Recent Advances in Power Electronics and Drives: Shailendra Kumar Bhim Singh Arun Kumar Singh Editors

Lecture Notes in Electrical Engineering 852
Shailendra Kumar
Bhim Singh
Arun Kumar Singh Editors
Recent Advances
in Power
Electronics and
Drives
Select Proceedings of EPREC 2021
Lecture Notes in Electrical Engineering
Volume 852
Series Editors
Leopoldo Angrisani, Department of Electrical and Information Technologies Engineering, University of Napoli
Federico II, Naples, Italy
Marco Arteaga, Departament de Control y Robótica, Universidad Nacional Autónoma de México, Coyoacán,
Mexico
Bijaya Ketan Panigrahi, Electrical Engineering, Indian Institute of Technology Delhi, New Delhi, Delhi, India
Samarjit Chakraborty, Fakultät für Elektrotechnik und Informationstechnik, TU München, Munich, Germany
Jiming Chen, Zhejiang University, Hangzhou, Zhejiang, China
Shanben Chen, Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai, China
Tan Kay Chen, Department of Electrical and Computer Engineering, National University of Singapore,
Singapore, Singapore
Rüdiger Dillmann, Humanoids and Intelligent Systems Laboratory, Karlsruhe Institute for Technology,
Karlsruhe, Germany
Haibin Duan, Beijing University of Aeronautics and Astronautics, Beijing, China
Gianluigi Ferrari, Università di Parma, Parma, Italy
Manuel Ferre, Centre for Automation and Robotics CAR (UPM-CSIC), Universidad Politécnica de Madrid,
Madrid, Spain
Sandra Hirche, Department of Electrical Engineering and Information Science, Technische Universität
München, Munich, Germany
Faryar Jabbari, Department of Mechanical and Aerospace Engineering, University of California, Irvine, CA,
USA
Limin Jia, State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, China
Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland
Alaa Khamis, German University in Egypt El Tagamoa El Khames, New Cairo City, Egypt
Torsten Kroeger, Stanford University, Stanford, CA, USA
Yong Li, Hunan University, Changsha, Hunan, China
Qilian Liang, Department of Electrical Engineering, University of Texas at Arlington, Arlington, TX, USA
Ferran Martín, Departament d’Enginyeria Electrònica, Universitat Autònoma de Barcelona, Bellaterra,
Barcelona, Spain
Tan Cher Ming, College of Engineering, Nanyang Technological University, Singapore, Singapore
Wolfgang Minker, Institute of Information Technology, University of Ulm, Ulm, Germany
Pradeep Misra, Department of Electrical Engineering, Wright State University, Dayton, OH, USA
Sebastian Möller, Quality and Usability Laboratory, TU Berlin, Berlin, Germany
Subhas Mukhopadhyay, School of Engineering & Advanced Technology, Massey University,
Palmerston North, Manawatu-Wanganui, New Zealand
Cun-Zheng Ning, Electrical Engineering, Arizona State University, Tempe, AZ, USA
Toyoaki Nishida, Graduate School of Informatics, Kyoto University, Kyoto, Japan
Federica Pascucci, Dipartimento di Ingegneria, Università degli Studi “Roma Tre”, Rome, Italy
Yong Qin, State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, China
Gan Woon Seng, School of Electrical & Electronic Engineering, Nanyang Technological University,
Singapore, Singapore
Joachim Speidel, Institute of Telecommunications, Universität Stuttgart, Stuttgart, Germany
Germano Veiga, Campus da FEUP, INESC Porto, Porto, Portugal
Haitao Wu, Academy of Opto-electronics, Chinese Academy of Sciences, Beijing, China
Walter Zamboni, DIEM - Università degli studi di Salerno, Fisciano, Salerno, Italy
Junjie James Zhang, Charlotte, NC, USA
The book series Lecture Notes in Electrical Engineering (LNEE) publishes the
latest developments in Electrical Engineering - quickly, informally and in high
quality. While original research reported in proceedings and monographs has
traditionally formed the core of LNEE, we also encourage authors to submit books
devoted to supporting student education and professional training in the various
fields and applications areas of electrical engineering. The series cover classical and
emerging topics concerning:
• Communication Engineering, Information Theory and Networks
• Electronics Engineering and Microelectronics
• Signal, Image and Speech Processing
• Wireless and Mobile Communication
• Circuits and Systems
• Energy Systems, Power Electronics and Electrical Machines
• Electro-optical Engineering
• Instrumentation Engineering
• Avionics Engineering
• Control Systems
• Internet-of-Things and Cybersecurity
• Biomedical Devices, MEMS and NEMS
For general information about this book series, comments or suggestions, please
contact leontina.dicecco@springer.com.
To submit a proposal or request further information, please contact the Publishing
Editor in your country:
China
Jasmine Dou, Editor (jasmine.dou@springer.com)
India, Japan, Rest of Asia
Swati Meherishi, Editorial Director (Swati.Meherishi@springer.com)
Southeast Asia, Australia, New Zealand
Ramesh Nath Premnath, Editor (ramesh.premnath@springernature.com)
USA, Canada:
Michael Luby, Senior Editor (michael.luby@springer.com)
All other Countries:
Leontina Di Cecco, Senior Editor (leontina.dicecco@springer.com)
** This series is indexed by EI Compendex and Scopus databases. **
More information about this series at https://link.springer.com/bookseries/7818

Shailendra Kumar · Bhim Singh ·
Arun Kumar Singh
Editors
Recent Advances in Power

Electronics and Drives
Select Proceedings of EPREC 2021
Editors
Shailendra Kumar Bhim Singh
Department of Electrical Engineering Department of Electrical Engineering
MANIT Indian Institute of Technology Delhi
Bhopal, India Delhi, India
Arun Kumar Singh

Department of Electrical Engineering
National Institute of Technology Jamshed
Jamshedpur, India
ISSN 1876-1100 ISSN 1876-1119 (electronic)

Lecture Notes in Electrical Engineering
ISBN 978-981-16-9238-3 ISBN 978-981-16-9239-0 (eBook)
https://doi.org/10.1007/978-981-16-9239-0
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Singapore Pte Ltd. 2022
This work is subject to copyright. All rights are solely and exclusively licensed 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, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Contents
Fuzzy Controlled Multi-converter-UPQC for Power Quality

Improvement in Distribution System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
S. Joyal Isac and S. Srinath
Modified BBO-Based PV Integrated Reduced Component
Multilevel Inverter for Mitigating THD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Debanjan Mukherjee and Sourav Mallick
Power Quality Improvement of an Interconnected Grid System
Using PWM Technique of D-STATCOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Manoj Kumar Kar, Sanjay Kumar, and Arun Kumar Singh
Speed Control of a Three-Phase IM with Closed-Loop Control
Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Rajkumar Yadav, Manoj Kumar Kar, and Arun Kumar Singh
A Review on Recent Technologies in Power Electronic Drives
for Hybrid Electric Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Rajakumar Sakile and Umesh Kumar Sinha
Comparative THD Analysis of Multilevel Inverter Using Different
Multicarrier PWM Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Shashi Shekhar Tripathi, Manoj Kumar Kar, and Arun Kumar Singh
Design and Implementation of Two-Phase Interleaved Boost
Converter with Voltage Multiplier for Fuel Cell Hybrid Electric
Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Melisa Miranda, M. Komal, and Shweta P. Hegde
A Unipolar Phase Disposition PWM Technique for Reduced
Switch Count Symmetrical Nine-Level Multilevel DC Link
Inverter Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Rajakumar Sakile, Marco Rivera, Kasoju Bharath Kumar,
Bandela Supriya, and A. Bhanuchandar
v
vi Contents
A Novel Current Control Scheme for Three-Phase Three-Level

Grid-Tied Neutral Point Clamped Inverter . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Kowstubha Palle and Bandela Supriya
Active and Reactive Power Control of Grid-Tied Asymmetrical
MLI-based PV System with Reduced Switching Frequency . . . . . . . . . . . . 119
Hareesh Myneni, A. Pranay Kumar, Somnath A. Mandale,
and Ganjikunta Siva Kumar
Modelling and Control Aspects of STATCOM Connected
to a Grid-Integrated PV System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Koneti Varalakshmi, R. L. Narasimham, and G. Tulasi Ramdas
Comparative Analysis Between Single Diode and Double Diode
Solar Topologies of Solar PV Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Sriparna Das and Kumari Namrata
Control of Three-Phase Grid-Connected Inverter Using dq Axis
Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Deepak Kumar Singh, Saibal Manna, and Ashok Kumar Akella
Active–Reactive Power Support with Optimal Allocation of DG
and DSTATCOM in Distribution System Using Flower Pollination
and Artificial Bee Colony Algorithm with Load Growth . . . . . . . . . . . . . . . 169
Nandola Maitrey Bharatbhai and Atma Ram Gupta
Analysis and Design Optimization of Six-Phase Induction Motor
by Metaheuristic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Pratyush Prasanna Das and S. N. Mahato
Implementation of Complete Vector Control for DFIG Based
Wind Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Rajesh Kumar, Ch. Sekhar, Saibal Manna, and A. K. Akella
TLO Based OPF with FACTS Devices for DC Link Placement
Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
B. Suresh Babu
Performance Comparison Analysis of Energy Management
Strategies for Hybrid Electric Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Jai Kumar Maherchandani, R. R. Joshi, Ritesh Tirole,
Raju Kumar Swami, and Bibhu Prasad Ganthia
Evaluative Study of Cascaded Multilevel Inverter Topologies . . . . . . . . . . 255
Hemant Gupta, Arvind Yadav, and Sanjay Kumar Maurya
Low Speed Performance Improvement of Dual VSI Fed Direct
Torque Controlled Five Phase Open-End Winding Induction Motor . . . . 267
C. Venkata Subba Reddy and Swati Devabhaktuni
Contents vii
Self-balanced Symmetric Source Configuration of Nine Level

Switched Capacitor-Based Grid Connected Inverter with LCL
Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Aratipamula Bhanuchandar and Bhagwan K. Murthy
Priority-Based Charging of Electric Vehicles to Prevent
Distribution Transformer Overloading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Arjun Visakh, T. Sornavel, and M. P. Selvan
Implementation of Three-Phase Hybrid Energy System Integrated
with UPQC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Shravan Kumar Yadav and Krishna Bihari Yadav
Design of Permanent Magnet Brushless DC Motor for Electric
Vehicle Traction Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
Sandeep Vuddanti, Sharankumar Shastri, and Surender Reddy Salkuti
D-FACTS-Based Power Quality Enhancement Using Power
Balance Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
Vikash Anand and S. K. Mallik
Design and Implementation of Electrical Dynamic Braking System
in Electric Car Using Buck Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Walid Alqaisi and Claude Ziad El-Bayeh
A Reduced Device Count of Self Balancing Five-Level Switched
Capacitor-Based Grid-Connected Inverter . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
Pitch Angle Control Using Fuzzy Logic for DFIG-Based WECS . . . . . . . . 373
Kirti Kaharwar and Bhavnesh Kumar
Class-E Power Amplifier-Based Wireless Power Transfer System . . . . . . 387
Jay Prakash Narayan, Anamika Das, and Ananyo Bhattacharya
Enhancement of Power Factor and DC-Link Voltage Stabilization
Using Multiple Techniques of HPFC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
Satya Venkata Kishore Pulavarthi, Jami Rajesh, Nakka Jayaram,
and Sukanta Halder
Design and Modeling of 3.3 kW GaN-Based HPFC Converter
for Onboard EV Charger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
Piyush Kumar, Amit Singh Tandon, Naveen Yalla,
and A. V. J. S. Praneeth
A Novel Thirteen-Level Two-Fold Gain Inverter Topology
with Reduced Voltage Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
M. S. H. Naidu, C. Sadanala, and S. Patnaik
viii Contents
PV-Array-Integrated UPQC for Power Quality Enhancement

and PV Power Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
Harsh D. Patel, Priyank Gandhi, and Pranav Darji
A Fuzzy-Based Buck-Boost Photovoltaic Inverter for Voltage
Stabilization During Mismatched Environmental Conditions . . . . . . . . . . 451
Varun Bhardwaj and Prerna Gaur
Modified Hysteresis Current Control Implementation
for Three-Phase Grid-Connected Inverter . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Mayank Arora and C. Vyjayanthi
A Quasi-Direct Buck-Boost Fuzzy Controlled Rectifier for Rapid
Charging of Electric Vehicle Battery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
Rohan Jain and M. S. Manisha
Comparative Analysis of Controlling Methods for Doubly Fed
Induction Generator Based Wind Energy System . . . . . . . . . . . . . . . . . . . . . 493
Aftab Ahmed Ansari and Giribabu Dyanamina
Adaptive Pitch Control of Wind Turbine Generator to Counter
Transient Contingencies in a Multi Machine System . . . . . . . . . . . . . . . . . . 509
Hailiya Ahsan, Abdul Waheed Kumar, and M. D. Mufti
Detection of Broken Rotor Bar Fault in an Induction Motor
Employing Motor Current Signature Analysis . . . . . . . . . . . . . . . . . . . . . . . 523
Alok Verma, Pratul Arvind, Somnath Sarangi, Jayendra Kumar,
and Anumeha
Performance Analysis of Photovoltaic-Grid Connected System
for Electric Vehicle Charging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531
Manoj Kumar Sharma, Karanbir Singh, and Satish Kansal
Development of the System of Initial Excitation of the Autonomous
Induction Generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
Volodymyr Chenchevoi, Valeriy Kuznetsov, Iurii Zachepa,
Oleksii Chornyi, Olga Chencheva, Vitaliy Kuznetsov,
Rostyslav Yatsiuk, and Olha Luhova
Single Phase Five Level Step-Up Switched Capacitor Based Grid
Connected Inverter with LCL Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
Single Stage Multiple Source Bidirectional Converter for Electric
Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567
Siddhant Gudhe and Sanjeev Singh
Indirect Current Vector Controlled Three-Level Inverter
for Induction Motor Driven Electric Vehicle . . . . . . . . . . . . . . . . . . . . . . . . . 575
Paramjeet Singh Jamwal, Sanjeev Singh, and Shailendra Jain
Contents ix
Comparative Analysis of Conventional and Sliding Mode Control

Techniques for DC-DC Boost Converter for PV System Under
Transient Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587
Pankaj Swarnkar, Suresh Kumar Gawre, and Gagnesh Akodiya
Development of Fault Indicators for Stator Inter-turn Fault
Diagnosis of a Synchronous Generator Using Kalman Filter . . . . . . . . . . . 601
P. V. Sunil Nag and C. Santhosh Kumar
Grid Integration of Wind Energy Conversion System Through
Parallel Inverter Topology for Power Quality Enhancement . . . . . . . . . . . 621
M. E. Aswathi, C. M. Nirmal Mukundan, M. V. Manoj Kumar,
and P. Jayaprakash
Improvement in Quality of Power of Grid-Connected Wind
Energy System Through STATCOM—A Review . . . . . . . . . . . . . . . . . . . . . 627
Manisha Priyadarsani Nayak, Suresh Kumar Gawre,
and Shailendra Kumar
A Computational Intelligence-Based Novel Bearing Defect
Detection Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
Arpana Singh, K. R. Satyajit, and Papia Ray
About the Editors
Dr. Shailendra Kumar is an Assistant Professor in the Department of Electrical

Engineering at MANIT Bhopal, India. He received his M.Tech. and a Ph.D. degree
from the Indian Institute of Technology (IIT) Delhi, India, in 2015 and 2019, respec-
tively. His research interest areas are power quality, grid integration, and microgrid.
Dr. Kumar received POSOCO Power System Awards (PPSA) in 2016 and 2019.
Dr. Kumar has got four patents (filed) and over 80 publications to his credit.
Prof. Bhim Singh is a Professor in the Department of Electrical Engineering at

IIT Delhi. He received his B.E. degree in electrical engineering from the University
of Roorkee (now IIT Roorkee), Roorkee, India, in 1977. He obtained his M.Tech.
degree in power apparatus and systems, and a Ph.D. degree in electrical engineering
from IIT Delhi, in 1979 and 1983, respectively. Prof. Singh has guided 84 Ph.D.
dissertations and 168 M.E./M.Tech./M.S.(R) theses. He has filed 58 patents and has
executed over 80 sponsored and consultancy projects. His areas of interest are solar
PV grid interface systems, microgrids, power quality monitoring, and mitigation,
solar PV water pumping systems, and improved power quality AC/DC converters.
Dr. Arun Kumar Singh is a Professor in the Department of Electrical Engineering,

National Institute of Technology Jamshedpur, India. He received his B.Sc. (Engg.)
from Kurukshetra University, M.Tech. from the Banaras Hindu University, Varanasi,
and a Ph.D. degree in electrical engineering from the IIT Kharagpur, India. He has
over 30 years of teaching experience and research experience. Dr. Singh’s research
areas are control systems, control system applications in different areas, and non-
conventional energy.
xi
Fuzzy Controlled Multi-converter-UPQC
for Power Quality Improvement
in Distribution System
S. Joyal Isac and S. Srinath
1 Introduction
A UPQC derives from the UPFC [1] that is present at the level of distribution. It
is composed of joined shunt and series converters for the consequent adjustment
of current and voltage unbalance available in a supply feeder [2]. It restricts the
reactive, unbalances, and harmonic power that is being demanded by the load. The
best solution to attain a healthy power distribution system is to use the MC-UPQC
[3].
The power electronics-oriented devices draw the reactive and harmonic power
from the supply owing to their inherent non-linearity. In the case of three-phase
systems, they create unbalance and produce a vast amount of neutral currents [4]. The
excessive neutral currents, unbalance, reactive power burden, and injected harmonics
create worst power factor and less system efficiency. Additionally, the power system
is given to several transients such as flickers, swells, and voltage sags. These transients
damage the voltage that is present at the distribution levels [5]. The excessive reactive
power in the loads enhances the transmission losses of the lines and also enhances
the producing capacity of generating stations. Therefore, it becomes necessary to
supply the reactive power at the load ends [6].
Power quality enhancement by means of Artificial Intelligence (AI)-controlled
strategies is the recent research field in the area of custom power devices that are
related to the electrical distribution systems. The intelligent control strategies, which
employ AI approaches like artificial bee colony algorithm, fuzzy logic, and Artificial
Neural Networks (anns) are the best substitutes for the traditional control strategies in
identifying the power quality enhancement in the traditional scenarios [7]. The power
S. Joyal Isac (B)

Department of EEE, Saveetha Engineering College, Chennai, India
e-mail: joyalisac@saveetha.ac.in
S. Srinath
Department of EEE, COE, Velammal Engineering College, Chennai, India
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 1
S. Kumar et al. (eds.), Recent Advances in Power Electronics and Drives, Lecture Notes
in Electrical Engineering 852, https://doi.org/10.1007/978-981-16-9239-0_1
2 S. Joyal Isac and S. Srinath
quality identification is composed of performing load balancing, eradicating the

harmonics that occur because of the non-linear loads, and handling the voltage quality
with respect to the constant frequency and load imperfections [3, 8, 9]. Nowadays,
due to the enhancement in the non-linear power electronic equipment usage, power
quality seems to be the main conflict, and hybrid AI approaches are in the area of
research for the UPQCS.
The major improvement of the paper is as follows.
• To manage the power quality in the distribution system by means of a novel
controller-oriented MC-UPQC.
• To implement the control schemes of MC-UPQC by means of the fuzzy theory
with SRF, thereby reducing the THD.
• To evaluate the effectiveness of the developed fuzzy controlled-MC-UPQC by
comparing it with the existing models.
The organization of the paper is shown as Sect. 1 provides the introduction of
the MC-UPQC. The literature related works of the MC-UPQC are described in
Sect. 2. The enhanced MC-UPQC for power quality improvement is explained in
Sect. 3. Section 4 provides the fuzzy controller-oriented MC-UPQC for power quality
improvement. The results and discussions are explained in Sect. 5. Section 6 ends
the paper.
2 Literature Survey
2.1 Related Works
In 2018, Nagireddy et al. [10] have proposed a model of the hybrid fuzzy backprop-
agation control strategy. The reference currents were described with the help of the
back-propagation algorithms that contain the load and source currents as the input
control parameters. This investigation was done in a multilevel UPQC. The outcomes
of the mitigation of the voltage sag, load balancing, dynamic performance, and total
harmonic distortion were examined by means of the MATLAB/Simulink.
In 2015, Boddepalli and Sangameswara [11] have addressed the control and design
operation of MC-UPQC. A series VSC was included in the nearby feeder. The device
was joined among multiple feeders that arise from distinct substations. The combined
system was used to reduce the voltage and the current fluctuations. The control
schemes used here were the p-q theory and d-q method that were composed of
extended mathematical modeling. The artificial intelligence techniques were used to
reduce the harmonics by means of the MATLAB/SIMULINK Software.
In 2013, Boddepalli and Raju [12] have developed a new structure for a three-
phase four-wire (3P4W) distribution system that employed the MC-UPQC. It also
proposed a novel control scheme for handling the unbalanced load currents. The
Fuzzy Controlled Multi-converter-UPQC for Power … 3
simulation outcomes revealed the efficiency of the MC-UPQC-oriented 3P4W distri-

bution system. The control schemes were based on Neuro-Fuzzy controller, FUZZY,
and PI of the MC-UPQC.
In 2015, Mahanty [13] has introduced a new MC-UPQC for continuous adjustment
of current and voltage in multifeeder/multibus distribution systems. A MC-UPQC
joined multiple apfs and one shunt APF. MC-UPQC compensated load current unbal-
ances and supply voltage on the main feeder and complete adjustment of supply
voltage unbalances on the remaining feeders. Here, a two-feeder MC-UPQC was
employed that was composed of two series apfs and one shunt APF. The entire apfs
shared a common DC-link capacitor. A new hybrid fuzzy-PI controller was modeled
that was non-linear and robust to parameter variations and was composed of a quick
dynamic response together with the superior steady-state response. The outcomes
showed that the MC-UPQC compensated load current and supply voltage imper-
fections on feeder 1 and completely protected the critical/sensitive load on feeder 2
against the interruption, swell/sag, and distortion.
In 2015, Gaikwad et al. [14] have proposed a MC-UPQC that has the capability
of consequent compensation for the current and voltage in multifeeder/multibus
systems. It consisted of multiple VSCS and one shunt VSC. The power compensated
for the interruption and swell/sag. It was simulated in MATLAB Simulink and power
transfer was compared among the two nearby feeders for the interruption compen-
sation. It returned minimum Total Harmonic Distortion (THD). The major problem
solved was the voltage dip.
2.2 Review
The MC-UPQC offers high current and voltage, minimizes the cost since series
transformer is not necessary, easier capacity expansion, redundancy, and working at
maximum rating, etc. But, it lacks from high conduction loss, centralized approach,
difficult capacity expansion, etc. These challenges must be handled quickly. Few
features with the challenges are shown in Table 1. Hybrid fuzzy-backpropagation
control [10] regulates the DC voltage without any undershoot or overshoot beneath
abnormal conditions and also estimates the gating signals for the shunt as well as
the series VSCS of the multilevel UPQCS. But, within distinct levels, voltage unbal-
ance occurs. NFC [11] achieves better compensation functions and also reduces
the THD with the current and the voltage. Still, it cannot be utilized in multilevel
operations. Power quality theory [12] frees the load currents from the distortion
and can be executed with the help of simple analog hardware. Yet, it returns less
efficiency near the nominal operating point. Hybrid Fuzzy-PI control [13] can be
extended to multifeeder/multibus distribution systems with the addition of several
series apfs and is robust to parameter variations, non-linear, and is composed of
quick dynamic response together with steady-state performance. But, high DC-link
losses are resulted. Novel multifeeder distribution system approach [14] shares the
power compensation capabilities among the two nearby feeders and also handles the
Table 1 Features and challenges of state-of-the-art MC-UPQC methods

Author [citation] Methodology Features Challenges
Nagireddy et al. Hybrid • The gating signals are • Voltage
[10] fuzzy-backpropagation estimated for the shunt as unbalance
control well as the series VSCS happens among
of the multilevel UPQCS distinct levels
• It regulates the DC
voltage without any
undershoot or overshoot
beneath abnormal
conditions
Boddepalli and NFC • It minimizes the THD • It cannot be
Sangameswara [11] with the current and the utilized in
voltage multilevel
• It achieves better operations
compensation functions
Boddepalli and Power quality theory • It can be executed by • Less efficiency
Raju [12] means of simple analog is returned near
hardware the nominal
• The load currents are free operating point
from the distortion
Mahanty [13] Hybrid Fuzzy-PI • It is robust to parameter • It results in high
control variations, non-linear, and dc-link losses
is composed of quick
dynamic response
together with steady-state
performance
• It can be extended to
multifeeder/multibus
distribution systems with
the addition of several
series apfs
Gaikwad et al. [14] Novel multifeeder • It handles the problem of • It produces
distribution system voltage dip heavy and bulky
approach • It shares the power dc inductor
compensation capabilities
among the two nearby
feeders
voltage dip problem. Still, heavy and bulky dc inductor is produced in some cases.
Hence, these challenges are acted as a motivation in developing a fuzzy controlled
MC-UPQC for power quality improvement.
Fig. 1 Diagrammatic illustration of MC-UPQC
3 Power Quality Improvement Using Improved MC-UPQC
3.1 MC-UPQC
The reactive current coming from the source is likely to be in-phase with the currently
present feeder voltages. The feeder 1 is linked to a non-linear load and feeder 2 is
linked to a linear load, hence both these feeders are joined in the MC-UPQC. This is
used to handle the power quality related problems. As non-linear load is present in
feeder 1, it gets affected by the unbalanced currents/voltage, harmonic distortions,
and the feeder sources. Feeder 2 is composed of linear load, and so it is not composed
of any interruption, current/voltage unbalance, harmonic distortion, swell, and sag.
Thus, both the feeders do not contain any effects. The two series VSCs are linked by
means of a series transformer. The switching harmonics get rejected by power RC
high pass filter that contains a communication reactor L with the complete VSCs.
It is done by verifying the single strategy of the MC-UPQC with the FLC. The
diagrammatic model of the MC-UPQC is portrayed in Fig. 1.
3.2 Control Strategies
The suggested method generates the reference signals for the shunt as well as series
voltage source converter. The current as well as the voltage unbalance, voltage sags,
load currents of feeders, reactive and harmonic component, swell and harmonics,
and source voltage distributions are extracted by the control technique.
The control algorithm that is associated with the shunt VSC block is explained
here. The shunt VSC is composed of the MSRF theory and Fuzzy Logic Classifier
(FLC) technique. Synchronization is achieved by a PLL with the help of the supply
voltage. At every phase, it calculates the 120° phase displacement. Depending on the
unit vector template, the shunt VSC follows the MSRF theory approach. The phase
angle
is extracted using three independent two-phase system that is denoted by lag or
2. The theory exists independently for the three-phase of the entire phase system.
It is displayed in Eqs. (1) and (2).
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
i n−d Id i n−a
⎣ i n−q ⎦ = ⎣ Iq ⎦ ⎣ i n−b ⎦ (1)
i n−0 I0 i n−c
⎡ ⎤ ⎡ ⎤⎡ ⎤
Id sin ωz sin ωz − 2π3
sin ωz + 2π
3
Ia
2
⎣ Iq ⎦ = ⎣ cos ωz cos ωz − 2π cos ωz + 2π ⎦ ⎣ Ib ⎦
3 3
(2)
3 1 1 1
I0 2 2 2
Ic
The primary direct axis component current is given to the DC as portrayed in

Eqs. (3) and (4).
r ce
i g−d = i nd + Idc (3)
r ce
i g−q = i n−q (4)
The power released from the DC-link capacitor minimizes the average value that
is present in the DC bus voltage. These are given to the FLC that is used to lessen the
error between the measured as well as the desired capacitor voltage. The controlling
signal output is used as input to the shunt VSC’s current control system. The power
is returned from the source and the DC capacitor voltage is stabilized with the help
of the shunt VSC. Hence, feeder 1 does not contain harmonic as well as reactive
component. This behavior is shown in Eq. (5).
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
r ce r ce
i g−a Ia i g−d
⎢ r ce ⎥ ⎣ ⎦ ⎢ r ce ⎥
⎣ i g−b ⎦ = Ib ⎣ i g−q ⎦ (5)
r ce r ce
i g−c Ic i g−0
The addition of shunt currents to the abc reference currents takes place. The
controlling of currents occurs by the sensation of the reference frame currents. The
current in the shunt VSC is given to the controller part. The series VSC is composed
of the MSRF theory as well as the enhanced PWM generator. The series VSC block
is operated by the MSRF theory. This performance is depicted in Eqs. (6) and (7).
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
vn−d vd vn−a
⎣ vn−q ⎦ = ⎣ vq ⎦ ⎣ vn−b ⎦ (6)
vn−0 v0 vn−c
⎡ ⎤ ⎡ ⎤⎡ ⎤
vd sin ωz sin ωz − 2π3
sin ωz + 2π
3
va
2
⎣ vq ⎦ = ⎣ cos ωz cos ωz − 2π cos ωz + 2π ⎦ ⎣ vb ⎦
3 3
(7)
3
v0 1
2
1
2
1
2
vc
The load voltage is kept sinusoidal with the stable amplitude. Equation (8) shows
the subtraction of the forecasted load synchronous reference dq0 voltages from the
Vm−dq0 . This is depicted in Eq. (9).
⎡ ⎤ ⎡ ⎤ ⎡ exp o ⎤
vg−d
r ce
vn−d v
⎢ r ce ⎥ ⎣ ⎦ ⎢ g−d
exp o ⎥
⎣ vg−q ⎦ = vn−q ⎣ vg−q ⎦ (8)
exp o
vg−0
r ce
vn−0 vg−0
⎡ ⎤ ⎡ ⎤−1 ⎡ ⎤
vtg−a
r ce
vd vg−d
r ce
⎢ r ce ⎥ ⎣ ⎦ ⎢ r ce ⎥
⎣ vtg−b ⎦ = vq ⎣ vg−q ⎦ (9)
vtg−c
r ce v0 vg−0
r ce
4 Fuzzy Controller-Based MC-UPQC for Power Quality

Improvement
4.1 Proposed Fuzzy Controller
The main concept of the developed technique is to design a MC-UPQC that can
perform the power quality issues present in the distribution system. This technique
works on the principle of the SRF theory. Normally, the MC-UPQC consists of two
series VSC, in which the power transformation from other feeders eliminates the
“voltage sag, swell, interruption, and transient response of the system”. The Fuzzy
logic controllers are very cheaper to develop. It is very useful for Natural language
processing. It is also used for the modern control systems. Here, the power quality is
managed by the novel controller-oriented MC-UPQC. The control strategies related
to the MC-UPQC are implemented by the fuzzy theory that contains the SRF. The
major aim of the fuzzy controlled-MC-UPQC is to reduce the THD. The effectiveness
is measured by comparing it over the existing models. The block diagram for fuzzy
controller is shown in Fig. 2.
FLC [15, 16] models a system by means of the MC-UPQC with the concept of
fuzzy theory. The logical variables are examined as input values that consider the
continuous variables from 0 to 1. The controllers offer the linguistic schemes by
Control Signal
Output signal
Controlled
system
Fuzzification
Defuzzificatio Decision interference
n interface Making logic
Knowledge
base
Fig. 2 Block diagram for fuzzy controller
means of the expert knowledge. FLC was introduced by Professor Lotfia Zadeh at
California University in the year 1965. The inaccurate data is processed here. The
applications cannot be revealed without the computers as well as the controllers.
The controller operation works on the principle of the fuzzy rules that are made by
the fuzzy set theory. Fuzzy controller compensates the power quality problem. The
phases available in the fuzzy controller are the “decision-making, defuzzification,
and fuzzification”. In fuzzification, the crisp value is altered to the fuzzy value. It is
got by several types of fuzzifier. The fuzzy set shapes exist as triangular, trapezoidal,
etc. The fuzzified process output is produced by the rule creation. The FLC model
returns the dynamic performance of the small and the large signal, but it is impossible
with the linear control approach. The FLC input represents the alteration of error as
well as the voltage error. The membership functions are assumed to be triangular.
The fuzzy approach is assumed to be the area center.
The knowledge base rules and the database output create the inference relation C S
as displayed in Eq. (10). The variables are assumed to consist of fuzzy set description.
The fuzzy memberships are shown in Eq. (10).
C S (Q O) = I F U1 is G H1 AN D U2 is G H2 · · · Uoi is G Hoi
(10)
then V is DS (Q O)
In the above equation, the term Q O = 1, 2, 3 · · · O I , where O I denotes the

rule count, G H1 , G H2 , · · · G Hoi defines the fuzzy sets, oi defines the fuzzy variable
count, U1 , U2 , · · · Uoi defines the input variable vector, and V defines the output
variable that is also called as control variable. The fuzzy controller computes the
input signals to define the efficient control action. The FLC is implemented by the
two input state variables as portrayed in Eq. (11).
verr f uzz = vdc − vdc

r ce
(11)
The fuzzy control rules are used for Vdc and Vdc . It regulates the voltage using
less real loss quantity that is being taken as the output of FLC. The MSRF-oriented
currents are provided to the relay and a sensing takes place in the shunt VSC control
circuit.
5 Results and Discussions
5.1 Simulation Setup
The proposed fuzzy controller MC-UPQC was performed in MATLAB 2019a and
the tests were carried out. The transmission of power from feeder 1 to feeder 2 was
done to eliminate the “voltage sag, swell, interruption and transient response of the
system”. The proposed method was compared with distinct models like “MC-UPQC,
FLC-MC-UPQC, and NN-MC-UPQC” for describing that the THD is reduced in the
developed method.
5.2 MC-UPQC Performance in Connection with Feeder 1

with Different Controllers
The proposed method was compared with the traditional models that are connected
with feeder 1 as portrayed in Fig. 3. The outcomes revealed the “bus 1 voltage, series
compensation voltage and load 1 voltage”. The source voltage is provided with sag
from 0.1 to 0.2 s and it tends to swell from 0.2 to 0.3 s. If fault happens in feeder
1, then there occurs an effect in feeder 1 with voltage across the linear load having
interruptions, swell, and sag. This problem is handled by connecting an FLC to the
shunt VSC. This is efficiently handled by the FLC than the remaining controllers.
5.3 MC-UPQC Performance in Connection with Feeder 2

with Different Controllers
The proposed method is compared over the state-of-the-art models that are linked
with the feeder 2 as in Fig. 4. It clearly portrays the “bus 1 voltage, series compensa-
tion voltage and load 1 voltage”. The disturbances and voltage sag swell are reduced
by joining the system with the two-feeder system. The source voltage is applied that
contains the sag from 0.1 to 0.2 s and swelling from 0.2 to 0.3 s. If fault occurs in
feeder 2, then the effect of voltage occurs along the linear load that contains the inter-
ruptions, sag, and swell. This problem is solved by linking the FLC with the shunt
Fig. 3 Analysis of proposed and existing MC-UPQC a General MC-UPQC system, b NN-MC-
UPQC, and c FLC-MC-UPQC for “Bus 1, series compensation, and load 1 voltages in feeder
1”
VSC. It is effectively portrayed in the load voltage that is solved by the proposed
method in an efficient manner than the various controllers.
5.4 Harmonic Analysis
The harmonic analysis of the proposed method over the traditional models that are
linked with the feeder 1 and feeder 2 is portrayed in Fig. 5, respectively. It portrays
the THD of the general MC-UPQC, FLC-MC-UPQC, and NN-MC-UPQC at distinct
seconds. Here, the THD of the developed method is minimized over the time sequence
when compared with the existing models. At 0.15 s, the THD of the proposed method
Fig. 4 Analysis of proposed and existing MC-UPQC a General MC-UPQC system, b NN-MC-
UPQC, and c FLC-MC-UPQC for “Bus 1, series compensation, and load 1 voltages in feeder
2”
is 33.33 and 20% better than general MC-UPQC and NN-MC-UPQC system. There-
fore, MC-UPQC reduced the THD in the developed method than the traditional
models when it is joined with the feeder 1 and feeder 2.
Fig. 5 Harmonic Analysis of proposed and existing MC-UPQC a General MC-UPQC system, b
NN-MC-UPQC, and c FLC-MC-UPQC for feeder 1 and feeder 2
6 Conclusion
This paper has managed the power quality in distribution system by means of a new
controller-oriented MC-UPQC. The fuzzy theory containing the SRF implemented
the control schemes of MC-UPQC. As a major objective, the fuzzy controlled-MC-
UPQC minimized the THD. MC-UPQC is used to mitigate voltage and current
harmonics and for improving the voltage regulation and to compensate reactive
power. The effectiveness was described by comparing it with traditional models.
From the analysis, at 0.15 s, the THD of the proposed method was 33.33% and
20% better than general MC-UPQC and NN-MC-UPQC system. Hence, MC-UPQC
reduced the THD in the proposed method than the traditional models when it was
combined with the feeder 1 and feeder 2.
References
1. Mohammadi HR, Varjani AY, Mokhtari H (2009) Multiconverter unified power-quality

conditioning system: MC-UPQC. IEEE Trans Power Deliv 24(3)
2. Rezaeipour R, Kazemi A (2008) Review of Novel control strategies for UPQC. Int J Electr
Power Eng 2(4):241–247
3. Basu M, Das SP, Dubey GK (2007) Comparative evaluation of two models of UPQC for suitable
interface to enhance power quality. Elect Power Syst Res 821–830
4. Khadkikar V, Chandra A (2009) A Noval structure for three-phase four-wire distribution system
utilizing unified power quality conditioner (UPQC). IEEE Trans Ind Appl 45(5):1897–1902
5. Abdelkhalek O, Benachaiba C, Gasbaoui B, Nasri A (2010) Using of ANFIS and FIS methods
to improve the UPQC performance. Int J Eng Sci Technol 2(12):6889–6901
6. Ghosh A, Ledwich G (2001) A unified power quality conditioner (UPQC) for simultaneous
voltage and current compensation. Elect Power Syst Res 55–63
7. Zerikat M, Chekroun S (2007) Design and implementation of a hybrid fuzzy controller for a
high perforance induction motor. World Acad Sci Eng and Technol 20:263
8. Sajedi S, Khalifeh F (2011) Modeling and application of UPQC to power quality improvement
considering loading of series and shunt converters. Aust J Basic Appl Sci 5(5):300–309
9. Aredes M, Heumann K, Watanabe EH (1998) An universal active power line conditioner. IEEE
Trans Power Del 13(2):545–551
10. Nagireddy VV, Kota VR, Kumar DVA (2018) Hybrid fuzzy back-propagation control scheme
for multilevel unified power quality conditioner. Ain Shams Eng J 9(4):2709–2724
11. Boddepalli R, Sangameswara P (2015) Performance analysis of MC-UPQC using artificial
intelligence. Int J Emerg Eng Res Technol 3(6):141–156
12. Boddepalli R, Raju PS (2013) Power quality improvement using multi converter unified power
quality conditioner with Pq theory. Int J Electr Electron Eng Res 3(2):185–201
13. Mahanty R (2015) Hybrid fuzzy-PI controller based MC-UPQC for power quality enhance-
ment. Int J Adv Res Electr Electron Instrum Eng 4(11)
14. Gaikwad SN, Singh H, Jha RK (2015) MC-UPQC-A novel approach in multifeeder distribution
system for improvement of power quality issues. IJARIIE 1(5):2395–4396
15. Rajani B, Raju PS (2013) Neuro-fuzzy controller based multi converter unified power quality
conditioner with PQ theory. Int Electr Eng J 4(1):926–938
16. Paduchuri CB, Dash SS, Chinnamuthu S (2014) A new control strategy based multi converter
UPQC using fuzzy logic controller to improve the power quality issues. Power Eng Electr Eng
12(2)
Modified BBO-Based PV Integrated
Reduced Component Multilevel Inverter
for Mitigating THD
Debanjan Mukherjee and Sourav Mallick
1 Introduction
Over the recent years, research fraternity in power electronics domain has been
engrossed toward the modification and extensive usage of multilevel inverters
(MLIs). Though the MLI was first developed almost three decades ago, it has gained
popularity over the time for its wide applications in various fields, viz., renew-
able energy source integration, electric drives, Flexible AcTransmission Systems
(FACTS) compensation, electric vehicles, medium or high-power industrial appli-
cations, etc. Total Harmonic Distortion (THD) in output voltage, electro-magnetic
interference (EMI), and total standing voltage (TSV) are high in case of the usual
two-stage inverter [1]. These disadvantages of such inverters have been alleviated
with the advent of MLI. There are mainly three kinds of MLI, i.e., Cascaded H-
Bridge (CHB) type, Neutral Point Clamped (NPC) type, and Flying Capacitor (FC)
type. A large number of component requirements of NPCMLI for producing higher
voltage level has restricted its usage. On the other hand, capacitor voltage tuning
problem increases complexity of circuitry of FCMLI, thereby, increasing the size
and cost. The CHBMLI has won the attention of the industry for such drawbacks of
the NPCMLI and FCMLI [2]. Furthermore, the provision of isolated DC sources in
CHBMLI topology is also advantageous for integrating the renewable energy sources
[3].
With the fast depletion of fossil fuels, usage of renewable sources is increasing to
satisfy the world’s growing demand of power. Solar energy, being the largest source
of renewable energy, is under the consideration. The solar irradiation when fall on
photovoltaic cell (PV cell) releases electrons which creates a potential difference and
D. Mukherjee (B) · S. Mallick

National Institute of Technology Sikkim, Ravangla, South Sikkim 737139, India
S. Mallick
e-mail: s.mallick@nitsikkim.ac.in
16 D. Mukherjee and S. Mallick
the solar energy is thus trapped in the form of electrical energy. But, as the solar irradi-
ation varies with time, constant power is not available. Hence, a battery is connected
with PV cell in such a fashion that the battery will get charged when solar radiation is
available and will deliver power in the absence of solar power. The delivery of power
from the battery can be possible if it is connected to electrical grid through inverter.
Hence, the PV systems are being integrated with the MLIs. Moreover, the traditional
MLI configuration having large number of switches becomes bulky and expensive
with low efficiency. The recent state-of-the-art literatures emphasized on various MLI
topologies with reduced number of devices to increase efficiency [4–10]. The recent
research trend also aims at improving the quality by reducing harmonics. To avoid
switching loss, low frequency switching is preferred over high frequency switching
[11]. To overcome the lacunae of available inverter topologies, a reduced component
count topology of 5-level MLI with integration of PV system is proposed in this
work. To increase the output power quality, suppression of the lower order harmonic
components is essential which can be aptly done by the Selective Harmonic Elimina-
tion (SHE) technique. The SHE technique can efficiently mitigate the selected lower
order harmonics by resolving the trigonometric transcendental equations which are
nonlinear in nature [12]. Traditional iterative methods like Newton Raphson (NR),
Resultant Theory (RT), etc. are not so good in solving the nonlinear problems as
they are sensitive to the initial guess and involve complex mathematical expressions
[13–15]. Hence, the SHE problem is generally solved by different bio-inspired meta-
heuristic techniques efficiently [16]. The modified biogeography-based optimization
(MBBO) technique is developed in this work for improving the output quality by opti-
mizing firing angles and its performance is compared by using popular metaheuristic
techniques like Butterfly Optimization Algorithm (BOA) [17], BBO [18], gravita-
tional search algorithm (GSA) [19], Particle Swarm Optimization (PSO) [20]. All the
simulations are experimentally validated using Hardware In Loop (HIL) setup which
has high security, fidelity, and effectiveness [21–23]. In this work, the PV system
has been designed according to the datasheet of commercially available SLP080-12
[24, 25].
In the following section of the paper, PV system and the proposed MLI topology
are discussed sequentially. Then, in the subsequent section, the SHE scheme is
conceptualized followed by the discussion on the BBO and the proposed techniques.
In the next sections, simulation and hardware in loop details are described consecu-
tively. At last, the conclusion section is presented after thorough discussions of the
available results in the result section.
2 Description of the Photovoltaic System
Advantages of photovoltaic system as a viable source of energy motivate this work

to utilize it. A commercial PV panel, viz., SLP080-12 [24, 25], is replicated for
simulation in MATLAB and HIL in this research work. The essential functional
parameters of that replicated PV system are presented in Table 1. In that modeled
Modified BBO-Based PV Integrated Reduced Component Multilevel … 17
Table 1 Description of the

Parameters SLP080-12
PV system parameters
Number of parallel strings 3
Modules in series per string 1
Open circuit voltage VOC (V) 21.5
Voltage at maximum power point VM (V) 17.2
Short circuit current ISC (A) 5.17
Current at maximum power point IM (A) 4.65
Temperature co-efficient of VOC (%/◦ C) −0.31
Temperature co-efficient of ISC (%/◦ C) 0.03
Cells per module 36
Input parameters are irradiance = 1000 w/m2 and temperature =
25 °C
PV system, 3mF DC link capacitor has the initial voltage 0f 21.5 V which is same
as the output voltage of PV panel.
In order to get good quality AC from the PV, proper MLI scheme is needed to
be used; otherwise, large number of harmonics will distort the output voltage and
current. There are several MLI topologies used to get proper output. The traditional
MLI suffers from the losses due to the presence of large number of switches and driver
circuits. Hence, the necessity of the reduced switch MLI has got the attention of the
researchers. In this work, the authors proposed to integrate the modeled PV system
with a reduced switch MLI topology without sacrificing the efficiency. Detailed
discussion regarding the proposed topology is presented in the following section.
3 Proposed Topology of Single Phase 5-Level MLI
To generate 5-level voltage output, the proposed MLI scheme requires 1 diode (D1 ),
5 switches (T1 to T5 ), an isolated DC source (VDC ) and a PV system with the twice
voltage output that of the DC source as shown in Fig. 1. D1 is used to make unidirec-
tional current flow. Presentation of switching states to generate 5-level output voltage
in the proposed topology is shown in Table 2 where 1 and 0 indicate ON state and
OFF state of the switches, respectively.
Thus, the proposed topology is able to produce the desired 5-level staircase output
voltage which is similar to that of the traditional MLI. But due to the switching
of MLIs, output voltage is contaminated with harmonics. Due to those harmonics,
voltage quality and system’s efficiency are dropped down. Therefore, harmonics
suppression is very essential to achieve the required efficiency. The above objective
can be fulfilled by utilizing optimized switching. Adaptation of the SHE scheme can
fulfill the above-mentioned goal.
Fig. 1 Proposed topology of single phase 5-level PV integrated MLI with 5 switches
Table 2 Switching states of the proposed MLI for producing 5-level output voltage
T1 T2 T3 T4 T5 D1 Vout
0 1 1 0 1 0 −2Vdc
0 1 1 0 0 1 −Vdc
1 0 1 0 0 0 0
1 0 0 1 0 1 Vdc
1 0 0 1 1 0 2Vdc
4 Selective Harmonic Elimination as an Optimization

Problem
Expression of output voltage by expanding through Fourier series is necessary for

realizing the theory of SHE [2] which is considering quarter wave symmetry as
presented in Eq. (1); whereas, the peak magnitude of voltage at kth harmonic is
expressed by Eq. (2).
∞

V (ωt) = Vk sin(kωt) (1)
k=1,3,5
4VDC
Vk = (cos(kα1 ) + cos(kα2 )) (2)
kπ
subjected to, 0 < α1 < α2 < 90◦ .
Where α1 , α2 are the firing angles for generating the 5-level output, VDC is the
voltage of single DC source in symmetric MLI. According to SHE, the transcendental
equations are used, presented in (3) and (4) to suppress the kth harmonic at a specific
Modulation Index (MI).
1
(cos α1 + cos α2 ) = M I (3)
2
(cos(kα1 ) + cos(kα2 )) = 0 (4)
where MI can be computed as Eq. (5).
π Vdesir ed
MI = (5)
4Nsour ce VDC
where Vdesir ed is the desired voltage magnitude at fundamental frequency.

