impose best practices to be urgently
implemented in the developing countries [1] as well as in Romania. Implementing
these best practices requires a fair and transparent financing of those measures that

890 A. Drosu and C. Cofaru
intends to reduce the road accidents, a real burden on the Romanian health system.
Also, for a fair assessment of the Romanian health system, an estimating of the crashes’
costs will be needed.
At national level, determining the road accidents costs leads to a correctly resources
planning. In this way, the road safety should benefit of a fair and transparent bud-
get allocation. Also, by determining the road accidents costs, the governments may
decide a better use of the investments in road safety and can assure that the best road
safety improvements are made in order to generate costs benefits related to their
implementation [5].
Nevertheless, not all the EU countries have proceeded to estimate the accidents
costs, except BG, DE, EE, IE, GR, ES, FR, IT, CY, LT, HU, NL, AT, PL, PT, SI, SK,
FI, SE and UK (Fig. 1). From this reason, there is no any possibility to calculate an
average of the accidents costs at European level [1].
Fig. 1. European countries that estimated their GDP lost due to road traffic accidents
In 2007, the costs of road accidents from Romania were estimated around 1.20
billion of Euros, but there are no details about the method of estimating them [6].
2 The Method
Estimating of the road accidents cost is quite difficult, as the traffic accident is a
complex phenomenon. In the literature, there are eight methods that are used to esti-
mate or calculate the road traffic accidents: “net output”, “gross output (Human Cap-
ital)”, “life insurance”, “costs awards”, “implicit public sector”, “cost of restitution”,
“cost of value of life” and “willingness to pay” [5]. Each of these methods may be
applied according the available data. Not all countries and governments make available
these data. Applying the various methods listed above leads to the determination or
estimation of different costs, because the methods are based on assumptions. According
to [5], the values of the road accidents estimated using these methods in various studies
Estimating the Costs Caused by Road Traffic Accidents in Romania 891
that have been carried out in developed countries during the period 1965–1978 ranged
from £ 1,500 to £ 20,000,000.
The only relevant methods used in determining or estimating the road accidents
costs are “Gross Output” (this is also called Human Capital - HC) and “Willingness to
Pay (WTP) [7, 8].
In the developed countries, WTP method has replaced the HC method. WTP is very
difficult to apply in developing countries (e.g. Romania) because it is based on complex
questionnaires about the perception of the risk for that population which is willing to
pay in order to improve it. As WTP is based mostly on HC [9], the method that will be
used in this paper will be HC.
All the data used to estimate the costs of the road accidents are related to 2013.
3 Costs Components
3.1 Classification of the Road Accidents and Injured Persons

In order to define the right cost that will be estimated on this paper, a classification of
the road accidents and injured people is needed. This classification will be based on the
significance of the injuries.
The road traffic accident that causes injuries it is that accident that implies at least
one car in motion on a public or private road where the people are allowed to travel and
in which at least one person is killed or injured. Suicides and suicide attempts are
excluded. The total number of accidents Nacc = Gacc + Sacc, where Gacc = the number
of serious (grave) accidents and Sacc = the number of slight accidents.
Fatalities (F): all the persons involved in a road traffic accident who died in
maximum 30 days from the date of the accident. Suicides and suicide attempts are
excluded.
Injured people (I): all the persons involved in road traffic accidents that did not die
in maximum 30 days from the accident date, but suffered injuries that required medical
treatment or hospitalization. Suicides and suicide attempts are excluded. I = Ig + Is,
where Ig is the number of serious injured people and I the number of slight injured
people [10].
In 2013, Nacc = 24.827 [10] and the number of serious accidents is Gacc = 8555
[11]. In this case, Sacc = Nacc- Gacc = 16.272.
The number of fatalities is F = 1861 [10]. In a report regarding the forensic activity,
another number of fatalities F = 1978 [16] was found and this latter number will be
used in this paper. The total number of injured people is I = 31.464 and Ig = 8158 [11].
Using these numbers, we can now calculate Is = I − Ig = 23.306 (Table 1).
3.2 Costs’ Structure

The Human Capital method includes two types of costs:
Direct costs are the costs related to the persons killed or injured in road traffic
accidents (Table 2).
Table 1. Number of victims of the road traffic accidents occurred in 2013.

Accidents 2013 Number
Number of accidents Nacc 24.827
Serious injured Gacc 8.555
Slight injured Sacc 16.272
Number of victims Number
Fatalities F 1.978
Injured I 31.464
Serious injures Ig 8.158
Slight injures Is 23.306
Table 2. Human Capital costs structure.

Costs structure
Direct costs Indirect costs
Medical and hospital costs (MHC) Lost outputs (LO)
Medical emergency services costs (MESC) Human Costs (HC)
Property damage costs (PDC)
Administrative costs (AC)
Other costs (OC)
Indirect costs are the costs related to the losses generated by the lost output of the
killed or injured persons and to the sufferance causes to relatives, friend, families,
communities, etc.
Direct Costs. The following components are included in the direct costs:
Medical and Hospital Costs (MHC). Medical and hospital costs are those costs that are
generated by the hospital and ambulatory care for the crashes’ victims. These costs
include the expenses for accommodation in hospitals, food, medical treatments, sur-
gery, various performed tests (included Imagistics) and other medical services provided
during the medical care processes.
In order to estimate an average for the daily medical and hospital costs, several
hospitals have been interviewed (surgery emergency hospitals, post surgery recovery
hospitals, emergency children hospitals), resulting an average daily cost of 276,87
lei/day. As the average exchange rate in 2013 is 1€ = 4,4190 lei [14], the average daily
cost ADC = 62,66 €/day.
As it was impossible to get the data related to the average number of hospitalization
days for serious injured (AIg) and slight injured (AIs) persons, we make the assumption
that these could be used from [19]. Consequently, AIg = 30 days and AIs = 2 days.
The average medical costs for the seriously injured persons (MCIg) and slight
injured persons (MCIs) are calculated as followings:
MCIg = ADC × AIg = 62,66 €/day × 30 days = 1.879,80 €
MCIs = ADC × AIs = 62,66 €/day × 2 days = 125,32 €
The total medical hospital costs for serious injured people are: MHCIg = Ig ×
MCIg = 8.158 × 1.879,80 € = 15.335.408,40 €
The total medical hospital costs for slight injured people are: MHCIs = Is ×
MCIs = 23.306 × 125,32 € = 2.920.707,92 €
The total medical hospital costs can now be calculated as MHC = MHCIg +
MHCIs = 18.256.116,32 €
Medical Emergency Services Costs (MESC). These are the costs associated with the
Medical Emergency Service for Reanimation and Extrication (SMURD) and ambu-
lance service. The total costs include the expenses for moving the extrication teams and
ambulances from the departure base to the place of accidents and return to the departure
places, the costs of emergency treatments for victims as well as the staff’s wages. This
category of costs includes also the costs for air ambulances.
During 2013, SMURD had 270.348 missions and 5 % of them (13.517,40) were for
traffic accidents [17]. The General Inspectorate for Emergency Situations (IGSU) which
manages SMURD, recorded 311.188 missions, with a budget of 1.367,249 million lei.
The average cost per IGSU intervention was 4.393,65 lei (994,27 €). The total costs for
SMURD missions is CSMURD = 994,27 € × 13.517,40 = 13.439.945,30 €.
When an accident occurs and there are victims, even slight injured persons, an
ambulance is sent out to the place of accident. In 2013, the national ambulance budget
was 672.022.000 lei [17]. The total number of ambulance interventions is 2.723.115.
Consequently, the average costs per ambulance mission is 246,78 lei (55,85 €). As
24.825 accidents have been recorded on 2013, the total ambulance costs are CAMB =
4.827 × 55,85 € = 1.386.587,95 €.
Regarding the air rescue missions deployed at the place of severe accidents, this
service had a budget of 81.041.000 lei for 5.790 flight hours (trainings and rescue
missions). The average cost of a flight hour is 2.128,46 lei/flight hour (481,66 €). The
number flight hours used for air rescue missions were 4.062. Therefore, the total costs
of air rescue missions is CARM = 1.956.502,92 €.
Property Damage Costs (PDC). These are the costs generated by the property damage
as result of road traffic accidents. They include the repair costs for the damaged cars
and all the costs of other damaged properties (roads, traffic signs, security fences,
private houses, etc.).
All these costs are covered by the third party liability insurance (RCA) but also by
CASCO. In 2013, all the costs of property damage in road traffic accidents were
PDC = 3.691.406.650 lei = 835.348.868,53 € [12].
Administrative Costs (AC). These are the costs that arise from the settlement of the
insurance claims. According to the Report of the National Financial Supervising
Authority [12], the total administrative costs are AC = 1.062.483.840 lei
(240.435.356,42 €).
Other Costs (OC). This category includes the funeral costs (FC), technical expertise
costs (TEC) that are compulsory at each crash with injured victims and all those costs
related to the forensic expertises (FC).
In order to estimate the funeral costs, we have to take into account the costs for
funeral ceremonies, the costs of the specialized companies that provides such services,
the cost for a grave and other costs related to the traditional funeral rituals. From the
discussions with several specialized funeral companies, the cost for one deceased
person is 7.000 lei/person (1.584,07 €). As the number of fatalities was 1.978, we can
calculate FC = 1.584,07 € × 1978 fatalities = 3.133.290,46 €.
After each serious road traffic accidents, a technical expertise has to be carried out.
On the market, the average of this such expertise is 1.500 lei (339,45 €). As the total
number of serious accidents is 8.555, TEC = 339,45 € × 8.555 = 2.903.994,75 €. The
National Institute for Forensic Expertise conducted 1.345 expertises during 2013 [18],
but only 498 for road traffic accidents. As the total budget of the institute was
7.428.694 lei (1.681.080,34 €), the average cost for an expertise is 1.249,88 € and the
total costs is FEC = 498 × 1.249,88 € = 622.440,24 €.
Indirect Costs (IC). The following components are included in the indirect costs:
Lost Output (LO). These costs represent the lost productivity by of those persons killed
or injured in road traffic accidents. Actually, they represent the amounts that would be
earned by the persons involved in road traffic accidents if they would not been killed or
injured. These costs depend on the age of the victims, the consequences caused by the
accidents (killed, serious or slight injured) and the average annual income per capita.
These costs must be also adjusted to a discount rate in order to bring these amounts to
the present value of money.
In order to calculate the lost output value of the persons who died in road traffic
accidents, the following parameters are needed: the average age of the victims, the
annual average income per capita and the average age of retirement. As the retirement
age is different for men and women, the calculations will be based on their gender.
For those victims who died in crashes, LO will be calculated using formula (1) [9]:
Xn w
LO ¼ F ð1Þ
0 ð1 þ rÞn
where:
F = the number of fatalities in 2013
w = the average monthly wage.
r = a discount rate.
n = the average number of years that a person could work if he/she would not been
killed a road traffic accident.
According to the Romanian law 6/2013, the 2013 average monthly wage was 2.223
lei/month. Therefore, for one year, a person could earn 2.223 lei/month × 12
months = 26.676 lei (6.036,65 €/year/person). The discount rate is r = 3,3 [13].
n = Ra − Aa, where Ra is the retirement age of the victims and Aa is the victims’
average age when they died in crashes. As the retirement age for women Raw = 63
years is different than retirement age for men Ram = 65 years, we can calculate LO as
the sum of LO of women (LOw) and LO of men (LOm): LO = LOw + LOm. The
women’s average age (Aaw) and men’s average age (Aam) need also to be calculated.
The distribution by age and gender of the victims who died in road traffic accidents
is shown on Table 3 [4].
Table 3. Distribution of the fatalities by age and gender