Thus, Eq. (6) provides the objective function (OF) derived using the SHE problem
and applied in the various optimization techniques used in this paper.
4 2
Vdesir ed − V1
13
1 Vk
O F = min 2×S× + S× (6)
Vdesir ed k=3
k V1
subject to, 0 < α1 < α2 < 90◦ and S is a positive non-zero integer and k to be an
odd integer, where V1 is the voltage magnitude of fundamental frequency.
In this work, the chosen value of S is 50. The first term of the above equation
ensures fundamental voltage magnitude with 1% error; whereas, the second term
helps to suppress the voltage magnitudes of lower order harmonics from as 3rd to
13th. In OF, (1/k) is the weight factor. Thus, Eq. (6) satisfies the IEEE-519 harmonic
standard. As Eqs. (1)–(4) contain nonlinear elements, metaheuristic techniques are
well-suited to find the optimal solution. In this work, biogeography-based optimiza-
tion (BBO) is initially used to find the optimal solution. The following subsection
describes the BBO technique.
5 Brief Discussion on Biogeography-Based Optimization
The biogeography-based optimization (BBO) technique was developed by D. Simon

in the year 2008. This metaheuristic technique is inspired from the natural way of
species accumulation phenomena in biogeography. Natural tendency of the species is
to move in the habitat which is favorable for inhabitants. The surrounding factors that
attract the species toward habitat are temperature, weather, humidity, etc. In BBO, the
said factors are termed as suitability index variables (SIVs); whereas, the particular
habitat quality is named as habitat suitability index (HSI). The species are moved
from one habitat to other according to migration process which involves the concepts
of emigration and immigration procedures. The species are further upgraded through
mutation process. The mutation phase helps to break the trap of local optima. The
flow chart, presented in Fig. 2, demonstrates the concept of BBO. This technique
may provide good result but still having limitation of local optima stuck problem
which have been reduced by modifying its mutation phase in the proposed technique.
Solution update
Random Set
according to emigration
initialization iteration
and immigration
of population count=1
processes
NO
Mutation phase
Maxmum Obtain the best
of solution YES
iteration solution
update
Fig. 2 Flow chart of BBO technique
6 The Proposed Technique
In the mutation phase of BBO technique, few solutions are updated randomly within
the search space but the rest are kept intact which may cause local trapping problem.
To get rid of such issue of BBO, modification of mutation phase is introduced in
this work. Firstly, the mutation rate (mu) is calculated for all species [18]. If the
mutation rate becomes less than a generated random number, then a new species will
be generated randomly within the search space. The authors propose here to use the
rest of the species (solutions) which will be updated with random walk surrounding
the best species. The step length of that random walk is derived from the exponential
distribution in Eq. (7).
S ∼ E x p(l) (7)
where l is the decay parameter.
Hi (S I V ) = η ∗ S*(Hi (S I V )−Hbest (S I V )) (8)
where Hbest (S I V ): The best species so far, η: scale factor to adjust the step length
according to the requirement of the problem.
Species updating approach using Eqs. (7, 8) explores the local search surrounding
the best species so far. In addition, the extended tail of exponential distribution curve
creates the probability of generating long step which can overcome the local trapping
issue. Thus, the balanced exploration and exploitation abilities are achieved which
helps to provide better solution and speed of convergence in the proposed MBBO
technique than the original one. Proposed mutation phase is realized by the pseudo-
code in Fig. 3. Here, N, D are the species number and dimension of each species,
respectively.
Fig. 3 Pseudo-code of the modified mutation phase in the MBBO technique
6.1 Flowchart for Firing Angle Optimization
The proposed and the other metaheuristic techniques are employed to optimize the
firing angles of 5-level MLI in different modulation indexes. The entire process is
depicted in the following flowchart (Fig. 4).
7 Hardware in Loop (HIL)
Hardware in loop is the real-time emulator to mimic the actual experimental setup.
Due to high security and fidelity, HIL is being used in recent research works to
validate the simulation [21–23]. This work employed Typhoon HIL [22] to vali-
date the simulated work. The PC is interfaced with the device (HIL-402) where the
simulations are modeled in “Typhoon HIL control center” which is an open-source
START
Identify best solution and
corresponding OF
Set the MI=0.1, set iteration count=1
Maximum
Randomly initialize the α1,α2
Increment by 0.01 iteration
Increment by 1 NO
YES
Compute the OF using equation (6)
Save best solution so far
and corresponding OF
Execute proposed, BBO, BOA, GSA

and PSO techniques to optimize α1,α2 MI=1
YES
NO
END
Fig. 4 Flow chart of entire process for firing angle optimization in variable MI
Fig. 5 Experimental setup of hardware in loop
software. The HIL SCADA in the HIL basically provides the scope of controlling,
monitoring, and capturing the simulation outcome. Figure 5 shows the experimental
setup of HIL.
8 Simulation Details
All the simulations in this work are done in the MATLAB R2013a software platform.
The firing angles of the proposed MLI are optimized by using the BOA, the BBO, the
GSA, and the PSO techniques and their optimizing performances are compared with
the proposed technique. For all the techniques, maximum iteration number is set as
25. And population size is taken as 50. The default values of control variables for
the BOA, the BBO, and the GSA are taken in this research as the authors suggested
in [17, 18] and [19], respectively. The control variables in the proposed technique
are same as in [18] but the additional parameters required for modification, i.e.,
decay parameter (l), scale factor (η) are assigned as 1 and 0.01, respectively. In the
PSO technique, chosen set of control parameters are inertia weight (w) = 0.5 and
constriction factors, i.e., c1 = c2 = 2.03.
This section arranges all the simulated and hardware in loop validated results.
Discussions on the obtained results are categorized in the subsequent sub-sections.
Table 3 Comparison of available single phase 5-level MLI topologies with proposed MLI
Design MLI 1 [5] MLI 2 [6] MLI 3 [7] MLI 4 [8] MLI 5 [9] MLI 6 [10] Proposed
parameters MLI
NSWITCH 8 8 8 7 7 6 5
NDRIVER 8 8 8 7 7 6 5
VLS 4VDC 4VDC 4VDC 3VDC 3VDC 2VDC VDC
TSV 12VDC 12VDC 12VDC 11VDC 11VDC 10VDC 9VDC
9.1 Comparison of Reduced Switch MLIs with the Proposed

MLI Topology
The comparisons among some of the available MLIs [5–10] and proposed MLI in
terms of component counts and total standing voltage (TSV) are shown in Table
3. TSV is the summation of the maximum withstand voltage appearing across
all switches. In all test cases, number of levels (NLEVEL = 5), number of sources
(NSOURCE = 2), and the sum of the withstand voltages across the switches of H-
bridge (VHS = 8VDC ) are kept at the same values. From the comparative study, it can
be said that the required number of switches (NSWITCH ), driver circuits (NDRIVER ),
sum of the withstand voltages across level producing switches (VLS ), and TSV are of
the least values in case of the proposed method. This clearly indicates the superiority
of the proposed MLI over available MLIs.
9.2 Comparison of Results of Different Optimization

Techniques with the Proposed MBBO
Figure 6a–e shows the comparison of the results of percentage reduction in THD
with change of MI from 0.1 to 1 obtained from the BBO, the BOA, the GSA, and
the PSO techniques with the proposed MBBO technique. Figure 6f shows that the
minimum THD is obtained when MI is 0.87. It also exhibits the supremacy of the
proposed technique in achieving the least value of THD among all the techniques.
Figure 7 portrays the variation of lower order harmonic components and total
percentage of THD with the change in MI. When MI equals to 0.87, the 3rd, 5th,
7th, and 9th harmonics are at their lower values. Therefore, minimum value of THD
is achieved. This clearly justifies the claim mentioned earlier. In Fig. 8, at MI = 0.87
convergence curves of the optimization techniques, used in this paper, are compared.
This comparison confirms the better accuracy of the MBBO over the other techniques.
Using the different optimization techniques and the proposed MBBO, the opti-
mized values of firing angles, i.e., α1 and α2 , are obtained for different values of
MI. For sake of brevity, only the variations of firing angles obtained by MBBO
are depicted in Fig. 9. The comparison of firing angles with %THD for different
Fig. 6 Comparison of MI versus THD for different techniques
Fig. 7 Frequency
components versus MI
techniques when MI is 0.87 is shown in Table 4. The results clearly indicate the
supremacy of the proposed MBBO technique among all the techniques (used in this
work) to mitigate THD.
The firing angles which are obtained from the proposed MBBO technique, corre-
sponding to MI at 0.87 are fed to the Simulink model of the proposed 5-level MLI.
The attained output voltage waveform and its corresponding FFT spectrum have been
presented in Fig. 10. It is observed from the figure that the lower order dominating
harmonics are being suppressed substantially. Moreover, the percentage of THD is
reduced to 16.74% which matches the aimed percentage THD, shown in Table 4. The
Fig. 8 Convergence curves

at MI = 0.87
Fig. 9 Firing angles versus

MI
Table 4 Comparison of
Techniques α1 α1 %THD
firing angles (in radian) with
%THD for different BBO 0.2287 0.7154 15.9127
techniques at MI = 0.87 BOA 0.2348 0.7109 15.9606
GSA 0.2257 0.7115 15.9221
PSO 0.2291 0.7152 16.0317
Proposed MBBO 0.2213 0.7154 15.897
Load current versus Time and the corresponding FFT spectrum, depicted in Fig. 11,
clearly points out to the reduction of percentage THD to 6.30%, when the load is
R-L type whose R = 24 , L = 20 mH.
In order to validate the MATLAB simulation results, the same parameters of load
along with the system model are applied in HIL.
Fig. 10 Output voltage with FFT spectrum
Fig. 11 Output current with FFT spectrum

9.3 Hardware in Loop (HIL) Validated Results
The test conditions for the experiments in HIL are kept the same as that of the
simulations. The obtained output voltage and current waveforms using HIL setup and
their corresponding fast Fourier transform (FFT) are presented in Figs. 12, 13, 14 and
15. In the FFT spectrum of voltage shown in Fig. 10 and Fig. 13, the 13th harmonic
component is the largest. Moreover, the %THD obtained from HIL (16.9%) is very
close to MATLAB simulation result (16.74%). Similar pattern can be seen for the
load current in Figs. 11 and 14. This clearly proves the advantages of proposed MLI.
Thus, it can be concluded that the simulation of the proposed MLI with optimized
firing angle using proposed MBBO technique is successfully validated through the
HIL setup.
Fig. 12 Output voltage

waveform of HIL for one
cycle
Fig. 13 FFT spectrum of

output voltage of HIL
Fig. 14 Output current

waveform of HIL for one
cycle
Fig. 15 FFT spectrum of

output current of HIL
10 Conclusions
In this article, the proposed MLI topology has successfully produced 5-level voltage
output with reduced switch count. The proposed topology is suitable for PV
system integration. Moreover, the proposed optimization technique namely modi-
fied biogeography-based optimization (MBBO) shows better performance than other
well-known techniques, i.e., the BBO, the BOA, the GSA, and the PSO techniques
in THD mitigation. All the simulated results are successfully validated through
hardware in loop setup.
References
1. Benedetto M, Lidozzi A, Solero L, Crescimbini F, Grbović PJ (2018) Five-level E-type inverter

for grid-connected applications. IEEE Trans Ind Appl 54(5):5536–5548
2. Panda KP, Panda G (2018) Application of swarm optimisation-based modified algorithm for
selective harmonic elimination in reduced switch count multilevel inverter. IET Power Electron
11(8):1472–1482
3. Panda KP, Lee SS, Panda G (2019) Reduced switch cascaded multilevel inverter with new
selective harmonic elimination control for standalone renewable energy system. IEEE Trans
Ind Appl 55(6):7561–7574
4. Gupta KK, Ranjan A, Bhatnagar P, Sahu LK, Jain S (2016) Multilevel inverter topologies with
reduced device count: a review. IEEE Trans Power Electron 31(1):135–151
5. Lee SS, Chu B, Idris NRN, Goh HH, Heng YE (2015) Switched-battery boost-multilevel
inverter with GA optimized SHEPWM for standalone application. IEEE Trans Industr Electron
63(4):2133–2142
6. Babaei E, Hosseini SH (2009) New cascaded multilevel inverter topology with minimum
number of switches. Energy Convers Manag 50(11):2761–2767
7. Najafi E, Yatim AHM (2011) Design and implementation of a new multilevel inverter topology.
IEEE Trans Industr Electron 59(11):4148–4154
8. Nedumgatt JJ, Kumar DV, Kirubakaran A, Umashankar S (2012) A multilevel inverter with
reduced number of switches. In: 2012 IEEE students’ conference on electrical, electronics and
computer science. IEEE, Bhopal, pp 1–4
9. Wang L, Wu QH, Tang W (2017) Novel cascaded switched-diode multilevel inverter for
renewable energy integration. IEEE Trans Energy Convers 32(4):1574–1582
10. Sen P, Bana PR, Panda KP (2019) Firefly assisted genetic algorithm for selective harmonic
elimination in PV interfacing reduced switch multilevel inverter. Int J Renew Energy Res
9(1):32–43
11. Rotella M, Peñailillo G, Pereda J, Dixon J (2008) PWM method to eliminate power sources in
a nonredundant 27-level inverter for machine drive applications. IEEE Trans Industr Electron
56(1):194–201
12. Bana PR, Panda KP, Naayagi RT, Siano P, Panda G (2019) Recently developed reduced
switch multilevel inverter for renewable energy integration and drives application: topologies,
comprehensive analysis and comparative evaluation. IEEE Access 7:54888–54909
13. Sun J, Grotstollen H (1992) Solving nonlinear equations for selective harmonic eliminated
PWM using predicted initial values. In: Proceedings of the 1992 international conference on
industrial electronics, control, instrumentation, and automation, vol 1. IEEE, San Diego, CA,
pp 259–264
14. Chiasson JN, Tolbert LM, McKenzie KJ, Du Z (2005) Elimination of harmonics in a multilevel
converter using the theory of symmetric polynomials and resultants. IEEE Trans Control Syst
Technol 13(2):216–223
15. Du Z, Tolbert LM, Chiasson JN (2006) Active harmonic elimination for multilevel converters.
IEEE Trans Power Electron 21(2):459–469
16. Memon MA, Mekhilef S, Mubin M, Aamir M (2018) Selective harmonic elimination in
inverters using bio-inspired intelligent algorithms for renewable energy conversion applica-
tions: a review. Renew Sustain Energy Rev 82(3):2235–2253
17. Arora S, Singh S (2019) Butterfly optimization algorithm: a novel approach for global
optimization. Soft Comput 23(3):715–734
18. Simon D (2008) Biogeography-based optimization. IEEE Trans Evol Comput 12(6):702–713
19. Rashedi E, Nezamabadi-Pour H, Saryazdi S (2009) GSA: a gravitational search algorithm. Inf
Sci 179(13):2232–2248
20. Eberhart R, Kennedy J (1995) A new optimizer using particle swarm theory. In: Proceedings
of the sixth international symposium on micro machine and human science. IEEE, Nagoya, pp
39–43
21. Li F, Wang Y, Wu F, Huang Y, Liu Y, Zhang X, Ma M (2020) Review of Real-time simulation
of power electronics. J Mod Power Syst Clean Energy 8(4):796–808
22. Typhoon HIL Homepage. https://www.typhoon-hil.com. Accessed 18 Jan 2021
23. Ahmad J, Pervez I, Sarwar A, Tariq M, Fahad M, Chakraborty RK, Ryan MJ (2020) Perfor-
mance analysis and hardware-in-the-loop (HIL) validation of single switch high voltage gain
DC-DC converters for MPP tracking in solar PV system. IEEE Access
24. Agwa AM, El-Fergany AA, Maksoud HA (2020) Electrical characterization of photovoltaic
modules using farmland fertility optimizer. Energy Convers Manag 217:1–12
25. SLP080-12 Datasheet. https://www.solarland.com.tr/main/media/poly/SLP080-12.pdf.
Accessed 18 Jan 2021
Power Quality Improvement
of an Interconnected Grid System Using
PWM Technique of D-STATCOM
Manoj Kumar Kar, Sanjay Kumar, and Arun Kumar Singh
1 Introduction
In general, power squality (PQ) is the maintenance of a sinusoidal voltage waveform

at nominal voltage and frequency. It is the power supply characteristics that allow the
equipment to operate properly. Any faults on transmission lines, capacitor switching,
switching of large loads, use of power electronic devices cause distortion due to its
non-linearity nature. At the generation point, the electric power waveform is strictly
sinusoidal and is free from distortion. The same power would be available in the load
side if there is no distortion. But, in practical case due to the presence of transmission
line reactance, the power to be transmitted gets affected. The current flowing through
the system’s impedance can cause a number of voltage disruptions. The voltage is
often skewed by distorted currents from harmonic production loads as they pass
through the impedance of the system. These distortions spread all over the network.
In the early days, the quality of power was mainly concerned with the continuity
of the supply of electricity at appropriate voltage and frequency. AC distribution
systems are faced with a number of problems in the quality of power. In particular,
because of the use of sensitive equipment in the majority of production, residential,
industrial, and traction applications. Not only can non-linear loads cause problems
with power efficiency, but they are also extremely susceptible to voltage deviations.
The ability to monitor both actual and reactive power exchanges can be effectively
used to damp power oscillation damping, and for providing uninterrupted power to
essential loads.
M. K. Kar (B) · S. Kumar · A. K. Singh

EE Department, NIT Jamshedpur, Jamshedpur, Jharkhand, India
S. Kumar
e-mail: sanjay.ee@nitjsr.ac.in
A. K. Singh
e-mail: aksingh.ee@nitjsr.ac.in
32 M. K. Kar et al.
Nonlinear loads cause distortions in currents and voltages in the supply mains,
leading to deterioration of power quality [1]. Different approaches have been applied
for compensating the effect of these non-linear loads [2]. The performance can be
improved with different algorithms [3, 4]. A new unified method was proposed
for the manipulation of power system waveform parameters such as magnitude,
phaseaangle, and ffrequency [5]. Though variation of load has effect on PQ, still
based on the different environmental conditions PQ disruptions are also classified
[6]. The power quality can be improved by using D-STATCOM near the load [7]. In
order to produce the reference load voltage for a voltage-controlled DSTATCOM, a
control algorithm was proposed [8]. In VSC-based DSTATCOM, an adaptive fuzzy
logic controller with IRPT control algorithm has been used [9]. A control algo-
rithm based on SRF theory is implemented to control DSTATCOM efficiently in
grid level PQ improvement [10]. A hybrid technique-based icosφ control algorithm
has been used for analyzing the performance [11]. Taking into consideration the
costs of DSTATCOM, the optimal location is chosen for loss minimization and
enhancement of voltage profile [12]. A DSTATCOM integrated with distribution
transformer is used for reactive power compensation [13]. An optimal allocation
model of DSTATCOM with multi-objective functions is described [14]. The methods
to protect critical loads from voltage-based PQ issues are presented [15]. The PCC
voltage and the reactive power are taken care by the DSTATCOM [16]. The miti-
gation of harmonics and power factor improvement is done by using D-STATCOM
[17]. The distorted voltage and hence, the power quality is improved [18].
2 Operation and Control of DSTATCOM
Static compensators connecting at distributed system and operating for mitigating of

multiple current power quality problems commonly known as distributed STATCOM
(DSTATCOM). It is typically used to boost power quality in low or medium voltage
distribution systems. It is often used in three-phase systems to control the terminal
vvoltage, dominate the voltage flicker and enhance voltage balance. The basic
purpose of DSTATCOM is to alleviate all current power-based problems such as
VAR, unbalanced currents, neutral currents, and harmonics. An IGBT dependent
current controlled VSC is used as the DSTATCOM with a DC bus capacitor to
provide sinusoidal balanced currents in the supply. The VSC uses PWM power, so
small ripple filters are needed to minimize switching ripples.
A control algorithm is generally used to control directly the reference currents
of the DSTATCOM. By using PWM current ccontrol, which results in an indirect
current control, the gating pulses to the DSTATCOM are produced. In all control
algorithms, the reactive power and the unbalanced current are being compensated
using DSTATCOM.
The main purpose of DSTATCOM’s control algorithm is to use feedback signals to
calculate the reference currents. In PWM current controllers, these reference currents
are used, along with corresponding sensed currents, to generate PWM gating signals
Power Quality Improvement of an Interconnected … 33
Fig. 1 Schematic diagram of the DSTATCOM
for VSC switching devices used as DSTATCOM. Figure 1 shows the reactive power
generation using the basic VSC scheme.
A coupling transformer is installed to match the voltage level because the device
rating may not match with system rating. The DC storage capacitor is used to keep
the instantaneous output and input powers equal. Semiconductor switches having
some losses in a practical converter. These losses are being compensated by the
stored energy in the dc capacitor. Thus, to replenish the internal losses, the converter
consumes a small amount of active power from the system, maintaining the desired
capacitor voltage.
The schematic diagram of the DSTATCOM is shown in Fig. 1. If the voltage
magnitude of the VSC is more than the system AC voltage VM , the reactive current
starts flowing from the converter to the ac system through the tie line reactance, and
the VSC produces reactive power for the ac system which is capacitive in nature.
If the output voltage amplitude is less than that of the ac system, then the current
flows from the AC system to the converter and the reactive power is absorbed by the
converter which is inductive in nature. If the amplitude of both is equal, then there
is no exchange of reactive power.
3 Modeling
The basic diagram of the DSTATCOM based on VSC is shown in Fig. 2.

The three-phase AC system voltage Vs lags the STATCOM output voltage by an
angle and is given by Eq. (1).
⎡ ⎤ ⎡ ⎤
Vsa (t) sin ωt
2
Vs,abc = ⎣ Vsb (t) ⎦ = Vs ⎣ sin(ωt − 2π 3
)⎦ (1)
3
Vsc (t) sin(ωt + 3 )
2π
34 M. K. Kar et al.
Fig. 2 Block diagram of the VSC-based DSTATCOM
The STATCOM currents are given by Eq. (2).

⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
i (t) i ca (t) Vsa (t) Voa (t)
d ⎣ ca ⎦
Ls i cb (t) = Rs ⎣ i cb (t) ⎦ + ⎣ Vsb (t) ⎦ − ⎣ Vob (t) ⎦ (2)
dt
i cc (t) i cc (t) Vsc (t) Voc (t)
The above voltage and currents are transformed into α-β axis as below:
⎡ ⎤
−1 −1 v
vα 2 1√ 2 √ 2 ⎣ a⎦
= vb (3.1)
vβ 3 0 3 2− 3 2
vc
⎡ ⎤
−1 −1 i
iα 2 1√ 2 √ 2 ⎣ a⎦
= ib (3.2)
iβ 3 0 3 2− 3 2
ic
where α and β are the orthogonal coordinates.

These currents can be transformed from α-β to d-q by using Park’s transformation
defined as

id cos θ sin θ iα
= (4)
iq − sin θ cos θ i β
The STATCOM model in d-q frame is given in Eq. (5).

Fig. 3 Transformation block diagram
⎡ ⎤ ⎡ −Rs −m
⎤⎡ ⎤ ⎡ ⎤
i cd (t) L ω L i cd (t) cos α
d⎣ ⎢ s s
⎥ Vs ⎣
i cq (t) ⎦ = ⎣ −ω −Rs L 0 ⎦⎣ i cq (t) ⎦ + − sin α ⎦ (5)
dt s L s
vdc (t) m
C 0 0 vdc (t) 0
The transformation block diagram is shown in Fig. 3.
4 Controller Design
Two controllers namely voltage and current controller are used to analyze the
performance.
Equation (5) can be rewritten as
−Rs
d i cq Ls −ω i cq 1 Vsq Voq
= −Rs + − (6)
dt i cd ω Ls i cd Ls Vsd Vod
4.1 Current Controller Design
The harmonic current is supplied by the d-axis controller and a small amount of
active current is drawn to meet the switching losses where as both harmonic and
reactive current are supplied by the q-axis controller.
∗
voq = − voq −ωL s i cd + vsq (7.1)
∗
vod = − vod −ωL s i cq + vsd (7.2)
36 M. K. Kar et al.
Applying the Laplace transformation and then rearranging, the transfer function
can be obtained as follows
Icq (s) Icd (s) 1
G i (s) = ∗
= ∗ = (8)
voq (s) vod (s) Rs + L s
4.2 Voltage Controller Design
The DC link voltage is maintained constant using voltage controller

1
Vdc = i dc (t) (9)
C
Vdc 1
G v (s) = = (10)
Idc sC
5 Simulation Results
The performance of the 3-φ grid-connected system with nonlinear load has been
studied with MATLAB/SIMULINK and the results have been presented. The 3-φ
source voltage waveform is given in Fig. 4. Due to the non-linearity nature of the load,
the current waveform gets distorted and is shown in Fig. 5. When the DSTATCOM
is enabled at t = 0.1 s, it injects the harmonic current and thus maintains sinusoidal
source current by compensating the distortion. The injected current waveform with
enabled DSTATCOM at t = 0.1 s is shown in Fig. 6. The power factor is found to be
unity after the DSTATCOM is enabled thus improves the power quality of the system
which is shown in Fig. 7. The THD of the grid current is found to be 31.08% and
2.67% without and with compensation which is given in Fig. 8 and Fig. 9 respectively.
6 Conclusion
The analysis of DSTATCOM and design of controllers are presented in this paper. A
DSTATCOM based on VSI is used to provide the load with both VAR and harmonic
current so that the source current becomes sinusoidal and has a unit power factor.
The controllers are designed on the basis of parameters of the STATCOM and time
constant. THD is reduced from 31.08 to 2.67% after connecting DSTATCOM. Thus,
the effectiveness of DSTATCOM with PI controller to improve the power factor
Fig. 4 Three-phase source voltage waveforms
Fig. 5 Load current waveform

38 M. K. Kar et al.
Fig. 6 Current injected by DSTATCOM
Fig. 7 UPF operation after using DSTATCOM

Fig. 8 THD before using DSTATCOM
is also investigated. Unity power factor operation is observed after 0.1 s when
the DSTATCOM is enabled. The proposed DSTATCOM model effectively reduces
the current harmonics injected by the load into the distribution network and hence
enhances the power quality of the system.
40 M. K. Kar et al.
Fig. 9 THD after using DSTATCOM
References
1. Padiyar KR (2008) FACTS controllers in power transmission and distribution. New Age
International, New Delhi
2. Herrera RS, Salmeron P, Kim H (2008) Instantaneous reactive power theory applied to active
power filter compensations: different approaches, assessment, and experimental results. IEEE
Trans Ind Electron 55(1):184–196
3. Masand D, Jain S, Agnihotri G (2008) Control strategies for distribution static compensator
for power quality improvement. IETE J Res 54(6):421–428
4. Singh B, Solanki J (2009) A comparison of control algorithms for DSTATCOM. IEEE Trans
Ind Electron 56(7):2738–2745
5. Dash P, Padhee M, Barik S. Estimation of power quality indices in distributed generation
systems during power islanding conditions. Int J Elect Power Energy Syst 36(1):18–30
6. Ray PK, Mohanty SR, Kishore N (2013) Classification of power quality disturbances due
to environmental characteristics in distributed generation system. IEEE Trans Sust Energy
4(2):302–313
7. Kar MK, Rout BD, Moharana JK (2014) Improvement of power factor of a grid connected
load system using a static compensator. J Found Appl Phys 1(1):5–10
8. Kumar C, Mishra MK (2014) A voltage controlled DSTATCOM for power quality improve-
ment. IEEE Trans Power Deliv 29(3)
9. Singh B, Dube SK, Arya SR (2015) An improved control algorithm of DSTATCOM for power
quality improvement. Electr Power Energy Syst 64:493–504
10. Mahela OP, Shaik AG (2016) Power quality improvement in distribution network using
DSTATCOM with battery energy storage system. Int J Electr Power Energy syst 83:229–240
11. Mangaraj M, Panda AK (2017) Performance analysis of DSTATCOM employing various

control algorithms. IET Gener Transm Distrib 11(10):2643–2653
12. Iqbal F, Khan MT, Siddiqui AS (2018) Optimal placement of DG and DSTATCOM for loss
reduction and voltage profile improvement. Alex Eng J 755–765
13. Chen Y, Wen M, Xianggen Yin EL, Lai J, Wang Z (2018) Passivity-based control of cascaded
multilevel converter based DSTATCOM integrated with distribution transformer. Electr Power
Syst Res 154:1–12
14. Pradhan M, Mishra MK (2019) Dual P-Q Theory based energy-optimized dynamic voltage
restorer for power quality improvement in a distribution system. IEEE Trans Ind Electron
66(4)
15. Zhang T, Yu L (2020) Optimal allocation of DSTATCOM considering the uncertainty of
photovoltaic systems. IEEJ Trans Electr Electron Eng 15:355–363
16. Sayahi K, Kadri A, Bacha F, Marzougui H (2020) Implementation of a DSTATCOM control
strategy based on direct power control method for grid connected wind turbine. Electr Power
Energy Syst 121:1–14
17. Rohouma W, Balog RS, Peerzada AA, Begovic MM (2020) D-STATCOM for harmonic miti-
gation in low voltage distribution network with high penetration of nonlinear loads. Renew
Energy 145:1449–1464
18. Abas N et al (2020) Power quality improvement using dynamic voltage restorer. IEEE Access
8:164325–164349
Speed Control of a Three-Phase IM
with Closed-Loop Control Scheme
Rajkumar Yadav, Manoj Kumar Kar, and Arun Kumar Singh
1 Introduction
Induction motor runs at rated speed, when it is connected to main supply. However,
depending on various applications, there is a need to regulate the speed of the IM.
Various methods are available to control the speed of IM from both sides either
stator side or rotor side. The voltage and frequency are varied simultaneously but
the v/f ratio is kept constant so that the air-gap flux remains constant and to avoid
the saturation. v/f control method reduces steady-state error as discussed in [1, 2].
A sensor less strategy with the help of direct torque control (DTC) was discussed in
[3], and it reduces the perturbations of speed, torque, and current around steady state.
An intelligent technique with PI controller deal with real-time estimation, improve
performance and decrease the sensitivity with changing in parameter and load [4].
An automated PLC system was used for better speed regulation in induction motor
drive with variable load [5]. The PLC control system is more efficient than open-
loop control of IM fed by inverter. The closed loop control scheme has been used to
control the speed of a permanent magnet synchronous machine [6] . The hybrid PWM
technique [7] reduces ripple present in torque with improving the harmonics distor-
tion in current. The hybrid PWM technique has good result in such areas as compared
to conventional SVPWM technique. Mathematical modeling and simulation of VSI-
fed IM with closed-loop control method is done by maintaining v/f as constant to
control speed of IM drive discussed in [8]. The same approach is implemented in
laboratory with the help of PC and add-on card. The idea of Z-source inverter (ZSI)
useful for the speed control of IM using v/f control method [9]. The ZSI diminishes
the stress voltage across such capacitor, hence this topology makes drive system
economical. The traditional model-based design replaced after the development of
model-based embedded system design in MATLAB is described in [10]. SPWM
control method is developed based on MATLAB platform and implemented using
R. Yadav (B) · M. K. Kar · A. K. Singh

NIT Jamshedpur, Jamshedpur 831014, India
44 R. Yadav et al.
model-based design which saves both effort and time. A sensor less approach is estab-
lished on the application of Lueberger observer for flux and rotor speed calculations
[11]. In case of open path errors take place in switches of inverter with sensor less
induction motor must be detected quickly and disconnect the healthy drive system
from faulty system. In SPWM, closed-loop control and open-loop control methods
converters are used which provide output waveforms inclosing some ripples, now to
eliminate this problem converters use PI-controlled elementary positive output lift
Luo converter (EPOLLC) [12] for speed control of IM. The whole drive systems
are assembled through field programmable gate array spartan-6 computer. The paper
[13] discusses a scheme where it proposes about vector organized squirrel cage IM
drive with synchronized SPWM providing wide range of speed variations suitable
for drive applications. The paper [14] explains SVPWM method for speed control of
TIM drive with open-loop and closed-loop control approach. In SVPWM technique
the output voltage almost similar as sinusoidal provides good concert and becomes
more efficient to the drive systems [15]. A new idea for speed control of IM known
as direct matrix converter (DMC) which is controlled by global synchronous PWM
(GSPWM) approach. The proposed idea is used to keep the input power factor as a
unity from rectifier side of DMC. In this way the drive system becomes very simple
and free from maintenance [16]. To control the speed of squirrel cage IM, the prac-
tical setup is done by v/f scalar control with two methods, one is open-loop control
and the other is closed-loop control. During experiment, it is essential to monitor the
behavior of current to escape from possible damages. In [17], various loss minimiza-
tion techniques are discussed, the losses arise due to the switching devices used in
Inverter-fed drive systems. The component-level loss minimization method does not
guarantee about minimization of total losses. But system-level loss method provides
a true minimum total loss in the motor drive systems. On the basis of performance,
the comparative examination for speed control of IM drive is done by SVM and
SPWM based VSI-fed motor drive having PI controller as in feedback path with
v/f ratio same. Analyzing the result [18, 19], it is found that SVPWM technique has
better performance compare to SPWM technique with closed-loop control approach.
A digital basis FPGA controller gives the signals for switching devices used in drive
[20]. The idea of SPWM used with FPGA controller into the application of IM drive
system and also enhance the routine in variable-voltage/variable-frequency (VVVF).
The closed-loop speed control method with PI controller limits the starting current
of motor within range. PI controller reduces steady-state error and provides smooth
speed but does not limit the dynamic behavior. The complete policy and estab-
lishment of close-loop and open-loop control method is done in [21] by using PI
controller. The digital signal controller (DSC) reduces the circuit complexity, hence
decrease the overall cost of the IM drive system. An algorithm is designed in [22]
for a digital signal PIC microcontroller in VSI-fed IM. A comparative study is done
in [23] to enhance the efficiency of three-phase induction motor (TIM). Another
advanced controller proposed with switching topology to monitor the harmonics,
flux, voltage, current, and frequency control during the speed control of TIM. The
paper [24] elaborates the difficulties that arise during speed control for IM through a
Speed Control of a Three-Phase IM with Closed-Loop … 45
literature survey. There are many techniques available for IM drive system, the trou-
bles faced by such techniques as variations in speed, harmonics, or ripples present in
voltage and current waveforms. By adjusting the duty cycles of switches, the stator
terminal voltage of VSI-fed IM is to be controlled [25]. Hence the speed and torque
can be adjusted easily. Without any speed control scheme, IM runs at its own speed
after connecting with supply. Among the various schemes, VVVF method is most
popular to find the variable speed for such motor drive system. The developed torque
is related to the flux which is associated with ratio of change in voltage and change in
frequency having the v/f ratio constant [26]. The paper [27] discusses a comparative
analysis among the various PWM approaches such as space vector PWM, sinusoidal
PWM, and multiple PWM having different power rating of machine. All these PWM
methods help to produce gate signals for VSI-fed IM drive. To reduce the THD
and find better quality of voltage sine waves at output terminals of Diode Clamped
Multilevel Inverter, different modulation schemes have been implemented [28, 29].
A 15-level diode clamped MLI was used to control the speed of TIM with improving
the performance characteristics.
In this work, the speed control of 3-F IM is achieved by adjusting the pulse width
in case of variable-voltage method whereas in case of VVVF method, by maintaining
the v/f ratio constant throughout the entire speed range, the machine performance
and hence efficiency is improved. In VVVF method, the motor reference speed is
compared with the actual speed, the error obtained produces a reference frequency
of PWM. Then, the motor speed can be changed according to the reference speed.
2 Sinusoidal PWM Technique
Among the various modulation approaches, SPWM is popular in industries to control

the speed of an IM. Three references sinusoidal waves (120° apart in phase) are
compared by a saw tooth wave signal used to produce gate pulses for three-phase
inverters. A squirrel cage IM with ratings 5.4 HP, 400 V, 50 Hz, 1430 RPM is used.
The Simulink block diagram of SPWM control scheme is depicted in Fig. 1. The gate
pulses are generated using SPWM technique. The inverter output is fed to the IM for
analyzing the performance. In this case, the speed control of the IM is achieved only
by varying the voltage.
The idea about on/off Inverter has six switches and a DC voltage source as an
input. Now after comparison between reference and carrier wave, six gate signals
(A1, A2, A3, A4, A5, A6) are generated by the SPWM technique. These pulses are
given to the switching devices shown in Fig. 2.
The above diagram gives switching pattern of the devices. Switch pair (S1-S4,
S3-S6, and S5-S6) from the same leg cannot conduct at the similar time to escape
the short circuit. In this case, the output voltage becomes zero and it may cause
explosion.
The speed and torque of the IM related with supply frequency and voltage are
represented as following:
46 R. Yadav et al.
Fig. 1 Simulink diagram with SPWM inverter
Fig. 2 Gate signals of switching devices
N = 120 ∗ f / p (1)
V = 4.44 ∗ ϕ ∗ f ∗ N p (2)

3 ∗ V 2 ∗ R2 /s
T = (3)
ωs ∗ (R1 + R2 /s)2 + (X 1 + X 2 )2
where N = speed in rpm, f&V are supply frequency and voltage respectively, &T
are flux per pole and torque (N-m) respectively, ωs = angular speed (rad/s), s = slip.
It is clear from the above equations, the torque and speed of TIM can be changed by
altering the frequency, voltage, or both at a time.
3 Simulation Result and Discussion for SPWM Scheme
Figure 3 shows the inverter line voltage achieved by SPWM method which is fed to
the induction motor for controlling the speed.
Figure 4 shows the rotor speed characteristic waveform w.r.t the time imple-
menting SPWM scheme. The settling time for rotor speed is 0.12 s. The rotor speed
changes between 1475 and 1425 rpm after steady state.
Figure 5 explains the behavior of torque w.r.t time implementing SPWM scheme.
The torque has transient nature at starting but it settles after 0.1 s. From Eq. 3, torque
is proportional to the square of voltage. Hence torque is also varied. Hence the torque
and voltage relation are satisfied. From speed and torque wave form, it is seen that
as speed is higher at the same time torque has a minimum value to make the drive
system under steady state.
During the speed control of IM, one of the important parameters is known as
stator. To avoid the damage due to excess heat produced by the current must have
the safe limit during drive application. IM draws the high amount of current just at
starting as shown in Fig. 6. The stator current has direct relation to the supply voltage
generated by SPWM inverter.
Fig. 3 Response of line voltage

48 R. Yadav et al.
Fig. 4 Response of rotor speed using SPWM scheme
Fig. 5 Response of electromagnetic torque using SPWM scheme

Fig. 6 Response of stator current using SPWM scheme
4 Advantages of Sinusoidal SPWM Scheme
• It is less costly. Because the result is found without need of any filter by keeping
frequency as the constant parameter.
• It is simple for design and implementation. There is no requirement of controller
and look-up table like SVPWM.
5 Speed Control by Closed-Loop V/F Approach
As the frequency changes, speed also changes in the same proportion which is clear
from Eq. 1. Since the flux has an inverse relationship with frequency, the frequency
cannot be changed beyond the limit to take care of air-gap flux. From Eq. 2, the voltage
is directly proportional to flux means supply voltage must vary within the limit for the
same reason. The slip speed being the difference between the synchronous speed and
rotor speed can’t be maintained. It leads to the unstable operation of Torque-Speed
characteristics. As a result, the stator current exceeds the rated value of current which
50 R. Yadav et al.
Fig. 7 Simulink model implementing v/f scheme with closed-loop control
may have adverse effect on the inverter-converter combination. The above problem
can be overcome by using an outer loop in the IM shown in Fig. 7.
The scheme consists of a PID controller, VSI, and speed sensing feedback. The
actual rotor speed is compared with the reference value to generate an error. This
error is passed through a controller and its output sets the inverter frequency. This
frequency generates the reference signal for the closed-loop control of the IM.
6 Result and Discussion for Closed-Loop V/F Control

Method
The frequency of supply voltage fed to the IM can be varied by varying the width of
pulses. The frequency has a direct relation with the speed from Eq. 1. Hence speed
of IM is regulated by controlling the supply frequency. Figure 8 shows the nature of
rotor speed w.r.t. time implementing closed loop v/f control scheme.
The speed-time graph settles after 2 s. Before that time, speed is under transient
condition. At starting difference between actual and reference speed is high. It shows
the larger error signal, the error is given to the controller. The PID controller has
advantage to limit speed difference until the error signal becomes zero. Variation in
speed is going to be controlled within the fixed range.
The electromagnetic torque w.r.t. time using closed loop v/f control scheme is
shown in Fig. 9. It shows the fluctuations just at starting. From the graph, it can be
seen that torque has more fluctuations between (0–0.8) s and after that it is settled
down between (0.8–1.8) s having less fluctuation. But for fractions of second (1.8–
2.2) again transient occurs and finally achieve the steady-state condition after 2.2 s.
v/f ratio is maintained as constant in closed-loop control so that flux setup to become
unchanged and torque would be free from the supply frequency. Now in such case
speed of IM allowed to vary with a larger range by changing the frequency. It does
not affect the torque operated under loaded condition.
The rotor angle (rad) graph is shown in Fig. 10. The rotor angle is related to the
stator current and slip. It can be adjusted from 0 to a fixed angle. It is a straight-line
Fig. 8 Response of rotor speed using closed loop v/f control scheme
Fig. 9 Response of electromagnetic torque using closed loop v/f control scheme
52 R. Yadav et al.
Fig. 10 Response of rotor angle
graph starting from zero. IM behaves as a load itself, so the rotor current drawn by
the motor should be under control during closed-loop v/f control.
The response of the stator current using closed loop v/f control scheme is presented
in Fig. 11. It is found that stator current has a direct relation with supply voltage. So,
at the time of starting favorable supply given to the drive system. The stator current
is found to be within the limit of the rated value. Distortion in current waveform is
improved as compared to the open-loop scheme.
7 Advantages of Closed-Loop V/F Control Method
• Maximum torque is constant throughout the operation due to flux unchanged

having voltage to frequency ratio constant.
• IM can run as greater than base speed, if increase the frequency only where voltage
cannot go higher than rated value.
8 Conclusion
The speed control of IM by variable voltage and by VVVF method is achieved

using MATLAB/SIMULINK. To analyze the performance by speed control of IM,
Fig. 11 Response of stator current using closed loop v/f control scheme
two cases are considered. The different motor parameters such as voltage, rotor
speed, electromagnetic torque, and stator current waveform are observed by each
method. The demerits of earlier method are compensated by the later discussed
method. It is clear from the obtained result that the application of closed-loop control
scheme enhances the performance as compared to open-loop control because the
motor parameters have less harmonics.
References
1. Koga K, Ueda R, Sonoda T (1992) Constitution of V/f control for reducing the steady-state
speed error to zero in induction motor drive system. IEEE Trans Ind Appl 28(2):463–471.
https://doi.org/10.1109/28.126757
2. Munoz-Garcia A, Lipo TA, Novotny DW (1997) A new induction motor open-loop speed
control capable of low frequency operation. In: IEEE Industry applications society annual
meeting, New Orleans, Louisiana, 5–9 October 1997
3. Lascu C, Boldea I, Blaabjerg F (2000) A modified direct torque control for induction motor
sensor less drive. IEEE Trans Ind Appl 36(1):122–130. https://doi.org/10.1109/28.821806
4. Chen T-C, Sheu T-T (2002) Model reference neural network controller for induction motor
speed control. IEEE Trans Energy Convers 17(2):157–163. https://doi.org/10.1109/TEC.2002.
1009462
5. Ioannides MG (2004) Design and implementation of PLC-based monitoring control for induc-
tion motor. IEEE Trans Energy Convers 19(3):469–476. https://doi.org/10.1109/TEC.2003.
822303
54 R. Yadav et al.
6. Yadav R, Kar MK, Singh AK (2021) Controlling speed of a permanent magnet synchronous
machine using closed loop control scheme. In: Emerging trends in industry 4.0 (ETI 4.0) pp
1-6. https://doi.org/10.1109/ETI4.051663.2021.9619448
7. Harsha Vardhan Reddy M, Jegathesan V (2011) Open loop V/F control of induction motor
based on hybrid PWM with reduced torque ripple. In: International conference on emerging
trends in engineering and technology ICETET-2011
8. Vijay Babu K, Magajikondi SJ, Raju AB (2011) Modeling, simulation and PC based
implementation of a closed lopp speed control of VSI fed induction motor drive. IEEE
9. Amudhavalli D, Narendran L (2012) Speed control of an induction motor by V/F method
using an improved Z source inverter. In: 2012 International conference on emerging trends
in electrical engineering and energy management (ICETEEEM), Chennai, 2012, pp 436–440.
https://doi.org/10.1109/ICETEEEM.2012.6494486
10. Li H, Li W, Cui D (2013) Automatic code generation of SPWM for single phase inverter by
model-based design. In: Proceedings of 2013 2nd international conference on measurement,
information and control, Harbin, pp 1035–1038. https://doi.org/10.1109/MIC.2013.6758136
11. Trabelsi M, Boussak M (2014) Sensorless speed control of VSI-FED induction motor drive
under IGBT open-switch damage: performances and fault tolerant analysis. In: 2014 Interna-
tional conference on electrical sciences and technologies in Maghreb (CISTEM), Tunis, 2014,
pp 1–8. https://doi.org/10.1109/CISTEM.2014.7368727
12. Elangovan P, Mohanty NK (2015) FPGA based V/F control of three phase induction motor
drives. Int J Power Electron Drives Syst (IJPEDS) 5(3):393–403. ISSN: 2088-8694
13. Sahoo SK, Bhattacharya T (2016) Rotor flux-oriented control of induction motor with synchro-
nized sinusoidal PWM for traction application. IEEE Trans Power Electron 31(6):4429–4439.
https://doi.org/10.1109/TPEL.2015.2475398
14. Abdul M, Biabani KA, Ali SM (2016) Control of induction motor drive using space vector
PWM. In: International conference on electrical electronics and optimization techniques
(ICEEOT)-2016
15. Singh AK, Kumar NM, Pattnaik S, Vinay Reddy K (2016) Speed control of 3-phase induction
motor fed through direct matrix converter using GSPWM technique with unity input power
factor. In: International conference on electrical power and energy systems (ICEPES), 14–16
Dec 2016
16. Pena JM, Diaz EV (2016) Implementation of V/F scalar control for speed regulation of a
three-phase induction motor. IEEE
17. Sridharan S, Krein PT (2017) Minimization of system-level losses in VSI-based induction
motor drives: offline strategies. IEEE Trans Ind Appl 53(2):1096–1105. https://doi.org/10.
1109/TIA.2016.2631513
18. Alizadeh M, Masoumi M, Ebrahim E (2017) Closed loop speed control of induction motor
using constant V/F applying SPWM and SVM based inverter. Int J Eng Adv Technol (IJEAT)
6(5). ISSN: 2249-8958
19. Shukla NK, Srivastava R (2017) SVPWM and SPWM controller based performance evaluation
of three phase induction motor. Int J Eng Sci Res Technol (IJESRT)
20. Selvabharathi P, Kamatchi Kannan V, Sathish Kumar S (2018) SPWM base FPGA controller.
Int J Pure Appl Math (IJPAM) 119(18)
21. Naveed I, Sabir A (2018) Design and implementation of open and close loop speed control of
three phase induction motor using PI controller. Int J Eng Res Gen Sci 6(2)
22. Moundekar PC, Goyal GN, Hatwar PS (2018) Generation of SPWM control pulses for voltage
source inverter sing dcPIC 33EP256MU 810. In: International conference on smart electric
drives and power system
23. Hannan MA, Ali JA, Jern KP, Mohamed A, Hossain Lipu MS, Hussain A (2018) Switching
techniques and intelligent controllers for induction motor drives: issues and recommendations.
IEEE Access
24. Agrawal A, Lodhi RS, Nema P (2018) A review of speed control methods of induction motor.
IOSR J Electr Electron Eng (IOSR-JEEE) 13(3):09–14. e-ISSN: 2278-1676, p-ISSN: 2320-
3331
25. Bhagat B, Singh G (2019) Performance and evaluation of voltage source inverter feed induction
motor drives. Int Res J Eng Technol (IRJET) 6(2)
26. Gawade O, Utekar O, Rajane S, Raut A, Sonune R (2019) V/F speed control technique of three
phase induction motor. Int J Innov Res Electr Electron Instrum Control Eng 7(1)
27. Divyam, Saxena A, Rai JN, Singh B (2020) Comparative analysis of PWM techniques and
SVPWM operated V/F control of induction motor having different power ratings. In: Fifth
international conference on communication and electronics systems (ICCES)
28. Kar MK, Giri P, Gupta NK, Singh AK (2021) Control of a three-phase diode clamped multilevel
inverter using phase disposition modulation scheme. In: Kumar J, Jena P (eds) Recent advances
in power electronics and drives. Lecture notes in electrical engineering, vol 707. Springer,
Singapore. https://doi.org/10.1007/978-981-15-8586-9_6
29. Sumathi S, Bhavani C (2020) Sinusoidal pulse width modulation based speed control of induc-
tion motor using fifteen level diode clamped multilevel inverter. Int J Innov Technol Explor
Eng (IJITEE)
A Review on Recent Technologies
in Power Electronic Drives for Hybrid
Electric Vehicles
Rajakumar Sakile and Umesh Kumar Sinha
1 Introduction
Nowadays, fuel cost is gradually increasing and it will create more burden on the
human and the ecosystem [1]. To overcome the above-said disadvantages electrical
vehicles are a promising solution for transportation [2]. Due to the lack of fossil fuel,
alternate energy sources are required, the oil price index is gradually increasing day
by day and International Energy Agency conducted a survey on the oil price index
as shown in Fig. 1. When electrical vehicles are introduced then the consumption
of crude oil is predominantly reduced. Therefore, air pollution is also reduced [3].
To reduce global warming all countries are encouraged by automobile industries to
develop zero-emission vehicles and also provide incentives such as road tax-free and
zero insurance. Some countries started fast-charging stations to attract customers [4].
(i) Plug in electrical vehicle (PEV); the construction of PEV is very easy, it
consists of a battery, converter, motor, and gear system. But it has one major
disadvantage is that the battery is discharged suddenly, and therefore for long
transportation it is not suitable [5].
(ii) Hybrid electrical vehicle (HEV); the HEV has three types.
(1) Series HEV: It is the extension of Battery Electrical Vehicle (BEV) by adding
the Internal Combustion Engine (ICE) in series with the system. In the urban
area, the ICE can be OFF mode to reduce Greenhouse Gas (GHG) emission.
When the battery energy is low then turn ON the ICE, mostly on the highways.
To maintain the higher efficiency the ICE is operated as a constant output
power source [6]. Therefore, the ICE can be turned ON or OFF there is an
improvement in the fuel economy. The power rating of the ICE is depending
R. Sakile (B) · U. K. Sinha

Electrical Engineering Department, NIT Jamshedpur, Jamshedpur, Jharkhand, India
U. K. Sinha
e-mail: uksinha.ee@nitjsr.ac.in
58 R. Sakile and U. K. Sinha
Oil price Index

250
200
200
150
160
140
100 115 125
100
50 70 80
50 60
40
0
oil price
2000 2005 2010 2015 2020 2025
2030 2035 2040 2045 2050
Fig. 1 Oil price index (Source International Energy Agency)
upon the characteristics of the series vehicle. Series HEV needs three rotating
parts and its cost is high and construction is easy. Series HEV required a bigger
internal combustion engine compared to parallel HEV. According to driving
performance, the acceleration of the series electrical vehicle is poor. The fuel
consumption of series EV is more. An example of a series hybrid electrical
vehicle is the Chevrolet volt [7]. The series HEV is shown in Fig. 2.
(2) Parallel HEV: In the parallel HEV, the ICE is always in ON mode and oper-
ated at maximum efficiency point. Therefore, it will greatly improve the fuel
economy. Parallel one needs only two rotating parts and its cost is lesser than
the series HEV. The internal combustion engine size should be less and there-
fore the fuel consumption of parallel HEV is less [8, 9]. The parallel HEV is
shown in Fig. 3.
(3) Dual mode HEV/series–parallel HEV: To overcome the drawbacks of both
series, parallel HEV, the dual-mode HEV is introduced. To achieve more
Fuel
Tank
ICE Battery
Trans
mission
Power Electric
Generator
Converter Motor
Fig. 2 Series HEV

A Review on Recent Technologies in Power Electronic Drives … 59
Fuel ICE
Tank
Trans
mission
Battery Power Electric
Converter Motor
Fig. 3 Parallel HEV
power/torque a flywheel may be added to the output of the motor. It can be

operated for both series/parallel HEV. Therefore, using dual-mode HEV will
get more flexible operation, the construction is difficult and the cost is also
high [10–12]. The dual-mode HEV is shown in Fig. 4 and the comparison of
series, parallel and dual-mode HEV is discussed in Table 1.
In this paper, Sect. 1 is discussing the introduction of Hybrid Electric Vehicles

(HEV) and different types of HEVs. Section 2 discusses about the different types
of motor drives and power electronic interface circuits in hybrid electric vehicles
Fuel
ICE
Tan k
Trans
Generator
mission
Power Electric
Battery
Converter Motor
Fig. 4 Series–parallel HEV
Table 1 Comparison table for hybrid electrical vehicles [9]

Classification Fuel economy improvement Driving performance
of HEVs Idling Energy High Total Acceleration Continuous
stop recovery efficiency efficiency high output
operation
control
Series ● ∇ ● ●
Parallel ● ● ● ●
Series/parallel ∇ ∇ ∇ ∇ ● ●
∇ Excellent ● Superior Somewhat unfavorable
are discussed. Section 3 discusses about the electrical configuration of the HEV,
and Sects. 4 and 5 discuss about the electrical propulsion system of the Evs and the
conclusion of the paper.
2 Types of Motors and Power Electronic Interface Circuits
Based on the vehicle requirements both AC and DC motors can be used in HEV.
DC motors can be powered directly from the battery. Whereas AC motors need
converters to convert DC voltage into AC voltage. Due to the presence of brushes
and commutators, DC motors have been rarely used in HEV. Although brushless DC
motors can be used in HEV, compared to DC motors AC motors are used in HEV
because AC motors are more reliable, lightweight, and high efficiency [13, 14].
Recently three popular motors are used in HEV, those are permanent magnet
synchronous motor (PMSM), induction motor (IM), and switched reluctance motor
(SRM). PMSM is similar to the brushless DC motor. It has higher power density, high
starting torque, and high efficiency under operation. In [3], the detailed comparison
of various motors used in HEV is discussed in the below table at a various speed of
operation (Table 2).
Power electronics will play a vital role in HEVs for the conversion of DC voltage
to AC voltage and vice versa. Normally power electronics has several advantages and
those are low cost, low weight, and high reliability. Nowadays, HEVs configurations
required many converters, like Toyota Prius there exists a DC/DC converter between
the battery and inverter. The schematic diagram of the power train in HEVs is shown
in Fig. 5. HEVs required two inverters; inverter 1 is connected to the motor to drive
the ICE. During regenerative braking, it also acts as a generator to recharge the
battery. Inverter 2 is connected to the generator which will transfer the power from
ICE to battery [8, 15–17].
Speed 1500 rpm
various electrical
specifications at different Various specifications PMSM IM SRM
speeds [10] Torque (Nm) 303 297 294
Iron loss (W) 198 148 404
Copper loss (W) 4328 8591 7653
Efficiency (%) 91.3 83.1 85.2
Speed 6000 rpm
Various specifications PMSM IM SRM
Torque (Nm) 45.6 50.8 52.1
Iron loss (W) 953 439 4074
Copper loss (W) 219 730 306
Efficiency (%) 96.1 95.2 88.2
Fig. 5 Schematic of power train in HEVs
Inverter 1 Inverter 2
PMSM
Stator
Winding
C1 a'
a
Battery1 b b'
(Vbat1 )
c c'
C2
Fig. 6 Dual-inverter PMSM drive with single source [12]
Figure 6 shows the power electronic interface in HEVs. The diagram shows that
the power flows from the battery to the motor and also grid to the battery to charge
battery [18, 19]. Figures 6 and 7 are different types of drives used in hybrid electric
vehicles. The DC to DC boost converter used increases the output voltage. If the
motor is operating with high voltage then it means batteries which do not achieve
the high voltage for that boost converters will be used to step up the voltage level up
to the motor voltage level.
If AC motor is used in an HEV then it means AC drives will be used. In Figs. 6 and
7, dual-inverter PMSM drive with single and two sources is used for the operation
of AC drives. DC source is connected to the inverter, and then it will be converted as
an AC by using the above configuration. Recently, dual-inverter PMSM drives with
two source AC drive is mostly used because of many advantages [20].
Inverter 1 Inverter 2
PMSM
Stator
Winding
Battery2
C1 a' C'1
a (Vbat 2 )
Battery1 b b'
(Vbat1 )
c Battery2 c'
(Vbat2 )
C2 C' 2
Fig. 7 Dual-inverter PMSM drive with two sources
3 Configuration of the EVs
Battery Management System (BMS) plays a key role in the electric vehicles control
and operation of the battery. The battery remaining power is obtained by the state of
charge (SOC) and an accurate SOC value gives the best performance of the EV. The
SOC is obtained by considering the battery voltage, temperature, and other param-
eters such as polarization and concentrated effects and discharge current [21, 22].
The obtained SOC is sent to the Hybrid Control Unit (HCU) by the communication
lines. From the SOC information, the HCU operates and controls the motor with the
best performance and safe operation of the battery. The configuration of the EVs is
shown in Fig. 8.
4 Electric Propulsion System of HEVs
An electric propulsion system is the heart of Hybrid Electrical Vehicles (HEV). It

consists of an electric motor, power converter, electric controller, and energy storage
system. Electric motor converts electrical energy into mechanical energy to drive the
system [23]. Power converters will give reliable voltage and current to the electric
motor, and lastly the electronic controller gives commands to the power converter to
give fruitful power, torque, and speed to the transmission system. During regenerative
braking, the battery is charged with Battery Management System (BMS), and in that
time vehicle is operated with Internal Combustion Engine (ICE). If the battery is
Fig. 8 Configuration of the electric vehicle
charged fully then the vehicle is converted into battery running mode to achieve
maximum speed [24].
Drives used in HEVs:
Many motor drives are used in EVs and HEVs to drive the vehicle with frequent starts
and stops and also provide a wide speed range of operation. There are two types of
drives which are utilized in HEVs, those are commutator drives and commutator-
less drives. Commutator drives are normally called DC drives and commutator-less
drives are known as AC drives.
(A) DC Drives:
Due to the presence of a commutator sparking will take place at brush contacts,
which leads to motor damage [25]. Therefore, the DC drives are not suitable for
HEVs applications. To increase efficiency and higher power density, AC drives are
preferred. AC drives are free from the commutator, and therefore AC drives cost is
less, more reliable, and maintenance-free as compared to DC drives.
(B) Induction Motor Drives:
For the electric propulsion purpose, in Electric Vehicles (EVs) and Hybrid Electric
Vehicle (HEVs), the induction motor drives are used. In the present scenario, these
types of motor drives employed many advantages compared to commutator motor
drives. Basically, induction motor drives are two types and those are wound-rotor
and squirrel-cage motors. Due to the lack of applications, wound-rotor induction
DSP
Controller
BLDC Power DC Power

Machine Converter Supply
Fig. 9 BLDC motor
motor drives are not suitable for HEV. Therefore, squirrel-cage induction motors are
themselves treated as induction motor drive [26].
(C) BLDC Motor Drives:

A BLDC motor drive mainly consists of brushless DC machine, DSP controller,
and power converter, As shown in Fig. 9, permanent magnet brushless DC motor
drive can be designed with the higher power density and high operation efficiency.
Due to this advantage the BLDC motor drives are most suitable for EV and HEV
applications (Fig. 10).
Advantages of BLDC Motor Drives:
• BLDC motors are more efficient machines compared to all-electric motors. Due
to the absence of a mechanical commutator and brushes, low frictional losses will
have occurred in the machine which leads to higher efficiency.
• BLDC motors are achieving very high flux densities because of the recent
introduction of rare-earth magnets.
• Easy to control, the control variables are constant throughout the working
operation of the motor.
• Less noise will be present at commutation.
• More reliable and less maintenance.
BLDC motor drives also suffer from some drawbacks which are of more cost,
limited constant power, less safety, demagnetization, and fault tolerance. To avoid
the above-said disadvantages replace the BLDC motors with SRM drives [27, 28].
(D) SRM Drive:
Switched Reluctance Motor (SRM) drive is the best suitable candidate for EV or
HEV applications due to its simple structure, rugged structure, low cost, and wide
range of speed control compared to all discussed electrical drives [29, 30]. It does not
have any permanent magnets on its rotor side and therefore capable of high-speed
operation. It consists of SRM; various sensors like voltage, current, and position
sensing units; power converter; and DSP-related control peripherals are connected
as shown in Fig. 11. The total estimation cost is explained in Table 3.
Battery
Supercapacitor
Fuel cell
Flywheel
Energy
Storage
Electric Power Electric

Controller Converter Motor Trans
mission
Micro SCR DC Motor

Controller MOSFET IM
Micro IGBT PMSM
Processor ICT BLDC
DSP Chopper SRM
PWM
Inverter
Fig. 10 Block diagram of electric propulsion system
Electric Power SRM Load

Supply Converter
Current Position
Sen sin g Sen sin g
Unit Unit
Command
Controller
Signal
Fig. 11 SRM motor drive
5 Conclusion
In this paper, the working operation and different motor drives used in HEVs are
discussed. According to motor drives, the SRM has many advantages compared to
IM, PMSM, and BLDC motor drives. It shows clearly that the SRM motor drives
Table 3 Comparison of special electrical motors

Items Maximum BLDC motor PMSM SRM
Power density 10 9 10 8
Overload 10 7 7 8
High-speed range 20 9 10 8
Control 20 15 15 16
Noise 10 8 8 6
Size and weight 10 6 8 5
Ruggedness 10 8 9 7
Torque ripple 20 14 12 18
Maintenance 10 8 8 9
Manufacturing 20 14 12 18
Cost 30 20 18 26
Total 180 128 135 146
are highest in number compared to all-electric motor drives. The power electronic
converters are used for the conversion and control of HEV, and dual-inverter PMSM
drive with two sources has given better dynamic characteristics compared to a single
source. The electrical configuration of the HEVs is explained with a neat block
diagram and the electric propulsion system of HEVs was discussed with various
drives.
References
1. Hoque MM, Hannan MA, Mohamed A, Ayob A (2017) Battery charge equalization controller
in electric vehicle applications: a review. Renew Sustain Energy Rev 75:1363–1385
2. Hannan MA, Lipu MSH, Hussain A, Mohamed A (2017) A review of lithium-ion battery state
of charge estimation and management system in electric vehicle applications: challenges and
recommendations. Renew Sustain Energy Rev 78:834–854
3. Fotouhi A, Auger DJ, Propp (2016) A review on electric vehicle battery modelling: from
lithium-ion toward lithium–sulphur. Renew Sustain Energy Rev 56(2016):1008–1021
4. Sakile R, Sinha UK (2020) Estimation of SOC (state-of-charge) and SOH (state-of health) of
lithium-ion battery by using adaptive techniques. IJITEE J 9(3)
5. Maggetto G, Van Mierlo J (2011) Electrical hybrid vehicle technology; a survey. InTech
6. Ehsani M, Gao Y, Gay SE, Emadi A (2004) Modern electric, hybrid electric and fuel cell
vehicles book. CRC Press
7. Mehmet Ugras Cuma (2015) Tahsin Koroglu, A comprehensive review on estimation strategies
used in hybrid and battery electric vehicles. Renew Sustain Energy Rev 42:517–531
8. Vasebi A, Bathaee SMT, Partovibakhsh M (2008) Predicting state of charge of lead-acid
batteries for hybrid electric vehicles by extended Kalman filter. Energy Convers Manag
49:75–82
9. Edward WCLO (2009) Review on the configurations of Hybrid Electric vehicles. In: IEEE
conference on power electronics systems and applications
10. Liu D, Li L, Song Y, Wu L, Peng Y (2019) Hybrid state of charge estimation for lithium-ion
battery under dynamic operating conditions. Electr Power Energy Syst 0142-0615
11. Lulhe AM, Date TN (2015) A technology review paper for drives used in electrical vehicle
(EV) & hybrid electrical vehicles (HEV). In: 2015 International conference on control,
instrumentation, communication and computational technologies (ICCICCT)
12. Alagarsamy T, Moulik B (2018) A review on optimal design of hybrid electric vehicles and
electric vehicles. In: 2018 3rd international conference for convergence in technology (I2CT),
The Gateway Hotel, XION Complex, Wakad Road, Pune, India, 06–08 Apr 2018
13. Yao LW, Aziz JA, Kong PY, Idris NRN. Modelling of lithium-ıon battery using
MATLAB/Simulink. In: IEEE conference. 978-1-4799-0224-8
14. Kim S (2008) Nonlinear state of charge estimator for hybrid electric vehicle battery. IEEE
Trans Power Electron 23(4)
15. Yang L, Cai Y, Yang Y, Deng Z (2019) Supervisory long-term prediction of state of available
power for lithium-ion batteries in electric vehicles. Appl Energy 0306-2619
16. Shen Y (2009) Adaptive online state-of-charge determination based on neuro-controller and
neural network. Energy Convers Manag 0196-8904
17. Cuma MU, Koroglu T (2015) A comprehensive review on estimation strategies used in hybrid
and battery electric vehicles. Renew Sustain Energy Rev 42:517–531
18. Ma Z, Murgovski N, Cui S (2019) Predictive energy management for electric variable
transmission HEV. IFAC Papers Online 52–5:417–422
19. Sakile R, Kumar P, Sinha UK (2020) Modelling of lithium-ion battery using matlab/simulink
for electric vehicle applications. Recent Adv Power Electron Drives 291–299
20. Messier P, Nguyen B-H et al (2020) Disturbance observer-based state-of-charge estimation for
Li-ion battery used in light electric vehicles. J Energy Storage 27:101144
21. Chen Q, Jiang J, Raun H, Zhang C (2017) Simply designed and universal sliding mode observer
for the SOC estimation of lithium-ion batteries. IET Power Electron 10(6):697–705
22. Sakile R, Sinha UK (2020) Estimation of SOC (state- of-charge) and SOH (state-of-health) of
lithium-ion battery by using adaptive techniques. IJITEE J 9(3)
23. Spurgeon SK (2008) Sliding mode observers: a survey. Int J Syst Sci 39(8):751–764. https://
doi.org/10.1080/00207720701847638
24. Hannan MA, Lipu MSH, Mohamed HA (2017) A review of lithium-ion battery state of
charge estimation and management system in electric vehicle applications: challenges and
recommendations. Renew Sustain Energy Rev 2017(78):834–854
25. Hoque MM, Hannan MA, Mohamed AA (2017) Battery charge equalization controller in
electric vehicle applications: a review. Renew Sustain Energy Rev 2017(75):1363–1385
26. Yan W, Zhang B, et al (2018) Battery management system with Lebesgue sampling-based
extended Kalman filter. IEEE Trans Industr Electron. https://doi.org/10.1109/TIE.2018.284
2782
27. Tang X, Wang Y, Chen Z (2015) A method for state-of-charge estimation of LiFePO4 batteries
based on a dual-circuit state observer. J Power Sources 296(2015):23–29
28. Tian Y, Chen C, Xia B et al (2014) An adaptive gain nonlinear observer for state of charge
estimation of lithium-ıon batteries in electric vehicles. Energies 7:5995–6012. https://doi.org/
10.3390/en7095995
29. Zou Y, Hu X, Ma H, Li SE (2017) Combined state of charge and state of health estimation over
lithium-ion battery cell cycle lifespan for electric vehicles. J Power Sources 273(2015):793–803
30. Zhang ZL, Cheng X, Lu ZY, Gu DJ (2017) SOC estimation of lithium-ion batteries with AEKF
and wavelet transform matrix. IEEE Trans Power Electron 2017(32):7626–7634
Comparative THD Analysis of Multilevel
Inverter Using Different Multicarrier
PWM Schemes
Shashi Shekhar Tripathi, Manoj Kumar Kar, and Arun Kumar Singh
1 Introduction
Modified MLI topology seven-level pack u-cell inverter for seven-level output voltage
with lower harmonic distortion for photovoltaic (PV) application is discussed in [1].
Based on input DC voltage sources cascaded H-bridge MLI has two configurations
symmetric (equal magnitude) and asymmetric (unequal magnitude). For same higher
levels, output asymmetric configuration has less THD as per IEEE standard than
symmetric configuration [2]. High-frequency magnetic link (HFML)-based cascaded
H-bridge MLI has advantages of reduced device count, size, and cost than conven-
tional invertors is discussed in [3]. The cascaded H-bridge topology of MLI is more
competent than diode clamped and capacitor clamped MLI. With increase in number
of levels in output waveform of CHBMLI the harmonic distortion and voltage stress
dv/dt reduce [4]. A new cascaded switch-ladder MLI topologies gives more number
of output voltage levels with decreased switch count and voltage rating of components
which is discussed in [5]. Different MLI topologies and its application are discussed
in [6]. Three-phase CHBMLI topology with reduced power electronic components
and hybrid control technique to produce gate pulses for switches are discussed in
[7]. Cascaded H-bridge MLI topology with different control techniques for output
voltage level generates less THD which is discussed in [8]. A self-balanced step-up
MLI topology to reduce total standing voltage and PIV for switches is discussed in
[9]. Different reduced switch MLI topologies give higher voltage levels to improve
power quality and reduce filter requirements which are discussed in [10]. A new
three-phase modular multilevel inverter (MMLI) topology using both SPWM and
staircase modulation techniques generates output voltage levels with reduced switch
count, voltage stress, and harmonic distortion than other MLI topologies which
are discussed in [11]. New cascade MLI topology generating both even and odd
voltage levels at output with reduced power electronic components and cost by using
S. S. Tripathi (B) · M. K. Kar · A. K. Singh

EE Department, NIT Jamshedpur, Jamshedpur, Jharkhand, India
70 S. S. Tripathi et al.
series connection of extended sub-multilevel units is discussed in [12]. A novel

MLI topology using sinusoidal pulse width modulation techniques to generate gate
pulses and filters to reduce THD is discussed in [13]. The advantages of multilevel
inverter and various compromises for reducing power semiconductor components
are discussed in [14]. The novel unified power flow controller (UPFC) with cascaded
H-bridge MLI offers many advantages and applications than conventional configu-
ration which is discussed in [15]. For large power and medium voltage application,
a modified three-phase MLI configuration using PWM technique is discussed in
[16]. Cascaded MLI giving positive and negative voltage levels at output by using
different algorithm is discussed in [17]. Symmetric MLI with reduced switch count,
cost, and loss compared to cascaded H-bridge MLI generates higher levels of output
voltage which are discussed in [18]. In MLI, SHEPWM techniques diminish more
THD value related to ordinary PWM techniques which are discussed in [19]. A new
MLI topology using in-phase disposition (IPD) SPWM techniques for high-voltage
and high-power application is discussed in [20, 21]. Modern advancement in MLI
topologies, modulation, and control techniques is discussed in [22]. Symmetric and
asymmetric topologies of MLI producing all levels of output voltage with reduc-
tion in switches, cost, and size are discussed in [23]. A current control algorithm
with discrete-time model for CHBMLI generating all levels of output voltage is
discussed in [24]. Particle swarm optimization (PSO) method to solve harmonic elim-
ination problems with unequal DC sources in cascaded H-bridge MLI is discussed
in [25]. Series connection of CHB cells with photovoltaic (PV) module for MPPT
is discussed in [26]. The three-phase configuration of CHBMLI circuits and various
forms of multi-carrier sinusoidal pulse width modulation (MCSPWM) techniques
are discussed in [27]. Emerging topologies, the most important methods of control
and modulation, and recent MLI applications are discussed in [28].
2 Cascaded H-bridge MLI
A seven-level CHBMLI is depicted in Fig. 1. It consists of two stages: (i) DC-DC

converter and (ii) inverter. Every H-bridge cell has four switches. Single phase full
bridge inverters with separate DC source are connected in series to form CHBMLI.
The objective of series connection is to increase the number of levels of output voltage
and to reduce THD. When ‘K’ H-bridges are linked in series, a maximum KE output
voltage of 2K + 1 separate output voltage level is obtained, where E is input voltage
of single H-bridge. When K = 3, the output voltage has seven levels 3E, −3E, 2E,
−2E, E, −E, and 0, respectively. Switching state of seven-level output waveform
which is given in Table 1.
Comparative THD Analysis of Multilevel Inverter Using … 71
Fig. 1 Circuit for seven-level CHBMLI

Table 1 Switching table for seven-level CHBMLI

Voltages Q11 Q2 Q33 Q44 Q55 Q66 Q77 Q88 Q99 Q110 Q111 Q112
3*E 1 0 0 1 1 0 0 1 1 0 0 1
2*E 1 0 0 1 1 0 0 1 1 0 1 0
E 1 0 0 1 1 0 1 0 1 0 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0
−E 0 1 1 0 1 0 1 0 1 0 1 0
−2*E 0 1 1 0 0 1 1 0 1 0 1 0
−3*E 0 1 1 0 0 1 1 0 0 1 1 0
3 Problem Definition
The inverter has attracted a great interest to be used in heavy duty industries and
large power applications, but it is difficult to connect a single power semiconductor
switch directly to medium voltage grids, even the two-level inverter with high-voltage
grids. Two-level inverter has greater harmonics distortion that requires a compli-
cated filtering circuit to obtain the sinusoidal waveform. To overcome these draw-
backs of two-level inverter, multilevel inverter is used. MLI has lesser harmonic
distortion and does not need a filtering circuit to produce smoother sinusoidal wave-
form. Cascaded H-bridge topology of MLI is preferred compared to flying capacitor
and diode clamped topologies because it includes small number of semiconductors
components. Therefore, the weight and price of this type of topology are lower than
that of other two topologies.
In this study, out of different carrier-based PWM techniques, a pure sine wave
is obtained by providing appropriate carrier frequency. The model is designed with
minimum number of switches to reduce the complexity, cost, and switching losses.
The performance of the proposed technique is analyzed using IPD, POD, and APOD
techniques. The proposed technique also gives lesser torque ripples which is clear
from the FFT analysis.
4 Sinusoidal PWM Scheme for Inverter
PWM (pulse width modulation) is the inverter’s internal control tool. A sinusoidal
(reference) wave of low frequency is compared with a triangular (carrier) wave of
high frequency in the PWM scheme to produce gate pulses for switches.
The output pulses are not identical, i.e., they are having the variable width of
pulses. The width of pulse varies in accordance with the magnitude of sinusoidal
waveform. Pulses are generated using SPWM scheme and provided to switches for
different levels of output voltage. THD percentage for 3, 5, 7, 9, 11, and 13 levels of
output voltage is discussed and compared in table below.
Time(s)
Fig. 2 IPD modulation technique
5 Multicarrier PWM (MCPWM) Scheme for Inverter
To get output waveform with reduced harmonic contents of MLI, it is necessary

to implement a proper modulation technique. MCPWM scheme is one of the best
methods of pulse generation for switches due to simple architecture and simpler
implementation. In MCPWM scheme, the reference wave is a sinusoidal signal and
the carrier wave is a triangular signal. ‘m − 1’ numbers of triangular carriers are
required for an MLI having ‘m’ number of voltage levels. For all carrier signals,
the frequency and the Vp-p remain same. The signal of reference and carrier are
compared and therefore the gate pulses are generated.
MCPWM schemes are categorized as ‘level shifted’, ‘phase shifted’, and ‘hybrid’
PWM. Level-shifted PWM is of three types: in-phase disposition (IPD), phasekop-
positekdisposition (POD), and alternative phase opposite disposition (APOD). In
IPD, signal is in phase for all carriers (Fig. 2).
The carriers above zero reference in POD are in phase but are in reverse to those
present below the null reference. All carriers in APOD are alternatively in reverse
disposition (Figs. 3 and 4).
6 Benefits of IPD Modulation Technique
In the IPD modulation technique, the voltage (L-L) is reduced because the fcarrier
harmonics are made as common mode voltage of fcarrier and thus cancelled out
in voltage (L-L) [21]. It is advantageous over POD and APOD techniques. This
approach is valid for all types of multilevel inverters.
Time(s)
Fig. 3 POD modulation technique
Time (s)
Fig. 4 APOD modulation technique
7 Simulation Result
The simulation result using MATLAB technique is discussed in this paper. The
frequency of carrier signal is 1 kHz. The frequency of reference signal is 50 Hz. The
reference signal (sinusoidal) is compared with carrier signal (triangular) to produce
gate pulses for the switches of CHBMLI. The voltage output of 13-level using IPD,
POD and APOD schemes are shown in Figs. 5, 7 and 9 respectively. The THD% of
13-level CHBMLI using IPD, POD, and APOD PWM schemes are shown in Figs. 6,
8 and 10 respectively. In IPD scheme, THD is 8.91%. In POD scheme, THD is 9.37%.
For E = 50 V, voltage output waveform of 13-level CHBMLI using APOD scheme
is depicted in Fig. 9. In APOD scheme, THD is found to be 9.67%.
The value of modulation index is always kept one.
Fig. 5 Voltage output of 13-level using IPD scheme
Fig. 6 THD of 13-level CHBMLI using IPD scheme

Fig. 7 Voltage output of 13-level using POD scheme
Fig. 8 THD of 13-level CHBMLI using POD scheme

Fig. 9 Voltage output of 13-level using APOD scheme
Fig. 10 THD of 13-level CHBMLI using APOD scheme

Table 2 THD percentage for

No. of levels THD percentage (%)
different levels of output
voltage 3 55.88
5 26.89
7 16.43
9 16.79
11 14.04
13 8.58
Table 3 Comparative
No. of levels THD%
analysis of THD
percentage after using IPD POD APOD
different schemes for different 3 53.21 55.87 54.03
levels of output voltage
5 24.98 26.89 28.16
7 16.49 18.07 18.98
9 12.94 17.59 15.37
11 10.31 10.55 12.16
13 8.91 9.37 9.67
8 Conclusion
IPD, POD, and APOD schemes which are based on MCPWM have been effectively
applied for 3, 5, 7, 9, 11, and 13 levels of output voltage in CHBMLI. THD percentage
for these different levels is depicted in Table 3. THD percentage for different levels
of output voltage using SPWM schemes is depicted in Table 2. From Tables 2 and
3, it is obvious that as the number of levels enhances THD% reduces. Among all
MCPWM schemes, IPD has the lowest THD% for same level output voltage.
References
1. Vahedi H, Sharifzadeh M, Al-Haddad K (2018) Modified seven level pack u-cell inverter for
photovoltaic application. IEEE Trans Emerg Sel Topics Power Electron 6(3):1508–1516
2. Dhanamjayulum C, Meikandasivam S (2018) Implementation and comparison of symmetric
and asymmetric multilevel inverters for dynamic loads. IEEE Access 6:738–746
3. Hasan M, Abu-siada A, Islam SM, Dahidah M (2018) A new cascaded multilevel inverter
topology with galvanic isolation. IEEE Trans Ind Appl
4. Singh G, Garg VK (2017) THD analysis of cascaded h-bridge multilevel inverter. In: 4th IEEE
International conference on signal proccessing, computing, and control, 21–23 Sept 2017
5. Alishah RS, Hosseini SH, Babaei E, Sabahi M (2017) Optimal design of new cascaded switch-
ladder multilevel inverter structure. IEEE Trans Ind Electron 64(3):2072–2080
6. Akagi H (2017) Multilevel converters: fundamental circuits and systems. Proc IEEE
105(11):2048–2065
7. Sabayasachi S, Borghate VB, Karasani RR, Suryawanshi HM (2017) Hybrid control technique
based three-phase cascaded multilevel inverter topology. IEEE Access 5:26912–26921
8. Kannan C, Mohanty NK, Selvarasu R (2017) A new topology for cascaded h-bridge multilevel
inverter with PI and fuzzy control. In: 1st International conference on power engineering,
computing and control, pp 917–926
9. Taghvaie A, Adabi J, Rezanejad M (2017) A self-balanced step-up multilevel inverter based
on switched-capacitor structure. IEEE Trans Power Electron 33(1):199–209
10. Natarajan P, Kaliannah P (2017) A comprehensive review on reduced switch multilevel inverter
topologies, modulation techniques and applications. Renew Sustain Energy Rev 76:1248–1282
11. Salem A, Ahmad M, Orabi M, Ahmed EM (2016) Study and analysis of new three-phase
modular multilevel inverter. IEEE Trans Ind Electron
12. Alishah RS, Hosseini SH, Babaei E, Sabahi M, Gharehkoushan AZ (2016) Optimal design
of new cascaded multilevel converter topology based on series connection of extended sub-
multilevel units. IET Power Electron 9(7):1341–1349
13. Hsieh C, Liang T, Chen S, Tsai S (2016) Design and implementation of a novel multilevel
dc-ac inverter. IEEE Trans Ind Electron
15. Peng F, Lui Y, Yang S, Zhang S, Gunasekaran D, Karki U (2015) Transformer-less unified
power flow controller using the cascaded multilevel inverter. IEEE Trans Power Electron
16. Sanjeevikumar P, Blaabjerg F, Wheeler PW, Olorunfemi J (2015) Three-phase multilevel
inverter configuration for open winding high-power application. IEEE
17. Babaei E, Laali S (2015) Optimum structures of proposed new cascaded multilevel inverter
with reduced number of components. IEEE Trans Ind Electron 62(11):6887–6895
18. Oskuee MJ, Karimi M, Ravadanegh SN, Gharehpetian GB (2015) An innovative scheme of
symmetric multilevel voltage source inverter with lower number of circuit devices. IEEE Trans
Ind Electron
19. Ashok B, Rajendran A (2013) Selective harmonic elimination of multilevel inverter using
SHEPWM technique. Int J Soft Comput Eng 3:79–82
20. Najafi E, Yatim AH (2012) Design and implementation of a new multilevel inverter topology.
IEEE Trans Ind Electron 59(11):4148–4154
21. Kar MK, Giri P, Gupta NK, Singh AK (2021) Control of a three-phase diode clamped multilevel
ınverter using phase disposition modulation scheme. In: Kumar J, Jena P (eds) Recent advances
in power electronics and drives. Lecture notes in electrical engineering, vol 707. Springer,
Singapore. https://doi.org/10.1007/978-981-15-8586-9_6
22. Kouro S, Malinowski M, Pou J, Lenon JI, Franquelo LG, Perez A (2010) Recent advances and
industrial applications of multilevel converters. IEEE Trans Ind Electron 57(8):2553–2580
23. Ahmed RA, Mekhilef S, Ping H (2010) New multilevel inverter topology with minimum no.
of switches. In: IEEE TENCON, pp 1862–1867
24. Cortes P, Wilson A, Rodriguez J, Abu-Rub H (2010) Model predictive control of multilevel
cascaded h-bridge inverters. IEEE Trans Ind Electron 57:2691–2699
25. Taghizadeh H, Tarafdar M (2010) Harmonic elimination of cascade multilevel inverters with
nonequal dc sources using particle swarm optimization. IEEE Trans Ind Electron 57(11):3678–
3684
26. Villanneva E, Correa P, Rodriguez J, Pacas M (2009) Control of a single-phase cascaded h-
bridge multilevel inverter for grid connected photo-voltaic systems. IEEE Trans Ind Electron
56(11):4399–4406
27. Kar MK, Mansoori M, Kumar S, Gupta SK (2021) Harmonic elimination of a T-type multilevel
ınverter based on multistate switching cell. In: Kumar J, Jena P (eds) Recent advances in power
electronics and drives. Lecture notes in electrical engineering, vol 707. Springer, Singapore.
https://doi.org/10.1007/978-981-15-8586-9_9
28. Rodriguez J, Slai J, Peng FZ (2002) Multilevel inverters: a survey of topologies, control, and
applications. IEEE Trans Ind Electron 49(4):724–738
Design and Implementation
of Two-Phase Interleaved Boost
Converter with Voltage Multiplier
for Fuel Cell Hybrid Electric Vehicles
Melisa Miranda, M. Komal, and Shweta P. Hegde
1 Introduction
Fuel Cell (FC) technologies can be considered an imminent power source for auto-
mobile applications because they are clean, highly efficient, and highly reliable. The
power capability of FC systems is considerably good while operating in steady state.
But the instantaneous power demands of FCs are relatively slow. Therefore, FCs are
incorporated with Energy Storage Systems (ESS) like batteries and supercapacitors
to meet the demands of the Hybrid Electric Vehicles (HEV). And also, fuel cells
generally have low voltage ratings. But higher voltage level is required for driving
the motors. So, there is a need for boosting the power and the voltage by using boost
converter like a DC system of multiple DC-DC converters to interface between FC
or ESS and the DC bus in the drivetrain of FCHEVs [1].
In this application, maintaining high efficiency using the traditional DC-DC boost
converter becomes a major challenge. Simultaneously, for high-power applications
such as electric vehicles, handling the large input current and large output voltage
becomes difficult [2]. An effective way is to use a two-phase Interleaved Boost
Converter (IBC) with a Voltage Multiplier (VM) cell.
Various converter circuits with which a higher gain in voltage can be achieved
are discussed in [3, 4]. The major drawbacks of these circuits are the large number
of components, the complexity, and the remarkably high input current ripple. In this
paper, we are adding a voltage multiplier to overcome these disadvantages.
Powertrain system of FCHEV comprises a DC-DC boost converter linking the
FC system and DC-AC inverter, and also a bidirectional DC-DC boost converter that
joins the ESS to the DC-AC inverter as shown in Fig. 1.
M. Miranda · M. Komal (B) · S. P. Hegde

Department of Electronics and Communication Engineering, PES University, Bengaluru, India
M. Miranda
e-mail: melisamiranda@pes.edu
82 M. Miranda et al.
Fig. 1 Block diagram of the drivetrain system in FCHEV
Two conventional DC-DC boost converters connected in parallel with each other
form an IBC. The proposed structure of two-phase IBC is composed of two parallel
phases each having a diode and a power switch. With this approach, the passive
component size can be scaled down easily. This structure also improves the transient
response and efficiency of the FCHEV system [5].
Voltage multiplier circuits are simple circuits consisting of switched capacitors
and diodes. They are efficient and have a low cost. They are designed in such a
way so as to acquire large output DC voltage. The distribution of the current in
the voltage multiplier lowers the effective current decreasing the conduction losses.
Also, the currents through the diodes becomes zero before turning off alleviating the
diode recovery losses. There are two ways of connecting the VM cells to the DC-DC
converter circuit.
• The VM cell can be connected in the middle of the circuit preferably after the
power switch, so as to scale down the voltage stress.
• The VM cell can be connected in between the output of the FC system, i.e., at the
input stage of the proposed converter design and the input inductors, in order to
rectify pulsating DC or AC voltages.
The design of a two-phase IBC incorporated with a VM cell in the middle of the
circuit is implemented in this paper. Adding VM cell reduces input current ripple
and also a high voltage gain can be obtained without adding to the complexity. This
is one of the main advantages of adding a VM cell. A comprehensive simulation
analysis is conducted by using MATLAB-SIMULINK to test the performance of the
system.
2 Modeling of the Converter
The converter put forward in this paper comprises a two-phase IBC integrated with
VM cell as shown in Fig. 2.
Design and Implementation of Two-Phase Interleaved Boost … 83
Fig. 2 Two-phase IBC with VM cell
The proposed structure is inferred from the DC-DC boost converter and the struc-
ture has two input inductors (L1, L2), one for each parallel phase. The inductor is
designed in the same way as that of traditional DC-DC boost converter. The ripple in
the input current can be scaled down by increasing the number of stages that operate
in different phases. The proposed converter is composed of two parallel stages, i.e.,
S = 2 and one VM cell, i.e., N = 1. The number of VM cells and parallel stages
is obtained by analyzing the interdependence of input and output voltages and also
the peak voltage of the power switch and current stress. There is a major difference
between the conventional low-frequency voltage multipliers that usually operate at
low-frequency range and the voltage multipliers incorporated with DC-DC boost
converters put forward in this paper. Increasing the number of cells in the traditional
VM increases the total number of diodes connected in series and also increases the
losses occurred due to conduction. Hence, the efficiency of the system gets reduced
with the increment in the number of VM cell. But this does not occur in the proposed
VM cell, since the current gets distributed between the diodes in the VM cell. Thus,
the addition in the number of VM cells does not result in large conduction losses of
the diodes.
3 Design and Working of the Converter
From Fig. 4, it can be deduced that the two power stages have a phase difference by
180°, which results in the reduction of peak-to-peak ripple current which is shown
in Fig. 3.
The individual power stages of the converter operate at the same frequency of
about 20 kHz as that of single-phase design. But input and output ripple frequencies
are effective at 40 kHz.
Fig. 3 Inductor currents iL1 and iL2
Fig. 4 Currents through the power switches of the converter
The proposed interleaved design calculation uses half the frequency, that is, 20 kHz
and the ripple current is twice as that of the single-phase model. This scales down
the inductance value by half, which can be verified from (15). Also, the total input
inductor current of the single phase gets divided in the two-phase IBC. This reduces
the inductor specifications such as current and power ratings that are mentioned in
[6]. The losses occurred due to conduction are also substantially low compared to
the conventional single-phase boost converter.
Discontinuous Conduction Mode (DCM) can be used in low-power applications
when the energy requirement of the load is significantly less. Since the proposed
model deals with high-power applications, Continuous Conduction Mode (CCM) is
considered for the analysis and operation. The CCM design of proposed two-phase
IBC is identical to the traditional single-phase boost converter. But the duty cycle
formula and inductance value to be chosen are different. DCM design is compar-
atively different from the CCM design. Also, the duty cycle changes according to
the load in DCM, but duty cycle value is independent of load in CCM design as
mentioned in [7].
In the proposed circuit, the inductor currents are out of phase by 180 degrees as
observed from the graph in Fig. 3. Thus, it can be concluded that when input inductor
L1 is charging, the other inductor L2 will get discharged and vice versa. The inductor
currents have an average of 12A. The ripple observed in each inductor current is 2 A.
The switching frequency of the inductor currents is observed to be twice as that of the
single-phase design and the ripple is also significantly scaled down. The observed
value of the input current ripple is 1.5 A which indicates that the input current ripple
can be cancelled naturally by using interleaving technique.
4 Operation of the Converter
The analysis of operation of two-phase IBC with VM cell is done by using state-
space averaging technique. The functioning of the converter is illustrated with the
help of four switching stages of the converter. The equations for each stage can be
put forth in the state-space representation of the form:
ẋ = Ax + Bu (1)
y = C x + Du (2)
where
⎡ ⎤
i L1
x = ⎣ i L2 ⎦ and u = Vin
VC
The functioning of the two-phase IBC with VM cell in CCM with k > 0.5 occurs
in four stages as explained below.
4.1 First Stage: t1 -t0
During this stage, both the power switches S1 and S2 conduct as shown in Fig. 5.
Since the switches act as a short circuit, the diodes will have no significance and
will be in OFF state. The input inductors L1 and L2 store energy supplied from the
FC system. Hence, the inductor current increases and reaches peak.
Fig. 5 Circuit operation during first stage with k > 0.5
Considering the inductor currents, iL1 and iL2 , and output capacitor voltage VC as
state variables, the equations for the state space can be written as
di L1 V in
= (3)
dt L1
di L2 V in
= (4)
dt L2
dv c Vc
= (5)
dt C1 ∗ R 1
⎡ ⎤ ⎡ ⎤
1
0 0 0
⎢ ⎥ ⎢ L11 ⎥
A1 = ⎣ 0 0 0 ⎦ and B1 = ⎣ L ⎦
2
0 0 R−1
1 C1
0
4.2 Second Stage: t2 -t1
Here, in this stage, the power switch S2 is in OFF state. The power switch S1 conducts.
Diodes DS1 and DM2 conduct and corresponding topologies of the converter are
designed as represented in Fig. 6. The output capacitor C1 gets charged due to the
energy dissipated from the input inductors through DS2 and also through CM1 and
DM2 .
Figure 6 shows that the VM cell capacitors CM1 and CM2 are connected in series
through the diode DM2 . Also, each capacitor in VM cell is connected in parallel with
the output capacitor by the diode DS2. Hence, the output capacitor voltage is double
the multiplier capacitor voltage, i.e., (VC = 2VCM ). Here, the node and loop equations
Fig. 6 Circuit operation during second stage with k > 0.5
can be written as
di L2 V in − VC M1
= (6)
dt L2
di L1 V in
= (7)
dt L1
di L2 V in − V c − VC M2
= (8)
dt L2
From (6) and (8),
VC M1 = VC + VC M2 (9)
With the help of these equations, A2 and B2 matrices can be obtained as

⎡ ⎤ ⎡ ⎤
1
00 0
⎢ ⎥ ⎢ L1
⎥
A2 = ⎣ 0 0 −1
L 2 ⎦ and B2 = ⎣
1
L2
− VC M2
L 2 ∗Vin ⎦
00 0 0
4.3 Third Stage: t3 -t2
During the third stage, the power switches S1 and S2 will be in ON state as seen in
Fig. 7.
The energy is stored in both the input inductors as in the first stage. Hence, the
equations of first stage are applicable in this stage. Therefore, A3 = A1 and B3 = B1 .
Fig. 7 Circuit operation during third stage with k > 0.5
4.4 Fourth Stage: t4 -t3
Here, the switch S1 is in OFF state as shown in Fig. 8.