Male Female
fatalities fatalities
Age No. Age No.
0–14 45 0–14 31
15–17 22 15–17 18
18–24 166 18–24 54
25–49 579 25–49 125
50–64 318 50–64 79
65–85 243 65–85 179
Total 1373 Total 486
In order to calculate the average victim’s wage for women and men, the weighted
arithmetic average method with relative frequencies will be used [15]. This can be
calculated using the following formula:
Xn
Am ¼ i¼1
xifi ð2Þ
The average age for male is Amm = 45 years (Table 4).
Table 4. Age ranges, frequencies, relative frequencies and average age for the male fatalities.
Male fatalities
Age Frequencies Mid range Relative frequencies xifi*
xi fi*
0–14 45 7 3.28 % 0.23
15–17 22 16 1.60 % 0.26
18–24 166 21 12.09 % 2.54
25–49 579 37 42.17 % 15.60
50–64 318 57 23.16 % 13.20
65–85 243 75 17.70 % 13.27
Total 1373 Weighted average 45 years
The average age for females is Amw = 50 years (Table 5). The average number of
years that a man could work if he would not been killed a road traffic accident is:
nm = Ram − Amm = 65 years – 45 years = 20 years.
The average number of years that a woman could work if she would not been killed
a road traffic accident is: nw = Raw − Amw = 63 years – 50 years = 13 years.
Using formula (1),
P we can calculate the lost output for men (LOm) and women
(LOw): LOm ¼ Fm n0 ð1 þw rÞn , where Fm = 1.373 is the number of male fatalities and
n = nm = 20. TheP result is LOm = 119.957.637 €.
LOw ¼ Fw n0 ð1 þw rÞn ; where Fw = 486 is the number of female fatalities and
n = nw = 13. The result is LOw = 30.610.710 €. Therefore the final lost output is
LO = LOw + LOm = 150.568.347 €.
Table 5. Age ranges, frequencies, relative frequencies and average age for the female fatalities.
Female fatalities
Age Frequencies Mid range Relative frequencies xi*fi
xi fi*
0–14 31 7 6.38 % 0.45
15–17 18 16 3.70 % 0.59
18–24 54 21 11.11 % 2.33
25–49 125 37 25.72 % 9.52
50–64 79 57 16.26 % 9.27
65–85 179 75 36.83 % 27.62
Total 486 Weighted average 50 years
Pain Grief and Sufferance Costs (PGSC). The road traffic accidents are affecting not
only the quality of life, but also produce pain, grief and suffering to their families,
relatives and communities which they belong to. In order to quantify these sufferings,
RL (1995) and Silcock and TRL (2003) recommend to add 8 % from the total direct
costs and lost output value in order to cover the lost quality of life. Consequently,
PGSC = 101.444.115,99 €.
Now, the total indirect costs are IC = LO + PGSC = 252.012.462,99 €.
3.3 Total Costs

Every cost that was estimated in this paper and also the total cost are presented in
Table 6:
Table 6. Breakdown of road traffic accidents costs and the total costs estimated in this paper.
Direct costs Medical and hospital costs (MHC) 18,256,116.32 €
Medical emergency services costs (MESC)
Ambulance costs (CAMB) 1,386,587.95 €
SMURD (CSMURD) 13,439,945.30 €
Air rescue missions costs (CARM) 1,956,502.92 €
Property damage costs (PDC) 835,348,868.53 €
Administrative costs (AC) 240,435,356.42 €
Other costs(OC)
Funeral costs (FC) 3,133,290.46 €
Technical expertise costs (TEC) 2,903,994.75 €
Forensic expertise costs (FEC) 622,440.24 €
Total direct costs (DC) 1,117,483,102.89 €
Indirect costs Lost Output (LO) 150,568,347.00 €
Pain grief and suffering 8 % (PGSC) 101,444,115.99 €
Total indirect costs (IC) 252,012,462.99 €
Total road traffic accidents costs Romania 2013 1,369,495,565.88 €
4 Conclusions
In 2013, the costs of the traffic accidents are 0,95 % from the Romanian Gross
Domestic Product. The largest share from the total costs is represented by the direct
costs 81,60 %. From the direct costs, the highest value belongs to Property Damage
Costs, followed by insurance Administrative Costs.
The road traffic accidents in Romania are a burden for the insurance and health
systems. Along with the huge costs generated every year, the road accidents cause
sufferings to the victims’ relatives, families and their communities. Therefore, reducing
accidents costs should be a priority in Romania from now on.
Further researches should focus on conduct costs – benefits analyses, in order to
determine what kind of investments should be done in road safety and how these could
reduce the accidents costs.
Also, the system of collecting data about accidents costs should be improved. The
hospital must have all the information about the accidents costs and further researches
have also to be conducted in this area. Also, the traffic police should be able to estimate
their administrative costs.
Acknowledgments. The authors would like to thank to the Romanian Automobile Club
(ACR) and Emergency Situations Department (DSU) for their support in carry out this research.
References
1. Global Status Report on Road Safety. World Health Organization (2015)
2. Global health estimates. Geneva: World Health Organization (2014). http://www.who.int/
healthinfo/global_burden_disease/estimates/en/index1.html, http://www.who.int/healthinfo/
global_burden_disease/projections/en
3. The Global Burden of Disease: 2004 Update. World Health Organization, Geneva (2008)
4. Road Safety Vademecum: Road Safety Trends, Statistics and Challenges in the EU 2010–
2013; European Commission DG for Mobility and Transport, March 2014. http://ec.europa.
eu/transport/road_safety/pdf/vademecum_2014.pdf
5. Costing Road Accidents in Developing Countries. Overseas Transport Research Laboratory,
UK (1995)
6. European Status Report on Road Safety. World Health Organization (2009)
7. The Casualty Cost of Slight Motorcycle Injury in Surabaya. Hera Widyastuti and Corinne
Muller, Indonesia (2005)
8. Hills, P.J., Jones-Lee, M.W.: The costs of traffic accidents and evaluation of accident
prevention in developing countries. In: PTRC, Summer Annual Meeting. University of
Warwick, 13–16 July 1981 (1981)
9. Silcock, R.: Guidelines for Estimating the Cost of Road Crashes in Developing Countries.
Department of International Development Project R7780, TRL Limited (2003)
10. National Institute for Statistics, Section: Transports
11. Road Traffic Police, National Institute for Prevention: Road Safety Report (2013)
12. Annual Report 2013, The National Authority for Financial Supervising, Bucharest (2014)
13. http://www.oecd.org/tad/xcred/2013-ddr-rev.pdf
14. http://www.bnr.ro/Cursul-de-schimb-3544.aspx
15. Duguleană, L.: Statistics Research. Transilvania University of Brasov, Doctoral School
16. www.legmed.ro, 2013 Forensic Report Activity
17. http://www.igsu.ro/documente/informare_publica/evaluari/evaluare-DSU-2013.pdf
18. http://www.inec.ro/RA2013.pdf
19. Boris, A., et.al: Estimation of the Traffic Accidents Costs in Serbia by Using Dominant Costs
Model. Taylor & Francis (2011). ISSN:1648–4142