The output capacitor C1 and also the capacitor in the VM cell, i.e., CM2 gets
charged due to the energy dissipated from the inductor L1 . The state-space equations
of this stage and corresponding equations of the second stage are alike and can be
obtained by swapping the subscripts 1 and 2. Thus the equations are
di L1 V in − VC M2
= (10)
dt L1
di L2 V in
= (11)
dt L2
Fig. 8 Circuit operation during fourth stage with k > 0.5

di L1 V in − V c − VC M1
= (12)
dt L1
VC M2 = VC + VC M1 (13)
With the help of these equations, A4 and B4 matrices can be obtained as

⎡ ⎤ ⎡ ⎤
1
00 0
⎢ ⎥ ⎢ L2
⎥
A4 = ⎣ 0 0 −1
L 1 ⎦ and B4 = ⎣
1
L1
− VC M1
L 1 ∗Vin ⎦
00 0 0
The operation in CCM with k < 0.5 is represented in Figs. 9, 10, 11 and 12. The
duty cycle of both the switches is kept same.
Fig. 9 Circuit operation during first stage with k < 0.5
Fig. 10 Circuit operation during second stage with k < 0.5

Fig. 11 Circuit operation during third stage with k < 0.5
Fig. 12 Circuit operation during fourth stage with k < 0.5
4.5 First Stage: t1 -t0
The switch S1 is in ON state and inductor L1 stores the energy supplied from the
power source. The output capacitor gets charged due to the energy dissipated from
the inductor L2 through the diode DS2. Similarly, the capacitor in VM cell, CM1 gets
charged through diode DM2 .
4.6 Second Stage: t2 -t1
Here, all the switches will be in OFF state. Also, all the diodes in the VM cell will be
in OFF state. Thus, the output capacitor C1 gets charged due to the energy dissipated
from the input inductors L1 and L2 through the diode DS1 . The diode DS2 continues
to be in ON state and conducts the current iL2 .
The switch S2 is in ON state. The energy supplied from the power source gets stored
in the input inductor L2 . The diode DM1 goes to ON state and starts conducting and
also charges the capacitor in the VM cell, CM2 .
This stage is identical to the second stage.
5 Proposed Converter Design
5.1 Static Gain of the Boost Converter
As mentioned in [8], the static gain of the two-phase IBC with VM cell working in
CCM can be represented as (14). This equation is also applicable for the functioning
of the converter with k < 0.5.
Vo (N + 1)
G= = (14)
V in (1 − k)
where N is the number of VM cells and k is the duty cycle of the power switch.
5.2 Magnitude of the Passive Components
Determining the value of input inductance is similar to that of conventional boost

converter as mentioned in [9]. Considering the peak input inductor current ripple as
15% of the average input inductor current, the inductance values can be evaluated
from the equation
k ∗ V in
L1 = L2 = (15)
f ∗ S ∗ i L
where iL is the input inductor current ripple.

Fig. 13 Ripple in the output voltage
The equations for calculating the capacitance values of voltage multiplier stage
and output capacitor are presented in (16) and (17), respectively.
(1 − k) ∗ I in
C M1 = C M2 = (16)
S ∗ (N + 1) ∗ f ∗ Vc
k ∗ Vc
C1 = (17)
f ∗ R1 ∗ Vc
where Vc is the output voltage ripple.

Rearranging (17), the equation for output voltage ripple can be given as
k ∗ Vin
Vc = (18)
C 1 ∗ f ∗ R1
Using this technique, the ripple frequency can be doubled. This results in scaling
down of the output voltage ripple which is represented in Fig. 13.
6 Experimental Results
The experimental model of two-phase IBC with VM cell working in discontinuous

conduction mode [10] designed using MATLAB-SIMULINK is depicted in Fig. 14
(Table 1).
It can be confirmed from Fig. 3 that the design works in DCM mode, since the
value of inductor current reaches zero for a short time period during each cycle. The
analysis is performed under open-loop condition. The output voltage is observed to
be 1000 V for an input of 150 V as depicted in Fig. 15 and also reduces output voltage
ripple from 200 to 180 mV. The experimental voltage gain ratio can be verified with
the theoretical voltage gain G mentioned in (14).
Fig. 14 Experimental circuit of IBC with VM cell
Table 1 Design metrics of

Design metrics Values
the converter
Input voltage 150 V
Output capacitor voltage 1000 V
Switching frequency 20 kHz
Output voltage ripple 180 mV
Multiplier capacitances 40 µF
Output capacitance 500 µF
Inductance 480 µH
Resistive load 850
7 Conclusion
In this paper, the functioning of two-phase IBC with VM cell in continuous conduc-
tion mode has been analyzed using state-space averaging technique. The proposed
converter design is simulated in discontinuous conduction mode using MATLAB-
SIMULINK and the behavior of the design like high voltage gain, reduction of ripples
in current, and voltage values are verified. The input voltage can be boosted by a
factor of 3.33 using interleaving technique. The addition of VM cell to the two-phase
IBC doubles the output voltage by a factor of 6.66. The design significantly reduces
the ripple in the output voltage from 200 to 180 mV. The design proves that the
Fig. 15 Voltage across the output capacitor or the load
two-phase interleaving technique significantly reduces the inductance value than the
single-phase design by dividing the input current into two inductor currents of the
two-phase design.
References
1. Xu H, Wen X, Qiao E, Guo X, Kong L (2005) High power interleaved boost converter in fuel
cell hybrid electric vehicle. In: 2005 IEEE international conference on electric machines and
drives, pp 1814–1819, 15 May 2005
2. Wen J, Jin T, Smedley K (2006) A new interleaved isolated boost converter for high power
applications. In: Applied power electronics conference and exposition, 2006. APEC ‘06.
Twenty-First Annual IEEE, vol 6, pp 19–23, March 2006
3. Pietkiewicz A, Cuk S. A three-switch high-voltage converter. IEEE Trans Power Electron
14(1):177–183
4. Berkovich Y, Axelrod B (2012) High step-up DC-DC converter based on the switched-coupled-
inductor boost converter and diode-capacitor multiplier. In: 6th IET international conference
power electronics, machines and drives (PEMD 2012), pp P43–P43
5. Hegazy O, Mierlo JV, Lataire P (2012) Analysis, modeling, and implementation of a multidevice
interleaved DC/DC converter for fuel cell hybrid electric vehicles. IEEE Trans Power Electron
27(11):4445–4458. https://doi.org/10.1109/TPEL.2012.2183148
6. Betten J, Kollman R. Interleaving DC-DC converters boost efficiency and voltage. Application
report, Texas Instruments
7. Zhang L, Shen G, Chen M, Ioinovici A, Xu D (2012) Two-phase interleaved boost converter
with voltage multiplier under APS control method for fuel cell power system. In: Proceedings
of the 7th international power electronics and motion control conference, Harbin, 2012, pp
963–967. https://doi.org/10.1109/IPEMC.2012.6258925
8. Franco LC, Pfitscher LL, Gules R (2003) A new high static gain nonisolated DC-DC converter.
In: IEEE 34th annual conference on power electronics specialist, 2003. PESC ‘03, Acapulco,
Mexico, vol 3, pp 1367–1372. https://doi.org/10.1109/PESC.2003.1216786
9. Lloyd A, Dyer C (1999) The future of fuel cells. Sci Am 281:64–75
10. Altimania M, Nia MSS, Ferdowsi M, Shamsi P (2019) Analysis and modeling of a non-isolated
two-phase interleaved boost converter with diode-capacitor cells in the DCM. In: 2019 North
American power symposium (NAPS). https://doi.org/10.1109/naps46351.2019.9000260
A Unipolar Phase Disposition PWM
Technique for Reduced Switch Count
Symmetrical Nine-Level Multilevel DC
Link Inverter Topology
Rajakumar Sakile, Marco Rivera, Kasoju Bharath Kumar,

Bandela Supriya, and A. Bhanuchandar
1 Introduction
DC-AC power conversion plays a vital role in the area of generation, transmission,
distribution, and utilization of the electrical energy such as power quality, HVDC
transmission, flexible AC transmission (FACTS), and renewable energy applications
(solar PV, fuel cell, wind) [1]. By nature of the output, DC-AC power converters
(inverter) are classified as square wave inverter, quasi-square wave inverter, two-
level PWM inverter, and multilevel inverters (MLIs) [2]. Due to the low power
quality output, requirement of filters before the load, the two-level PWM inverter
and quasi-square wave inverters are not suitable for the inversion operation [3]. The
MLIs are introduced to overcome above-said drawbacks. MLIs are widely used in
scientific communication and many industries, due to the output of MLIs which are
imitating a sine wave [4]. The MLIs offer many advantages which are
(i) MLIs maintain power quality in high-power applications.
(ii) MLIs have high modularity.
(iii) MLIs has a capability of transformerless operation.
(iv) MLIs have high fault tolerance capability.
R. Sakile (B)
Electrical Engineering Department, NIT Jamshedpur, Jamshedpur, Jharkhand, India
e-mail: 2018rsee006@nitjsr.ac.in
M. Rivera
Faculty of Engineering, Universidad de Talca, Maule, Chile
K. B. Kumar
EEE Department, Mahatma Gandhi Institute of Technology, Hyderabad, Telangana, India
B. Supriya
EEE Department, Chaitanya Bharathi Institute of Technology (A), Hyderabad, Telangana, India
A. Bhanuchandar
Electrical Engineering Department, NIT Warangal, Warangal, Telangana, India
96 R. Sakile et al.
(v) MLIs draw very low input current distortions.

(vi) Filter requirements have been greatly brought down or even eliminated for
higher level of operations.
(vii) MLIs produce low or NO common mode voltages, and thus the over stress
of multilevel motor drive can be eliminated.
Due to the above-said abilities, the MLIs have been impacted in several areas of
industries, such as electric traction, electric vehicles, marine propulsion drives, high-
power conveyors, and also power quality applications like FACTS, HVDC, and static
var compensation [5]. Various basic MLIs topologies are present for the analysis,
and out of all the basic or classical topologies which are used in many industries are
1. Diode clamped MLI (DCMLI).
2. Flying capacitor MLI (FCMLI).
3. Cascaded H-bridge MLI (CHB).
CHB and DCMLI required only one DC source in the operation, whereas FCMLI
required multiple DC sources [6]. The most important property of MLI topology
is modularity. The modularity of CHB structure is high compared to DCMLI and
FCMLI. Practically, DCMLI, FCMLI generates up to five-level output beyond that
the structure becomes complex and device count has been increased. So far CHB is
best for the higher levels and same power rating [7]. When CHB is implemented for
the MLIs topology many number of semiconductor devices are required and design
of control strategies is more complex [8]. Therefore, the new MLI has proposed
to eliminate or reduce semiconductor devices which is known as reduced switch
count MLI or multilevel DC link inverter (MLDCL). For K level inverter, number of
switches required for classical topologies are 2 × (K − 1) and (K + 3) switches are
required for MLDCL topologies [9]. Therefore, the MLDCL topology is better than
classical or existing topologies.
In this paper, Sect. 2 discusses about the description of MLDCL topology, Sect. 3
represents the proposed control scheme used in the circuit, and Sects. 4 and 5 discuss
the simulation results and conclusion.
2 Description of MLDCL Topology
Generally, MLDCL inverters are operated with 100kw load for low inductance motor
drives. MLDCL inverters reduce the capacitors, clamping diodes, and reduce the
switch count [10]. For every switch, one protection circuit, gate driver circuit, heat
sink, and control unit are required, and therefore the initial cost has been increased. By
using MLDCL inverters the cost and volume of the system have drastically reduced.
The single-phase MLDCL inverter with RL load is shown in Fig. 1. It consists of
series connection of symmetrical DC sources and H-bridge [11]. In order to generate
multilevel outputs, MLDCL inverters are separated with level generator and polarity
generator. The level generator is used to generate the multilevel output and polarity
generator is used to change the polarity of the output as per the control strategy.
A Unipolar Phase Disposition PWM Technique … 97
Fig. 1 Single-phase nine-level multilevel DC link inverter with RL load
In polarity generator side, one full H-bridge is required to decide the polarity, and
it consists of four switches named as H1, H2, H3, and H4. In level generator side, four
half-bridge circuits are connected in cascaded manner with eight switches [12, 13].
To get the 0Vdc output, the combination of conducted switches in level generator is
S2, S4, S6, S8 and polarity generator is H1, H4. Similarly to get remaining outputs,
switching sequence is discussed in Table 1. In level generator side, four switches are
operated every time at high switching frequency, whereas in polarity generator side
two switches are operated every time with low or fundamental frequency [14]. The
MLDCL inverter produces required voltage levels with less switching and conduction
losses [15–18]. Three-phase MLDCL inverter with RL load is shown in Fig. 2.
Table 1 Switching sequence

Level number Output voltage ON state switches
for single-phase nine-level
MLDCL inverter 1 4 Vdc S1-S3-S5-S7-H1-H4
2 3 Vdc S1-S3-S5-S8-H1-H4
3 2 Vdc S1-S3-S6-S8-H1-H4
4 1 Vdc S1-S4-S6-S8-H1-H4
5 0 S2-S4-S6-S8-H1-H4
6 −1 Vdc S1-S4-S6-S8-H2-H3
7 −2 Vdc S1-S3-S6-S8-H2-H3
8 −3 Vdc S1-S3-S5-S8-H2-H3
9 −4 Vdc S1-S3-S5-S7-H2-H3
98 R. Sakile et al.
Fig. 2 Three-phase multilevel DC link inverter with RL load
3 Proposed Control Scheme
In order to generate valid gate pulses, control scheme development is very impor-
tant. Generally, control schemes are broadly classified into Pulse Width Modula-
tion (PWM) technique and stepped low-frequency technique. Among the traditional
modulation schemes, the space vector method gives feasible solution with better DC
bus utilization. However, this control technique can be easily applicable up to two-
or three- or five-level inverter topologies according to number of switching states. As
going to higher levels, the SVM technique becomes more complex. Therefore, the
researchers or developers choose the carrier-based modulation technique. In carrier
based, the most famous technique is Phase Disposition (PD), and it generates the
level-shifted carriers within the same phase to generate K level output and it requires
(K − 1) carriers. In this paper, a Unipolar Phase Disposition (UPD) technique has
been proposed. In order to generate K level output, UPD requires only (K − 1)/2
carriers. Almost half of the carriers are reduced as compared with the conventional
PD technique. The proposed control technique is well suitable for generalized struc-
ture of MLDCL topology to any level count. The proposed control scheme is shown
in Fig. 3.
Modulating signal or reference signal is applied to both polarity and level gener-
ators. The modulating signal is compared with the zero, and then it will generate
the gate pulses or switching pulses for the polarity generator from H1 to H4. In
level generator case, the modulating signal is given to the absolute function and it
Fig. 3 Proposed control scheme
Table 2 Carrier count

Control scheme name Carrier level Carrier count
comparison
PD K−1 8
Proposed UPD (K − 1)/2 4
generates unipolar modulating signal. Due to absolute function, the number of carrier
count can be reduced by half. Now the unipolar modulating signal is compared with
the high switching frequency carriers and then it generates gate pulses from S1 to
S8. Finally, this proposed control scheme gives the nine-level output for the inverter
topology. If asymmetrical configurations are increased (1:2:3:4 and 1:2:4:8) with the
same number of switches, the level count can be increased (Table 2).
4 Simulation Results and Discussion
In the proposed control scheme, the modulating signal is given to absolute function,
and then unipolar modulating signal is generated with the reduced carrier count. The
reduced carrier strategy PWM is shown in Fig. 4, and it is clearly observed that the
carrier count has become reduced to four carriers.
From the proposed UPD control technique, nine-level output is generated, and the
nine-level phase voltage is shown in Fig. 6. In level generator side, all the switches
100 R. Sakile et al.
Fig. 4 Reduced carrier strategy PWM scheme level generator side
S1 to S8 are operated with high switching frequency, whereas in polarity generator

side, the switches H1 to H4 are operated with low-frequency mode. The switching
pulses of S1 to S8 and H1 to H4 are shown in Fig. 5. The load current depends on
the RL load.
According to the Total Hormonic Distortion (THD), the phase voltage, line
voltage, and load current have been analyzed. At fundamental frequency, the peak
voltage of the phase voltage is 393.2 V and THD = 14.36% and line voltage at
fundamental frequency is 679.7 V and THD = 12.76%. Similarly, the load current
of the inverter peak current at fundamental frequency is 3.923A and THD = 4.18%.
The phase, line voltages, and load currents are shown in Figs. 7, 8 and 9.
Simulation Parameters
See Table 3.
A Unipolar Phase Disposition PWM Technique …
Fig. 5 Simulation waveforms for the single-phase RL load, both level and polarity generators
101
Fig. 6 Line voltage for three-phase RL load
Fig. 7 FFT analysis of phase voltage
5 Conclusion
In this paper, the single-phase MLDCL inverter with UPD PWM control technique
has been discussed. Basically this topology has the advantages like high modularity,
redundant structure, even power distribution, equal utilization of the DC sources,
great fault-tolerant capability, and equal device blocking voltages in level generator
side. By taking asymmetrical configuration with the same number of switches, it is
Fig. 8 FFT analysis of line voltage
Fig. 9 FFT analysis of load current
possible to generate more number of levels. The proposed UPD control technique
produces nine-level output voltage with reduced carrier strategy. Due to the reduc-
tion of carrier signals, the control complexity and burden on the processor can be
reduced as compared with conventional PD technique. This control scheme is greatly
applicable for any level count of MLDCL inverter topology and it is well suitable
for photovoltaic-based grid-connected applications.
Table 3 Simulation
Parameter Parameter value
parameters
Vdc 100 V
R 100
L 22 mH
Fundamental frequency 50 Hz
Modulation index 0.98
Total blocking voltage 6 p.u
Voltage gain 1
References
1. Omer P, Kumar J, Surjan BS (2020) A review on reduced switch count multilevel inverter
topologies. IEEE Access 8:22281–22302. https://doi.org/10.1109/Access.2020.2969551
2. Kumawat RK, Palwalia DK (2018) A comprehensive analysis of reduced switch count
multilevel inverter. Aust J Electr Electron Eng. https://doi.org/10.1080/1448837X.2019.169
3884
3. Su G-J (2005) Multilevel DC-link inverter. IEEE Trans Ind Appl 41(3):848–854. https://doi.
org/10.1109/TIA.2005.847306
4. Rao SN, Kumar DVA, Babu CS (2013) New multilevel inverter topology with reduced number
of switches using advanced modulation strategies. In: 2013 international conference on power,
energy and control (ICPEC), Sri Rangalatchum Dindigul, pp 693–699. https://doi.org/10.1109/
ICPEC.2013.6527745
5. Zeb K, Khan I, Uddin W, Khan MA, Sathishkumar P, Busarello TDC, Ahmad I, Kim HJ
(2018) A review on recent advances and future trends of transformer less inverter structures
for single-phase grid-connected photovoltaic systems. Energies
6. Rodriguez J, Lai J, Peng FZ (2002) Multilevel inverters: a survey of topologies, controls,
and applications. IEEE Trans Ind Electron 49(4):724–738. https://doi.org/10.1109/TIE.2002.
801052
7. Belekar RA, Nakade RP et al (2013) Single phase 9 level MLDCL inverter with half bridge
cell topology: performance analysis. In: 2013 international conference on power, energy and
control (ICPEC)
8. Siddique MD, Rawa M et al (2021) A new cascaded asymmetrical multilevel inverter based
on switched dc voltage sources. Int J Electr Power Energy Syst 128:106730
9. Selvakumar K, Patel PD (2021) Single phase five-level inverter topology with reduced number
of switches for PV application. Mater Today Proc
10. Maheswari KT, Bharanikumar R et al (2020) A comprehensive review on cascaded H-bridge
multilevel inverter for medium voltage high power applications. Mater Today Proc
11. Eguchi K, Shibata A et al (2020) An inductor-less step-up/step-down multilevel inverter with
a single input source. Energy Rep 6(Supplement 2):146–152
12. Palle K, Bhanuchandar A (2020) A novel modified voltage oriented control of an active of
an active front-end rectifier used for PMSG based wind turbine systems. IJITEE 9(8). ISSN-
2278–3075
13. Su GJ, Adams DJ (2001) Multilevel DC link inverter for brushlesspermanent magnet motors
with very low inductance. İn: Conference on REC 2001 IEEE-IAS annual meeting, vol 2, pp
829–834
14. Lai JS, Peng FZ (1995) Multilevel converters—a new breed of power converters. İn: Conference
on REC 2001 IEEE-IAS annual meeting, pp 2348–2356
15. Nabae A, Takahashi I, Akagi H (1981) A new neutral-point-clamped PWM inverter. IEEE
Trans Ind Appl 17(5):518–523
16. Kumar KB, Bhanuchandar A, Mahesh C (2021) A novel control scheme for symmetric seven
level reduced device count multi-level DC link (MLDCL) inverter. In: 2021 international confer-
ence on sustainable energy and future electric transportation (SEFET), Hyderabad, India, 2021,
pp 1–4. https://doi.org/10.1109/SeFet48154.2021.9375714
17. Bhanuchandar A, Murthy BK (2020) Single phase nine level switched capacitor based grid
connected inverter with LCL filter. In: 2020 3rd international conference on energy, power and
environment: towards clean energy technologies, 2021, pp 1–5. https://doi.org/10.1109/ICE
PE50861.2021.9404491
18. Sathisha D (2021) A simplified control approach of AC-DC-TLBC for bi-polar DC microgrid
applications. Turk J Comput Math Educ (TURCOMAT) 12(2):818–824. https://doi.org/10.
17762/turcomat.v12i2.1090
A Novel Current Control Scheme
for Three-Phase Three-Level Grid-Tied
Neutral Point Clamped Inverter
Kowstubha Palle and Bandela Supriya
1 Introduction
The generation of clean energy is becoming a challenge due to shortage of the

resources, geographical constraints along with the drastic increase in power demand
by population [1]. Therefore, there is a need to explore alternative means for the
generation of electricity like solar and wind energy systems [2, 3]. In modern power
conversion systems, the grid converters (GC) play a vital role as they are much
flexible in converting different renewable sources such as solar, wind energy into
corresponding DC or AC systems [4, 5]. In comparison to the last decade, renewable
resources-based electricity generation has increased nowadays.
Different innovative control methods are being constantly employed on photo-
voltaic panels and wind turbines so that they can perform fully as an energy source
[6]. These control methods are employed on inverters, which are responsible for grid
synchronization and energy conversion [7, 8]. The performance of these inverter
topologies may vary from off-grid to grid connected and also with/without an
isolation transformer along with its ability to operate in closed-loop configuration
[9, 10].
These inverter topologies have different levels with an influence on the switching
pattern and resolution of the inverter [10, 11]. In addition, these selected topologies
determine the efficiency, utilization, size, weight, cost, etc. that affects the final
product. In this paper, an emphasis on neutral point clamped (NPC) inverter is carried
out and a simple and novel control scheme is proposed. In this scheme, PI controller
and phase-locked loop (PLL) are not used unlike that of conventional space vector
modulation (SVM) [12, 13]. Therefore, the computational burden on the processor
reduces in comparison with the conventional SVM technique and is simple to execute.
NPC inverter topology is widely used among various types of multilevel inverter
topologies [14, 15]. In NPC inverter, a combination of clamping diodes and cascaded
K. Palle (B) · B. Supriya

EEE Department, Chaitanya Bharathi Institute of Technology (A), Hyderabad, Telangana, India
108 K. Palle and B. Supriya
Fig. 1 Different modulation schemes—multilevel inverter topologies
DC capacitors is used to produce AC voltage waveforms with multiple levels [16,

17]. These inverters are used widely in large-power medium-voltage (MV) drives [18,
19] and grid-connected applications. In comparison with other multilevel inverters,
NPC inverter will provide reduced dv/dt stress and total harmonic distortion (THD)
in its AC output voltage. In this topology, the two cascaded DC link capacitors with
midpoint as the neutral point makes the DC input voltage to split into two equal
voltages.
In the field of power electronics, the development of control schemes for different
power converters is a biggest thrust area for research. Summarized details of the
most established control schemes for different power electronic converters are given
in Fig. 1. Conventionally, different modulation schemes are available for multilevel
inverter topologies and are depicted in Fig. 1.
Many pulse width modulation (PWM) techniques are developed for inverter
circuits to control the fundamental component of output voltage with reduced THD
[20, 21]. In the recent past, many topologies like a seven-level multilevel DC
link inverter and a nine-level switched capacitor-based inverter were proposed as
compared to grid-connected systems [22–24].
Among various PWM techniques, SVM is widely implemented for multilevel
inverters. But in this technique, mathematical complexity for every switching state is
predominant [25]. In this paper, a simple and novel current control scheme for three-
phase three-level grid-connected NPC inverter is proposed. This proposed scheme
is very much suitable for a renewable energy grid-connected system that is shown in
Fig. 2.
In this paper, the main focus of work is carried out in the shaded portion of Fig. 2
where it is assumed that DC link capacitors are balanced by DC link chopper circuit.
A Novel Current Control Scheme for Three-Phase … 109
Fig. 2 Grid-connected wind energy conversion system
This paper is organized as follows, after the introduction, the description of the
NPC inverter and its operation are explained in Sect. 2, the proposed control technique
is presented in Sect. 3. Section 4 analyzes the simulation results which is followed
by the conclusion in Sect. 5.
2 Description of NPC Inverter and Its Operation
Figure 3 represents a three-level three-phase NPC inverter circuit with 3-ϕ grid. In
this diagram, leg A consists of four switching devices from ga1 to ga4 with diodes
D1 to D4 connected in anti-parallel and clamping diodes Da1 and Da2 (Table 1).
The capacitors Cd1 and Cd2 are called as equalizing capacitors used to equalize
the voltages as E or Vdc /2 taken from Vdc . The number of components used for
Fig. 3 Three-phase three-level NPC inverter circuit with 3-ϕ grid

Table 1 Generalized expressions for selecting the number of components used for one leg of
three-phase k-level NPC inverter
Power component NPC
Main switching devices 2 (k − 1)
Anti-parallel diodes 2 (k − 1)
Clamping diodes (k – 1) (k − 2)
DC bus capacitors (k − 1)
Flying capacitors zero
Table 2 Switching pattern for three-phase three-level NPC inverter circuit

Switching state Device switching status (Phase A) Pole voltage, VAZ
ga1 /S1 ga2 /S2 ga3 /S3 ga4 /S4
1 (or) + On On Off Off E or Vdc /2
0 (or) 0 Off On On Off 0
−1 (or) − Off Off On On −E or −Vdc /2
three-phase p-level NPC inverter for one leg is given in Table 2 with generalized
expressions. The switching pattern for three-level three-phase NPC inverter circuit
with IGBTs as switches is given in Table 2.
From Table 2, it is understood that ga1 and ga3 are complementary and similarly
ga2 and ga4 are complimentary. From Table 1, it is clear that the switching transition
from “1” to “−1” makes inverter pole voltage VAZ to go from +E, 0 to −E directly.
This transition causes a high amount of dv/dt stress on switches, unequal dynamic
voltage sharing on switches, and doubling of switching loss. Therefore, the switching
transition from “1” to “−1” is always avoided for NPC inverter. Generally, switching
transitions are taken as [1] to [0], [−1] to [0] or vice versa giving rise to equal
dynamic voltage sharing during the commutation from each switching state. Three-
level three-phase NPC inverter has reduced THD and dv/dt in comparison with two-
level inverters for the same operating specifications. The static voltage equalization
is obtained in NPC, only when the leakage current flow of switches ga1 and ga4 in a
leg is lower than that of switches ga2 and ga3 .
Space vector is a mathematical concept that is useful for visualizing the effect of
three-phase variables in space. The division of sectors and the regions for three-phase
three-level NPC inverter circuit is given in Fig. 4.
Depending on the magnitude of the voltage vectors, they are classified as four
groups and have a total of 33 = 27 switching states. Description of these switching
states (stationary space vectors) along with their vector classification and magnitude
is given in Table 3.
Fig. 4 Division of sectors and regions for three-phase three-level NPC inverter circuit
3 Proposed Control Technique
Figure 5 represents the block diagram of the proposed current and the control scheme
of three-phase three-level grid-tied NPC inverter is given in Fig. 3.
In this block diagram of the control technique, SVM is basically reflected with a
new digital switching approach. The entire explanation of block diagram is presented
as two parts, namely, part A and part B. The working procedure of part A is given as
follows. In this process, three-phase load currents and three-phase injected reference
grid currents are initially converted into αβ frames. These converted αβ frame currents
are compared and the errors so obtained will undergo a conversion process, resulting
in two outputs, namely, magnitude and phase angle as given in Fig. 5. Magnitude
output vector undergoes relational comparisons with a set of a simple error band
signal (EB1 ) or with the summation of two error band signals (EB1 and EB2 ). The
outputs so obtained are multiplied with gains and finally added to obtain the control
port signal—CP1 . This control port signal CP1 can be interpreted with reference
control vector of switching states of SVM.
The other output (angle) obtained from the conversion process is considered for
the remaining operation. The angle gives the exact location of the reference vector
of SVM with its presence in any one of the sector areas as given in Fig. 4. Therefore,
this output is considered for different blocks with different sector areas as mentioned
in Table 3 as small, medium, and large vectors and is used in the block diagram given
in Fig. 5 Data port 2 (DP2 ) is taken from the block consisting of small vector (V1
to V6 ) with multiplicities. The switching states [1 0 0], [1 0 −1] are automatically
obtained from sector selection of medium and large by taking 4 A (peak current)
of reference active current injecting to the grid. Data port 3 (DP3 ) is taken from the
Table 3 Space vectors-27 switching states

Space vector Switching state Vector classification Vector magnitude
V0 [+ + + ][0 0 0][− − −] Zero vector 0
−
→
v + Type − Type Small vector 1
1 3 Vd
−
→
v 1+ [+ 0 0]
−
→
v [0 − −]
1−
−
→
v2 −
→
v 2+ [+ + 0]
−
→
v [0 0 −]
2−
−
→
v3 −
→
v 3+ [0 + 0]
−
→
v [− 0 −]
3−
−
→
v4 −
→
v 4+ [0 + + ]
−
→
v [− 0 0]
4−
−
→
v5 −
→
v 5+ [0 0 + ]
−
→
v [− − 0]
5−
−
→
v6 −
→
v 6+ [+ 0 + ]
−
→
v [0 − 0]
6−
√
−
→
v7 [+ 0 −] Medium vector 3
3 Vd
−
→
v [0 + −]
8
−
→
v9 [− + 0]
−
→
v [− 0 + ]
10
−
→
v 11 [0 − + ]
−
→
v [+ − 0]
12
−
→
v 13 [+ − −] Large vector 2
3 Vd
−
→
v [+ + −]
14
−
→
v 15 [− + −]
−
→
v [− + + ]
16
−
→
v 17 [− − + ]
−
→
v [+ − + ]
18
summation of sector area comprising of medium and large vectors (V7 -V12 and V13 -
V18 ). The explanation for part B is given as follows. The block truncated value (first
input) which is given in part B just works like a digital multiple port digital switch.
The generated CP1 , DP2 , DP3 , and additional zero vector V19 (DP1 ) are given to
truncated block and this block will select proper data port signal (DP3 ) according to
CP1 . The output of this block CP1 is given with saturation limits and the obtained
value CP2 is once again given to truncated block along with all other input vectors
(V1 to V19 ) as data ports. According to CP2 , the respective data port V7 is selected
[1 0 −1] automatically. To this value of V7 , a constant value k is added to generate
desired CPs for abc phases labeled in the block diagram as CP3 , CP4 , and CP5 .In
each phase, these control port values will undergo truncation once again with the
Fig. 5 Block diagram of proposed current control algorithm of three-phase three-level grid-tied
NPC inverter
switching action x-[0 0 1 1], y-[0 1 1 0], z-[1 1 0 0] as data ports giving rise to gate
pulses for three phases according to the injecting reference active grid current 4 A
(peak).
The advantages of the proposed control technique are: there is no need of PI
controller (tuning is not required), phase-locked loop (PLL), and calculation of
dwell time. Finally, this technique reduces the computational burden on processor in
comparison with conventional SVM technique and is simple to execute.
4 Simulation Results and Their Analysis
The specifications considered for current control technique of three-phase three-level

grid-tied NPC inverter are given in Table 4.
For the given specifications, a simulation current control scheme is developed on
MATLAB/Simulink platform. The simulation is run for a period of 5 s and the results
are presented as follows.
The NPC inverter terminal/pole voltage on incorporating the proposed control
scheme for three-phase three-level grid-tied NPC inverter is shown in Fig. 6. It is
understood that three levels (+300 V, 0, −300 V) are obtained on each leg of the
Table 4 Specifications of three-phase three-level NPC grid-tied inverter

Parameter Value
DC link voltage (Vdc1 = Vdc2 ) 300 V
Injected grid current reference value (ia_ref = ib_ref = ic_ref ) in amperes 4A
Grid voltage (peak amplitude) 339 V
Grid frequency 50 Hz
Sampling time (Ts ) 10e−6 s
Filter values (R, L) 0.031 , 3.02 mH
Fig. 6 Inverter terminal/pole voltage (V) for one phase
inverter. Figure 7 shows the NPC inverter terminal three-phase line-to-line voltages
Vab , Vbc ,Vca with a maximum value of 600 V each. It is concluded that five levels
of voltage are obtained at the output of the inverter from Fig. 7.
Figure 8 shows the three-phase grid currents in amperes for an injected maximum
reference current of 4 A. It is understood from this figure that the maximum of 4 A
is obtained for each phase.
Figure 9 shows injected reference phase-A current and grid current in amperes.
In this figure, it is understood that the grid current follows the reference current.
Fig. 7 Line voltage (five level) in volts

Fig. 8 Three-phase grid currents in amperes
Fig. 9 Injected reference phase-A current and grid current in amperes
In this proposed control scheme, if the reference current changes, automatically the
switching states will be changed for giving the gate pulses to the switches.
Figure 10 shows harmonic spectrum for grid current obtained through FFT anal-
ysis. From the figure, it is observed that the THD is 3.51% and this value is meeting
the standards of IEEE-519. Figure 11 represents the three-phase grid line voltages
with a peak value of 587.9 V at fundamental frequency.
Fig. 10 Harmonic spectrum—grid current

Fig. 11 Three-phase grid line voltages
5 Conclusion
In this paper, a simple and novel current control scheme for three-phase three-level
grid-tied NPC inverter is proposed. This scheme is simple and intuitive because of
the elimination of all dwell time calculations in comparison with the conventional
modulation schemes. For a reference injected grid current with a maximum value of
4 A, it is understood that the grid current follows same as reference by the proposed
control scheme and grid current THD is obtained as 3.51%. Based on the reference
injected grid current, the region of switching states is automatically selected with this
new digital switching control approach. This scheme reduces the control complexity
and burden on the processor. The NPC with proposed current control scheme is well
suitable to wind energy conversion and photovoltaic-based systems.
References

Trans Ind Appl IA-17(5):518–523. https://doi.org/10.1109/TIA.1981.4503992
2. Layate Z, Bahi T, Abadlia I, Bouzeria H, Lekhchine S (2015) Reactive power compensation
control for three phase grid-connected photovoltaic generator. Int J Hydrog Energy 40
3. MCU SAE Team (2012) Center-aligned SVPWM realization for 3-phase, 3-level inverter. Texas
Instruments
4. Palle K, Bhanuchandar A (2020) A novel modified voltage oriented control of an active front-
end rectifier used for PMSG based wind turbine systems. Int J Innov Technol Explor Eng
(IJITEE) 9(8)
5. Rodríguez P, Teodorescu R, Candela I, Timbus AV, Liserre M, Blaabjerg F (2006) New positive-
sequence voltage detector for grid synchronization of power converters under faulty grid
conditions. In: IEEE annual power electronics specialists conference
6. Ross D, Theys J, Bowling S (2004) Using the dsPIC30F for vector control of an ACIM.
Microchip
7. Sharma A, Singh A, Yadav P (2015) Analysis of 3 level SVPWM based open loop and closed
loop V/F control of induction motor. Int J Eng Res Technol 4(4)
8. Solar Choice Staff (2010) Conventional and transformerless inverters. Solar Choice
9. Teodorescu R, Liserre M, Rodríguez P (2011) Grid converters for photovoltaic and wind power
systems, 1st edn. Wiley
10. Tolunay B (2012) Space vector pulse width modulation for three-level converters. Uppsala
Universitet
11. Zhu R, Wu X, Tang Y (2015) Duty cycle-based three-level space vector pulse-width modulation
with overmodulation and neutral-point balancing capabilities for 3-phase neutral-point clamped
inverters. IET Power Electron 8(10)
12. Celanovic N, Boroyevich D (2001) A fast space-vector modulation algorithm for multilevel
three-phase converters. IEEE Trans Ind Appl 37(2):637–641
13. Wu B (2006) High-power converters and AC drives, 1st edn. Wiley-IEEE Press. Wiley
14. Aneesh M, Gopinath A, Baiju MR (2009) A simple space vector PWM generation scheme for
any general n – level inverter. IEEE Trans Ind Electron 56(5):1649–1656
15. Leon J, Vazquez S, Portillo R, Franquelo L, Carrasco J, Wheeler P, Watson A (2009)
Three-dimensional feedforward space vector modulation applied to multilevel diode-clamped
converters. IEEE Trans Ind Electron 56(1):101–109
16. Mekhilef S, Abdul Kadir M (2010) Voltage control of three-stage hybrid multilevel inverter
using vector transformation. IEEE Trans. Power Electron 25(10):2599–2606
17. Gopinath A, Mohamed A, Baiju MR (2009) Fractal based space vector PWM for multilevel
inverters—a novel approach. IEEE Trans Ind Electron 56(4):1230–1237
18. Saeedifard M, Bakhshai A, Joos G (2005) Low switching frequency space vector modulators
for high power multimodule converters. IEEE Trans Power Electron 20(6):1310–1318
19. Fernando Silva J, Rodrigues N, Costa J (2000) Space vector alpha-beta sliding mode
current controllers for three-phase multilevel inverters. In: EEE Power Electronics Specialists
Conference (PESC), vol 1, pp 133–138
20. Kouro S, Rebolledo J, Rodriguez J (2007) Reduced switching-frequency-modulation algorithm
for high-power multilevel inverters. IEEE Trans Ind Electron 54(5):2894–2901
21. Saeedifard M, Nikkhajoei H, Iravani R (2007) A space vector modulated STATCOM based on
a three-level neutral point clamped converter. IEEE Trans Power Del 22(2):1029–1039
22. Kumar KB, Bhanuchandar A, Mahesh C (2021) A novel control scheme for symmetric seven
level reduced device count multi-level DC link (MLDCL) inverter. In: 2021 international confer-
ence on sustainable energy and future electric transportation (SEFET), pp 1-4. https://doi.org/
10.1109/SeFet48154.2021.9375714
connected inverter with LCL filter. In: 2020 3rd international conference on energy, power
and environment: towards clean energy technologies, pp 1–5.https://doi.org/10.1109/ICEPE5
0861.2021.9404491
24. Sathisha D, Bhanuchandar A, Harshaa D, Srisailam C (2021) A simplified control approach of
AC-DC-TLBC for bi-polar DC microgrid applications. Turk J Comput Math Educ 12(2):818–
824
25. Rodriguez J, Bernet S, Steimer PK, Lizama IE (2010) A survey on neutral-point-clamped
inverters. IEEE Trans Industr Electron 57(7):2219–2230
Active and Reactive Power Control
of Grid-Tied Asymmetrical MLI-based
PV System with Reduced Switching
Frequency
Hareesh Myneni, A. Pranay Kumar, Somnath A. Mandale,

and Ganjikunta Siva Kumar
1 Introduction
The use of renewable energy resources is growing rapidly in recent days. Renewable
energy resources exist worldwide, in contrast to fossil fuels, which are insufficient
for the present energy requirements. The rapid development of renewable energy
resources plays a vital role in electricity generation. Solar, wind, geothermal, etc.,
are the abundant sources of renewable energy. Electricity production from renewable
energy is free from pollution and its cost also less. The energy extraction from the
sun is increasing, because of abundantly available solar energy. Multilevel inverter
with solar PV system is widely used in applications like stand-alone systems and
grid- connected system [1]. There are two types of inverters which are commonly
used for solar application. One is a conventional two-level and three-level inverter
used in a solar PV system for stand-alone and grid-tied mode. The drawback of the
conventional inverter is having high Total Harmonic Distortion (THD) and excess
voltage stress on the inverter switches. Hence, for the minimization of THD in the
inverter current, required high switching frequency operated system [2, 3]. These
higher switching frequencies further leads to higher switching losses.
In recent years, various MLI topologies are using grid-tied PV systems because
of having numerous advantages over conventional inverters. The significant features
of the MLI are low voltage stress on the inverter switching devices, lower THD,
H. Myneni (B)
National Institute of Technology Srinagar (NITSRI), Srinagar, India
A. Pranay Kumar
KITS Warangal, Warangal, India
S. A. Mandale
Toshiba Mitsubishi-Electric Industrial Systems Corporation (TMEIC), Bengaluru, India
G. Siva Kumar
National Institute of Technology Warangal (NITW), Warangal, India
120 H. Myneni et al.
lesser Electro-Magnetic Interference (EMI), and reduce switching frequency [4].

The MLI name itself says that multiple number of voltage levels which is more than
conventional (i.e., greater than three). The requirement of switching devices also
increases to generate additional voltage levels. The increase in switching devices
increases the size and complexity of the inverter. This drawback can be overcome
by using reduced switching devices count MLI. There are numerous MLI topologies
which are having reduced switching devices count [4, 5]. In this work, a grid-tied solar
PV system is implemented with the reduced switching device count asymmetrical
MLI [6, 7]. In the past decade, several MLI topologies with reduced switch count are
used in power electronic applications [4, 5]. In the asymmetrical MLI configuration,
both unidirectional and bidirectional switching devices are used. These switching
devices increase the voltage levels of the seven-level inverter output voltage by giving
proper switching pulses. More voltage levels will be generated at reduced switching
device count using asymmetrical MLI configuration [6]. The asymmetrical MLI
configuration also minimizes the systems’ cost.
In literature, various PWM techniques are proposed to control current injection,
reduce switching losses, and THD [8]. Among these techniques, the sinusoidal PWM
technique is the most commonly used technique and further it is classified into level
shifted PWM and phase shifted PWM techniques [9, 10]. In this paper, level shifted
PWM (LS-PWM) technique for seven-level asymmetrical MLI is used. The grid
current injection and maintaining power factor at unity is done with a proportional
integral (PI) controller-based LS-PWM technique [6]. The comparative study of LS-
PWM with PI controller and Model Predictive Control (MPC) algorithm is studied
and simulation results are verified. The application of MPC techniques to control the
power electronic applications is a versatile method [11]. In this technique, variables’
future behavior will be predicted to analyze the control variable. Depending on the
control variable, cost function is defined in MPC. In this method, MPC is basically
used to control the injection of grid current and maintain the unity power factor with
reduced switching frequency [12, 13]. The load current reactive component is also
controlled by the MPC algorithm. Solar photovoltaic (PV) system with multilevel
inverter is widely used in applications like grid-connected and stand-alone systems.
The grid-connected multilevel inverter has the drawback of power quality, so for
improving the power quality, the controlled reactive power injection is needed. The
power quality issue is minimized by the proposed model predictive active and reactive
power control method [14, 15].
In this paper, Sect. 2, describes the detailed analysis of asymmetrical MLI with
LS-PWM technique and also an explanation of the PI controller-based closed-loop
algorithm to control the required variables. In Sect. 3, the study of MPC and control
algorithm for maintaining unity power factor and to control the injected grid current
at the reduced switching frequency. Section 4 describes, the controlling of active
and reactive power at reduced switching frequency. The simulation results of the
proposed control algorithm are discussed in Sect. 5. At the end, the conclusion is
presented in Sect. 6.
Active and Reactive Power Control of Grid … 121
2 Grid-tied PV System with Asymmetrical Multi Level

Inverter
An asymmetrical MLI topology with a reduced switching device count is shown in

Fig. 1. It consists of four unidirectional switches (T1 , T2 , T3 , T4 ) and two bidirectional
switches (S1 , S2 ) to produce seven-level waveform. This MLI topology having two
DC-link voltage sources which are not equal in magnitude. The voltage magnitude
of DC-link sources is in the proportion such as 1:2 (i.e., Vdc1 : Vdc2 = Vdc : 2Vdc ).
For achieving all possible combination of the voltage levels in an asymmetrical
MLI, a proper switching of devices is mandatory. The switching status of each switch
for the generation of seven voltage levels of magnitude 3Vdc = Vdc1 + Vdc2 , 2Vdc =
Vdc2 , Vdc = Vdc1 , 0, −Vdc = −Vdc1 , −2Vdc = −Vdc2 , and −3Vdc = −(Vdc1 + Vdc2 )
are shown in Table 1. The voltage level of 3Vdc = Vdc1 + Vdc2 will be achieved by an
additive series combination of two dc-link voltage sources Vdc1 and Vdc2 . To attain
these voltage level, switches T1 and T4 should be ON. Similarly, for the generation of
the 2Vdc voltage level, switch T1 and S2 should be ON. In case if the switches T4 and
S1 are ON, the Vdc voltage level will be obtained. Fundamental positive half cycle
of the inverter output voltage is generated by conducting T1 and T4 . Similarly, the
switches T2 and T3 are responsible for the generation of the fundamental negative half
cycle of the inverter output volatge. Application of LS-PWM technique to generate
the switching states using LS-PWM is explained below.
Fig. 1 Grid-tied PV system with seven-level asymmetrical MLI configuration

Table 1 Voltage levels corresponding to switching devices status in the inverter with ‘1’ as ON,
‘0’ as OFF
Voltage Switching status
level
T1 T2 T3 T4 S1 S2
3Vdc 1 0 0 1 0 0
2Vdc 1 0 0 0 0 1
0 0 0 1 1 0
Vdc
0 0 0 1 1 0 0
1 1 0 0 0 0
0 0 0 0 1 1
−Vdc 0 0 1 0 0 1
−2Vdc 0 1 0 0 1 0
−3Vdc 0 1 1 0 0 0
2.1 Carrier-Based LS-PWM Technique
Carrier-based LS-PWM technique has been implemented for the seven-level grid-
tied asymmetrical MLI. In carrier-based LS-PWM technique, for the generation of
n-level inverter, (n − 1) high-frequency triangular carriers will be required. Hence,
six high-frequency triangular carrier signals should be required to generate the gale
pulses for seven-level inverter. Each carrier is compared with a reference signal to
produce the switching state pulses.
The amplitude modulation index of fixed magnitude of the carrier-based LS-PWM
technique is defined as,
mr
ma = (1)
mc
where, m a is defined as the amplitude modulation index which is the ratio of the peak
value of reference sinusoidal signal (m r ) to the peak value of high-frequency carrier
signal (m c ).
2.2 Control Strategy for Grid-Tied Asymmetrical MLI Based

Solar Photo Voltaic System
The grid current reference magnitude depends on the magnitude of available PV

power. It is required to track the frequency of the grid voltage to maintain power
factor as unity and to generate the grid current reference (i g∗ ). To achieve this, the
proposed technique is incorporated with phase locked loop (PLL). The difference
between reference (i g∗ ) and actual measured grid current (i g ) is passed through a
proportional integral (PI) controller. The gains of PI controller K p , K c are tuned
to produce switching signals for MLI [16]. In this way, the unity power factor is
maintained on the grid side and the PV power is injected to the grid. There are two
types of control algorithms that are implemented and explained below.
3 Principle of Model Predictive Controller
There are so many control techniques that have been developed, among these control
techniques, MPC is more efficient than the conventional techniques [17, 18]. The
MPC has many advantages, such as being easy to understand, implement and multiple
variables can be controlled. Future behavior of the control variables is predicted using
model of the system. These are achieved by minimizing the cost function. The cost
function basically consists of the difference between the predicted and reference
control variable. A discrete state-space model is used in MPC, which is as follows:
X (K + 1) = AX (K ) + BU (K )
Y (K + 1) = C X (K ) + DU (K ) (2)
For getting discrete time model, discretization methods are used. Euler forward
method is used to approximate the derivative for first-order system
dy x(k + 1) − x(k)
= (3)
dx Ts
where, Ts is the sampling time. This Euler forward method to approximate the deriva-
tive. The desired behavior of system is expressed in cost function as follows:
J = f (x(k), u(k), ....u(k + N )) (4)
The value of the weighing factor is decided according to the priority of each term in
the cost function and it is used to control the variable. In MPC, the behavior of these
control variables is predicted for all possible combination switching states and the
value of cost function is calculated for all switching states, and finally optimal state
is selected which minimizes the cost function. That switching state is called optimal
switching state which is applied to the switches of MLI.
4 Proposed Control Algorithm
Figure 2 shows the control block diagram of the proposed MPC with an asymmetrical
seven-level MLI. MPC consists of future value prediction block that will generate
future behavior of controlled variables i(k + 1) signal using measured actual con-
trolled variables i(k) signal. The MPC will select optimal switching state from avail-
able states (S) by evaluating the minimized cost function (S). The optimal switching
state (S) is applied to control the behavior of the controlled variables.
4.1 Predictive Grid Current Tracking and Maintain UPF
In the proposed predictive current control technique, the behavior of the control
variables is predicted for seven switching states of inverter model of the system.
Evaluation of the minimized cost function is done and the selected optimal switching
state is applied to the inverter. The primary objective of the proposed current control
is to track the reference grid current and maintaining power factor as unity. The cost
function is defined for achieving this objective as follows:
g = |ir∗ (k + 1) − i p (k + 1)|. (5)
Fig. 2 Proposed model predictive control block diagram

where, ir∗ (k + 1) represents the grid current reference values and i p (k + 1) is the
predicted value of grid current. For the calculation of the i p (k + 1) i.e., predicted
value of grid current from the grid voltage and measured grid current, discrete time
model will be used. Forward Euler approximation is used to approximate grid current
derivative di p /dt and is given as
di p i p (k + 1) − i p (k)
= . (6)
dt Ts
By applying KVL to the grid-connected inverter model
di p
v(k) = L + Ri p (k) + e(k). (7)
dt
By solving, an expression for the prediction of future grid current for every seven
switching states by considering the voltage vectors of the inverter v(k) and grid e(k)
is
(v(k) − e(k))Ts RTs
i p (k + 1) = + i p (k) 1 − . (8)
L L
From the above expression, prediction of future grid current is evaluated and which
is used to evaluate the cost function consists of the difference between reference
and predicted grid current vector for each seven switching states. From the available
seven switching states, the optimal switching state gives the minimized cost function
and it is applied to the inverter model.
4.2 Controlling of Active and Reactive Powers Using MPC
The flow of active and reactive powers in a PV supported grid-connected system is

controlled in the proposed MPC. In MPC technique, the load currents are used to
extract the reference current. In general, the load current i L (t) can be expressed as
i L (t) = Idc + Im1 sin(wt + φ1 ) + Im2 sin(2wt + φ2 )

+Im3 sin(3wt + φ3 ) + .....Imn sin(nwt + φn ). (9)
In the above equation ‘mn’ indicates the nth harmonics peak value and φn indicates
phase angle of the load current i L (t). The power consumed is given below.
Ps = (Vm1 Ims cosφs )/2 (10)

where, Vm1 represents the fundamental peak value of the supply voltage and Ims is
the reference source currents peak value. The peak value of reference source current
(Ims ) from (10) is given as
Ims = 2Ps /(Vm1 cosφs ) (11)
The time domain reference source current is expressed by
i s∗ (t) = Ims sin(wt − φs ) (12)
Finally, from (9) and (12), for controlling the reactive component of load current the
reference current is expressed as
i ∗f (t) = i L (t) − i s∗ (t) (13)
The injecting current equal to reference current (i ∗f ) by MLI with model predictive
control, unity power factor is achieved on grid side. In order to inject active power to
the grid, first, calculate the reference current corresponding to PV power. Suppose,
the reference current corresponding to PV power is i ∗pv , the resultant reference for
MLI is i ∗pv + i ∗f . This resultant reference current controls active and reactive powers
which are injected into the grid.
4.3 Switching Frequency Reduction
In this method, switching frequency reduction is one of the control variables along
with current control and power control. In which, the error between present and
previous switching state is considered as cost function to achieve reduced switching
frequency. For all seven switching states (S), the cost function has been calculated
and the switching state that gives minimal cost function value is called as optimal
cost function. The optimal cost function is applied to the seven-level inverter model.
In this paper, the control technique is implemented to fulfill power quality enhance-
ment and PV power injection into the grid. To achieve these aspects, a model predic-
tive controller is implemented. This method gives the advantage of reduced switching
frequency when compared to conventional methods. The simulation results of LS-
PWM technique with PI type current controller performance are compared with the
proposed MPC technique.
The proposed reduced switching frequency model predictive control algorithm of

asymmetrical MLI is simulated in MATLAB/Simulink and corresponding simulation
parameters are mentioned in Table 2. In simulation analysis, an asymmetrical grid-
Table 2 Simulation parameters

Parameters Values
Grid voltage 240 V rms
Load resistance (R ) 14
Load inductance (L ) 22 mH
Interfacing inductance (L f ) 12 mH
dc-link capacitance (C1 , C2 ) 3600 µF
connected MLI with R L load connected in parallel with the grid consisting of two
DC voltage sources in the proportion of 1:2 is considered. In the simulation results,
mainly focusing on two control algorithm, i.e., controlling of active and reactive
powers using MPC technique and second is the reduction of switching frequency.
5.1 Only Reactive Power Control Technique
Reactive power control technique has been implemented to reduce the power quality
issues. In this control technique, the main control variable is reactive power injected
by the inverter into the load. In this case, the active component of load current is
supplied by the grid. It implies that the grid current (i g ) is in phase with grid voltage
(vg ), such that the unity power factor will be maintained.
The inverter voltage (vinv ), inverter current (i inv ), grid current (i g ), grid voltage
(vg ), and load current (iload ) in case of only reactive power injection are shown
in Fig. 3. It is observed that the inverter current (i inv ) supplied by the inverter is
phase shifted by 90◦ to the inverter output voltage (vinv ). It implies that the reactive
component of load current is injected by the inverter and the active component is
injected by the grid.
Figure 4a shows the simulation results of current tracking of actual current and
the reference sinusoidal current under only reactive power control. It is observed
from zoomed part of Fig. 4a that, the tracking is perfect with the proposed method.
Figure 4b shows the real and reactive powers injected by the inverter (Pinv and Q inv ),
and the grid real and reactive powers (Pg and Q g ) to the load. The reactive power
required for load is supplied by the inverter, that’s why grid reactive power (Q g ) is
zero. At the same time, the inverter is not injecting any real power. However, it draws
a small amount of real power from the grid to supply losses in the inverter, that’s
why Pinv value is negative as shown in Fig. 4b.
Fig. 3 During only reactive power control: seven-level inverter voltage (vinv ), grid voltage (vg ),
grid current (i g ), injected current by the inverter (i inv ) and load current (iload ) under reactive power
control; (scale: X-axis in seconds, Y-axis voltage in volts, current in ampere)
5.2 Real and Reactive Power Control Technique
A real-reactive power control means controlling of total power supplied by the

inverter which is required for the load. In this control algorithm, the requirement
of load power has been supplied by the inverter, which implies that the inverter is
extracting maximum power. In this case, the power supplied by the grid will be zero
or negligible. In case if the inverter is not capable to supply the total amount of
required load power, the leftover power is supplied by the grid. In the model pre-
dictive direct power control algorithm, only the required amount of total load power
will be extracted from the inverter.
Fig. 4 a Current tracking under reactive power control only (scale: X-axis in seconds, Y-axis in
Ampere); b real and reactive powers of grid and inverter
Figure 5 shows, the simulation results of the inverter voltage (vinv ), grid voltage
(vg ), grid current (i g ), inverter current (i inv ) and load current (iload ), respectively.
It is observed from Fig. 5 that grid current is very small in magnitude, which
implies that the power supplied by the grid to the load is also negligible. In this
case, the total power required for the load is injected by the inverter only. Therefore,
the inverter current (i inv ) and load current (iload ) are equal in magnitude, it can be
observed from Fig. 5.
Figure 6a shows the simulation results of current tracking of actual current and
the reference current under the real-reactive power control. It is observed from Fig.
6a that, the tracking of reference with the proposed method is well and good. Figure
6b shows the power injected by the inverter and grid to the load. In this case, the
active and reactive power injected by the inverter (Pinv ) and (Q inv ) is equal to the
required load power. Therefore, real power (Pg ) and reactive power (Q g ) supplied
by the grid to the load is zero as shown in Fig. 6b.
Fig. 5 During real and reactive power control: sevel-level inverter voltage (vinv ), grid voltage (vg ),
grid current (i g ), injected current by the inverter (i inv ) and load current (iload ) under real-reactive
power control; (scale: X-axis in seconds, Y-axis voltage in volts, current in ampere)
5.3 Load Variation
The main motivation of load variation is to analyze the dynamic performance of the
proposed model predictive control under different powers injection. The proposed
model predictive control requires less time to get stable when compared to conven-
tional methods. Figure 7 shows the simulation results under load variation at t = 1 s
and t = 1.05 s. It is observed that for both the instants of load variation within one
and a half cycle, the grid current and inverter current reaches a steady state.
Fig. 6 a Current tracking under real-reactive power control (scale: X-axis in seconds, Y-axis in
Ampere); b real and reactive powers of grid and inverter
Under the variable of load condition, the reference current (ir e f ) and actual current
(i actual ) of inverter are shown in Fig. 8a. It is observed that the current tracking
capability of the proposed control technique is excellent, which implies the faster
dynamics response of the proposed predictive control algorithm. The real and reactive
powers of inverter and grid under variable load conditions are shown in Fig. 8b.
5.4 Reduction of Switching Frequency
The proposed reduced switching frequency model predictive control algorithm of

asymmetrical MLI is simulated. Table 3 presents the reduced switching frequencies
(Hz) for each switching device. Table 3 shows the switching frequency in Hz for
all six switching devices with and without switching frequency constraints. The
switching frequency is reduced with adding frequency constraint when compared to
without constraint. Even though, the THD of the grid current increases, it is less than
the specified limit of 5% (IEEE-519 standard).
Fig. 7 Simulation results under load variation: seven-level inverter voltage (vinv ), grid voltage (vg ),
grid current (i g ), injected current by the inverter (i inv ), and load current (iload ); (scale: X-axis in
seconds, Y-axis voltage in volts, current in ampere)
6 Conclusion
In the proposed work, the performance of seven-level asymmetrical Multi-Level

Inverter (MLI) under the controlled injection of the grid current with accurate track-
ing capability is analyzed. Also, the control of active power and reactive power by
using a model predictive controller is discussed at reduced switching frequency.
The performance of the proposed model predictive control (MPC) is analyzed under
the variable weighting factor. The controlled active and reactive power injection at
reduced switching frequency to maintain the Unity Power Factor (UPF) on the grid
side is analyzed and the discussion on simulation results is done in this paper. It has
been concluded that the variation of the weighting factor leads to the weightage of
Fig. 8 a Current tracking under load variation (scale: X-axis in seconds, Y-axis in Ampere); b real
and reactive powers of grid and inverter
Table 3 Switching frequency (Hz) and grid current THD (%) without and with adding switching
frequency constraint
Number of Switching frequency (Hz) Grid
constraints current
THD
T1 T2 T3 T4 S1 S2
Without 3050 3050 10900 11050 13093 14100 0.34%
switching
frequency
constraint
With 1411 1383 3397 3478 4156 4205 2.63%
adding
switching
frequency
constraint
the controlled variable implies the reduction of switching frequency with accurate
current tracking capability. The proposed model predictive control was also tested
under the variable load condition to analyze the dynamic performance of the MPC.
References
1. El-Hawary ME (2014) The smart grid-state-of-the-art and future trends. Electr Power Compon
Syst 42(3–4):239–250
2. Alepuz S, Busquets-Monge S, Bordonau J, Gago J, González D, Balcells J (2006) Interfacing
renewable energy sources to the utility grid using a three-level inverter. IEEE Trans Ind Electron
53(5):1504–1511
3. Patrao I, Garcerá G, Figueres E, González-Medina R (2014) Grid-tie inverter topology with
maximum power extraction from two photovoltaic arrays. IET Renew Power Gener 8(6):638–
648
5. Babaei E (2008) A cascade multilevel converter topology with reduced number of switches.
6. Das MK, Jana KC, Sinha A (2017) Performance evaluation of an asymmetrical reduced
switched multi-level inverter for a grid-connected PV system. IET Renew Power Gener
12(2):252–263
7. Rahim NA, Chaniago K, Selvaraj J (2010) Single-phase seven-level grid-connected inverter
for photovoltaic system. IEEE Trans Ind Electron 58(6):2435–2443
8. Bhavsar T, Narayanan G (2009) Harmonic analysis of advanced bus-clamping PWM tech-
niques. IEEE Trans Power Electron 24(10):2347–2352
9. McGrath BP, Holmes DG (2002) Multicarrier PWM strategies for multilevel inverters. IEEE
10. Chavarria J, Biel D, Guinjoan F, Meza C, Negroni JJ (2012) Energy-balance control of PV
cascaded multilevel grid-connected inverters under level-shifted and phase-shifted PWMs.
11. Kouro S, Cortés P, Vargas R, Ammann U, Rodríguez J (2008) Model predictive control—a
simple and powerful method to control power converters. IEEE Trans Ind Electron 56(6):1826–
1838
12. Kwak S, Kim SE, Park JC (2015) Predictive current control methods with reduced current errors
and ripples for single-phase voltage source inverters. IEEE Trans Ind Electron 11(5):1006–1016
13. Mathew D, Shukla A, Bandyopadhyay S (2016) Modified predictive current control of neutral-
point clamped converter with reduced switching frequency. In: IEEE applied power electronics
conference and exposition (APEC) 2016, pp 3286–3290. IEEE
14. Bozorgi AM, Gholami-Khesht H, Farasat M, Mehraeen S, Monfared M (2018) Model predictive
direct power control of three-phase grid-connected converters with fuzzy-based duty cycle
modulation. IEEE Trans Ind Appl 54(5):4875–4885
15. Latifi M, Sabzehgar R, Rasouli M, Fajri P (2019) Active and reactive power compensation
using a PEV-based three-level capacitor clamped inverter. In: IEEE Texas power and energy
conference (TPEC) 2019, pp 1–6
16. Myneni H, Ganjikunta SK (2021) Design, analysis, and switching losses reduction of hybrid
shunt compensator with adaptive control scheme. Int Trans Electr Energ Syst 31
17. Garcia CE, Prett DM, Morari M (1989) Model predictive control: theory and practice-a survey.
Automatica 25(3):335–348
18. Kumar AP, Kumar GS, Sreenivasarao D, Myneni H (2019) Model predictive current control
of DSTATCOM with simplified weighting factor selection using VIKOR method for power
quality improvement. IET Gener, Transm Distrib 13(16):3649–3660
Modelling and Control Aspects
of STATCOM Connected
to a Grid-Integrated PV System
Koneti Varalakshmi, R. L. Narasimham, and G. Tulasi Ramdas
1 Introduction
In an AC microgrid system, different reactive as well as nonlinear loads consume

reactive power. Adequate control of reactive power can solve many different prob-
lems which arise in an AC power system. Very good review articles on reactive power
compensation system can be found in [1–4]. Reactive power compensation has two
different aspects such as (a) load compensation and (b) voltage support. STATCOM
is a shunt-connected CPD which is capable of providing compensating effect to cur-
rent associated with power quality events. STATCOM is generally a voltage source
inverter (VSI) which comprises of DC source and different power electronic switches
(IGBT). The inclusion of STATCOM has numerous benefits in microgrid system [5].
Nowadays a lot of research is carried out to design efficient topology and control
mechanism for STATCOM. Integration of distributed energy resources (DERs) pro-
vide a significant benefit to the existing power system in terms of voltage profile.
Grid-connected PV system provides significant benefit, but grid-connected PV sys-
tem is only operational at day time and remains idle at night time which reduces
the efficiency and utilization aspect of the system. STATCOM is combined with
PV system to increase the utilization and efficiency of the system as well as pro-
vide reactive power compensation to the grid. Generally, classical 2-level inverter is
used as STATCOM and coupling transformer or inductor is used for connecting the
K. Varalakshmi (B)
Department of Electrical and Electronics Engineering, Aditya College of Engineering and
Technology, Surampalem, Andhra Pradesh, India
R. L. Narasimham
Department of Electrical Engineering, Andhra University, Visakhapatnam, India
G. Tulasi Ramdas
Department of Electrical and Electronics Engineering, Jawaharlal Nehru Technological University
Hyderabad, Hyderabad, Telangana, India
136 K. Varalakshmi et al.
grid with STATCOM. There are different topologies, as well as control schemes, for
STATCOM. The DC-link voltage of STATCOM is controlled using different con-
trol scheme such as PID controller, vector controller, model predictive controller,
fractional controller and intelligent controller.
This paper provides the mathematical modelling and analysis of grid-integrated
PV system along with STATCOM. Different modes of STATCOM and the operation
of STATCOM in an AC microgrid have been discussed in this paper. Detailed math-
ematical modelling and controller design has been done for the system where the
current control scheme is used for the control of STATCOM and the MPPT scheme
is used to extract maximum power from PV panel. Simulation analyses have been
provided to validate the theoretical explanation.
The paper is organized as follows. Section 2 provides the operation of STATCOM
in a microgrid. Section 3 presents the overall system description of grid-integrated
PV system along with STATCOM. Section 4 presents the control scheme for PV-fed
boost converter as well as STATCOM. Section 5 presents the simulation results and
Sect. 6 provides the concluding remarks.
2 STATCOM Operations in Microgrid
The different operating modes of STATCOM are as follows [6]:

– Partial STATCOM mode: This mode is used during the daytime when the inverter
provides reactive power to the grid. The most priority is given to the real power
during partial STATCOM mode.
– Full STATCOM mode: Full STATCOM mode is required when there is some sort
of faults, disturbances, etc., and there is an urgent need for reactive power sup-
port. In this mode, the PV system discontinues the generation and real power and
provides reactive power via the inverter. Full STATCOM mode is completely used
during nighttime. The discontinuation of real power during daytime via STATCOM
is done by a solid state relay that disconnects the PV panel.
– Full PV mode: Solar system generates real power without the requirement of
reactive power.
3 System Description
The complete circuit diagram and control scheme of grid-integrated PV system along
with STATCOM has been shown in Fig. 1. Figure 1 comprises of an array of PV
cells, DC-DC boost converter. As PV voltage is variable to change in solar irradiance
and ambient temperature, maximum power point tracking (MPPT) unit along with
a voltage regulation unit is placed to extract maximum power from the PV module.
Boost converter is used to step-up the voltage of PV module and switching signal to
Modelling and Control Aspects of STATCOM … 137
Fig. 1 Circuit diagram and control scheme of PV-STATCOM
boost converter is provided by the MPPT unit. DC-link capacitors (split capacitors)
are connected in the output terminal of the boost converter to maintain the constant
DC-link voltage. STATCOM is modelled as 3-φ-2-level,2-leg, six-pulse IGBT power
semiconductors with antiparallel freewheeling diode based DC-AC voltage source
inverter operating at a switching frequency f sw . Output current of STATCOM is
represented as i as , i bs and i cs . STATCOM is connected to the middle point of trans-
mission line at point of common coupling (PCC) via interfacing filter L f . To suppress
any additional ripple of STATCOM, ripple filters can also be placed. The gate-pulse
of the STATCOM is provided by the control unit of STATCOM. 3-φ-4-wire AC
source represented as Vsa , Vsb and Vsc having an impedance of Z is connected to
linear as well as nonlinear load via PCC. AC-DC bridge rectifier with capacitive filter
is considered as the nonlinear load. The voltage at PCC is represented as Val , Vbl
and Vcl .
3.1 PV Array
The dynamics of a practical PV cell can be represented as

V + Rs I V + Rs I
I = I pv − Io exp −1 − (1)
Vt a Rp
where PV current is represented by I pv , saturation current is represented by Io ,

parallel resistance of PV cell is R p , series resistance of PV cell is Rs and thermal
voltage of PV cell is Vt
3.2 VSI Based STATCOM
The 3-φ dynamic model of STATCOM can be represented as
Lf di La
dt
= −R f i sta + Vas − Val
Lf di Lc
dt
= −R f i stb + Vbs − Vbl (2)
Lf di Lc
dt
= −R f i stc + Vcs − Vcl

d 1 V2
2
Cdc Vdc (t) = − (Vas i as + Vbs i bs + Vcs i cs ) − dc (3)
dt 2 Rc
Using abc to dq transformation

⎡ ⎤ ⎡ R ⎤⎡ ⎤
i ds − L ff ω K
Lf
cos α i ds
d ⎣ ⎢ ⎥
i qs ⎦ = ⎣ −ω
R
− L ff K
Lf
sin α ⎦ ⎣ i qs ⎦ (4)
dt V Vdc
dc − 2Cdc cos α − 2Cdc sin α − Rc1Cdc
3K 3K
The design of AC interfacing inductor L f is as follows:

√
3mVdc
Lf = (5)
12a f sw i cr ( p− p)
where m is the modulation index, a is the overloading factor, Vdc is the DC-link
voltage, i cr ( p− p) is 15%, f sw is the switching frequency.
4 Control Scheme for PV-Integrated STATCOM
4.1 MPPT and DC-Link Control
Based on voltage and current measurement at each step, power is computed. The
P&O MPPT steps the panel voltage along the P–V curve such that voltage at MPP
can be reached. The step-size of P&O MPPT is large enough due to settling issues
of converter dynamics. P&O MPPT with PID controller is one of the widely used
technique which provides a relatively stable results for power point tracking of PV
module [7, 8].
At MPP
d Ppv d I pv d V pv
= 0 ⇒ V pv + I pv =0 (6)
d V pv dt dt
Equation 6 can be re-written as
d I pv
V pv + I pv = 0 (7)
dt
At any point other than MPP
d I pv
V pv + I pv = ψ; ψ = 0 (8)
dt
Dynamic resistance of PV panel can be represented as
d V pv
r pv = − (9)
d I pv
Using dynamic resistance of PV panel, we get
V pv
I pv − =ψ (10)
r pv
Figure 2 provides the control scheme of STATCOM connected to a grid-integrated

PV system.
4.2 Generation of Reference Current Source
The reference compensating current are calculated by subtracting load current from
reference source current
140
Fig. 2 Control scheme of STATCOM connected to a grid-integrated PV system

K. Varalakshmi et al.
∞
i L a = I La1 sin (ωt − φa1 ) + I Lan sin (nωt − φan )
n=2
∞
i L b = I Lb1 sin (ωt − φb1 ) + I Lbn sin (nωt − φbn ) (11)
n=2
∞
i L c = I Lc1 sin (ωt − φc1 ) + I Lcn sin (nωt − φcn )
n=2
The active component of load current are
Isam = |I La1 | cos φ1

Isbm = |I Lb1 | cos φ1 (12)
Iscm = |I Lc1 | cos φ1
DC-link voltage is measured

and passed through a low pass filter with transfer
function G (s) = 1+τ f s is used. Error detector is used to get the error signal which
1
is then provided to the proportional-integral (PI) controller which can be represented

as
u (s) Ki
G c (s) = = Kp + (13)
e (s) s
The output of PI controller is added to the calculation of peak magnitude of the

reference source current. For unity power factor, the reference source current should
be in phase with source voltage. Therefore, reference source current supplied by
compensator is to be multiplied by unit sinusoidal template which are in phase with
source voltage.
The simulation parameters of grid-tied PV system with STATCOM are summarized

in Table 1. MATLAB-Simulink has been used for simulation purpose.
Figure 3a presents the PV module characteristics with varying solar irradiance.
Figure 3b illustrates the PV array characteristics (including 5 series modules and 66
parallel strings) with varying solar irradiance.
As the voltage of solar array is variable with respect to solar irradiance and temper-
ature, to extract maximum power from the solar array, voltage-based MPPT scheme
is used. Classical P&O MPPT is used for extracting maximum power from array.
Figure 4a illustrates the DC-link voltage with uniform solar irradiance. Figure 4b
illustrates the DC-link voltage with ramp increase in solar irradiance and Fig. 4c
presents the DC-link voltage with non-uniform solar irradiance.
Table 1 Simulation parameters for grid-tied PV system with STATCOM

PV module Power at standard test PST C 305 W
condiion
Voltage at MPP Vmp 54.7 V
Current at MPP Imp 5.58 A
Short circuit current Isc 5.98 A
Open circuit voltage Voc 64.2 V
PV array Number of series Ns 5
connected module
Number of parallel Np 66
connected module
DC–DC converter Input capacitor Cin 100 µF
Peak value of V peak 100 V
capacitor
Inductor L 5 mH
RMS value of current Ir ms 3A
DC-link capacitor Cdc 24 mF
Peak value of V peak 500 V
capacitor
Switching frequency f sw 5 kHz
Peak voltage of diode V peak 500 V
RMS current of diode Ir ms 5A
Peak voltage of V peak 500 V
MOSFET
RMS current of Ir ms 5A
MOSFET
3-φ DC–AC inverter Switching frequency f sw 20 kHz
Lf Filter inductance Lf 3 mH
RMS current of Ir ms 3A
inductance
Source voltage 3-φ 230 V, 50 Hz
Source impedance Rs = 0.1, L s = 0.1 mH
Nonlinear load R L = 100, L = 10 mH
Figure 5a provides the graph for 3-φ–4-wire source voltage and source current.
Figure 5b illustrates the load voltage and load current due to different nonlinear load.
In this work, the nonlinear load is modelled as a 3-φ rectifier with passive load.
Figure 5c presents the voltage and current at PCC. Figure 5d illustrates the inverter
current injected to PCC. In this work, the inverter is modelled as STATCOM.
Fig. 3 Characteristics of PV panel with varying solar irradiance a PV module, b PV array
Fig. 4 Simulation results of DC-link voltage at different solar irradiance a uniform irradiance, b
ramp increase in solar irradiance, c non-uniform solar irradiance
Fig. 5 Simulation results a 3-φ–4-wire source voltage and source current, b load voltage and load
current due to nonlinear loads, c voltage and current at PCC, d output inverter current to be injected
to PCC
6 Conclusion
This paper provides mathematical modelling and control of grid-integrated PV sys-

tem with STATCOM, where STATCOM is used for reactive power compensation.
Mathematical modelling and control structure of different converter and STATCOM
has been discussed. MPPT scheme is used for PV-fed boost converter and current
control scheme is used for controlling the STATCOM. Simulation analyses have been
presented for the AC microgrid.
References
1. Ismail B, Wahab NIA, Othman ML, Radzi MAM, Vijyakumar KN, Naain MNM (2020) A
comprehensive review on optimal location and sizing of reactive power compensation using
hybrid-based approaches for power loss reduction, voltage stability improvement, voltage profile
enhancement and loadability enhancement. IEEE Access 8:222733–222765
2. Dixon J, Moran L, Rodriguez J, Domke R (2005) Reactive power compensation technologies:
state-of-the-art review. Proc IEEE 93(12):2144–2164
3. Majumder R (2012) Reactive power compensation in single-phase operation of microgrid. IEEE
4. Barr J, Majumder R (2014) Integration of distributed generation in the volt/var management
system for active distribution networks. IEEE Trans Smart Grid 6(2):576–586
5. Singh B, Jayaprakash P, Kothari DP, Chandra A, Haddad KA (2014) Comprehensive study of
DSTATCOM configurations. IEEE Trans Ind Inform 10(2):854–870
6. Varma RK, Siavashi EM (2018) PV-STATCOM: a new smart inverter for voltage control in
distribution systems. IEEE Trans Sustain Energy 9(4):1681–1691
7. Subudhi B, Pradhan R (2018) A new adaptive maximum power point controller for a photovoltaic
system. IEEE Trans Sustain Energy 10(4):1625–1632
8. De Brito MAG, Galotto L, Sampaio LP, de Azevedo e Melo G, Canesin CA (2012) Evaluation of
the main MPPT techniques for photovoltaic applications. IEEE Trans Ind Electron 60(3):1156–
1167
Comparative Analysis Between Single
Diode and Double Diode Solar Topologies
of Solar PV Cell
Sriparna Das and Kumari Namrata
1 Introduction
The demand of using renewable energy has improved significantly. Nowadays, a

greater number of solar-, hydro-, wind-, etc., based plants are being set up for
increasing in power production [1]. Out of all these existing plants, solar or sun’s
rays is easily available in most of the places throughout the year. So, solar-based
generation is more preferred than other ones. About 40.4% of the total energy comes
from solar generation [2]. Thus, the designing and performance analysis is impor-
tant for solar photovoltaic cell [3, 4]. In this paper, two topologies of solar modelling
are considered, i.e., single and double diode schemes. A photovoltaic cell named
Solarex: MSX60 is considered as test parameters for making the comparison between
the schemes. The modeling is first shown followed by their circuit diagrams and the
inputs are taken from the parameters of the existing solar cells and the outputs of
the various schemes are compared. The yearly production of power for single and
double diode schemes [5, 6] is plotted to get how the variation of generated power
occurs. The variation of power is compared w.r.t single cell used. To obtain desired
power, a greater number of cells can be connected either in series or parallel fashion.
2 Modeling of Single Diode Scheme
The ideal circuital representation of photovoltaic cell is shown in Fig. 1. From Fig. 1,
applying Kirchhoff’s current division rule, we can get [4]
S. Das (B) · K. Namrata

Department of Electrical Engineering, NIT Jamshedpur, Jamshedpur, Jharkhand 831014, India
K. Namrata
e-mail: namrata.ee@nitjsr.ac.in
148 S. Das and K. Namrata
Fig. 1 Representation of ideal solar cell
Iout = I ph1−I d1 (1)
The general diode equation can be expressed as

q1 Voc1
Id1 = Is1 exp −1 (2)
Ns1 K 1 A1 T0
Placing the value of diode current from Eqs. (2) to (1), the final output current
can be expressed as

q1 voc1
Iout = I ph1 − Is1 exp −1 (3)
Ns1 K 1 A1 To
These above equations are valid for an ideal system. But in practical scenario,
there are some losses [5] that will take place into the system. So, the Fig. 1 [7] is
modified into Fig. 2 with the addition of some series and parallel resistances [8]. The
loss that happens due to flowing of current in metal grid, semiconductor material,
collecting bus, and terminal is denoted by adding some series drop into the circuit.
The parallel drop is due to current flowing through cell thickness and surface effect.
So, the diode current and output current expression gets changed and are represented
as
Fig. 2 Single diode scheme solar cell representation

Comparative Analysis Between Single Diode … 149

q1 (V1 + I1 Rs1
Id1 = Is1 exp −1 (4)
Ns1 K 1 A1 T0

q1 (V1 + I1 Rs1 )
Iout = I ph1 − Is1 exp −1 (5)
Ns1 K 1 A1 T0
There are certain parameters those are needed to calculate the output current of
the system. But they are not described in the above equations. They are photo current
(I ph1 ), reverse saturation current (Ir s1 ), and the saturation current (Is1 ), which are
expressed as follows:
G1
I ph1 = [Isc1 + K i1 (T01 − Tr 1 )] ∗ (6)
Gr e f 1
Isc1
Ir s1 = (7)
exp Ns1qK1 V1oc1
A1 T0
−1

T01 3 q1 E g1 1 1
Is1 = Ir s1 [ ] exp − (8)
Tr 1 A1 K 1 Tr 1 T01
This is the entire process of designing a solar photovoltaic cell according to the
single diode configuration.
3 Modeling of Double Diode Scheme
The double diode scheme is slightly different than single diode representation. Here,
the use of more than one diode takes place. The circuit representation is shown in
Fig. 3. The modeling is same as that of single diode, only the output current will
vary. The output current is expressed as
Fig. 3 Double diode scheme solar cell representaion

Table 1 Paramerets of
S. L. no Parameters Denoted name Values used
Solarex: MSX 60 Solar cell
[9] 1 Max. power Pm1 60 W
2 Max. voltage Vm1 17.1 V
3 Max power current Im1 3.5 A
4 O.C. voltage Voc1 21.1 V
5 S.C. current Isc1 3.8 A
6 No. of series cells Ns1 36
7 No. of parallel cell Np1 1
8 Ideality factor of A1 1.3
diode
9 Temp. coefficient of Ki1 0.003/°C
cell
10 Ref. temp Tref1 25 °C
11 Solar irradiance Gref1 1000 at STC

q1 (Vout1 + Iout1 Rs1
Iout = N p1 ∗ I ph1 − N p1 ∗ Is1 exp −1
Ns1 K 1 A1 T0

q2 (Vout1 + Iout1 Rs1
− N p2 ∗ Is2 exp −1 (9)
Ns1 K 1 A1 T0
The advantage of this model is in case of failure of a diode it can behave as single
diode approach, so the power generation will not get stopped. Due to connection of
critical loads that need to get supply throughout the time, stopping of generation is
not a proper solution thus this model can be considered.
The parameters of solar cell that has been used to test the performance of the
models are shown in Table 1. Basically, the temperature and solar irradiance are
obtained for the place of New Delhi and these along with the parameters of Solarex:
MSX 60 are fed as an input to the model of single and double diode topologies
and their performance are being compared throughout the year. In the table, the
parameters to design the solar cell such as O.C. current, S.C. voltage, and other
various constants are described that will help to design the cell easily.
4 Simulation and Comparative Analysis
The simulated model of single diode modeling of solar cell is shown in Fig. 4 and
the double diode representation is shown in Fig. 5.
The above models are designed in Simulink and the inputs to this model are
temperature and irradiance that are taken for the place of New Delhi. The other input
is system voltage that is taken from the datasheet of the practical existing solar cell
[10, 11]. The output is power generated calculated for entire year by varying the
Fig. 4 Single diode simulated model
Fig. 5 Double diode simulated model
temperature and irradiance as they vary with change in months for a certain place.
The variation in power by using both the topologies are shown in Table 2.
From the above Table 2, the variation of monthly power can be seen. The maximum
power that can be produced while using Solarex: MSX 60 is 60 W. So, to get the
Table 2 Power and current variation for single and double diode schemes in case of Solarex: MSX
60 solar cell for the place New Delhi
S. L. Month Temp G (MJ/m2 month) S.D Pmp S.D D.D D.D
No (°C) (W) Imp (A) Pmp (W) Imp (A)
1 Jan 14.11 412.92 24.80 1.37 23.80 1.39
2 Feb 17.05 459.76 27.34 1.51 26.59 1.56
3 Mar 22.59 639.84 37.31 2.06 37.26 2.18
4 Apr 29.66 722.10 36.29 2.01 39.63 2.32
5 May 33.38 757.33 32.22 1.78 38.87 2.27
6 Jun 33.50 676.20 26.66 1.47 33.71 1.97
7 Jul 31.01 591.17 25.60 1.41 30.49 1.78
8 Aug 29.86 551.49 24.65 1.36 28.75 1.68
9 Sept 29.40 567.00 26.30 1.45 30.00 1.75
10 Oct 26.24 520.80 26.64 1.47 28.57 1.67
11 Nov 20.77 423.90 23.62 1.31 23.83 1.39
12 Dec 15.26 369.83 21.66 1.20 20.98 1.23
desired amount of power, more number of solar cell can be connected in various
fashion as required. The variation of monthly power is represented in Fig. 6.
From Fig. 5, clear comparison of power production between the two schemes
can be observed. With the change in temperature and solar radiation, the power
production get effected. With the increase in solar radiation the power production
will gets increased while keeping temperature constant. But in case of vice versa,
while keeping solar radiation constant and the temperature increase, it is observed
the power production gets decreased slightly.
In case of single diode scheme, the maximum power is generated, i.e., 36.29 W in
the month of April due to maximum deviation in the figure as seen and high amount of
radiation occurs whereas lowest power production of 21.66 W takes place in month
of December as minimum amount of solar radiation, i.e., 369.83 MJ/m2 month takes
place. The yearly production when the solar cell is acting as single diode scheme is
333.09 W, whereas the monthly average production is 27.75 W. This is the production
from a single solar cell. By adding more cells in number, the production can be done
as per requirements. The maximum production of solar cell takes place when the
solar irradiance is 1000 MJ/m2 month and the temperature is 25 °C. Thus, the error
of production is 0.1% that is quite less.
In case of double diode scheme, the maximum power is generated, i.e., 39.63 W
in the month of April due to increase in solar radiation of 722.10 MJ/m2 month
and the lowest production takes place for the month of December, i.e., 20.98 W
due to less radiation of 369.83 MJ/m2 month. The yearly production when the solar
cell is acting as double diode scheme is 362.48 W, whereas the monthly average
production is 30.21 W which is high enough than single diode scheme. Here, the
maximum production takes place when the solar irradiance is 1000 MJ/m2 month
Comparative Analysis Between Single Diode …
Fig. 6 Schematic representation of power variation throughout the year for various schemes
153
and temperature is 25 °C which is 59.50 W having an error of 0.83%. Thus, the

increase in production proves better result for double diode topology than single
diode scheme. The system errors are also within limits thus this model can represent
a solar PV cell.
5 Conclusion
A clear comparison is presented between the various approaches of modeling [12–14]

of solar PV cell. Due to higher demand of energy production from renewable energy
sources, the designing of PV cell is an important aspect of production purpose. Double
diode scheme produces higher amount of production than single diode scheme. Due
to higher production and more advantage this modeling can be preferred than single
diode scheme. Here, for the place New Delhi double diode modeling suits with
better production. As the Government of India is giving more emphasis on the use
of renewable energy sources [15], this paper will give a clear view of designing and
production of solar rays for a certain area.
References
1. Valencia-Calvo J, Oliver Tost G, Garcia Ortega M (2020) Model and simulation of a renew-
able energy market: integration of renewable energy sources with the conventional genera-
tion system. In: IEEE 2020, 15th Iberian conference on information system and technologies
(CISTI), pp 1–7
2. Ministry of New and Renewable Energy, Government of India, Physical Progress (Achieve-
ments), Retrieved 18th July, 2018
3. Perara PU, He L (2018) A simplified mathematical model of PV cell simulation. In: IEEE
2018, 7th world conference on photovltaic energy conversion (WCPEC), pp 3204–3207
4. Morsy KA, Abouelatta MA, El-Banna MM, Elsaid MK (2018) Modelling and simula-
tion of photovoltaic module for space applications using simulink/MATLAB. In: 2018 first
ınternational workshop on deep and representation learning (IWDRL), pp 55–61
5. Nascimento AJ, Cavalcanti MC, Bradaschia F, Silva EA, Michels L, Pietta LP (2017) Parameter
estimation technique for double diode model of photovoltaic module. In: 2017 Brazilian power
electronics conference (COBEP), pp 978–984
6. Vinod RK, Singh SK (2018) Solar photovoltaic modelling and simulation: as a renewable energy
solution. Energy report, Research Paper in Elsevier, Available in ScienceDirect, pp-710–712
7. Selmi T, Abdul-Niby M, Devis L, Devis A (2014) P&O MPPT ımplementation using
MATLAB/simulink. In: 2014 9th ınternational conference on ecological vehicles and renewable
energies (EVER), pp 1–8
8. Ahmad T, Sobhan S, Faysal Nayan Md (2016) Comparative analysis between single diode and
double diode method of PV cell: concentrate different parameters effect on its efficiency. J
Power Energy Eng 31–46
9. Bana S, Saini RP (2016) A mathematical modelling framework to evaluate the performance of
single diode and double diode based SPV systems. Energy report. Research Paper in Elsevier,
Available in ScienceDirect, pp 171–187
10. Rodrigues EMG, Godina R, Pouresmaeil E, Porto JPS (2017) Simulation study of a photovoltaic
cell with ıncreasing level of model complexity. In: 2017 IEEE ınternational conference on
environmental and electrical engineering and 2017 IEEE ındustrial and commercial power
systems europe (EEEIC/I&CPS Europe), pp 1–4
11. Louzazni M, Khouya A, Amechnoue K, Gandelli A, Mussetta M, Craciunescu A (2018) Meta-
heuristic algorithm for photovoltaic parameters: comparative study and prediction with a firefly
algorithm. MDPI Appl Sci 1–22
12. Ibharim MM, Jiya JJ, Idakwo HO (2016) Characterization of PV-cell using real data. Int J Eng
Dev Res 4(2):1696–1703
13. Chandel TA, Yasin MY, Mallick A (2019) Modelling and simulation of photovoltaic cell using
single diode solar cell and double diode solar cell model. Int J Innov Technol Explor Eng
8(10):558–565.
14. abed-Ali RA, Yasin NM (2019) Simulation of solar cell and sinusoidal pulse width modulation
inverter using MATLAB and proteus. In 4th Scientific International Conference-Najaf-IRAQ-
2019, pp 86–91
15. Sahu P, Verma D, Nema S (2016) Physical design and modelling of boost converter for
maximum power point tracking in solar PV systems. İn: 2016 ınternational conference on
electrical power and energy systems (ICEPES), pp 10–15
Control of Three-Phase Grid-Connected
Inverter Using dq Axis Theory
Deepak Kumar Singh, Saibal Manna, and Ashok Kumar Akella
1 Introduction
Energy demand is rising every day as urban development and population increase.
This increased energy demand contributes to a rapid decrease of fossil fuels and thus
promotes the distributed generations (DG) like wind power, hydro power, and PV
system, etc. All these are clean sources of energy. DG sources are connected to the
grid for the reliability of the system [1]. Energy is transferred from DG to the utility
grid using power electronic converters. Various power electronic converters are used
for DG, depending on the energy source. DC/DC converter is used to transform
DC energy provided by the PV panels to constant DC voltage, which provides the
inverter’s DC-link. An additional DC/DC converter for boosting DC voltage is used
in some applications. At last, an inverter is used for transferring energy to the grid
from DC-link [2, 3]. The main concern with inverter connected to grid system is THD
of grid current and the system’s power factor. The grid current has a THD value of
less than 5% and power factor should be nearly unity. 3-F voltages and currents must
be synchronized with each other [4].
Different methods, including dq theory, power balance control theory and pq
theory are mentioned in the literature for control of the grid converters. The dq axis
theory is used here as it is easy to implement, active and reactive current can be
controlled separately. One more reason to use this theory is all control variable are
in DC frame, so by using simple PI controller, the complete control algorithm can
be implemented [5]. SPWM technique is used to provide the gate signal to the 3-F
GCI. It is a conventional switching scheme. The carrier wave, which is the required
frequency, is compared with the reference wave (Sine wave). Further, SPWM has
two switching schemes such as bipolar and unipolar switching. Unipolar switching
has complex control, less fluctuation, and high switching frequency from bipolar
scheme [6]. Use of power converters causes a slew of power quality issues like poor
D. K. Singh (B) · S. Manna · A. K. Akella

National Institute of Technology, Jamshedpur 831014, Jharkhand, India
158 D. K. Singh et al.
power factor, interruptions, flickers, and noise. The use of filters resolves this issue
[7].
Synchronization is essential for the inverter connected to grid. Here, synchro-
nization means same frequency, phase angle, and voltage level. If it does not meet
then circulating current will flow in the system and which leads to the damage of
the system. A reference current must be found to achieve synchronization, which
will be injected into the system through inverter. Various synchronization techniques
exist, including zero crossing, Atan function, and PLL [8, 9]. The simplest technique
for grid synchronization is zero crossing. Each half cycle, grid voltage zero crossing
points are detected and grid frequency is defined. This technique, on the other hand,
does not have a quick dynamic response to changes in frequency [10]. Grid voltages
are calculated then converted into a reference frame and filtered using the Atan func-
tion technique to eliminate voltage notches effect. To increase performance, delay
cancellation methods are employed as signal delays result from filters [11].
PLL is most popular synchronization technique and it is widely used. To send
active power in the grid, first mark the grid side voltage. Now, the current which
has to be sent should be in phase with this voltage. To send this current, a reference
signal must be produced, which must be in phase with the actual voltage and have a
magnitude between −1 and 1. The signal is thus produced using PLL and used as a
reference signal in an inverter linked to the grid to execute current controller. In the
same way, PLL is used to produce a signal 90 degree out of phase with the actual
voltage to transfer reactive power to the grid [12, 13]. The control and simulation of
3-F GCI is presented. Section 2 describes the operation of 3-F GCI and dq control
scheme is explained in Sect. 3. Section 4 explains the detail implementation of
proposed control strategy. Simulation outputs are placed in Sects. 5 and 6 summarizes
conclusion of the article.
2 Operation of 3-F GCI
Figure 1 shows the basic block representation of 3-F GCI. DC source is connected
at the input with the bus capacitance across it. It is then connected to the inverter
bridge, which can be MOSFET/IGBT. The inverter output is then plugged into the
LCL filter.
The current produced by the inverter contains lots of harmonics. When this current
is directly inserted to the grid, then it deteriorates grid voltage and hence causes lots
of power qualities issue. In order to avoid all these issues, filter is used at the inverter
INVERTER LCL
Vdc GRID
FILTER
Fig. 1 Block representation of 3-F GCI

Control of Three-Phase Grid-Connected Inverter … 159
Fig. 2 Internal structure of 3-F GCI
output. After connecting the filter, smooth sinusoidal current is obtained without any
harmonics.
Figure 2 shows basic schematic of the 3-F GCI. Vdc is DC voltage source at
the input. S1 –S6 are the power electronics switches (MOSFET/IGBT) which form
the inverter bridge. Inductors (L1 , L2 ) and capacitors (C) form LCL filter. Finally,
inverter is connected to the grid. Vga , Vgb, and Vgc are grid voltage of a, b, and c
phases, respectively. Similarly, ia , ib, and ic are grid current.
3 dq Axis Theory
Direct-quadrature theory or dq is a transformation principle in which AC sinusoidal

signals are converted into DC signal and then operate upon the DC values and then
back again transform it into the AC signals. So, this process of transformation from
AC domain to DC domain, DC domain to AC domain is popularly known as dq axis
theory. This theory is generally used to design controller and analysis of 3-F grid
connected system. There are two transformations in the dq axis theory, i.e., forward
and reverse transformation. Forward transformation is AC to DC transformation
while reverse transformation is DC to AC transformation.
Figure 3 shows abc–αβ–dq coordinate system phasor representation, where ρ is
the angle between αβ and dq coordinates system. Clarke and Park transform are
used in abc–αβ–dq transformation. The abc to αβ transformation, i.e., three phase
system in time domain to orthogonal stationary frame is done using Clarke transform.
The αβ to dq (orthogonal rotating reference frame) transformation is done using
Park transform. The consecutive implementation of these two transforms simplifies
Fig. 3 abc–αβ–dq
coordinate axis [15]
calculations with AC current and voltage waveform transformation into DC signal