Contact 2016

Uploaded by

Copyright:

Available Formats

Contact 2016

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Contact 2016

Uploaded by

Copyright:

Available Formats

What are some of the costs considered in estimating road traffic accidents?

What are some of the costs considered in estimating road traffic accidents?

What percentage of the total costs of road traffic accidents in Romania are direct costs?

What percentage of the total costs of road traffic accidents in Romania are direct costs?

Anghel Chiru · Nicolae Ispas Editors

CONAT 2016 International

ISBN 978-3-319-45446-7 ISBN 978-3-319-45447-4 (eBook)

© Springer International Publishing Switzerland 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The initiation, implementation and completion of activities for preparing, collecting

President of the Scientiﬁc Committee

Coordinators of the Congress Sections

Scientiﬁc Committee Members

Adolfo Senatore Federico II University of Naples, Italy

Deﬁnitions, Acronyms, Abbreviations

Advanced Engineering Methods

Numerical and Experimental Investigations of a Twin Sheet Thermoplastic

Expanded Polypropylene Finite Elements Analysis in Transport

Functional Description and Design Aspects of Hybrid Transmissions. . . . . . . 36

Comparative Study on the Performances of Aerodynamic Devices Used

Analysis of the Influence of the Drive Force Distribution Between Axles

About the Steering Kinematics of the Road Trains . . . . . . . . . . . . . . . . . . . 64

Evaluation by Experimental Methods of the Parameters that Influence

Evaluation of the Behavior in Cornering for Different Classes of Passenger

Structural Dynamic Applications Using Principal Component Analysis

Weight Optimization of Existing Transmission Housing for Heavy Duty

New Approaches in Designing a Race Car Chassis for an Engineering

Presentation of Tests with a Dynamometer Related to the Determination

Validation Procedure for Worldwide Harmonized Light Vehicles Test

Experimental Investigation of a Vehicle Behavior Using Different Complex

Validation of Kinematic Simulation of Sprocket Contacts of Chain Links

Belt Losses Evaluation for a Push-Belt CVT . . . . . . . . . . . . . . . . . . . . . . . 152

Kinematic and Dynamic Study of a Mechanism for a Vehicle Front

Methods for Modeling an Elastic System with Permanent Contour Coupling

Reinvestigation of the Electromagnetic Valve Train (EMVT) Technology

Study on Mixture Formation at an Experimental Spark Ignition Engine . . . . . 187

Aspects Concerning Modeling of Combustion for Compression Ignition

Comparative Study on the Performances of the One-Way Hydraulic Valves . . . 203

Estimation of Boiling Points of Brake Fluids . . . . . . . . . . . . . . . . . . . . . . . 209

Modelling and Optimization of Constructive Form for GO-NOGO Exterior

Systems Engineering Requires Digital Twins of Machine Elements . . . . . . . . 227

Collaborative Design and Use of Interactive Simulations: Boost

Experimental Researches on the Magneto-Rheological Dampers Response

Innovative Solution for Motor Vehicles

Contributions in Experimental Research Concerning Diesel Fuel Supply

Experimental Research Regarding the Possibility of Biofuel Fumigation

Study on a Stratified Charge Spark Ignition Automotive Engine . . . . . . . . . . 274

Researches on Cooling Air Flow Control Devices Using on Cars

Modeling the Energy Evaluation for an Electric Machine . . . . . . . . . . . . . . . 295

Balancing of a Single Stage Reciprocating Compressor with Elastic

The Calculation Algorithm for the Determination of the Temperature

Diagnosing the Operation of a Locomotive Diesel Engine Based on the

Heavy and Special Vehicles

Simultaneous Influences of Tyre Pressure and Steering Geometry upon the

Aspects Regarding the Kinematic Optimization of a Tracked Military

Automobile and Environment

Research on Gasoline Homogenous Charge Compression Ignition (HCCI)

Study of the Diesel Engine Cycle Variability at LPG Fuelling . . . . . . . . . . . 371