[14]. Consider a space vector R having amplitude r at an angle θ.
3.1 AC to DC Transformation
This is also called forward transformation. abc to αβ transformation
⎡ r ⎤
rα 1 − 1
−√21 a
⎣ rb ⎦
= √2 (1)
rβ 0 23 − 23
rc
αβ to dq transformation

rd Cosρ Sinρ rα
= (2)
rq −Sinρ Cosρ rβ
3.2 DC to AC Transformation
This is also called reverse transformation. dq to αβ transformation

rα Cosρ −Sinρ rd
= (3)
rβ Sinρ Cosρ rq
αβ to abc transformation
⎡ ⎤ ⎡ ⎤
ra
2
3
0
⎣ rb ⎦ = ⎢
⎣ −3 3 ⎦
1 √1 ⎥ rα
(4)
rβ
rc − 13 − √13
The same procedure is applied to transform current or voltage.
4 Control Strategy
Figure 4 demonstrates the control strategy of 3-F GCI. To set up the controller, first
voltages need to be sensed. These 3-F abc voltages are then transformed to 2-phase
αβ voltages. Using these αβ voltages, PLL is implemented. The αβ voltages are then
converted to dq voltages. Now, currents need to be sensed for controller implemen-
tation. Either grid current or inverter side current is used for the implementation.
Here, inverter side current is being sensed. These currents are then transformed to
αβ domain and then it is transformed into dq domain.
Id and Iq correspond to active and reactive currents, respectively. To find the error,
Id and Iq are subtracted from the reference current. Error is being given to the PI
Fig. 4 Control block of 3-F GCI

controller to find out the voltages Ud and Uq .
Vd = Ud + E d + Lw Iq (5)
Vq = Uq + E q − Lw Id (6)
Here, w and L are grid frequency and inductance of the filter, respectively. For
SPWM, the relation between modulation index and inverter voltage is given by
Vdc
Vd = Md ∗ (7)
2
Vdc
Vq = Mq ∗ (8)
2
Therefore, V d and V q are multiplied with (2/V dc ) to obtain E d and E q . Finally, it
is transformed to abc voltages to get the reference for PWM generation.
5 Simulation Result
MATLAB/SIMULINK is used to simulate the system. The circuit’s parameters are

described in Table 1. The complete topology can be analyzed by varying the circuit
parameter. The results are taken into consideration for verification of the model.
Set active current reference (I d(ref) ) to 200A and I q(ref) to zero to pass only active
power to the utility grid. 3-F grid voltages and currents are displayed in Fig. 5
and Fig. 6, respectively. It is clearly recognized that the actual current is same as the
reference current. Figure 7 exhibits that grid current and voltage are in phase. Figure 8
depicts the grid current harmonic spectrum. The grid current’s THD is 2.55%. The
system has nearly UPF.
To check the effectiveness of the controller, set new reference current value (I d(ref)
= 150 A). It is clearly acknowledged that the grid current has changed to new
Table 1 Parameters of the

Parameters Values
circuit
DC input voltage 800 V
Grid voltage 415 V
Filter inductance 0.5 mH
Filter capacitance 0.1 mF
Fig. 5 3-F grid voltages
Fig. 6 3-F grid currents at I d(ref) = 200 A
reference value, i.e., 150 A as shown in Fig. 9. 1-F current and voltage are still in
phase as illustrated in Fig. 10.
To transfer only reactive power to the grid, set I q(ref) = 200 A and I d(ref) = 0 A.
It is clearly validated that current is same as reference current. Since, only reactive
Fig. 7 1-F grid voltage and current at I d(ref) = 200 A
Frequency (Hz)
Fig. 8 Grid current harmonic spectrum
current is being sent to the grid, then there is 90° phase difference between voltage
and current as depicted in Fig. 11.
6 Conclusion
In this study, control of 3-F GCI using dq axis theory is presented. LCL filter is
employed as it has superior filtering performance. 3-F inverter is regulated with
SPWM switching scheme at switching frequency 10 kHz. It has been seen that active
and reactive current are easily controlled and output is followed the reference value.
Fig. 9 3-F grid currents at I d(ref) = 150 A
Fig. 10 1-F grid voltage and current at I d(ref) = 150 A
The output frequency and voltage magnitude of the inverter has been regulated to
track the grid frequency and voltage in such a way that nearly UPF is always main-
tained. The grid current THD is 2.55%, which is less than 5%. The simulation output
exhibits that the proposed control scheme is effective. All the analysis, controller
design, and simulation results are verified.
Fig. 11 1-F grid voltage and current at I q(ref) = 200 A
References
1. Khanna A, Garg A, Singh A (2019) Maintenance of DC link voltage using synchronous refer-
ence frame theory. In: 2019 3rd international conference on recent developments in control,
automation & power engineering (RDCAPE), Noida, India, pp 17–21. https://doi.org/10.1109/
RDCAPE47089.2019.8979040
2. Isen E, Bakan AF (2016) Development of 10 kW three-phase grid connected inverter.
Automatika 57(2):319–328. https://doi.org/10.7305/automatika.2016.10.1081
3. Isen E, Bakan AF (2012) 10 kW grid-connected three-phase inverter system: control, simulation
and experimental results. In: 2012 3rd IEEE international symposium on power electronics for
distributed generation systems (PEDG), Aalborg, Denmark, pp 836–840. https://doi.org/10.
1109/PEDG.2012.6254098
4. Mejía-Ruiz GE, Rodríguez JR, Paternina MRA, Muñoz-Galeano N, Zamora A (2019) Grid-
connected three-phase inverter system with LCL filter: model, control and experimental results.
In: 2019 IEEE PES innovative smart grid technologies conference—Latin America (ISGT Latin
America), Gramado, Brazil, pp 1–6. https://doi.org/10.1109/ISGT-LA.2019.8895017
5. Yazdani A, Iravani R (2012) Voltage-sourced converters in power systems: modeling, control,
and applications. IEEE Press, Piscataway, NJ, USA
6. Bahrami A, Narimani M (2019) A sinusoidal pulsewidth modulation (SPWM) technique for
capacitor voltage balancing of a nested T-type four-level inverter. IEEE Trans Power Electron
7. Handoko S, Hadi SP, Suharyanto, Firmansyah E (2014) The performance of three-phase four-
wire grid-connected inverter with enhanced power quality. In: 2014 6th international conference
on information technology and electrical engineering (ICITEE), Yogyakarta, Indonesia, pp 1–5.
https://doi.org/10.1109/ICITEED.2014.7007956
8. Blaabjerg F, Teodorescu R, Liserre M, Timbus AV (2006) Overview of control and grid synchro-
nization for distributed power generation systems. IEEE Trans Ind Electron 53(5):1398–1409
9. Shaikh F, Joseph B (2017) Simulation of synchronous reference frame PLL for grid synchro-
nization using Simulink. In: 2017 international conference on advances in computing, commu-
nication and control (ICAC3), Mumbai, India, pp 1–6. https://doi.org/10.1109/ICAC3.2017.
8318790
10. Trinh Q, Choo FH, Wang P (2017) Control strategy to eliminate impact of voltage measurement
errors on grid current performance of three-phase grid-connected inverters. IEEE Trans Ind
Electron 64(9):7508–7519. https://doi.org/10.1109/TIE.2017.2682029
11. Klumpner C, Liserre M, Blaabjerg F (2004) Improved control of an active-front-end adjustable
speed drive with a small DC-link capacitor under real grid conditions. In: Proceedings of IEEE
35th annual power electronics specialists conference, Aachen, Germany, pp 1156–1162
12. Guo XQ, Wu WY (2013) Simple synchronisation technique for three-phase grid-connected
distributed generation systems. IET Renew Power Gener 7:55–62
13. Chung SK (2000) A phase tracking system for three phase utility interface inverters. IEEE
Trans Power Electron 15:431–438
14. https://in.mathworks.com/solutions/power-electronics-control/clarke-and-park-transforms.
html?s_tid=srchtitle
15. https://in.mathworks.com/help/physmod/sps/ref/clarketoparkangletransform.html?s_tid=src
htitle
Active–Reactive Power Support
with Optimal Allocation of DG
and DSTATCOM in Distribution System
Using Flower Pollination and Artificial
Bee Colony Algorithm with Load Growth
Nandola Maitrey Bharatbhai and Atma Ram Gupta
1 Introduction
The power grid network is made up of generation, transmission, and distribution.

Power is given from generation station to the transmission network with higher
voltage and transmission network transfers power to very long distance at load
center with lower voltage. So power reaches to load center via transmission and
DN. Transmission network requires large land space as compared to DN and also
they differ to each other via voltage ratings. It is also fact that power received
at load station is always lower than supplied by generation station due to various
losses in power system network. Losses are also dependent on equipment efficiency,
aging of the network, and rating of the network. Generally, power system losses are
majorly considering as T&D losses (transmission and distribution losses) because
as compared to other losses in the network, T&D losses are dominant. Due to high
R/X ratio in DN, high losses have been seen as compared to transmission system.
As per the report by Government of India in 2019, T&D losses in India are 21.04%
which is almost double than average T&D losses of the world [1]. As of August 2020,
with a gross installed capacity of 372,693 MW, India is the world’s third largest
producer and third largest consumer of electricity [2]. Mode wise installed capacity
is shown in Fig. 1, and mode wise gross generation is shown in Fig. 2. India is the
third largest country in consumption of electricity in this 41.16% of total electricity
consumed by industrial consumer, 24.76% by residential consumer, and 17.69% by
agricultural consumer. This consumption rate is gradually increasing year by year.
Per capita consumption in India is increased from 16 KWh in 1947 to 1181 KWh in
2019 as shown in Fig. 3.
N. M. Bharatbhai (B) · A. R. Gupta

Department of Electrical Engineering, National Institute of Technology Kurukshetra,
Kurukshetra, Haryana, India
e-mail: nandola_31904219@nitkkr.ac.in
170 N. M. Bharatbhai and A. R. Gupta
Coal/Lignite
13% Gas
22% Diesel
56%
2%
7% Nuclear
<1%
RES
Hydro
Fig. 1 All India installed capacity as on 31–03–2019 [2]
Coal/Lignite
10%
9% Gas
3%
<1% 4% Diesel
74% Nuclear
RES
Hydro
Fig. 2 Mode wise gross electricity generation (2018–19) [2]
Fig. 3 Per capita electricity consumption in India [2]

Active–Reactive Power Support with Optimal Allocation … 171
If we compare the growth in generation and growth in demand, then load growth
dominates and it is sensitive factor. For reliable operation, generation and demand
should equal and it is also true that if generation is increasing and if generated power
cannot be transferred through transmission and distribution network due to its lower
power rating, then power utility takes some steps like, transmission and distribution
network restructuring, substation expansion or new substation installation, and their
associated new feeders or both [3]. Study in [4] presented that many countries in the
world has almost 40-year-old power system network, so at this load, growth condition
power system network may not capable to operate in reliable manner. Also cost of
restructuring is too high and lack of investment in transmission network makes very
hard to operate power system in healthy manner. In the operation of the power grid,
the distribution system plays an important role since it is directly linked to the load
center. After deregulation of power system these problems becomes larger. Due to
limited reactive power sources, distribution system faces unwanted issues in voltage
like, voltage fluctuation in voltage, and voltage decay. Due to these problems, some
other power quality issues like increased power loss and decreased power flow limits.
The reason of limited reactive power sources is many power utilities work under stress
because of strict economic constraints and are as regards to the regulatory limits [5].
In order to earn benefit rather than pay a penalty, distribution firms typically aim to
keep the real power losses as low as possible.
The requirement of extra active and reactive power or increment in demand can
be fulfilled by DG and DSTATCOM, respectively. We can get benefits like reduction
in power loss, improved voltage profile, load balancing, and many more by using
DG and DSTATCOM.
There are serval definitions of DG based on voltage level (kV), DG scale (MW),
technology based on renewable and hybrid. According to CIGRE, DG is a system
which does not organized centrally and not dispatched centrally, and smaller than
50–100 MW and connected to the distribution grid [6].
The reasons for implementing DGs are shorter construction time, lower cost,
modular generation site, lower space requirement, diversification of energy sources,
energy efficiency, competition policy, etc. [3]. DG are often broadly classified in two
types renewable and non-renewable. Solar energy plants, wind generation plants,
power plants with fuel cells, power plants with biogas, etc., are counted in renewable
DG and non-renewable DG contains, duel-fed induction generator, ICE, etc. DG will
be graded on the idea of the active/reactive power supplied/absorbed as follows [7].
• Type-1: If DG supplies real and reactive power to the network at unity power
factor (pf) then it is known as type 1 DG. Wind, tidal, and wave energy sources
are the examples.
• Type-2: If DG supplies only real power to the network with leading pf of 0.8–0.99
then it is known as type 2 DG like solar energy plants, photovoltaic systems, fuel
cell, etc.
• Type-3: If DG supplies only reactive power to the network with zero pf, then it is
defined as type 3 DG, at zero power factor. Like, synchronous motor in overexcited
mode, synchronous condenser.
• Type-4: If DG provides real power and absorb reactive power from the system
at lagging pf of 0.8–0.99 then it is known as type 4 DG like duel-fed induction
generator.
In recent years, the energy sector has undergone many changes following the
implementation of smart grids. It affects not only utilities but consumers too. India
also moves faster towards smart grid and electric vehicles. With introduction of
electrical vehicles, electricity demand increases in remarkable way. To archive this,
government also promotes DG. GOI promotes DG via various policies and subsidies
like Rajiv Gandhi Grameen Vidhyutikaran Yojna (RGGVY) [8]. Since distribution
generation installed at the consumer side, consumer becomes prosumer (producer +
consumer).
Another device known as STATCOM which is power electronics-based voltage–
source converter. STATCOM may be operate as either source or sink of reactive
power. In some operations, it also supplies active power if connected as a source.
This device initially developed for transmission systems for operations like faster and
continuous inductive or capacitive compensation. Similarly, for distribution system, a
tool D-STATCOM is employed. D-STATCOM can inject leading or lagging current
as per the requirement of compensation. Here, the benefit is the total demand of
specified load meets utility connection specification because DSTATCOM is related
to a specified load [9, 10].
In actual scenario, loads are mixtures of varied load types like combination of
residential, industrial, and commercial loads. So, looking on the character of the
realm being supplied, each bus of the system includes a various type of load connected
thereto. Distribution system planners normally provide their various customer groups
with dedicated feeders, each of which have its distinctive load pattern that varies
hourly and seasonally [11]. As we have seen, there are four types of DG based on
active/reactive power supplied/absorbed, so as per the requirement DG couples with a
specific node which helps to fulfil some percentage of active/reactive power demand
at that node. Similarly, D-STATCOM couples with a particular load, which controls
bus voltage for controlling the power component. DG changes voltage at that bus and
branch current to get minimum power loss. There are many techniques for finding
optimal location and size of DG and DSTATCOM with objective of reducing power
loss in the system, (i) methods based on analysis, (ii) numerical methods, and (iii)
intelligence based methods [12]. Sandeep Kaur et al. [12] used mixed integer liner
programming-based numerical to find optimal placement of multiple DG. Shukla T.
N. et al. [3] used genetic algorithm. Ke-Yan Liu [14] used tie sequence characteristics
of loads and DGs to find optimal siting and sizing of DG. Yuvraj T. et al. [15]
proposed Cuckoo search method. S. Devi et al. [15] used particle swarm optimization
algorithm.
A review of DG planning and various optimization techniques have been carried
out in [16] and changing in power system load flow after DG implementation is
carried out in [17]. In this article, we compare two separate approaches, FPA and
ABC algorithm, to decide the best size and position of DG and DSTATCOM for
reducing losses and optimizing voltage profiles. Yang and Deb [18] suggested a
metaheuristic search algorithm called FPA. Alyasseri et al. [19] offers a descrip-
tion of FPA varieties. The FPA is focused on flower pollination methods used by
birds, insects, bats, and other species. In [20], Levy flights are used to characterize
the behavior of birds and insects. ABC algorithm is also a metaheuristic approach
which is proposed by Karaboga for solving multidimensional optimization prob-
lems [21]. Honey bee swarms’ intelligent foraging behavior motivated the ABC
algorithm. Seker and Hocaoglu [22] finds the optimum size and position of DG
using the ABC algorithm. IEEE 33 bus RDS is used to compare the two approaches
by evaluating the decrease in power loss and improvements in voltage profile after
DG and DSTATCOM are applied in the device. Load and bus data for IEEE 33 are
taken from [24].
In this analysis, type 2 DG is taken which can deliver only active power to the
system and DSTATCOM which can only deliver reactive power to the system. The
formulation of this paper is as follows, Sect. 2 covers problem formulation, Sects.
3 and 4 cover FPA and ABC algorithm, respectively, Sect. 5 includes results, and,
finally, conclusion is covered in Sect. 6.
2 Problem Formulation
2.1 Load Flow Analysis
In this paper, the direct approach based load flow analysis method is used. This
method first proposed by J. H. Teng [25]. By this, voltage profile at each node and
power loss at each branch are calculated. To understand this algorithm, here let’s
take small example of six bus distribution system (radial) as shown in Fig. 4.
4 5
Bus 1 B4
2 3
B3
6
B1 B2 IL4 IL5
S/S
B5
IL2 IL3
IL6
Fig. 4 Six bus RDS

2.1.1 Computation of BCBV (Branch Current to Bus Voltage Matrix)

and BIBC (Bus Injection to Branch Current Matrix) Matrix
For node voltages,

V2 = V1 – Z12 *I12 .
V3 = V2 – Z23 *I23 = V1 – Z12 *I12 – Z23 *I23 .
V4 = V3 – Z34 *I34 = V1 – Z12 *I12 – Z23 *I23 – Z34 *I34 .
V5 = V4 – Z45 *I45 = V1 – Z12 *I12 – Z23 *I23 – Z34 *I34 – Z45 *I45 .
V6 = V3 – Z36 *I36 = V1 – Z12 *I12 – Z23 *I23 – Z36 *I36
⎡ ⎤ ⎡ ⎤⎡ ⎤
V1 − V2 Z12 0 0 0 0 I12
⎢ V1 − V3 ⎥ ⎢ Z12 Z23 0 0 0 ⎥⎢ I23 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ V1 − V4 ⎥ = ⎢ Z12 Z23 Z34 0 0 ⎥⎢ I34 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎣ V1 − V5 ⎦ ⎣ Z12 Z23 Z34 Z45 0 ⎦⎣ I45 ⎦
V1 − V6 Z12 Z23 0 0 Z36 I36
[dV] = [BCBV][Ibranch] (1)
From the above figure,

I45 = I5 .
I34 = I45 + I4 = I5 + I4 .
I23 = I34 + I36 + I3 = I6 + I5 + I4 + I3 .
I12 =⎡I23 +⎤I2 = I⎡
6 + I5 + I4 + I⎤
3⎡ I36 = I6 .
+ I2, ⎤
I12 11111 I2
⎢ I23 ⎥ ⎢ 0 1 1 1 1 ⎥⎢ I3 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
So,⎢ I34 ⎥ = ⎢ 0 0 1 1 0 ⎥⎢ I4 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎣ I45 ⎦ ⎣ 0 0 0 1 0 ⎦⎣ I5 ⎦
I36 00001 I6
[Ibranch] = [BIBC][Ibus] (2)
2.1.2 Computation of Bus Voltage and Branch Currents
From Eqs. (1) and (2),
[Ibranch] = [BIBC][Ibus]
[dV] = [BIBC][Ibranch]
[dV] = [BIBC][BIBC][Ibus]
So, [dV] = [DLF][Ibus] (3)
where DLF is direct load flow matrix.
2.1.3 The Active and Reactive Power Loss for Branch J Can Be Found
by
P(j) = I2 (j) ∗ r(j) (4)
Q(j) = I2 (j) ∗ x(j) (5)
To perform this method in any number of bus system, general steps are given here,
for this consider Fig. 5.
Step-1 In this step, we have to initialize all bus voltages as 1 p.u. Initialization of
voltages, V (i) = 1p.u. for i = 1, 2, 3, …, N
Step-2 In this step, we have to read load data and bus data.
Step-3 Compute BCBV and BIBC matrices.
Step-4 Now initialize iteration count k = 0
Step-5 Calculate load current at all buses except slack bus which is generally
numbered bus 1.
IL(i) = PL(I)−jQL(i)
V(i)∗
for i = 2, 3, …, N
Step-6 Now calculate current through each branch,
[Ibranch] = [BIBC][Ibus]
Step-7 Find DLF matrix and compute bus voltages
[dV] = [DLF][Ibus]
Step-8 Calculate difference between new voltages and previous voltage values and
update voltages

eik = Vik − Vik−1 for i = 2, 3, …, N.V i = V i k
1 2 3 j m n N
Fig. 5 N bus distribution system

Step-9 Check the tolerance condition, emaxk = max(e1k , e2k , . . . , eNk )

If emax <tol, then move to nest step otherwise update the iteration count k = k+1
and go to step 5.
Step-10 Calculate line losses and line flows.
Step-11 Print the results.
Step-12 Stop
2.2 Objective Function (oF)
OF is helpful to find best solution of any problem using mathematical approaches.

In our analysis, we have used OF which contains only one component which is Total
Active Power Loss (TPL).
OF = minimum(TPL) (6)
N
TPL = I2 (j) ∗ r(j)
j=1
where N is the total number of branches.

The best position for DG and DSTATCOM based on the constraints mentioned
below.
2.2.1 Active Power Constraint
DGi < PDGi <PDGi

Pmin max
(7)
where PDGi max and PDGi min are the maximum and minimum sizes of DG at bus i.
2.2.2 Reactive Power Constraint
DSi < QDSi <QDSi

Qmin max
(8)
where QDSi min and QDSi max are the minimum and maximum sizes of DSTATCOM at
bus.
2.3 Investment Cost (IC)
For the cost analysis of our solution, we have to consider investment cost. For calcu-
lating investment expense of DG, consider cost characteristics of DG given in Eq. (9)
[26], DSTATCOM’s annual investment expense can be estimated using Eq. (10) [26].
Cost of DG = (aPdg2 + bPdg + c)$/MWh (9)
where Pdg is the active power supplied by DG in MW. In this paper, values of
confidents are taken as, a = 0, b = 20, c = 0.25
(1 + B)n ∗ B
C(DSTATCOM)per year = C(DSTATCOM) (10)
(1 + B)n − 1
where C(DSTATCOM)/year is the DSTATCOM annual cost in $ and C(DSTATCOM)

is the DSTATCOM investment cost in dollars per kVAr at the time of installation.
DSTATCOM’s asset rate of return and durability are B and n, respectively. Here,
C(DSTATCOM) = 50 $/kVAr, B = 0.1, and n = 30 used in this article.
2.4 Cost of Energy Saving (CE)
Energy saving (ES) is difference between energy loss before installation of DG and
DSTATCOM and energy loss after installation of DG and DSTATCOM [26]. It can
be formulated as (11)
ES = (EL)TkWh (11)
where EL is the difference between TPL in kW before installation and TPL’ in kW

after installation. T is the time duration in hours.
2.5 Annual Cost Saving (ACS)
ACS is the difference between cost of energy saving and cost of investment. It can
be calculated using (12).
ACS = Ke(ES) − IC$ (12)

where Ke is the energy rate in per kWh. In this paper, we have used Ke = 0.06 /KWh,
T = 8760 (hours),
3 Flower Pollination Algorithm (FPA)
The FPA is focused on flower pollination methods used by birds, insects, bats, and
other species. Yang and Deb [18] suggested a metaheuristic search algorithm called
FPA. FPA is metaheuristic search algorithm. Pollination are often divided in two
forms, organic and inorganic. Generally, 90% pollination is finished by pollens, so
it’s called as organic, and 10% pollination is inorganic means no pollen required.
Wind and water diffusion in this inorganic pollination allows the pollination of
certain plants. Organic pollination is done in two ways, pollination by self and
cross-pollination.
Organic and cross-pollination may have counted in global pollination because
in this, pollination can be done by long distance traveling of bees, birds, and other
animals. Study in [20] says that jump or distance covered by pollinators can be formu-
lated as Levy distribution. Hence, birds, bees may behave as Levy flight behavior.
Steps in this distribution can be found using similarity or difference of two flowers.
Based on this, global pollination can be mathematically as [18]
Pit+1 = Pit + λL(γ)(G − Pit ) (13)
where the solution vector of Pi at iteration t + 1 is Pit + 1. At iteration t + 1, G is the

best solution, λ is a scaling factor, L(λ) is the pollination power, which is effectively
a step size since pollinators travel in steps over a long distance. From [19], to imitate
these functions, we can use Levy travel (L), which is L > 0 from the Levy distribution.
λ
λ(γ)sin pi ∗ 1
L(γ) = 2
∗ s s0 > 0 (14)
pi s1+γ
This distribution is used for larger steps s > 0 and (γ) is the standard gamma
function. Here, (γ) = 1.5 is used based on review in [20].
For the local pollination,
Pit+1 = Pit + ε(Pjt − Pk t ) (15)
where Pjt and Pkt are pollen produced from separate flowers of the same genus of
plants. E is uniformly distributed in [0, 1]. A detailed review of variants of FPA is
given in [19]. In Fig. 6, flow chart of FPA is shown.
Start
Initialize population solution xi

(xi = xi1, xi2, xi3, xi4, …, xin)
Run power flow
Find the best solution at the initial population

and store in G and iteration (j) = 0
j=j+1
Y
Y rand < pswitch N
Global pollination Local pollination
Evaluate the new best solution using equation

N
(9) and (11)
New solution < G
Y
N
Update new best population position x(i)
j > max iteration
Optimal solution Xbest
Stop
Fig. 6 Flow chart for FPA [19]
4 Artificial Bee Colony (ABC) Algorithm
The behavior of real honey bees looking for food influenced the ABC algorithm. The
introduction of ABC is first given by Karaboga in 2005 [21]. ABC is an optimization
technique that offers a population-based search procedure in which artificial bees
change individual food locations over time, with the aim of the bees being to locate
the places of food supply with the most nectar and recognize it as the one with the
greatest nectar. Employed bees, onlooker bees, and scout bees are the three groups
of bees that make up a honey bee swarm [23].
• Employed bees: During this, randomly searching for food source locations (solu-
tions) as per number of communities such that the information is exchanged is
done. The dance cycle depends on the amount of honey from the food supply
(fitness value).
• Onlooker bees: in step with the standard of food source from searching by assorted
employed bees, best food source position could even be found.
• Scout bees: An overlooked employed bee source turns into a scout and continues
to fly in search of a specific food source [23].
Phase 1: Initialization
The initial food source or populations can be found by the expression,
xi = Lb + rand(−1, 1)(Ub − Lb) (16)
where Ub and Lb are the upper and lower bounds of the solution X.
Phase 2: Employed phase
The following expression is used to calculate the location of the nearest food source:
vi, j = xi, j + u(xi, j − xk, j) (17)
where j is the randomly selected parameters index. U is the random choice that exists
among [−1, 1] xk, j is a randomly selected food source. This formula is used to
measure fitness, after which greedy selection is added between xi and vi,j,
1
Fitm(X m) = i f Fm(X m) > 0 (18)
1 + Fm(X m)
Fitm(X m) = 1 + Fm(X m)i f Fm(X m) < 0 (19)
where Fm(Xm) is the objective function value of Xm.

Phase 3: Onlooker phase
The profitability of a food source influences its quality, and profitability is measured
by expression.
f itnessi
Pi = N (20)
i=0 f itnessi
where fitnessi denotes the solution’s fitness value and N denotes the number of food
source positions (populations). Onlooker bees use the following expression to find a
local food source:
vi, j = xi, j + u(xi, j − xk, j) (21)
Phase 4: Scout phase

When the location of food is discarded, employed bees become scouts. As follows,
Scout generates new food positions:
xi, j = Lb, j + rand(−1, 1)(Ubj − Lb, j) (22)
At initialization step random sources is generated using Eq. (12). After phase 1
completed, phase 2 to phase 4 proceed, this cycle repeats till it reaches maximum
cycle limit. In every cycle, new fitness value is computed and new fitness value is
greater than previous one then bee replace this new fitness value with old one. In
every cycle, employed bees first finds nectar information and its location then shares
information with onlooker bees. Onlooker bees utilize this information and try to
find best location of source which gives highest nectar quantity. Figure 7 Illustrates
the ABC algorithm flow chart.
As quantity of nectar increases, onlooker bees trying to prefer that location. The
location of highest nectar is found by onlooker bees using Eq. (16). Scout bees are the
pioneers but haven’t any direction when trying to find food, so it generates probability
of sources which has the lowest and the highest nectar quantity. The selection in scout
bee (phase 4) is regulated by the parameter "limit". When the number of trails reaches
the limit, the employed bee abandons the source and becomes a scout [28].
In this section, we will see results obtain by ABC and FPA optimization methods.
As we have seen FPA relies on flower pollination process. Number of populations N,
and switch probability p, are the sole parameters that makes FPA simple, and when
N is fixed, the values of p controls the pollinators. For the calculation, values of
n = 20 and p = 0.8 are taken here, and DG and DSTATCOM range and position are
found for IEEE 33 bus RDS. The ABC algorithm solution depends on the digestive
tracts of actual honey bees. This algorithm has three control parameters, the colony
size (CS) (which is summation of employed bees and onlooker bees), limit value,
and maximum cycle number MAXCYCLE. Table 1 lists the parameters required for
the ABC and FPA optimization methods.
FPA and ABC algorithms are coded in MATLAB R2015a and simulation results
for the IEEE 33 bus RDS are performed in Asus laptop with Intel core i5 @3.4 GHz.
The following are the results for IEEE 33 bus test systems.
Fig. 7 Flow chart for ABC Start

algorithm [22]
Set known parameter of ABC [limit, MCN, CS]
Read line and bus data
Run power flow
Inisiate food source position X(i)

where i = 1, 2, … CS/2
Set trial = 0 and iteration (j) = 0
j=j+1
j > MCN ?
NO
Compute fitness value each employed bees using
(17)
Is better position
derived for Y Set trial(i) = 0
employed bee?
Let trial(i) = trial(i) + 1
Compute fitness value each onlooker bees using

(21)
Yes
Is better position
derived for Y Set trial(i) = 0
onlooker bee?
Let trial(i) = trial(i) + 1
Reinitialize ith
Trial(i) > X(i) ? Y employed bee
using (16)
Remember best position till now
Print best position
End
Table 1 Parameters of FPA

FPA ABC
and ABC
No. of populations = 20 CS = 40
Probability = 0.8 MAXCYCLE(MCN) = 30
Max. iterations = 1500 Failure limit = 3
Here, optimal locations of DG and DSTATCOM are searched between bus

numbers 2 and 33. Results separated based on various cases like case 1 is base
case, case 2 is for only DG installation, case 3 is for only DSTATCOM installa-
tion, and in case 4, both DG and DSTATCOM have implemented in the system. To
find best solution of various cases, we have used two metaheuristic methods named
ABC and FPA. Initial parameters for the calculation of both methods as shown in
Table 1. For the load flow parameters, 100 MVA as a base power and 12.66 kV as a
base voltage are taken. For the base load scenario, the total active and reactive power
demand is 3.715 MW and 2.30 MVAr, respectively. Results for the optimization
problem is shown in Tables 2 and 3.
Table 2 DG and DSTATCOM optimum positioning and sizing using ABC

Parameters Case 1 Case 2 Case 3 Case 4
0.8 PF lagging UPF
Location – 8 7 30 7, 30
Size – 2000 kW 2000 kW 1258.8 kVAr 2000, 1258.8
TPL kW 210.98 119.09 115.19 151.37 61.69
Vmin in p.u 0.9038 0.9405 0.9364 0.9165 0.9486
EL in KWh 1,848,205.8 1,043,228.4 1,009,064.4 1,326,026.6 540,447.3
ES in KWh – 796,371.6 839,141.4 522,179.2 1,307,758.5
EC in $ – 47,782.29 50,348.48 31,330.75 78,465.51
ACS in $ – 7782.05 10,098.48 9227.70 16,109.98
Table 3 DG and DSTATCOM optimum positioning and sizing using FPA

0.8 PF lagging UPF
Location – 8 7 30 7, 30
Size – 2000 KW 2000 1258.00 2000,1258.00
TPL kW 210.98 119.09 115.98 151.37 61.69
Vmin in p.u 0.9038 0.9405 0.9364 0.9165 0.9486
EL in KWh 1,848,205.8 1,043,228.4 1,016,070.6 1,326,026.6 540,447.3
ES in KWh – 796,371.6 832,135.2 522,179.2 1,307,758.5
EC in $ – 47,782.296 49,928.11 31,330.75 78,465.51
ACS in $ – 7782.05 9678.11 9225.22 16,109.98
Table 2 shows results obtain using ABC algorithm. Without DG and DSTATCOM,
active power loss is 210.98 kW, for case 2, it is 119.09 kW (for 0.8 pf lagging), and
115.19 kW (for Upf), for case 3, it is 151.37 kW, and, for case 4, it becomes 61.69 kW
which is lowest among all other cases.
Loss reduction in case 2 is 45%, for case 3 is 28.3%, and for case 4, it is 70.7%.
Also the annual cost saving for case 2 is 9678 $, for case 3, it is 9225$, and, for case
4, it is 16109.98 $. In case 4, loss reduction is highest and cost saving is also highest
(Tables 4 and 5).
Minimum bus voltage in p.u. before installation of DG and DSTATCOM is
0.9038 p.u. which is lower than permissible limit 9.5 p.u. Now as we install DG
and DSTATCOM, it improves and, for case 4, it becomes 0.9486 p.u. for DG size of
2000 kW and for DSTATCOM size of 1258 kVAr. If size of DG and DSTATCOM
increases, then voltage profile also improves. Here, we find results for maximum
Table 4 Comparison with existing methods for case 2

Parameters MINLP [12] CPLSM [28] IVM [28] VSI [29] ABC FPA
location, size 6, 2590 8, 1800 30, 1550 16, 1000 7, 2000 7, 2000
(KW)
TPL without 211.176 210.976 210.976 210.976 210.982 210.982
DG
(KW)
TPL with 111.10 118.13 125.16 136.75 115.19 115.98
DG (KW)
Loss 47.39 44.01 40.68 35.18 45.40 45.03
reduction %
Vmin in p.u – 0.9433 0.9275 0.9318 0.9364 0.9364
EC in $ – 48,800.3 45,109.4 39,011.5 50,348.5 49,928.1
Cost of Pdg – 36.25 31.25 20.25 40.25 40.25
$/MWh
ACS in $ – 12,550.3 13,859.4 18,761.5 10,098.5 9678.1
Table 5 Comparison with existing methods for case 3

Parameters VSI [30] PSI [30] BA [31] ABC FPA
location, size 30, 1250 29, 1300 30, 1150 30, 1258.78 30, 1258.00
(KVAr)
TPL without 210.98 210.98 202.67 210.98 210.98
DG (KW)
TPL with DG (KW) 151.37 153.07 143.97 151.37 151.37
Loss reduction % 28.25 27.44 28.97 28.25 28.25
ES KWh 522,183.6 507,291.6 514,212.0 522,179.2 522,179.2
Vmin in p.u – – 0.9244 0.9165 0.9165
EC in $ – – 24,768.00 31,330.75 31,330.75
Fig. 8 Voltage in p.u. at each bus for different cases
DG rating as 2000 kW and maximum DSTATCOM rating as 2000 kVAr. Voltage

changes at each node in every cases and it is shown in Fig. 8.
Results in Tables 2 and 3 are only for constant loading. Tables 6 and 7 shows
optimization results for 20% under loading, 20% overloading, and 7% load growth
for 5 years. The active power demand at base loading is 3.715 MW, while the reactive
power demand is 2.30 MVAr. For 20% under loading, it becomes 2.972 MW and
1.84 MVAr, for 20% overloading, it becomes 4.458 MW and 2.76 MVAr. If the
load increases by 7% per year for 5 years, the power rises to 5.210 MW and 3.23
MVAr. Tables 4 and 5 demonstrate how the results correspond to all other methods
for deciding the best size and position of DG and DSTATCOM.
The value of branch currents changes as DG and DSTATCOM are installed in
the device and, as a result, active power loss (Ibr2 * r) decreases. Figure 9 shows
reduction in branch currents for all cases. From the graph, we can see the reduction
in branch currents as we install DG and DSTATCOM in the system.
Figure 10 illustrates voltage profile changes for different loading conditions, while
Fig. 11 illustrates active power loss for various loading conditions and cases. As the
load on each node increases, the voltage drops below the minimum permissible
voltage limit, and the load on each node decreases, the voltage attempts to reach the
maximum permissible voltage limit.
Annual Energy loss cost and annual cost savings for different cases are shown
in Fig. 12. Annual energy loss cost for case 1 is 110892.3 $ and it is decreased to
32,426.84 $ in case 4. Annual cost savings is also highest in case 4 by 16,109.98 $.
Table 6 Results for IEEE 33 bus RDS using ABC

With constant load
Location – 7 30 7, 30
Size – 2000 kW 1258.8 kVAr 2000, 1258.8
TPL in kW 210.98 115.19 151.37 61.69
Vmin in p.u 0.9038 0.9364 0.9165 0.9486
EL in KWh 1,848,205.8 1,009,064.4 1,326,026.6 540,447.3
ES in KWh – 839,141.42 522,179.22 1,307,758.50
EC in $ – 50,348.48 31,330.75 78,465.51
ACS in $ – 10,098.48 9227.7 16,109.98
20% load decrease at each bus
Location – 6 30 6, 30
Size – 2000 kW 1002.31 kVAr 2000, 1002.31
TPL in kW 130.72 69.85 94.55 36.97
Vmin in p.u 0.9244 0.9536 0.9342 0.9630
20% load increase at each bus
Location – 8 30 8, 30
Size – 2000 kW 1519.76 kVAr 2000, 1519.76
TPL in kW 314.49 174.36 223.59 394.20
Vmin in p.u 0.8822 0.9287 0.8982 0.9542
7% load growth for 5 years
Location – 8 30 8, 30
Size – 2000 kW 1791.2 kVAr 2000, 1791.2
TPL in kW 445.94 250.96 313.87 136.59
Table 7 Results for IEEE 33 bus RDS using FPA

With constant load
Location – 7 30 7, 30
Size – 2000 kW 1258.00 kVAr 2000, 1258.00
TPL in kW 210.98 115.98 151.38 61.69
Vmin in p.u 0.9038 0.9364 0.9165 0.9486
EL in KWh 1,848,205.8 1,016,070.6 1,326,026.6 540,447.3
ES in KWh – 832,135.18 522,179.22 1,307,758.50
EC in $ – 49,928.11 31,330.75 78,465.51
ACS in $ – 9678.11 9225.22 16,109.98
20% load decrease at each bus
Location – 6 30 6, 30
Size – 2000 kW 1000.41 kVAr 2000, 1000.41
TPL in kW 130.72 69.85 94.55 36.97
Vmin in p.u 0.9244 0.9536 0.9342 0.9630
20% load increase at each bus
Location – 8 30 8, 30
Size – 2000 kW 1519.22 kVAr 2000, 1519.22
TPL in kW 314.49 174.36 223.59 94.20
Vmin in p.u 0.8822 0.9287 0.8962 0.9542
7% load growth for 5 years
Location – 8 30 6, 30
Size – 2000 kW 1787.88 kVAr 2000, 1787.88
TPL in kW 445.94 250.96 313.88 136.58
Fig. 9 Branch currents in p.u. for each branch

Fig. 10 Voltage in p.u. at each bus for different cases
Fig. 11 Total active power loss for different cases and different loadings
Fig. 12 Annual energy loss cost and annual cost savings for different cases
6 Conclusions
In different cases, the FPA and ABC algorithms are used to determine the best size
and position of the DG and DSTATCOM in order to improve the voltage profile and
minimize active power loss and of the system. The comparison of the results for
IEEE 33 bus RDS with existing methods is shown. ABC algorithm requires very less
computation time as compared to FPA for solving these kind of optimization problem.
Solution using ABC requires 1.2–2 s, while FPA requires 27–30 s. Comparison of
base load results with various load growth cases is done. From the results, we can
conclude that,
• After installing DG and DSTATCOM, voltage profile of system improves. Results
with both DG and DSTATCOM is better as compared to alone DG or DSTATCOM.
For case 2, voltage improvement in case 2 is 3.6%, in case 3 is 1.4%, and in case
4 is 4.5%.
• The annual cost savings and loss reduction are greatest when both DG and
DSTATCOM are installed. Annual cost saving in case 2 is 9678.11 $, for case 3
is 9225.22 $, and in case 4 is 16109.98 $.
• As the load increases, the system’s minimum bus voltage changes. For base case,
minimum bus voltage is 0.9038 p.u., for 20% decrease in load, it becomes 0.9244
p.u., for 20% load increment, minimum bus voltage falls to 0.8822 p.u.
• As system loading changes, voltage profile of the system changes and, for over-
loading condition, it goes below the minimum voltage limit and for under loading,
it goes above maximum voltage limit.
Improved voltage profile of the system, reduced active power loss, and cost savings
are all advantages of using a network with both DG and DSTATCOM.
References
1. https://En.Wikipedia.Org/Wiki/Electricity_Sector_In_India
2. Growth of Electricity Sector In India From 1947–2019, Central Electricity Authority. August
2019
3. Shukla TN, Singh SP, Naik KB (2008) Allocation of optimal distributed generation using GA
for minimum system losses. NPSC, Dec 2008
4. Root CE (2006) The future beckons [Electric Power Industry]. IEEE Power Energy Mag 4:24–
31
5. Joscow P (2006) Patterns of transmission investments in competitive electricity markets and
sustainability. 131
6. Acharya N, Mahat P, Mithulananthan N (2006) An analytical approach for DG allocation in
primary distribution network. Int J Electr Power Energy Syst 28:669–678
7. Singh B, Mishra DK (2018) A survey on power performance by optimally placed DG in
distribution network. Energy Rep 4:129–132
8. https://Powermin.Nic.In/En/Content/Decentralized-Distributed-Generation-Ddg
9. Chen B-S, Hsu Y-Y (2008) A minimal harmonic controller for a STATCOM. IEEE Trans Ind
Electron 55(2):655–664
10. Devabalaji KR, Ravi K (2014) Power quality improvement in wind farm connected to grid
using STATCOM. In: Proceedings of the international conference on advances in electrical
engineering (ICAEE’14), pp 1–5, Unnao, India, Jan 2014
11. Kanwar N, Gupta N, Niazi KR, Swarnkar A (2015) Optimal distributed generation allocation
in radial distribution systems considering customer-wise dedicated feeders and load patterns.
J Mod Power Syst Clean Energy 3(4):475–484
12. Kaur S, Kumbhar G, Sharma J (2014) A MINLP technique for optimal placement of multiple
DG units in distribution systems. Electr Power Energy Syst 63:609–617
13. Yuvaraj T, Devabalaji KR, Ravi K (2015) Optimal allocation of DG and DSTATCOM in radial
distribution system using cuckoo search optimization algorithm. Hindawi, 2017
14. Liu K-Y, Sheng W, Liu Y, Meng X, Liu Y (2015) Optimal sitting and sizing of DGs in distribu-
tion system considering time sequence characteristics of loads and DGs. Electr Power Energy
Syst 69:430–440
15. Devi S, Geethanjali M (2014) Optimal location and sizing determination of distributed gener-
ation and DSTATCOM using particle swarm optimization algorithm. Electr Power Energy
Syst
16. Theo WL, Lim JS, Ho WS, Hashim H, Lee CT (2017) Review of distributed generation (DG)
system planning and optimisation techniques: comparison of numerical and mathematical
modelling methods. Renew Sustain Energy Rev 67:531–573
17. Singh B, Mishra DK (2018) A survey on enhancement of power system performances by
optimally placed DG in distribution networks. Energy Rep 4:129–158
18. Yang X-S (2012) Flower pollination algorithm for global optimization. Unconventional
computation and natural computation, vol 9. Springer, p 240
19. Alyasseri ZAA, Khader AT, Al-Betar MA, Awadallah MA, Yang X-S (2018) Variants of the
flower pollination algorithm: a review. Springer
20. Pavlyukevich I (2007) LÉvy flights, non-local search and simulated annealing. J Comput Phys
226:1830–1844
21. Karaboga D (2005) An idea based on honey bee swarm for numerical optimization. Technical
report TR06. Computer Engineering Department, Erciyes University, Turkey
22. Seker AA, Hocaoglu MH (2013) Artificial bee colony algorithm for optimal placement and
sizing of distributed generation. IEEE ELECO 2013
23. D. Karaboga, B. Akay: A Modified Artificial Bee Colony (ABC) Algorithm For Constrained
Optimization Problems. Applied Soft Computing, Volume 11, Issue 3, April 2011, Pages 3021–
3031, ISSN 1568–4946.
24. Kashem MA, Ganapathy V, Jasmon GB, Buhari MI (2000) A novel method for loss minimiza-
tion in distribution networks. In: International conference on electric utility deregulation and
restructuring and power technologies 2000. IEEE
25. Teng JH (2000) A network-topology based three-phase load flow for distribution systems. Proc
Natl Sci Council ROC (A) 24(4):259–264
26. Kumar A, Vijay Babu1 P, Murty VVSN (2017) Distributed generators allocation in radial
distribution systems with load growth using loss sensitivity approach. J Inst Eng India Ser B
98(3):275–287
27. Gupta AR, Kumar A (2019) Reactive power deployment and cost benefit analysis in DNO
operated distribution electricity markets with D-STATCOM. Front Energy 13:86–98
28. Murty VVSN, Kumar A (2013) Comparison of optimal DG allocation methods in radial
distribution systems based on sensitivity approaches. Electr Power Energy Syst 53:450–467
29. Kejun Q, Chengke Z, Malcolm A, Yue Y (2011) Effect of load models on assessment of energy
losses in distribution generation planning. Electr Power Res 2:1243–1250
30. Guptaa AR, Kumar A (2015) Energy savings using D-STATCOM placement in radial
distribution system. Proc Comput Sci 70:558–564
31. Yuvaraj T, Ravi K, Devabalaji KR (2015) DSTATCOM allocation in distribution networks
considering load variations using bat algorithm. Ain Shams Eng J
Analysis and Design Optimization
of Six-Phase Induction Motor
by Metaheuristic Methods
Pratyush Prasanna Das and S. N. Mahato
1 Introduction
Induction motor is the most popular machine because of its low cost, little mainte-
nance, easy availability, and robust construction. Nowadays, multi-phase induction
machines find many applications in high-power drives as these machines reduce the
torque pulsation frequency and amplitude. Also, it reduces harmonic currents that
flow in rotor circuits. The DC link current harmonics are reduced and reliability is
increased with the increase of power. These multi-phase machines have applications
in electric hybrid vehicles, in aerospace applications and ship propulsion [1]. The
modeling of six-phase induction motor under open-phase fault condition has been
performed in [2]. The genetic algorithm-based design optimization has been reported
by Fazlipour et al. [3] and the results have been compared with that of traditional
design method.
Digital synchronous frame current regulation for six-phase induction motor has
been reported by Bojoi et al. [4]. The high phase order induction motor theory has
been presented in [5]. Nanoty and Chudasama presented the control and design
of six-phase induction motor [6]. The testing of six-phase induction motor has been
discussed by Gregor et al. [7]. Using third harmonic current injection for the increase
in torque density in case of six-phase induction motor has been reported in [8]. Modi
and Chaturvedi [9] discussed the theory of six-phase induction motor design. Mondal
[10] analyzed the performance of a six-phase induction motor.
In case of Novel Bat Algorithm (NBA) with habitat selection and Doppler effect
in echos, a self-adaptive local search has been proposed by Meng et al. [11]. The
authors have developed the way of habitat selection based on quantum and mechan-
ical behavior of a bat. Cuckoo search algorithm facilitates both local and global
search capabilities and uses Levy flights as a way to reach the global optima [12].
P. P. Das (B) · S. N. Mahato

Department of Electrical Engineering, NIT Durgapur, Durgapur, India
e-mail: ppd.16EE1503@phd.nitdgp.ac.in
192 P. P. Das and S. N. Mahato
The Levy flight is the straight line trajectory flight with sudden twist of 90° within
the flight [13]. In Cuckoo search algorithm, each pattern is attributed to both a nest
and to an egg [14]. Many complex multi-objective engineering problems like spring
design optimization and welded beam design have been solved with Cuckoo search
algorithm [15]. Optimal design of PID controller for speed control of BLDC motor
have been done by Flower Pollination Algorithm [16].
The main contributions of this paper are:
• Modification of Flower Pollination optimization algorithm.
• Use of modified Flower Pollination optimization algorithm (mFPA), NBA,
Cuckoo Search Algorithm, and traditional design method for optimized design of
a six-phase induction motor.
• Study and comparison of the performances of the motor using these opti-
mized parameters of different algorithms in MATLAB/SIMULINK simulation
environment.
2 Optimization Algorithms
2.1 Modified Flower Pollination Algorithm
In case of flower pollination algorithm (FPA), pollens are pollinated to the flowers
by insects or any living being. This type of pollination is done by selecting the best
pollen and making the desired flower pollinated. This concept has the Levy flight
involved with the movement of the insects.
The modification is done in this mFPA by incorporating a matrix manipulation that
leads to more closure to the best-valued matrix or the position of the flower having
best pollen. The Levy step size is also applied each time of matrix management. The
updated position of the flower is as follows:
xi (t + 1) = xi (t) + L × {xi (t) − gbest}
Here, the evolutionary matrix manipulation has been done in order to get much
more convergent position of the optimal best flower, associated with globally best
position of the pollen. The matrices added here for convergence are as follows:
λ1 = L × (xi (t)/best)

λ2 = L × best/1.4142 /best
λ3 = L × ((1/best) × ((1 − xi (t))/(1 + xi (t))))

Analysis and Design Optimization of Six-Phase Induction … 193
λ4 = L × ((xi (t) − best)/2 + (best − xi (t))/2)
λ5 = L × (xi (t)/2) + L × (xi (t)/2)
λ6 = L × (|(xi (t) − best)/best|)
λ7 = L × (|(xi (t) − best)/(xi (t) + best)|)
λ8 = L × (best − xi (t)) × 1.414
λ9 = L × (xi (t) − best)
λ10 = xi (t)
The correction vector is the function of aforesaid vectors, i.e., equals to f (λ1 ,
λ2 …. λ10 ).
The equation for updated flower position for biotic pollination is given as
xi (t + 1) = xi (t) + f (λ1 , λ2 . . . ..λ10 )
2.2 Novel Bat Algorithm
A short-waved modulated frequency called echolocation is used by each bat to avoid

obstacles, which controls the velocity of the bat in the Bat algorithm. An inertia weight
factor, contraction and expansion coefficients, and a compensation rate of Doppler
effect involved for echoes are incorporated in positional and velocity equations in
order to increase the convergence of the optimization algorithm and thus called NBA.
In this paper, taking losses of six-phase induction motor as objective function, the
optimized design of the SPIM is obtained.
2.3 Cuckoo Search Algorithm
In Cuckoo search algorithm, cuckoos lay eggs to a nest selected by the cuckoos. The
host bird finds the best nest with eggs or could throw the egg away to build a new
nest instead, with a probability of 0–1. In this paper, both the rotor and stator copper
losses are minimized to get the optimized design of six-phase induction motor.
3 Design Problem Formulation of Six-Phase Induction

Motor
In order to formulate the design problem for optimization of the six-phase induction
motor, some design parameters are used as design variables to be optimized. The
objective functions involved here for optimization are stator and rotor copper losses.
If f1 (xi ) and f2 (xi ) are two objective functions, where xi are the design variables
to be optimized, then minimization of f1 (xi ) and f2 (xi ) gives optimized values of xi .
4 Objective Functions and Design Equations
The rating of six-phase induction motor used in traction application in Indian

scenario, which is intended to be optimally designed, is given as follows:
1150 kW, 50 Hz, 2180 V, 6 Pole, Delta-connected six-phase induction motor,
number of stator slots = 72, and number of rotor slots = 54.
The control parameters for mFPA, NBA, and Cuckoo search algorithm, respec-
tively, are given as follows:
Control parameters for Modified Flower Pollination Algorithm:
Size of population = 20
Value of switch probability = 0.8
Iteration number in total = 2000
Search variables dimension = 3 in case of rotor copper loss

= 1 in case of stator copper loss
Control parameters in Novel Bat Algorithm (NBA):
Rate of pulse emission and loudness updating frequency = 10

Selection ofhabitatprobability, maximum = 0.9
Selection ofhabitatprobability, minimum = 0.6
Co - efficient of expansion, maximum = 1
Co - efficient of contraction, minimum = 0.5
Weight of inertia, maximum = 0.9
Weight of inertia, maximum = 0.5
Doppler effect echos compensation rate, maximum = 0.9
Doppler effect echos compensation rate, minimum = 0.1
Search variables dimension = 3 in case of rotor copper loss

Control parameters for Cuckoo Search Algorithm:
Number of nests or solutions = 3

Rate of discovery of alienated eggs or solutions = 0.25
Iteration Number time = 4
Dimension of search variables = 3 in case of rotor copper loss

The design variables and the design constraints are given as follows:
There are seven design variables for six-phase induction motor intended to be
designed optimally. They are as follows:
Core length which is equal to rotor and stator = x1

Stator inner diameter = x2
Depth of stator slot = x3
Width of stator teeth = x4
Depth of stator core = x5
Efficiency, η = x6
Power factor = x7
Design Constraints are as follows:
Current density in the rotor bar is between 4 and 7 A/mm2

Ratio of pole length to pole face length, L/τ = 1.5
Average flux density, Bav = 0.45 Wb/m2
Kw = 0.955
Space factor = 0.4
Maximum allowable flux density in the stator slot = 1.7 Wb/m2 .
Flux density in the stator core = 0.8 Wb/m2
The maximum and minimum values design variables for mFPA, NBA, and Cuckoo
search algorithm are given as follows:
For Novel Bat algorithm:
0.3 < Core length, x1 < 0.54
0.68 < Stator inner diameter < 0.8
0.8 < Efficiency < 0.99
0.8 < Power factor < 0.99.

For modified Flower Pollination Algorithm:
0.3 < Core length, x1 < 0.54
0.99 < Efficiency < 1
For cuckoo search algorithm:
0.3 < Core length, x1 < 0.54
0.8 < Efficiency < 0.99
Rotor copper loss in six-phase cage rotor induction motor, in terms of bar current,
diameter of end ring can be written as

Pr cl = rr sr Ib2 /ab × ((L r + 2De )/P) = 6.8Ib (x1 + De /3)
In case of six-phase induction motor, power = 6Vph Iph cosØ and therefore
Is = 87.9/( p. f × η) A.
and the rotor bar current is given as
((2m s kws Ts )/Sr ) × (87.9/(cos∅ × η)) × cos∅
Putting the value of ms, this expression becomes as
1057.8 × kws (Ts /(Sr × η))
The expression for maximum flux density is (Bav L2 )/1.5 and in terms of the design
variables, it becomes as 0.29 × x12 .
The stator turns per phase = 2180/ 4.44 × 50 × 0.29 × x12 × 0.955 =
35.46/x12
Bar current = 4.8kws (Ts /(Sr η)) = 2.9/ x12 × η
The end ring current becomes(Sr Ib )/π p and, after putting the values of variables,
the expression becomes as 8.41/ x12 × η A. Considering
the value of current density
as 6 A/mm2 , the area of end ring will be 1.4/ x12 × η mm2 . Now from the table of
recommended sizes of copper strip IS 1897–1962, the value of t is taken as 0.47w.
The outside and insidediameters of the end ring interms of the design variables
are given as x22 − 0.006 m and x22 − 0.006 − 0.1/ x1 × η(1/2) m, respectively.
Therefore, the mean diameter of ring becomes x2 − 0.0006 − 0.1/ x1 × η(1/2) m.
The rotor copper loss = 6.8Ib (x1 + 0.33De )
1/2
= 19.72/ x12 × x6 × x1 + 0.33x2 − 0.0002 − 0.033/ x1 × x6 watts.
4.1 Novel Bat Algorithm

2
The rotor copper loss = 19.72/ x1 × x6 × x1 + 0.33x2 − 0.0002 − 0.033/ x1 ×
1/2
x6 watts.
As the stator current is 88.8/x7 , the area of stator conductor will be 14.8/x7 . The
length of mean turn is taken as (2L + 2.3 + 0.24) = 5x1 + 0.24.
The stator resistance = x7 (0.18x1 + 1.2)/x12 ×0.001 = x7 (0.18x1 + 1.2)/x12 ×
0.001 and the stator copper loss becomes as 210.473/x7 watts.
4.2 Modified Flower Pollination Algorithm
Similarly, as per above calculation, the rotor copper loss for

mFPA is.
2 1/2
19.72/ x1 × x6 × x1 + 0.33x2 − 0.0002 − 0.033/ x1 × x6 watts.
The stator copper loss = 659.223/x7 watts.
4.3 Cuckoo Search Algorithm

The rotor copper loss = 19.72/ x12 ×x6 × x1 +0.33x2 −0.0002−0.033/ x1 ×
1/2
x6 watts and the stator copper loss = 298.899/x7 watts.
The optimization of the parameters of the six-phase induction machine has been
carried out using the above-mentioned algorithms and also by traditional design
method. The results obtained have been presented in Table 1. It is observed from
this table that the new mFPA algorithm gives better efficiency of 82.828% and the
optimized value of power factor is 0.98, in comparison to the other optimization
algorithms and the traditional method of design. The modified version of FPA is the
most energy efficient method. It is also found that, in case of the mFPA algorithm
of optimization, the size of the motor is highly reduced as compared to the other
optimization methods, as stator stack length is reduced to 0.3 m. Hence, the cost of
material also gets reduced. As the flux density in the stator core and the maximum
flux density in the stator teeth are reduced in case of modified FPA, the specific iron
loss is reduced and in turn, the stator iron loss gets reduced.
In the same way, the rotor-specific iron loss and thus the rotor iron loss are also
reduced. As the magnetizing reactance, in case of mFPA, is increased to a value of
151.0655 than the other design methods, the short-circuit current is minimum in
this case. The mFPA provides reduced stator and rotor resistances, hence the stator
and rotor copper losses are also lowered. As the iron loss is proportional to the weight
of the stator and rotor, the weight of the motor gets reduced in case of mFPA. As the
stator and rotor losses, which are the main sources of temperature rise, are decreased
in case mFPA method, the temperature rise of the motor is also reduced. As the
maximum flux density of the rotor core, stator teeth, rotor teeth are lowest in value
in case of mFPA design method than other design methods, so the iron loss also gets
reduced in case of mFPA design method (Table 1).
Figure 1 shows the efficiency curve of different design methods. It is seen that the
efficiency of mFPA is the highest among all the design methods. The total losses as
obtained using different design methods are presented in Fig. 2, which shows that
the total loss in mFPA method is the lowest among the different methods of design.
Also, it is observed from Figure 3 that the stator stack length is the minimum in case
of mFPA method of design.
From Table 2, it is found that the time taken to reach steady operating condition
in case of mFPA design method is very much less and it is only 0.4 secs, whereas in
traditional design method it is 3.3 secs and other methods had not reached the steady
operating condition. The time taken to reach 500 RPM on 28 Nm traction load is
0.4 secs in case of mFPA design method but other methods had produced more time.
The mFPA design method yields less time taken to reach 500 RPM on no load and
that is 0.6 secs. Reduced times in both the cases contribute to reduced heating and
reduced power loss for mFPA design method. Also, results of Table 2 show mFPA
design method produces more stable motor than others.
Table 3 shows comparison of mFPA design method of this paper with that of
GA-based design method which is done by Fazlipur et al. and is referenced hereby
at [3]. From the Table-3, it is found that mFPA design method yields much more
improved power factor, increased efficiency, reduced total loss, and reduced stator
stack length. Although, motor rating for mFPA-based design and GA-based design
(under paper in Ref. [3]) are different but the mFPA design method is much more
good results.
5.1 Validation of Proposed Algorithm
From Table 4, it is found that the proposed mFPA design method is a confident
method of design as for both the cases of design, viz., for 1150 KW motor which
have been done here and for 0.26 KW motor which have been done but not shown
here, yields 12.4 and 4% increase in full-load efficiency, total loss decreased 41.94
and 19.3%, as well as stator stack length decreased 44.44 and 6.67% for 1150 and
0.26 KW motors, respectively. Other methods of designs like NBA and Cuckoo search
produces infidel results as in those cases efficiency at full load, total loss, and stator
Table 1 Parametric values of different design methods

Design parameters Novel Bat Modified FPA Cuckoo search Traditional
algorithm design
Power factor 0.99 0.98 0.9548 0.86
Efficiency at FL (%) 78.4711 82.828 70.3965 73.685
Total loss (W) 3.1551 × 105 2.3842 × 105 4.836 × 105 4.107 × 105
No-load power factor 0.1 0.1 0.1 0.1
Stator resistance () 0.4 0.3 0.4 0.4
Rotor resistance () 6 8 6 6
Equivalent resistance 150 138 150 150
()
Equivalent Reactance 3 5.7 5.2 5.2
()
Magnetizing Reactance 120.5096 151.0665 117.1576 120.5096
()
Stator stack length (m) 0.54 0.3 0.4955 0.54
Stator inner diameter 0.68 0.68 0.7887 0.68
(m)
Rotor inner diameter 0.399 0.499 0.5268 0.3996
(m)
Width of stator teeth (m) 0.006 0.0049 0.008 0.0088
Depth of stator slot (m) 0.015 0.015 0.015 0.015
Stator slot width (m) 0.0869 0.0869 0.0869 0.0869
Depth of stator core (m) 0.112 0.0625 0.103 0.0057
Flux density of stator 0.8 0.8 0.8 0.8
core (T)
Flux density of rotor 0.7397 0.4109 0.5501 0.7397
core (T)
Maximum stator teeth 3.4 1.8889 3.1198 3.4
flux density (T)
Maximum rotor teeth 0.8551 0.4751 0.6359 0.8551
flux density (T)
stack length decreased, increased, and increased, respectively, in case designing 0.26
KW motor. Those characteristics has been shown in bar chart Fig. 4a–c.
6 Development of Simulink Model
The electromagnetic torque developed by the six-phase induction motor is given as

(Eq. 1)
Fig. 1 Efficiency bar chart 85

of different design methods
Efficiency in percentage,%
NBA
80
mFPA
75
70 Cuckoo
Search
65 TradiƟonal
Design
60
Different Design methods
Fig. 2 Total loss bar chart of 600000

different design methods
NBA
500000
Total Loss in Watts
400000 mFPA
300000
Cuckoo
200000 Search
TradiƟonal
100000
Design
0
Fig. 3 Stator stack length 0.6

bar chart of different design
Stator Stack Length in metre
methods 0.5
0.4 NBA
mFPA
0.3
Cuckoo Search
0.2 TradiƟonal Design
0.1
0
Te = (3/2)(P/2)(L m /L lr )

(i dr L lr + i dr L m + i ds L m + i dso L m ) − i qr L lr + i qr L m + i qs L m + i dso L m
(1)
and the expression of speed can be written as (Eq. 2).

Table 2 Comparison of different design method’s parametric values

Design parameters Novel Modified Cuckoo Traditional
Bat FPA search design
algorithm
Maximum torque on 28 Nm traction load (Nm) 510 420 600 600
Time taken to reach maximum torque on 28 Nm 0.01 0.01 0.01 0.01
traction load (s)
Time taken to reach steady operating condition on Not 0.4 Not 3.3
28 Nm traction load (s) reached reached
Maximum torque on no load (Nm) 510 330 600 600
Time taken to reach maximum torque on no load 0.01 0.01 0.01 0.01
(s)
Time taken to reach 500 RPM on 28 Nm traction 15.4 0.4 20 11.1
load (s)
Time taken to reach 500 RPM on no load (s) 13.3 0.6 19.2 10
Table 3 Comparison of mFPA design method with that of GA-based design method
Design Percentage increase or Percentage increase or Remarks
parameters decrease (%) for decrease (%) for mFPA
GA-based design to design to traditional
traditional design (for design
paper in Ref. [3])
Power factor Same 13.95 (increase) Power factor improved
much more in case of
mFPA design
Efficiency at 7.5 (increase) 12.4 (increase) Efficiency increased
FL much more in case of
mFPA design
Total loss 22.3 (decrease) 41.94 (decrease) Total loss decreased
much more in case of
mFPA design
No-load power 18.8 (increase) Same No-load Power factor
factor remains same in of
mFPA design
Stator 71.45 (decrease) 25 (decrease) Stator resistance
resistance decreased in case of
mFPA design
Magnetizing 18.2 (increase) 24.79 (increase) Magnetic reactance
Reactance increased in case of
mFPA design
Stator stack 36 (increase) 44.44 (decrease) Stator stack length
length decreased much more in
case of mFPA design
Table 4 Validation data for different algorithms

Motor 1150 KW 0.26 KW
ratings
Design Percentage Percentage Percentage Percentage Percentage Percentage
parameters increase or increase or increase or increase or increase or increase or
decrease decrease decrease decrease decrease decrease
(%) for (%) for (%) for (%) for (%) for (%) for
mFPA NBA Cuckoo mFPA NBA Cuckoo
design to design to search design to design to search
traditional traditional design to traditional traditional design to
design design traditional design design traditional
design design
Efficiency 12.4 6.5 4.46 4 (increase) 6.5 14
at FL (increase) (increase) (increase) (decrease) (decrease)
Total loss 41.94 23.2 17.75 19.3 36 84
(decrease) (decrease) (increase) (decrease) (increase) (increase)
Stator stack 44.44 No Change 8.24 6.67 85.2 67
length (decrease) (decrease) (decrease) (increase) (increase)
ωr = (P/2J )(Te − TL ) dt. (2)
The zero sequence current does contribute to the torque and hence for the speed.
Accordingly, it is taken into consideration.
The direct and quadrature axes voltage equations of stator and rotor in terms
of currents and flux linkages and stator and rotor resistances, are as follows (from
Eqs. 3–6):
dψds
vds = Rs i ds + − ωe ψqs (3)
dt
dψqs
vqs = Rs i qs + + ωe ψds (4)
dt
dψdr
vdr = Rr i dr + + (ωe − ωr )ψqr (5)
dt
dψqr
vqr = Rr i dr + + (ωe − ωr )ψdr (6)
dt
The stator and rotor flux linkage equations in terms of magnetizing inductances,
stator and rotor inductances are as follows (from Eqs. 7–8):
ψds = L ls i ds + L m (i ds + i dr ) (7)
ψdr = L lr i dr + L m (i ds + i dr ) (8)
ValidaƟon of mFPA Design method

20
eff_ FL_mFPA_1150KW_%
10
eff_FL_mFPA_0.26KW_%
0
1
Total
-10
Loss_mFPA_1150KW_%
Total
-20
Loss_mFPA_0.26KW_%
-30 Stator Stack

Length_mFPA_1150KW_%
-40 Stator Stack

Length_mFPA_0.26KW_%
-50
ValidaƟon of NBA Design method

100
eff_ FL_NBA_1150KW_%
80
eff_FL_NBA_0.26KW_%
60
Total Loss_NBA_1150KW_%
40
20 Total Loss_NBA_0.26KW_%
0 Stator Stack
1 Length_NBA_1150KW_%
-20 Stator Stack
Length_NBA_0.26KW_%
-40
ValidaƟon of Cuckoosearch Design method

100
eff_
FL_Cuckoosearch_1150KW_%
80
eff_FL_Cuckoosearch_0.26KW_
%
60
Total
Loss_Cuckoosearch_1150KW_%
40
Total
Loss_Cuckoosearch_0.26KW_%
20
Stator Stack
Length_Cuckoosearch_1150KW
_%
0 Stator Stack
1 Length_Cuckoosearch_0.26KW_
%
-20
Fig. 4 a Validation bar chart of mFPA design method. b Validation bar chart of NBA design method.
c Validation bar chart of Cuckoo search design method

ψqs = L ls i ds + L m i qs − i qr (9)

ψqr = L lr i dr + L m i qs + i qr (10)
An arbitrary reference frame with a rotating speed we is considered here, to which

the machine equations are transformed. With the equations of voltage of stator and
rotor sides in terms of flux linkages and currents, the SIMULINK/MATLAB model
has been prepared for simulation.
7 Performance Analysis of the SPIM for Different Design

Methods
Simulation has been carried out in arbitrary reference frame (Figs. 5, 6, 7, 8, 9, 10,
11, 12, 13, 14). The parameters, as obtained from different optimization algorithms
and traditional method of design, have been incorporated in simulation. Simulation
has been done for 28 Nm traction load and no-load condition. Traction speed has
been taken as 500 RPM. From Fig. 15, it is found that, in case of mFPA design
no-load current is 2.4 A and, from Fig. 16, it is found that the full load current is
400
U
300 U
U
X1
200
U
Y1
U ,U ,U X 1 ,U Y 1 [V ]
100
-100
-200
-300
-400
0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2
Time [s]
Fig. 5 Alpha, beta, x, y component voltage curve of Cuckoo search design method
400
U
a
300 U
b
U
c
200
U
x
U
100 y
U
z
U [V ]
-100
-200
-300
-400
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 6 Six-phase input voltage curve of Cuckoo search design method
50
Ia
40
Ib
30 Ic
Ix
20
Iy
10
Iz
Iabc [A]
-10
-20
-30
-40
-50
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 7 Six-phase input current curve of Cuckoo search design method on no-load
50
Ia
40
Ib
30 Ic
Ix
20
Iy
10
Iz
Iabc [A]
-10
-20
-30
-40
-50
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 8 Six-phase input current waveform of Cuckoo search design method on 28 Nm traction load
1.5
0.5
[Wb]
-0.5
-1
-1.5
-1.5 -1 -0.5 0 0.5 1 1.5 2
[Wb]
Fig. 9 Flux distribution curve of alpha–beta axes of Cuckoo search design method on no-load
0.5
0
[Wb]
-0.5
-1
-1.5
-1.5 -1 -0.5 0 0.5 1 1.5 2
[Wb]
Fig. 10 Flux distribution curve of alpha–beta axes of Cuckoo search design method on 28 Nm
traction load
-18
x 10
5
1
[Wb]
0
Y1
-1
-2
-3
-4
-5
-4 -3 -2 -1 0 1 2 3 4
X1
[Wb] -18
x 10
Fig. 11 Flux distribution curve of X–Y component of Cuckoo search design method on no load
-18
x 10
5
1
[Wb]
0
Y1
-1
-2
-3
-4
-5
-4 -3 -2 -1 0 1 2 3 4
X1
[Wb] -18
x 10
Fig. 12 Flux distribution curve of X–Y component of Cuckoo search design method on 28 Nm
traction load
700
600
500
400
T orque [N m ]
300
200
100
-100
-10 0 10 20 30 40 50
Speed [rpm]
Fig. 13 Torque versus speed curve of Cuckoo search design method at no load
700
600
500
400
Torque [Nm]
300
200
100
-100
0 10 20 30 40 50 60
Speed [rpm]
Fig. 14 Torque versus speed curve of Cuckoo search design method on 28 Nm traction load
2.5
I
a
2
I
b
1.5 I
c
I
1 x
I
y
0.5
I
Iabc [A]
-0.5
-1
-1.5
-2
-2.5
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 15 Six-phase input current curve of mFPA design method on no load

4
Ia
3 Ib
Ic
2
Ix
Iy
1
Iz
Iabc [A]
-1
-2
-3
-4
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 16 Six-phase input current waveform of mFPA design method on 28 Nm traction load
3.8 A. So, change of current from no load to full load of 28 Nm is small. Figures 7,
8, 23, 24, 27, and 28 are the current curves of Cuckoo search, NBA, and traditional
design methods, respectively, and these curves show that no change of current from
no load to full load. Figures 17, 18 of mFPA design method show flux distribution
curve of alpha–beta components and Figs. 19, 20 show flux distribution curve of X–Y
components and those curves show that the flux distributions are uniformly rotating
for Alpha–Beta components, flux linkages are intense for X–Y components. Flux
distribution curves of other design methods of Alpha–Beta and X–Y components are
as in figures of Cuckoo search design method, which show that this method produces
less uniformly rotating and less intense flux distribution. Other design methods like
NBA and traditional method produce same magnetic properties like Cuckoo search
design method, so only Cuckoo search methods’ magnetic behavior regarding Alpha–
Beta and X–Y axes at no load and full load have been given in Figs. 9, 10, 11, and
12, respectively. Voltage curves of Alpha, Beta, X, Y components are at Fig. 5, input
six-phase voltage curves are at Fig. 5. These curves are for Cuckoo search design
method. The voltage curves are similar in nature for other design methods, so only
one design method of Cuckoo search is shown (Figs. 21, 22, 23, 24, 25, 26, 27 and
28). Figures 13, 14, 21, 22, 25, 26, 29, and 30 are the Torque–speed characteristics
curves on 28 Nm Traction load and on no load of Cuckoo search, mFPA, NBA,
and traditional design methods, respectively. From these curves, it is found that only
mFPA design method has stable operating region.
1.5
0.5
[Wb]
-0.5
-1
-1.5
-1.5 -1 -0.5 0 0.5 1 1.5 2
[Wb]
Fig. 17 Flux distribution curve of alpha–beta component of mFPA design method on no load
1.5
0.5
[Wb]
-0.5
-1
-1.5
-2
-1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
[Wb]
Fig. 18 Flux distribution curve of alpha–beta component of mFPA design method on 28 Nm

traction load
-18
x 10
5
1
[Wb]
0
Y1
-1
-2
-3
-4
-5
-4 -3 -2 -1 0 1 2 3 4
X1
[Wb] -18
x 10
Fig. 19 Flux distribution curve of X–Y component of mFPA design method on No load
-18
x 10
5
1
[Wb]
0
Y1
-1
-2
-3
-4
-5
-4 -3 -2 -1 0 1 2 3 4
X1
[Wb] -18
x 10
Fig. 20 Flux distribution curve of X–Y component of mFPA design method on 28 Nm traction
load
500
400
300
Torque [Nm]
200
100
-100
-200
-100 0 100 200 300 400 500
Speed [rpm]
Fig. 21 Torque versus speed curve of mFPA design method on 28 Nm traction load
350
300
250
200
Torque [Nm]
150
100
50
-50
-100
-100 0 100 200 300 400 500 600
Speed [rpm]
Fig. 22 Torque versus speed curve of mFPA design method on no load

40
Ia
30 Ib
Ic
20
Ix
Iy
10
Iz
Iabc [A]
-10
-20
-30
-40
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 23 Six-phase input current waveform of NBA design method on 28 Nm load
40
Ia
30 Ib
Ic
20
Ix
Iy
10
Iz
Iabc [A]
-10
-20
-30
-40
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 24 Six-phase input current waveform of NBA design method on no load
8 Conclusions
A new modified version of Flower Pollination algorithm, NBA with habitat selection
and Doppler effect in echos, Cuckoo search algorithm, and also traditional method of
600
500
400
Torque [N m ]
300
200
100
-100
-10 0 10 20 30 40 50 60 70
Speed [rpm]
Fig. 25 Torque–speed characteristics curve of NBA design method on 28 Nm traction load
600
500
400
T o rq u e [N m ]
300
200
100
-100
0 10 20 30 40 50 60 70 80
Speed [rpm]
Fig. 26 Torque–speed characteristics curve of NBA design method on no load

40
I
a
30 I
b
I
c
20
I
x
I
10 y
I
Iabc [A ]
-10
-20
-30
-40
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 27 Input current waveform of traditional design method on 28 Nm traction load
40
I
a
30 I
b
I
c
20
I
x
I
10 y
I
Iabc [A]
-10
-20
-30
-40
0.9 0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1
Time [s]
Fig. 28 Input current waveform of traditional design method on no load

700
600
500
400
Torque [N m ]
300
200
100
-100
-10 0 10 20 30 40 50 60 70 80 90
Speed [rpm]
Fig. 29 Torque versus speed curve of traditional design method on 28 Nm traction load
700
600
500
400
Torque [Nm]
300
200
100
-100
0 10 20 30 40 50 60 70 80 90 100
Speed [rpm]
Fig. 29 Torque versus speed curve of traditional design method on no load

design have been presented for design of a six-phase induction motor. Simulation has
been carried out using the optimized parameters obtained from the different methods
in MATLAB/SIMULINK environment. It is observed that the mFPA method of
optimization gives the better performance as given below:
• The full-load efficiency of the six-phase induction motor is increased under mFPA
design method.
• The size of the six-phase induction motor is reduced under mFPA design method.
• The cost of material is reduced as the size of the motor is reduced.
• The maximum torque generated is the highest in case of mFPA design.
The total losses are greatly reduced in case of mFPA design and it is the most
energy-efficient among these methods.
9 List of Symbols Used
The list of symbols used here for different equations and expressions are as follows:
xi (t + 1) = the updated position of a pollen or an insect carrying pollen.
xi (t) = any initial position of a pollen or insect carrying pollen.
gbest = It is the position of best flower within a different plant having that flower.
L = Levy flight or essentially a step size.
best = position of the best flower of different plant situated elsewhere.
av = (xi (t) + best)/2.
rr = 0.021 /m = resistivity of copper conductor at 75 °C.
sr = No. of rotor slots.
ms = No. of phases.
P = No. of poles.
Ib = Rotor bar current.
ab = Rotor bar x-sectional area.
De = Mean end ring diameter.
Lr = Length of the core.
I ph = I s = Stator current per phase.
Vph = Vs = Stator voltage per phase.
p.f. = cos Ø = power factor.
η = efficiency.
t = Thickness of copper strip.
w = width of copper strip.
J = Moment of inertia of SPIM.
Te = Electromagnetic torque developed by six-phase induction motor.
TL = Load torque.
Kws = Space factor.
dt = A very small amount of time.
iqs = Quadrature axis component of stator current.
idr = Direct axis component of rotor current.
ids = Direct axis component of stator current.

iqr = Quadrature axis component of rotor current.
Lm = Magnetizing inductance.
Rs , Rr = Resistances of stator and rotor.
Lls , Llr = Inductances of stator and rotor.
iqs0 = Zero sequence component of quadrature axis component of stator current.
ids0 = Zero sequence component of direct axis component of stator current.
vds , vdr = Direct axis voltages of stator and rotor, respectively.
vqs , vqr = Quadrature axis voltages of stator and rotor.
ds , dr = Direct axis flux linkages of stator and rotor.
qs , qr = Quadrature axis flux linkages of stator and rotor.
References
1. Nanoty A (2012) Control of designed developed six phase induction motor. Int J Electromagn
Appl 2(5):77–84
2. Miranda RS, Gomes EC (2012) Analysis and modeling of six-phase induction motor under
open phase fault. Anais do XIX Congresso Brasileiro de Automática, CBA
3. Fazlipour Z, Kianinezhad R, Razaz M (2015) Genetic algorithm based design optimization of
a six phase induction motor. J Electr Eng Technol 10(3):1007–1014
4. Bojoi R, Profumo F, Tenconi A (2003) Digital synchronous frame current regulation for dual
three-phase induction motor drives. In: Conf. Rec. IEEE-PESC’03, Acapulco (Mexico), pp
1475–1480
5. Klingshirn EA (1983) High phase order induction motors part-1, Description and theoretical
consideration. IEEE Trans PAS 102(1):47–54
6. Nanoty A, Chudasama AR (2012) Control of designed developed six phase induction motor.
Int J Electromagn Appl 2(5):77–84
7. Gregor R, Barrero F, Toral S, Durán MJ (2008) Realization of an asynchronous six-phase
induction motor drive test-rig, RE&PQJ 1(6)
8. Lyra ROC, Lipo TA (2002) Torque density improvement in a six-phase induction motor with
third harmonic current injection. In: IEEE Trans Ind Appl 38(5)
9. Modi SC, Chaturvedi RK (2015) Theory & manage of designed developed prototype six phase
induction motor. Int J Eng Res Manag Technol 2(Issue-6)
10. Mandal S (2015) Performance analysis of six-phase induction motor. Int J Eng Res Technol
(IJERT) 4(Issue 02)
11. Meng X-S, Gao XZ, Liu Y, Zhang H (2015) A novel bat algorithm with habitat selection and
Doppler effect in echoes for optimization. Expert Syst Appl 42(Issues 17–18):6350–6364
12. Fister I Jr, Fister D, Fister I (2013) A comprehensive review of cuckoo search: variants and
hybrids. Int J Math Model Numer Optim 4(4):2013
13. Yang X-S, Deb S (2009) Cuckoo search via Levy flights. In: 2009 world congress on nature &
biologically inspired computing (NaBIC 2009)
14. Civicioglu P, Besdok E (2011) ‘A conceptual comparison of the Cuckoo-search, particle swarm
optimization’, differential evolution and artificial bee colony algorithms. Springer Science +
Business Media B.V.
15. Yang X-S, Deb S (2010) Engineering optimization by Cuckoo search. Int J Math Model Numer
Optim 1(4):330–343
16. Potnurua D, Mary KA, Babuc ChS (2019) Experimental implementation of flower pollination
algorithm for speed controller of a BLDC motor. Ain Shams Eng J 10(Issue 2):287–295
Implementation of Complete Vector
Control for DFIG Based Wind Turbine
Rajesh Kumar, Ch. Sekhar, Saibal Manna, and A. K. Akella
1 Introduction
The need of wind turbines is nothing less than the demand of solar energy. With recent
developments and technologies, the off shore wind turbines are getting more interest
because of the percentage of availability of wind in offshore is more when compared
to onshore [1]. The challenging we can see in the off shore wind energy conversion
system are cost of installation, corrosion, erosion, frequent lightning strikes, and
others [2, 3]. By considering the measures of design or manufactures, the effect of
corrosion or lightnings can be minimized but the cost will be increased further. After
the installation the maintenance and transmission costs can be minimized by the
introduction of power electronic converters in the wind energy conversion. In 1990,
16 countries cumulatively generated electric power from wind around 3.58 billion
kWh and in 2017, electric power of 1.2 trillion kWh was generated from wind in a
whole by 129 countries. By recent report by power technology.com as on March 20,
2019, India ranks fourth position in the record of top countries from world in the area
of wind power capacity and India ranks second in Asian region with a total capacity
of 34.6 GW. In recent generation we can observe the preference of variable wind
speed turbines over the fixed speed wind turbines due to higher inertia of rotating
structure in the variable wind turbines and that helps to smoothen the output power
fluctuations [3, 4]. Furthermore, it also helps to reduce the mechanical stresses over
the drive train. The presence of power electronic converter systems is usually placed
between the generator system and the utility grid system. Power electronic converters
can be defined as the brain system for the wind turbine system because the control of
the system is carried by the designed power electronic converters in the system [5,
6]. It enables variant speed operation and sustained behavior of the turbine. For high
voltage transmission systems where large ocean-going wind farms are linked to each
other, there is a high need for FACTs devices like SVC, STATCOM for ensuring better
R. Kumar (B) · Ch. Sekhar · S. Manna · A. K. Akella

NIT Jamshedpur, Jamshedpur 831014, India
222 R. Kumar et al.
reliability and better power quality. For more dynamic compensation in modern large
turbines, thyristor switched capacitor (TSC) is used [7]. Not only for higher energy
yield but also for remote controllability and adjustability of power quality, power elec-
tronics are used widely. This availability feature of remote controllability provides
feasibility for offshore applications and the adjustability feature also provides fine
tuning for superior grid connections. The role of power electronic converters are
to match the wind turbine features with the grid featured requirements [6, 7]. The
characteristics are such as frequency, reactive power, voltage, power factor, active
power, and harmonics, etc. the wind turbine design system includes the components
like turbine rotor, generator, gearbox, power electronic system along with a trans-
former for connecting with the utility grid. In brief the rotor blades of the turbine
get rotated by the flow of wind and convert the effect into mechanical power, the
generator convert the mechanical power to electrical power, the gearbox maintains
the synchronism of turbine speed with the generator speed by some internal mech-
anism, the power electronic system takes care the process of power fed to the grid
from the wind power system, and the transformer along with circuit breakers ensures
the connection and power flow from wind system to grid [8]. In the context of wind
turbine design, power electronic devices play a crucial role in integrating the vari-
able speed (or fixed speed) wind power generation units with the main utility grid
for continuous power supply with better reliability and efficient performance.
The knowledge of simulating the mathematical models of wind energy, converter
systems, and their implementation with grid is well learned and designed in MATLAB
software tool. Simulation of controller related to both rotor side and grid side is carried
by the PI controllers and transformation blocks [9]. We correlate the theoretical
results with the obtained results from the designed model and analyze the practical
considerations to be taken care in the context of wind turbine design. A schematic
diagram of designed model is drawn in Fig. 1.
Fig. 1 Schematic diagram of DFIG linked wind turbine system implemented to grid
Implementation of Complete Vector Control for DFIG … 223
2 Wind Turbine
In general wind turbine consists of a group of blades (basically 3 blades) attached

to a hub. The hub is placed over a tower with a nacelle in which a gear box and a
generator is fixed as shown in Fig. 2. The wind movement on the blades produces
a force on the surface of blades and generate a rotational energy which is a form of
kinetic energy and is then transferred to the generator via gear box in the form of
mechanical energy and after that generator transforms the mechanical energy input
into an output form of electrical energy [10]. The gear box function is adjusting the
rotation of wind turbine with the rotation of the shaft of the generator. The general
equation for the wind power is expressed in Eq. 1.
1
P= ρ AC p ν 3 (1)
2
where P is the power in the wing when the blowing direction of wind is perpendicular,
ρ is the density of air flow, ν is the speed of the wind flow, and C p is the coefficient of
power extracted whose maximum value is 0.597 known as Betz limit. There are two
kinds of wind turbine designs, and they are vertical axis wind turbine and horizontal
axis wind turbine. In vertical axis wind turbine, the blades are connected in vertical
manner so that the wind will be facing the surface of the blades. In horizontal axis
Fig. 2 Block representation of components of wind turbine

224 R. Kumar et al.
wind turbine, the blades are connected in horizontal manner such that the wind falls
perpendicularly to the surface of the blade. The construction related to vertical axis
type is simpler and less expensive compared to horizontal axis type. The horizontal
type demands for a tower to mount the turbine in higher altitudes [10, 11]. Large wind
turbines are manufactured for rotation with a reference to the speed of wind. The
range of speed of wind is limited within the range of cut-in speed (lowest wind speed
required for given power generation in the name plate of the generator) and cut-off
speed (highest wind speed at which power flow from the generator stops transferring
to utility grid) [11].
3 Doubly Fed Induction Machine
Depending on the kind of energy form given to the machine, the machine is catego-
rized as generator and motor. If mechanical energy is given as input to the machine
and electrical energy is obtained as output energy then the machine can be named as
generator and if electrical energy is fed as input and mechanical energy is obtained as
output then the machine can be denoted as motor [12]. Depending on the type of elec-
trical supply, there are two types of generators viz. DC generator and AC generator.
With respect to the output speed, the AC generators are characterized into two types
which are synchronous generators and asynchronous generators. The asynchronous
generator is again categorized as squirrel cage induction generator (SIG) and wound
rotor induction generator (WIG). The WIG which converts the variable mechanical
energy to electrical energy with fixed frequency is termed as doubly fed induction
generator (DFIG). A DFIG is an induction generator with three phases where the
windings of stator and that of rotor are being fed with AC three phase signals [12,
13]. The construction of DFIG includes the multiphase windings and multi-phase
ring assembly for transferring power to the rotor. In DFIG rotor speed is controlled
by the means of back to back controllers which are named as rotor side controller
and grid side controller [13]. For electrical power generation using wind turbine, the
DFIG and permanent magnet synchronous generator (PMSG) are preferred mostly.
In this paper wind turbine linked with DFIG is simulated and the results are obtained.
The reasons for selecting DFIG are stated as follows:
a. Irrespective of variable wind speed, the supply of output signal with constant
frequency to the utility grid.
b. The rating of required power electronic devices in the system is of low power
rating and hence the cost of control system is low.
c. Almost unity power factor is maintained or controlled.
d. Even at low speed of wind, electrical power generation is possible.
e. The handle capacity of total load is partial i.e. 20–30% by the power electronic
converter [14].
4 Vector Control for Implemented DFIM
The aim of the control loop is to make the system respond to the requirement of
electrical energy and for that role sequential PWM converters are connected for
providing excitation to the rotor of the DFIG [15]. Many control methods exist as
current control methods from rotor side of DFIG but here vector control is considered
because of its extended version in large areas and appropriately the most stable one.
The space vector representation of DFIG is shown in Fig. 3 with respect to different
frames such as dq frame with reference to synchronous speed ωs , DQ with respect
to mechanical speed ωm and αβ with reference to plane [16].
The vector control for DFIG is simulated by taking the reference of the control
loops and is drawn for the rotor side and stator side controller for DFIG, and the
generated pulse signal is given as input to the universal bridges simulated for inter-
linking DFIG with the utility grid or main voltage source. The active output power
and reactive output power from the grid or source connected can be controlled up to
some extent by the voltage of the rotor controlled by the rotor side converter [16]. For
extraction of maximum output energy in wind turbine design system with variable
speed, the maximum power versus rotor speed plot is used widely. Two loops, viz.
control loop of rotor side and control loop of stator side (or referred as grid side),
are used to control the flow of active power and reactive power and to regulate the
shared voltage which is in DC type among the two converters. The vector control
for the DFIG is framed in a synchronously rotating dq plane reference to stator flux
space vector, where d represents direct axis and q represents quadrature axis. With
reference to the alignment, the direct axis component of current flowing in rotor
components (Idr ) is seen as proportional to the reactive power of stator (Qs ) and
the quadrature axis component of current in rotor (Iqr ), is seen as proportional to
the active power obtained from stator. Necessary transformation matrices are simu-
lated in the SIMULINK diagram and the angle is estimated for the reference plane
transformation. The required (dq to abc) transformation matrices and (abc to dq)
transformation matrices are expressed in the following equations.
Fig. 3 Space vector representation of DFIG with reference to different frames

226 R. Kumar et al.

Vα cos θ − sin θ Vd
= (2)
Vβ sin θ cos θ Vq
⎡ ⎤
⎡ ⎤ 1 0
Va √
⎢ ⎥
⎣ Vb ⎦ = ⎢ −1 2 3 2 ⎥ Vα (3)
⎣ √ ⎦ Vβ
Vc −1 2 − 3 2
⎡ ⎤
V
Vα 2 1 −1 2 −1 2 ⎣ a ⎦
= √ √ Vb (4)
Vβ 3 0 3 2− 3 2
Vc
T
Vα cos θ − sin θ Vα
= (5)
Vβ sin θ cos θ Vβ
where V d and V q are the dq frame space vectors of vector V, V α, and V β are the αβ
references of vector V and V a , V b , V c are phasor voltage representations of vector V.
If the DFIG used in the system provides different turns ratio at the side of rotor and
stator, then the effect must be considered at the control loop. Therefore, the stator
side referred rotor currents are taken as input for the control loop. A complete vector
control for DFIG is drawn in Fig. 4.
Fig. 4 Block diagram of complete vector control loop for DFIG

5 MATLAB Simulation
The MATLAB Simulink diagram for the block diagram shown in Fig. 1 is modeled
in Fig. 5. The subsystems of controller on rotor side and on stator side are shown
in Figs. 6 and 7 respectively to emphasis the input given to the controllers and the
output obtained from them. The parameter specifications of the DFIG and the grid
along with transmission lines are specified in Table 1.
Fig. 5 MATLAB block diagram of modeled wind turbine system
Fig. 6 Rotor side controller for the control of DFIG linked to wind turbine system
228 R. Kumar et al.
Fig. 7 Stator side controller for the control of DFIG linked to wind turbine system
Table 1 Parameter
DFIG parameters referred to stator side and grid parameters
specifications
Nominal power 2 MW
Supply frequency 50 Hz
Speed of synchronous 1500 rpm
Vs line to line 690 V
Stator current 1760 A
No of poles 2
Rstator 2.59 Mω
Lstator 2.577 Mh
Rrotor 2.9 Mω
Lrotor 2.587 mH
Maximum slip 0.33
Turns ratio 1/3
Electromagnetic torque 12,732 Nm
Switching frequency 4000 Hz
√
Bus voltage 230 2
Grid resistance per meter 20 μ
Grid inductance per meter 400 μH
Bus capacitance 80 mF
Gear box ratio 100
The Simulink model is examined for the runtime of 0.4 s or 20 cycles by applying
different wind speeds for different intervals of time. The results are auto scaled from
0 to 0.4 s for better understandings of results and wind speed available in this interval
is 10 m/s (Figs. 8, 9, 10 and 11).
The necessary output results are presented in this paper to show the function of
the vector control loop implemented to the DFIG which is linked to wind turbine
system. The stator current is not similar as voltage source due to the given values of
Fig. 8 Voltage source versus time for wind flow of 10 m/s
Fig. 9 Istator of DFIG versus time
Fig. 10 Rotor current of DFIG versus time

230 R. Kumar et al.
Fig. 11 D component reference of voltage versus time
speed reference, d-axis reference and q-axis reference. The control loop inputs are
the Qs available from the system and the speed of the machine implemented. We can
observe the symmetrical behavior in the system and the control loop takes care of the
changes in the measured values and input values. The rotor current in Fig. 10 and bus
voltage in Fig. 12 will be found to be satisfactory if compared with the ideal cases of
the system. The speed reference and speed waveform of the DFIG is simulated for
4 s as for the first 2 s, the speed of DFIG tends towards the reference value (Fig. 13).
Fig. 12 Voltage of the bus system and reference bus voltage versus time
Fig. 13 Speed of DFIG versus time

7 Conclusion
In this paper we can conclude that the vector control of DFIG gives appreciated
results when it is implemented with the wind turbine system and grid system. We
attempted to bring minor changes in the system but still the system acquires its
stability within a small interval of time. To the existing vector control concept, we
attempted some changes such as taking the reference value of q-axis as zero and
changing the speed reference with respect to percentage output. It helps to reduce
the power loss in the system when implemented practically. A lot of future scope is
available for the system discussed here because we have not implemented the pitch
control of the wind turbine system and assumed it as constant. Even with low speed
we can expect for stable power supply which is understood from the results. Further
hybrid system can also be designed by integrating a battery system or solar system
with the above system, then it will create a stable microgrid with continuous supply
of power.
References
1. Djeghloud H, Bentounsi A, Benalla H (2010) Simulation of a DFIG-based wind turbine with

active filtering function using MATLAB/Simulink. In: The XIX international conference on
electrical machines - ICEM 2010, Rome, Italy, pp 1–7. https://doi.org/10.1109/ICELMACH.
2010.5607885.
2. Puchalapalli S, Singh B (2020) A novel control scheme for wind turbine driven DFIG interfaced
to utility grid. IEEE Trans Indus Appl 56(3):2925–2937. https://doi.org/10.1109/TIA.2020.296
9400
3. Tang W, Hu J, Chang Y, Liu F (2018) Modeling of DFIG-based wind turbine for power system
transient response analysis in rotor speed control timescale. IEEE Trans Power Syst 33(6):6795–
6805. https://doi.org/10.1109/TPWRS.2018.2827402
4. Ouyang J, Tang T, Yao J, Li M (2019) Active voltage control for DFIG-based wind farm inte-
grated power system by coordinating active and reactive powers under wind speed variations.
IEEE Trans Energy Convers 34(3):1504–1511. https://doi.org/10.1109/TEC.2019.2905673
5. Ekanayake J, Jenkins N (2004) Comparison of the response of doubly fed and fixed-speed
induction generator wind turbines to changes in network frequency. IEEE Trans Energy Convers
19(4):800–802. https://doi.org/10.1109/TEC.2004.827712
6. Arani MFM, Mohamed YAI (2015) Analysis and impacts of implementing droop control
in DFIG-based wind turbines on microgrid/weak-grid stability. IEEE Trans Power Syst
30(1):385–396. https://doi.org/10.1109/TPWRS.2014.2321287
7. Li Y, Xu Z, Meng K (2017) Optimal power sharing control of wind turbines. IEEE Trans Power
Syst 32(1):824–825. https://doi.org/10.1109/TPWRS.2016.2549741
8. Li S, Haskew TA, Williams KA, Swatloski RP (2012) Control of DFIG wind turbine with
direct-current vector control configuration. IEEE Trans Sustain Energy 3(1):1–11. https://doi.
org/10.1109/TSTE.2011.2167001
9. Qiao W, Zhou W, Aller JM, Harley RG (2008) Wind speed estimation based sensor less
output maximization control for a wind turbine driving a DFIG. IEEE Trans Power Electron
10. Mohammadi J, VaezZadeh S, Afsharnia S, Daryabeigi E (2014) A combined vector and direct
power control for DFIG-based wind turbines. IEEE Trans Sustain Energy 5(3):767–775. https://
doi.org/10.1109/TSTE.2014.2301675
232 R. Kumar et al.
11. Ghosh S, Kamalasadan S, Senroy N, Enslin J (2016) Doubly fed induction generator (DFIG)-
based wind farm control framework for primary frequency and inertial response application.
IEEE Trans Power Syst 31(3):1861–1871. https://doi.org/10.1109/TPWRS.2015.2438861
12. Saruwatari M et al (2016) Design study of 15-MW fully superconducting generators for offshore
wind turbine. IEEE Trans Appl Supercond 26(4):1–5, Art no. 5206805. https://doi.org/10.1109/
TASC.2016.2535315
13. Nguyen N, Almasabi S, Mitra J (2018) Impact of correlation between wind speed and turbine
availability on wind farm reliability. In: 2018 IEEE industry applications society annual meeting
(IAS). Portland, OR, USA, pp 1–7. https://doi.org/10.1109/IAS.2018.8544643
14. Fischer K (2019) Reliability of power converters in wind turbines: exploratory analysis of failure
and operating data from a worldwide turbine fleet. IEEE Trans Power Electron 34(7):6332–6344
15. Abad G, López J, Rodríguez M, Marroyo L, Iwanski G (2011) Vector control strategies for
grid-connected DFIM wind turbines. In: Doubly fed induction machine: modeling and control
for wind energy generation applications. IEEE, pp 303–361. https://doi.org/10.1002/978111
8104965.ch7
16. Ma L, Zheng Y, Ma H (2011) Research and simulation of double-fed wind power gener-
ation rotor side control technology. In: 2011 International conference on electrical and
control engineering. Yichang, China, pp 2472–2475. https://doi.org/10.1109/ICECENG.2011.
6058038
TLO Based OPF with FACTS Devices
for DC Link Placement Problem
B. Suresh Babu
Nomenclature
Ai Bi Ci Fuel cost coefficients

Di E i Coefficients of valve point effects of the generator
TFC Total fuel cost
TRPL Total real power loss
NVSI Net voltage stability index
LVSI Largest voltage stability index
FACTS Flexible AC transmission systems
T ype f Integer number in the range of (1–3) denoting the type of f th FACTS
device, 1 represents SVC, 2 denote TCSC and 3 indicates UPFC
1 Introduction
The optimal power flow (OPF) has been widely used in power system operation and
planning since its introduction by Carpentier [1]. The existence of nonlinear power
flow constraints and the DC link equations make the problem non-convex even in
the absence of discrete control variables [2–5]. Over the years, numerous mathe-
matical programming techniques such as linear programming, nonlinear program-
ming [6–8]. The difficulties in solving OPF can be overcome by modern stochastic
algorithms such as particle swarm optimization (PSO), evolutionary programming
(EP), harmony search optimization (HSO) [9–14], Recently, teaching learning based
optimization (TLO) has been suggested for solving optimization problems [15].
B. S. Babu (B)
Electrical and Electronics Engineering, Shri Vishnu Engineering College for Women, West
Godavari District, Vishnupur, Bhimavaram, Andhra Pradesh 534202, India
234 B. S. Babu
The recent progress in power electronics have introduced Flexible AC Transmission

Systems (FACTS) [16]. TLO with a view of obtaining the global best solution and
demonstrates its performance through simulation results on IEEE 14 and 30 bus test
systems.
2 Problem Statement with FACTS Devices
Three FACTS devices, SVC, TCSC and UPFC, each with different characteristics
are chosen for placement in order to control reactive power injections and active
power flows in the proposed strategy.
Repair the i-th student, replace the transmission lines by DC links, place FACTS
devices and set the control parameters according to i-th student grade points and Run
AC/DC load flow.
Test 1: TFC
Minimize

1 (x, u) = A j P jG2 + B j P jG + C j + D j sin E j P jG (min) − P jG (1)
j∈
FACTS device constraints − 0.8 ≤ ηk ≤ 0.2 for TCSC and UPFC (2)
−100 M V A R ≤ Q iF ≤ +100 M V A R for SVC and UPFC (3)
Test 2: TRPL
Test 3: Enhancement of VS
Test 4: TFC + TRPL
Test 5: TFC + VS
Test 6: TRPL + VS
Test 7: TFC + TRPL + VS
Evaluate the AOF and performance function, select the best student with highest
Fi in the population as the optimal solution
1
Maximize f = (4)
1 + AO F
2.1 Representation of Decision Variables
The control variable in vector representation in TLO method as follows

TLO Based OPF with FACTS Devices for DC Link Placement Problem 235
Table 1 TLBO parameter

Parameter Value
No. of students 30
Max. iteration 300
Table 2 Number of devices

Test system 14 bus 30 bus
to be placed
Number of DC links 1 2
Number of FACTS devices 2 3

S = T ype f , Loc f , η f , Q Ff , PkG , V jG , Tv , L dc
p , Ip
dc
(5)
The TLO is tested on IEEE 14 and 30 bus test systems, and the data have been taken
from Ref. [17] (Tables 1, 2, 3, 4, 5 and 6).
It is observed from Tables 3 and 5 that TLO in this Test 1, Test 2 and Test 3 is the
reduce of the TFC,TRPL and LVSI for IEEE 14 and 30 bus test system. It is seen
from the results of Test 4, Test 5, Test 6 and Test 7 that the TLO and as well as the
PSO and HSO offer a compromised solution. The % savings of TFC, TRPL and VS
Enhancement for IEEE 14 and 30 bus test systems in Figs. 1 and 2.
4 Conclusion and Future Research
In this paper TLBO solution technique for OPFDC with FACTS placement are devel-
oped and tested on three IEEE14 and 30 bus test systems. The algorithms use sequen-
tial AC/DC load flow involving NR technique for computing the objective function
during search and are able to offer the global best solution. It can also be observed
that the proposed methods perform better with FACTS devices. The scope for future
research the other FACTS devices such as SSSC and IPFC may be considered in the
formulation of the problem.
236 B. S. Babu
Table 3 Results of IEEE 14 for Test 1–Test 7

Test Base Test IEEE 14
TLO PSO HSO
1 TFC 834.6716 814.8536 817.5147 815.7954
TRPL 8.9737 6.0321 7.0616 6.4411
NVSI 0.3724 0.3388 0.3265 0.3553
LVSI 0.0750 0.0718 0.0673 0.0727
2 TFC 834.6716 1020.7136 1021.2605 1015.0630
TRPL 8.9737 2.0437 2.3207 2.2245
NVSI 0.3724 0.3158 0.3451 0.3875
LVSI 0.0750 0.0600 0.0720 0.0780
3 TFC 834.6716 860.6775 874.4455 838.4990
TRPL 8.9737 8.5319 7.0419 8.6873
NVSI 0.3724 0.2157 0.2168 0.2177
LVSI 0.0750 0.0326 0.0379 0.0364
4 TFC 900.4238 954.0228 963.7662
TRPL 3.4375 2.7358 2.7435
NVSI 0.3305 0.3595 0.3492
LVSI 0.0674 0.0721 0.0738
5 TFC 842.5334 857.5641 834.3707
TRPL 8.7328 9.0501 10.1520
NVSI 0.2460 0.2827 0.3046
LVSI 0.0407 0.0542 0.0503
6 TFC 1003.7122 1001.2078 1005.2532
TRPL 2.3690 2.8254 2.4856
NVSI 0.2882 0.2757 0.3207
LVSI 0.0529 0.0593 0.0676
7 TFC 926.3375 943.0159 920.3096
TRPL 3.1284 3.3853 3.1100
NVSI 0.3129 0.2607 0.2994
LVSI 0.0570 0.0564 0.0665
Table 4 Solution of TLO for IEEE 14
Base Test Test 1 Test 2 Test 3 Test 6 Test 4 Test 5 Test 7
TFC TRPL VS TRPL + VS TFC + TRPL TFC + VS TFC + TRPL + VS
PG 188.974 208.901758 66.103032 162.465763 77.985842 127.562486 185.767811 112.111149
35.000 20.000000 80.000000 39.293198 80.000000 55.003782 26.093413 52.387035
20.000 16.130359 50.000000 18.018575 50.000000 36.083444 18.317468 40.330989
12.000 10.000000 35.000000 28.489782 23.824707 13.787810 18.240963 28.616944
12.000 10.000000 29.940624 19.264631 29.558413 30.000000 19.313102 28.682266
VG 1.060 1.091626 1.046741 1.044781 1.024141 1.038874 1.051549 1.016353
1.045 1.071827 1.038227 0.992240 1.008369 1.036558 1.040252 1.013070
1.010 1.047213 1.022883 0.950728 0.999669 1.019790 1.028669 0.992228
1.070 1.010819 1.006294 1.007576 1.038515 1.023449 1.067179 0.989859
1.090 1.013973 1.033028 1.018693 1.064612 1.044612 1.079894 1.037511
T 0.978 1.036260 1.035949 1.001334 0.973753 0.989459 0.977035 1.036637
0.969 1.089908 1.048395 1.046936 0.975872 0.981519 1.007912 1.006847
0.932 0.919960 1.022987 0.900000 0.977289 1.068977 0.984713 0.956953
L dc
p – 4 3 3 3 4 11 4
I pdc – 0.995070 0.525269 1.000000 0.613878 0.798408 0.334221 0.666732
Facts type 3 2 2 2 2 1 2
2 2 2 2 2 2 2
TLO Based OPF with FACTS Devices for DC Link Placement Problem
Facts location 7 17 11 12 11 12 11
12 11 17 17 12 17 17
VAR support 60.329 – – – – 55.346 –
– – – – – – –
Line com fact −0.138 −0.747 −0.800 −0.624 −0.260 – −0.712
0.176 −0.507 −0.797 −0.690 −0.468 −0.599 −0.638
237
238 B. S. Babu
Table 5 Results of IEEE 30 for Test 1–Test 7

Test Base test IEEE 30
TLO PSO HSO
1 TFC 813.6941 789.0199 793.9881 791.9971
TRPL 7.0990 5.7185 5.9928 6.7249
NVSI 1.6705 1.7005 1.5685 1.7032
LVSI 0.1336 0.1373 0.1295 0.1310
2 TFC 813.6941 964.8873 964.9062 964.9016
TRPL 7.0990 1.9381 1.9428 1.9412
NVSI 1.6705 1.4524 1.4538 1.4535
LVSI 0.1336 0.1068 0.1070 0.1069
3 TFC 813.6941 885.2828 863.1441 855.1094
TRPL 7.0990 11.5340 8.7832 8.8625
NVSI 1.6705 0.8189 0.9406 0.9125
LVSI 0.1336 0.0499 0.0625 0.0621
4 TFC 937.3572 944.7096 894.2697
TRPL 2.0710 2.0401 2.3679
NVSI 1.4344 1.4606 1.5078
LVSI 0.1049 0.1069 0.1226
5 TFC 794.7903 800.8039 806.0572
TRPL 7.6315 5.9798 9.8711
NVSI 1.2688 1.4024 1.2666
LVSI 0.0748 0.0878 0.0761
6 TFC 964.8502 964.8313 964.8999
TRPL 1.9419 1.9437 1.9424
NVSI 1.4515 1.4533 1.4523
LVSI 0.1065 0.1066 0.1068
7 TFC 922.3340 924.3042 894.9846
TRPL 2.1289 2.1326 2.3328
NVSI 1.4062 1.4180 1.4935
LVSI 0.1007 0.1025 0.1215
Table 6 Solution of TLO for IEEE 30
TFC TRPL VS TRPL + VS TFC + TRPL TFC + VS TFC + TRPL +
VS
PG 138.539 176.158866 50.340844 163.245678 50.367124 58.973346 180.653083 63.245366
57.560 48.197453 79.999982 20.163249 80.000000 77.643720 48.486956 76.867933
24.560 19.652961 49.999677 48.437909 50.000000 46.348710 19.479194 43.419597
35.000 20.299167 34.999930 26.593007 35.000000 34.944779 17.725836 34.997792
17.930 12.810086 30.000000 10.599845 29.974730 29.561586 12.005007 30.000000
16.910 12.000000 39.997704 25.894290 40.000000 37.998846 12.681471 36.998175
VG 1.050 1.038428 1.045526 1.054416 1.044582 1.047897 1.078516 1.049988
1.0338 1.019645 1.039370 1.041288 1.038140 1.040368 1.061678 1.041765
1.0058 0.999386 1.034382 1.013316 1.034043 1.034132 1.024740 1.034853
1.0230 1.003538 1.037476 1.053726 1.036359 1.039421 1.035812 1.039318
1.0913 1.043440 1.067559 1.054044 1.065253 1.068494 1.045030 1.071041
1.0883 1.047130 1.090400 1.081825 1.089889 1.091342 1.083279 1.092072
T 1.0155 1.082984 0.972115 1.050706 0.970561 0.968194 0.964225 0.963403
0.9629 0.947624 1.001305 0.997420 1.002034 0.999104 0.924463 0.994426
1.0129 1.034320 1.011821 0.900044 1.013126 1.010996 0.935517 1.002060
0.9581 0.965501 0.964632 1.020328 0.965851 0.964933 0.933096 0.962257
L dc
p – 5 6 40 6 6 11 6
2 5 3 5 5 5 5
TLO Based OPF with FACTS Devices for DC Link Placement Problem
I pdc – 0.681020 0.287488 0.569400 0.291466 0.306704 0.212152 0.296228

0.662481 0.401750 0.998841 0.407965 0.415457 0.609589 0.449165
Facts type 2 2 2 2 2 2 2
2 2 3 2 2 2 2
2 2 3 2 2 2 2
(continued)
239
Table 6 (continued)
240

TFC TRPL VS TRPL + VS TFC + TRPL TFC + VS TFC + TRPL +
VS
Facts 21 36 22 36 36 36 36
Location 39 25 36 25 25 38 25
28 31 12 31 31 25 31
VAR – – – – – – –
support – – 60.722 – – – –
– – 84.990 – – – –
Line com −0.100 −0.384 −0.788 −0.396 −0.404 −0.799 −0.467
fact 0.169 −0.312 −0.409 −0.309 −0.293 −0.580 −0.328
0.078 −0.381 −0.579 −0.368 −0.420 −0.754 −0.411
B. S. Babu
2.4 2.374
2.3 2.262
% FC Savings
2.2
2.1 2.056
2
1.9
1.8
TLO HSO PSO
IEEE 14
78 77.226
77
% RPL Savings
76 75.211
75
74.139
74
73
72
TLO HSO PSO
IEEE 14
58 56.533
56
% VS Enhancement
54
52 51.467
50 49.467
48
46
44
TLO HSO PSO
IEEE 14
Fig. 1 Percentage of savings TFC, TRPL and VS enhancement for IEEE 14

242 B. S. Babu
3.5
3.032
3 2.66
2.422
% FC Savings
2.5
2
1.5
1
0.5
0
TLO HSO PSO
IEEE 30
72.72
72.699
72.7
% RPL Savings
72.68
72.655
72.66
72.64 72.633
72.62
72.6
TLO HSO PSO
IEEE 30
65 62.65
% VS Enhancement
60
55 53.518 53.219
50
45
TLO HSO PSO
IEEE 30
Fig. 2 Percentage of savings TFC, TRPL and VS enhancement for IEEE 30
Acknowledgements The author gratefully acknowledge the authorities of Department of Electrical

and Electronics Engineering, Shri Vishnu Engineering College for Women for their continued
support, encouragement and the facilities provided to carry out this work.
References
1. Carpentier J (1962) Contribution a l’Etude du Dispatching Economique’. Bulletin de la Societe

Francaise des Electriciens 3:431–474

2. Momoh JA, El-Hawary ME, Adapa R (1999) A review of selected optimal power flow literature
to 1993 Part I: nonlinear and quadratic programming approaches. IEEE Trans Power Syst
14:96–104
3. Momoh JA, El-Hawary ME, Adapa R (1999) A review of selected optimal power flow literature
to 1993 Part II: Newton, linear programming and interior point methods. IEEE Trans Power
Syst 14:105–111
4. Nagsarkar TK, Sukhija MS (2007) Power system analysis. Oxford University Press
5. Arrillaga J, Watson NR (2013) Computer modelling of Electrical power systems. Wiley
6. Arrillaga J, Arnold C, Harker BJ (1983) Computer modelling of Electrical power systems.
Wiley
7. Lee KY, Park YM, Oritz JL (1984) Fuel-cost minimization for both real and reactive power
dispatches. IEE Proc C 131(3):85–93
8. Alsac O, Scott B (1974) Optimal load flow with steady state security. IEEE Trans Power
Apparatus Syst PAS 745–751
9. Abido MA (2002) Optimal power flow using particle swarm optimization’. Proc Int J Electr
Power Energy Syst 24(7):563–571
10. Abido MA, Al-Ali NA (2009) Multi-objective differential evolution for optimal power flow.
In: International conference on power engineering, energy and electrical drives, POWERENG
2009. Lisbon, Portugal, pp 101–106
11. Abou El Ela AA, Abido MA (2010) Optimal power flow using differential evolution algorithm.
Electr Power Syst Res 80(7):878–885
12. Naresh babu AV, Ramana T, Sivanagaraju S (2014) Analysis of optimal power flow problem
based on two stage initialization algorithm. Int J Electr Power Energy Syst 55:91–99
13. Niknam T, Narimani MR, Jabbari M, Malekpour AR (2011) A modified shuffle frog leaping
algorithm for multi-objective optimal power flow. Energy 36:6420–6432
14. Sivasubramani S, Swarup KS (2011) Multi-objective harmony search algorithm for optimal
power flow problem. Int J Electr Power Energy Syst 33(3):745–752
15. Rao RV, Savsani VJ, Vakharia DP (2012) Teaching-learning-based optimization: a novel
optimization method for continuous non-linear large scale problems. Inf Sci 183(1):1–15
16. Hingorani NG, Gyugy L (2001) Understanding FACTS concepts and technology of flexible
AC transmission systems. Standard Publishers Distributors, New Delhi
17. Test Archive, Systems. http://www.ee.washington.edu/research/pstca/. Accessed Dec 2012
Performance Comparison Analysis
of Energy Management Strategies
for Hybrid Electric Vehicles
Jai Kumar Maherchandani, R. R. Joshi, Ritesh Tirole, Raju Kumar Swami,

and Bibhu Prasad Ganthia
1 Introduction
Increasing pollution and the fast depletion of fossil fuels are the two major problems
of today’s world. Convention fuel-based transportation sector is one of the main
contributors to the pollution problem. The transport sector in the U.S. contributes
about 28% of the total greenhouse gas (CHG) emission and continuously increases
with a rapid growth rate [1]. In India, this sector is at third place in CHG emission
and growing exponentially. Approximately 70% of the total diesel and 95% of the
petrol consumption are caused by this sector in India [2].
Worldwide, continuous efforts are being made to shift from conventional fuel-
based vehicles to environmentally friendly electric vehicles (EV) or HEV. EV
consisting of a battery as a single source of power has a limited range with a single
charge. Combining two or more sources in HEV has many advantages, such as
increased range, reliability, and lifetime, in comparison to EV using a single source
of energy. HEV may consist of a combination of different types of energy sources
such as a battery, internal combustion engine (ICE), fuel cell (FC), and supercapacitor
(SC), in a hybrid manner.
Because of its various advantages, such as high performance, low operating
temperature, and almost zero-emission, the proton exchange membrane fuel cell
(PEMFC) is now commonly used as a primary energy source in electric vehicle
J. K. Maherchandani (B) · R. R. Joshi

College of Technology and Engineering, MPUAT, Udaipur, India
R. Tirole
Sir Padampat Singhania University, Udaipur, India
e-mail: ritesh.tirole@spsu.ac.in
R. K. Swami
Pacific University, Udaipur, Rajasthan, India
B. P. Ganthia
IGIT, Sarang, Dhenkanal, Odisha, India
246 J. K. Maherchandani et al.
(EV) applications [3, 4]. However, PEMFC’s sluggish dynamic is a major down-
side that prevents it from meeting fast-changing load demand. It is possible to solve
this problem by using additional energy sources with various characteristics. The
high dynamic efficiency of HEV is ensured by a combination of high energy density
sources such as battery and high-power density sources such as supercapacitor with
PEMFC [5–8].
A proper energy management strategy (EMS) is always needed to exploit the full
benefit of the hybridization of an HEV. Designing an effective EMS for an HEV
consisting of different energy sources is a difficult challenge [9, 10]. The main goal
of EMS is to break the instantaneous load demand among source components in
such a way that one or more performance objectives can be met, such as minimizing
fuel consumption, maximizing overall efficiency, putting the least amount of stress
on individual source components, reducing pollution, and achieving optimal vehicle
performance. Various methods for achieving optimum efficiency have been proposed
in the literature [11–15].
EMS’s may be classified into two broad categories, i.e., rule-based EMS and
optimization-based EMS. RB EMS is computationally efficient, easy to implement,
and can be applied in real-time. Designing rule-based EMS depends on experience,
engineering intuition, and characteristics of power train [16]. These types of strate-
gies do not involve any optimization techniques directly during the implementation
of EMS. Optimization-based EMS’s may be designed to meet one or more objectives.
They may be further classified as real-time and off-time EMS. Equivalent consump-
tion minimization (ECMS), model predictive control (MPC), and intelligent control
system are the example of few real-time EMS.
The present work focuses on the performance comparison of three popular EMS,
i.e., conventional PI, ECMS, and RB. HEV used in this work consists of FC, battery,
and SC. The performance comparison of all three strategies is carried out in terms
of efficiency, fuel consumption, and battery SOC. All three strategies are simulated
for identical conditions.
2 System Architecture
The system architecture of HEV used in this work is shown in Fig. 1. FC acts as a
primary source and is connected to the DC bus via a unidirectional DC-DC converter.
The battery may supply or absorb the power with the help of a bidirectional dc-
dc converter, depending on the mode of operation. SC being a high-power density
source helps in meeting high dynamic load demand or absorb the excess power
instantaneously.
The energy management controller decides power-sharing by the respective
source. It may have one or more objectives: ensure the operation of the respec-
tive source at an optimal level, maintain a high SOC level of the battery, achieve high
overall efficiency, minimize the fuel consumption, etc.
Dynamic modelling of the proposed HEV system components is presented below.
Performance Comparison Analysis of Energy … 247
DC Bus
Fuel Cell Stack DC-DC Converter
DC-AC Inverter Electric

Vehicle
BidirecƟonal DC-
BaƩery
DC Converter
Supercapacitor
Pack
Energy Management Controller
Fig. 1 System architecture
Modelling of fuel cell

FC stack voltage (V f c) and Nernst voltage (VN V ) relationship can be represented as
[17, 18]:
V f c = n cs VN V − VI V L (1)
where n cs , VI V L are series-connected cell and irreversible voltage loss, respectively.

The output power of FC is given as
Pf c = V f c I f c (2)
where I f c is FC output current.

Modelling of Battery Storage System
The lithium-ion battery model comprises a voltage source VB and its internal series
resistance Ri considering both are the function of battery SOC can be given by [19]:

VB (S OC) − VB2 (S OC) − 4 · Ri (S OC) · PB
IB = (3)
2 · Ri (S OC)
where PB and I B are the power supplied or absorbed by the battery and battery
current, respectively.
Table 1 Rating of HEV

HEV component Rating
Components
FC power 11 kW
Battery power 4 kW
Allowable battery SOC level 55–90%
DC bus voltage V
Supercapacitor Modelling
Here, the Stern model combines both the “Gouy–Chapman and Helmholtz models”
[20, 21] is considered. The capacitance may be expressed as
−1
1 1
C SC = + (4)
C GC CH
where C GC , C H are “Gouy–Chapman” and “Helmholtz” capacitance (in farads),

respectively.
Hybridization of FC with SC is done to enhance the dynamics of such a system to
efficiently meet the fast-varying load demand put on the electric vehicle. Bidirectional
DC-DC battery converter regulates the DC bus voltage as well as SC SOC level. The
value of different HEV components has been shown in Table 1.
3 Energy Management Strategies
This section briefly discusses widely used classical PI, optimization-based ECMS,
and RB EMS for the HEV system shown in Fig. 1.
3.1 Classical PI Control Strategy
This type of strategy relies on PI controllers to control key performance parameters

like battery SOC, SC voltage, and dc-bus voltage. The distribution of load power is
done in such a way that FC provides the steady-state load demand. In this type of
strategy, the PI controller may regulate the battery power depending upon the battery
current SOC level. Fuel cell reference power is decided after subtracting battery share
from load demand. Battery SOC level plays the main role in such a type of strategy.
Depending upon the battery SOC level above or below the reference value, the battery
power share increases or decreases, which in turn varies the fuel cell power share in
meeting the load demand. Conventional PI EMS is easier to implement, and the PI
gains may be tuned online for a better response.
3.2 Optimization Based ECMS
ECMS is the most widely used real-time EMS. ECMS is a cost function optimization
strategy to ensure FC optimal operation with maximum fuel economy [22, 23]. The
power distribution is determined by minimizing an instantaneous cost function that
includes the fuel consumption of the FC and the equivalent fuel consumption of
the battery’s electrical energy. The method proposed in [24] is used in this study,
where the battery SOC is regulated by the battery energy penalty coefficient. The
optimization problem is defined as follows.
Find FC power P f c , equivalence factor α and battery power Pbatt which minimizes
equivalent fuel cost
F = [P f c + α Pbatt ] · T (5)
under the equality constraints
Pload = P f c + Pbatt (6)
Within the boundary conditions
P f c_ min ≤ P f c ≤ P f c_ max
Pbatt_ min ≤ Pbatt ≤ Pbatt_ max (8)
0 ≤∝≤ 100
where, P f c , Pbatt , Pload are the FC, battery, and load power, respectively. α is the
equivalence factor, μ is a constant, T is the sampling time. P f c_min , P f c_max ,
Pbatt_min , Pbatt_max are the FC minimum, FC maximum power, battery maximum
and battery minimum power, respectively. S OC min and S OC max are the minimum
and maximum battery SOCs respectively.
Only FC and battery power are considered in optimization problems as SC is
connected to the DC bus and recharged from the battery directly.
3.3 RB Energy Management Strategy
RB control strategies are computationally efficient and can be applied in real-time.

Such a strategy is designed considering the characteristics of different sources, expe-
rience, driving cycle requirements, etc. This type of strategy may be further classified
broadly as a deterministic or fuzzy rule-based strategy. Here, a simple RB strategy
consisting of eight rules is taken for comparison purposes [25]. The RB strategy
implemented is shown in Table 2.
Table 2 Rule-based energy management strategy

Rule Operating conditions Fuel cell power
I If SOC High and Pload < P f c_min P ∗f c = P f c_min
II If SOC high and Pload ∈ [P f c_min , P f c_max ] P ∗f c = Pload
III If SOC high and Pload ≥ P f c_max P ∗f c = P f c_max
IV If SOC normal and Pload < P f c_opt P ∗f c = P f c_opt
V If SOC normal and Pload ∈ [P f c_opt , P f c_max ] P ∗f c = Pload
VI If SOC normal and Pload ≥ P f c_max P ∗f c = P f c_max
VII If SOC low and Pload < P f c_max P ∗f c = Pload + Pchar
VIII If SOC low and Pload ≥ P f c_max P f c_max
Here SOC High means OC > S OC max , SOC Normal means SOC is between 65
and 85%, and SOC Low means S OC < S OC min .
The FC power is decided based on the battery SOC range and load power Pload . The
comparison among EMSs is made using MATLAB software. Results are presented
and discussed in the next section.
4 Results and Discussion
This section presents a comparative analysis of the above discussed conventional

PI, ECMS, and RB EMS. Performance comparison in overall efficiency, battery
SOC, and hydrogen consumption is presented for these strategies. The same initial
conditions are used for all three EMS’s. The comparative results are given below.
Figure 2 shows the variable load demand of HEV.
Fig. 2 Load profile

Table 3 Comparative energy contribution (Wh)

S. No. Source name PI strategy ECFM strategy RB strategy
1 Fuel cell output 420.89 484.64 439.61
2 Battery output 82.45 18.62 56.07
3 Super-capacitor output −6.93 −6.85 0.750
4.1 Comparative Energy Contribution by Individual Power

Source
Table 3 show the energy contributed by individual energy source in the simulation
periods under different control strategy.
It is clear that FC energy contribution is maximum and minimum in the ECMS and
PI control strategy, respectively, and the reverse is true for battery energy contribution.
Rule-based control strategy shares the power in a more optimum way between the
fuel cell and battery.
4.2 Performance Comparison Among RB, PI and ECMS

Strategy
Overall Efficiency
Figure 3 shows overall efficiency comparison among RB, PI, and ECMS control
strategies. It can be concluded that the RB control strategy (83.01%) provide better
average efficiency in comparison with the ECMS control strategy (79.46%) and PI
control strategy (82.11%).
Fig. 3 Efficiency comparison of PI, ECMS and RB strategy

Fig. 4 Battery SOC comparison of PI, ECMS and RB strategy
Table 4 Performance
PI strategy RB strategy ECMS strategy
comparison of proposed, PI,
and EFCM strategies Overall 82.11 83.01 79.46
efficiency (%)
State of charge 58.15–65.07 59.93–65.07 61.59–65.87
(SOC)
H2 consumption 12.82 14.43 16.74
(gm)
Battery SOC
Figure 4 shows the comparative use of battery storage for RB, PI, and ECMS control
strategies. It can be revealed that the ECMS strategy uses the lesser battery power
and maintains a slightly higher SOC level compared to RB and PI control strategy,
as shown in Table 4.
Hydrogen Consumption
Figure 5 gives a comparison of hydrogen consumption for RB, PI, and ECMS control
strategies. ECMS has higher average fuel consumption compared to RB and PI
control strategies. The fuel consumption during the simulation period is given in
Table 4.
5 Conclusion
Performance comparison of commonly used PI, ECMS, and RB energy management

schemes for a fuel-cell hybrid energy vehicle has been presented. The same initial
conditions are used for all the schemes. Comparison among EMS’s the PI EMS
causes the lowest hydrogen consumption, ECMS causes the highest battery SOC,
and the RB EMS provided slightly better efficiency. It may be concluded that the
Fig. 5 Fuel consumption comparison of PI, ECMS and RB strategy
EMS suitable for different applications is selected based on a particular criterion to

prioritize. Proper tuning of PI controller for PI EMS, more expertise in designing RB
EMS rules, and efficient optimization algorithms for ECMS play a very important
role.
References
1. United States Environmental Protection Agency Department of Transportation and Air Quality
(2020) Fast FastFacts: U.S. transportation sector greenhouse gas emissions 1990–2018, p.
EPA-420-F-20-037
2. Singh N, Mishra T, Banerjee R (2019) Greenhouse gas emissions in India’s road transport sector.
Clim Chang Signals Response 197–209. https://doi.org/10.1007/978-981-13-0280-0_12
3. Geng B, Mills JK, Sun D (2012) Two-stage energy management control of fuel cell plug-in
hybrid electric vehicles considering fuel cell longevity. IEEE Trans Veh Technol 61(2):498–508.
https://doi.org/10.1109/TVT.2011.2177483
4. Erdinc O, Uzunoglu M (2010) Recent trends in PEM fuel cell-powered hybrid systems: inves-
tigation of application areas, design architectures and energy management approaches. Renew
Sustain Energy Rev 14(9):2874–2884. https://doi.org/10.1016/j.rser.2010.07.060
5. Emadi A, Rajashekara K, Williamson SS, Lukic SM (2005) Topological overview of hybrid
electric and fuel cell vehicular power system architectures and configurations. IEEE Trans Veh
Technol 54(3):763–770. https://doi.org/10.1109/TVT.2005.847445
6. Fadel A, Zhou B (2011) An experimental and analytical comparison study of power manage-
ment methodologies of fuel cell-battery hybrid vehicles. J Power Sources 196(6):3271–3279.
https://doi.org/10.1016/j.jpowsour.2010.11.114
7. Uzunoglu M, Onar OC, Alam MS (2009) Modeling, control and simulation of a PV/FC/UC
based hybrid power generation system for stand-alone applications. Renew Energy 34(3):509–
520. https://doi.org/10.1016/j.renene.2008.06.009
8. Erdinc O, Vural B, Uzunoglu M (2009) A wavelet-fuzzy logic based energy management
strategy for a fuel cell/battery/ultra-capacitor hybrid vehicular power system. J Power Sources
194(1):369–380. https://doi.org/10.1016/j.jpowsour.2009.04.072
9. Panday A, Bansal HO (2014) A review of optimal energy management strategies for hybrid
electric vehicle. Int J Veh Technol 2014. https://doi.org/10.1155/2014/160510
10. Wang Y et al (2019) A comprehensive analysis of energy management strategies for hybrid
electric vehicles based on bibliometrics. J Energy Storage 26(February):57–66. https://doi.org/
10.1109/TVT.2017.2660764
11. Lü X, et al (2020) Energy management of hybrid electric vehicles: a review of energy opti-
mization of fuel cell hybrid power system based on genetic algorithm. Energy Convers Manag
205. https://doi.org/10.1016/j.enconman.2020.112474
12. Zhou B, Burl JB, Rezaei A (2020) Equivalent consumption minimization strategy with consid-
eration of battery aging for parallel hybrid electric vehicles. IEEE Access 8. https://doi.org/10.
1109/ACCESS.2020.3036033
13. Wang Y, Wang W, Zhao Y, Yang L, Chen W (2016) A fuzzy-logic power management strategy
based on Markov random prediction for hybrid energy storage systems. Energies. https://doi.
org/10.3390/en9010025
14. Liu X, Qin D, Wang S (2019) Minimum energy management strategy of equivalent fuel
consumption of hybrid electric vehicle based on improved global optimization equivalent factor.
Energies 12(11). https://doi.org/10.3390/en12112076
15. Schmid R, Buerger J, Bajcinca N (2021) Energy management strategy for plug-in-hybrid
electric vehicles based on predictive PMP. IEEE Trans Control Syst Technol. https://doi.org/
10.1109/TCST.2020.3048129
16. Zhang P, Yan F, Du C (2015) A comprehensive analysis of energy management strategies for
hybrid electric vehicles based on bibliometrics. Renew Sustain Energy Rev 48(205):88–104.
https://doi.org/10.1016/j.rser.2015.03.093
17. Hajizadeh A, Golkar MA, Feliachi A (2010) Voltage control and active power management
of hybrid fuel-cell/energy-storage power conversion system under unbalanced voltage sag
conditions. IEEE Trans Energy Convers 25(4):1195–1208. https://doi.org/10.1109/TEC.2010.
2062516
18. Wang C, Nehrir MH, Shaw SR (2005) Dynamic models and model validation for PEM fuel
cells using electrical circuits. IEEE Trans Energy Convers. https://doi.org/10.1109/TEC.2004.
842357
19. Tang L, Rizzoni G, Onori S (2015) Energy management strategy for HEVs including battery
life optimization. IEEE Trans Transp Electrif. https://doi.org/10.1109/TTE.2015.2471180
20. Musolino V, Piegari L, Tironi E (2013) New full-frequency-range supercapacitor model with
easy identification procedure. IEEE Trans Ind Electron. https://doi.org/10.1109/TIE.2012.218
7412
21. Oldham KB (2008) A Gouy-Chapman-Stern model of the double layer at a (metal)/(ionic liquid)
interface. J Electroanal Chem 613(2):131–138. https://doi.org/10.1016/j.jelechem.2007.10.017
22. Rodatz P, Paganelli G, Sciarretta A, Guzzella L (2005) Optimal power management of an
experimental fuel cell/supercapacitor- powered hybrid vehicle. Control Eng Pract 13(1):41–53.
https://doi.org/10.1016/j.conengprac.2003.12.016
23. Xu L, Li J, Hua J, Li X, Ouyang M (2009) Optimal vehicle control strategy of a fuel cell/battery
hybrid city bus. Int J Hydrog Energy. https://doi.org/10.1016/j.ijhydene.2009.06.021
24. Garcia P, Fernandez LM, Garcia CA, Jurado F (2010) Energy management system of fuel-cell-
battery hybrid tramway. IEEE Trans Ind Electron 57(12):4013–4023. https://doi.org/10.1109/
TIE.2009.2034173
25. Njoya Motapon S, Dessaint LA, Al-Haddad K (2014) A comparative study of energy manage-
ment schemes for a fuel-cell hybrid emergency power system of more-electric aircraft. IEEE
Trans Ind Electron 61(3):1320–1334. https://doi.org/10.1109/TIE.2013.2257152
Evaluative Study of Cascaded Multilevel
Inverter Topologies
Hemant Gupta, Arvind Yadav, and Sanjay Kumar Maurya
1 Introduction
In all the high power applications of industry, multilevel inverter is more preferable
because it is tremendously advantageous due to its high quality of output voltage,
higher efficiency, less component of harmonics and its dynamic performance with
different loading conditions [1–4]. Multilevel inverters are very much capable to
synthesize approximated pure ac sinusoidal waveform by using separate dc input
sources [2]. In the classical structure of multilevel inverter it has many of topologies
out of which there are three basic and popular topologies named as flying capacitor
MI, neutral point clamped MI and cascaded H bridge MI [2, 5–7]. Among them
cascade H bridge based multilevel inverter is more preferable because of its config-
urable framework and minimized count number of switches in contrast with other
topologies [8]. The use of differing magnitudes in DC sources enables symmetric
and asymmetric configurations of cascaded multilevel inverter topologies. Asym-
metrical topologies can produce higher voltage levels by using the same number of
power semiconductor devices including DC sources of unequal magnitude, whereas
a symmetrical multilevel inverter needs DC sources of the same magnitude [9].
Hybrid modulation, Selective Harmonic Elimination (SHE), Nearest Level Modu-
lation (NLM), and Multicarrier PWM are few of the typical modulation methods
used in cascade multilevel inverters. Full-bridge inverter comprises switches that are
purposely designed to withstand voltage levels than other switches in some topologies
[10].
H. Gupta (B) · A. Yadav · S. K. Maurya

GLA University, Mathura, India
e-mail: hemant.gupta@gla.ac.in
A. Yadav
e-mail: arvind.yadav@gla.ac.in
S. K. Maurya
e-mail: sanjay.maurya@gla.ac.in
256 H. Gupta et al.
Literature reveals that first MLI was created at the beginning of 1975 and that
several forms and variations in multilevel converter topologies were documented
in literature with research and development. In recent decades, a good much of the
literature have shown a strong interest in the study of diode clamped, flying capacitor
and cascaded H-bridge topologies with regard to their respective pros and cons. The
main drawback with conventionally used multilevel inverter topologies is that the
quantity of power electronic components used increases as the levels increase [11]. In
recent years, research on CHB inverters has been more focused due to the modularity
of the circuit, balanced DC voltage, and more levels of inverter output voltage with
a smaller number of components without needing high ratings for each unit [12].
New structures have been developed to ensure the multilevel inverter’s excellent
performance. Recently the research work that has been done on multilevel inverter
relies on the required level count at the output side and dc source used at input.
And at the same time its switching losses and switching stress need to be considered.
Another study’s topology demonstrates the creation of output voltage having multiple
levels by configuring series and parallel connection of DC voltage sources. Besides
the conventional topologies of multilevel inverter, some advanced modular structure
of multilevel inverter also designed in which more attention has been paid towards
the number of switches count and dc source used [13].
2 Classical 11-Level CHB
Figure 1 presents the classical 11-level cascade multilevel inverter which comprises
five modules of H bridge, and each module is capable to provide three level of output
0 and ±Vdc [10]. Each H bridge contains one dc source and four switches S1, S2, S3,
S4 of low voltage rating, two switches in each leg of H bridge. A series-connected H
bridge string is used in a multilevel inverter to provide stepped output from multiple
dc sources. The numbers of levels achieved at the output side of inverter with the
number of H bridges used are related with equation as
Nlevel = 2N + 1 (1)
where N is the number of H bridges used. Here the output level and dc source required
has same count of number. On the basis of magnitude of dc voltage sources, two types
of configuration are available for multilevel inverter symmetrical and asymmetrical
[14]. If the magnitude of all dc sources is equal then it is called symmetrical otherwise
unsymmetrical. Here the used configuration is of symmetrical multilevel inverter and
the maximum voltage achieved ate the output is the sum of several the dc sources of
each H bridge as
Vo = Vdc1 + Vdc2 + Vdc3 + Vdc4 + Vdc5 (2)

Evaluative Study of Cascaded Multilevel Inverter Topologies 257
S1 S3
Vdc1
S4 S2
S5 S7
Vdc2
S8 S6
S9 S11 L
O
Vdc3 A
D
S12 S10
S13 S15
Vdc4
S16 S14
S17 S19
Vdc5
S20 S18
Fig. 1 Classical 11-level CHB
The biggest advantage of using symmetrical configuration is that it is even power

distribution and modular structure because of which its modulation technique can
be designed easily for getting the switched output voltage of the multilevel inverter
[15].
3 Modulation Strategy for Classical CHB
There are several modulation techniques for cascaded H bridge multilevel inverter
which is based upon the number of output level and amount of THD [16]. Among
the many recent modulation techniques multicarrier PWM technique has been most
commonly used because of its improved harmonic performance [17].
Here the carrier-based phase shifted PWM strategy is used for the 11-level
cascaded multilevel inverter in which N − 1 triangular carriers are used of same
magnitude and frequency where output voltage level count of inverter is typified as
N. The switched output voltage for the multilevel inverter is calculated by comparing
the reference sinusoidal signal with each carrier signal for each cell. In the midst of
the two neighbouring carrier waves the phase shift occurs of angle
360
cr = (3)
N −1
258 H. Gupta et al.
where N is typified as output voltage level count of inverter is typified. In this paper
ten carriers are used with a phase shift of 36° between two adjacent carrier signals for
the 11-level cascade multilevel inverter and there is an 180° phase shift between the
carrier-based gate triggered signals for upper and lower switch of cells of multilevel
inverter as shown in Fig. 2. All switching combination for getting the 11-level output
voltage for cascaded multilevel inverter of Fig. 1 is tabulated in Table 1.
>= 1 OR 0 UPPER
SWITCH
SIN
OF
WAVE
CELL
NOT 1 OR 0
+
<= 1 OR 0 LOWER
- SWITCH
OF
CELL
NOT 1 OR 0
GATING SIGNAL
Cr 2 (WITH 180˚ PHASE

SHIFT)
Cr 1
PHASE SHIFTED CARRIERS WAVE
Fig. 2 Phase shifted modulation pattern
Table 1 Switching table for classical 11-level CHB

S. No. Vo(t) Switches in on state
1 5Vdc S1, S2, S5, S6, S9, S10, S13, S14, S17, S18
2 4Vdc S1, S2, S5, S6, S9, S10, S13, S14, S18, S20
3 3Vdc S1, S2, S5, S6, S9, S10, S14, S16, S18, S20
4 2Vdc S1, S2, S5, S6, S10, S12, S14, S16, S18, S20
5 Vdc S1, S2, S6, S8, S10, S12, S14, S16, S18, S20
6 0 S1, S2, S6, S8, S10, S12, S14, S16, S18, S20
7 −Vdc S3, S4, S6, S8, S10, S12, S14, S16, S18, S20
8 −2Vdc S3, S4, S7, S8, S10, S12, S14, S16, S18, S20
9 −3Vdc S3, S4, S7, S8, S11, S12, S14, S16, S18, S20
10 −4Vdc S3, S4, S7, S8, S11, S12, S15, S16, S18, S20
11 −5Vdc S3, S4, S7, S8, S11, S12, S15, S16, S19, S20
4 Modulated 11-Level CHB
The modified structure of 11-level multilevel inverter is shown in Fig. 4 in which

ten semiconductor switches are used out of which six switches A, B, C, D, E, F are
unidirectional while other four switches G, H, I, J are bidirectional. The entire model
of this structure consists of two modules, each module contains two switches and
two dc sources. In this paper the symmetrical configuration is used for modulated
multilevel inverter in which magnitude of all the dc sources are equal [13]. Each
module of this inverter is capable to produce three voltage levels as 0, Vdc and 2Vdc
as shown in given Table 2 and the process of these voltage labels arrivals shown in
Fig. 3b–d.
Number of voltage level produced by the modified topology of MI can be decided
by the magnitude of external dc source as shown in Fig. 3. For N level of output
voltage, the number of dc sources required is
N −1
m= (4)
2
Table 2 Voltage levels of

Voltage level S1 S2
each module
2Vdc 1 0
Vdc 0 1
0 0 0
on
S1 off S1
Vdc Vdc
S2 Vdc S2 Vdc
(a) (b)
off off
off S1
S1 Vdc
Vdc
on
S2 Vdc S2 Vdc
(c) (d)
Fig. 3 a Module configuration b Equivalent circuit for the voltage 2Vdc c Equivalent circuit for
the voltage Vdc d Equivalent circuit for the voltage 0
260 H. Gupta et al.
A D
G H
Vdc1 Vdc3
Vs L
O
I J A
Vdc2 Vdc4
D
B
E
Vdc5
C F
Fig. 4 Modulated 11-level CHB
where N is the number of output voltage levels. In this paper for 11-level of output
voltage five dc sources are required because of which the value of Vs voltage source
in the proposed topology is zero. The number of modules used with respect to count
of dc source used is defined by the equation as
m−1
K= (5)
2
where m typifies the dc sources used in proposed topology. The number of output
voltage level with the maximum magnitude of output voltage and the magnitude of
dc source are related with the equation as
2Vo max
N= +1 (6)
Vdc
Where Vo max is the maximum magnitude of output voltage and Vdc is the magni-
tude of each dc source. In this configuration, the polarity of voltage can be easily
changed in both the direction by providing a suitable current path with the help
of unidirectional switches A, B, C, D, E and F where D, E, F switches working in
complimentary manner with respect to switches A, B and C respectively. In this paper
the modulation strategy has been designed for the modulated model of 11-level CHB
which is based upon the switching state for each voltage level as shown in Table 3 in
which 0 indicates the off position of switch while 1 indicates on position of switch
(Fig. 4).
Table 3 Switching table for modulated 11-level CHB

Vo A B C G H I J D E F
5Vdc 0 1 1 1 1 0 0 1 0 1
−5Vdc 1 0 1 1 1 0 0 0 1 0
4Vdc 0 1 0 0 1 1 0 1 0 1
−4Vdc 1 0 0 0 1 1 0 0 1 0
3Vdc 0 1 0 0 0 1 1 1 0 1
−3Vdc 1 0 0 0 0 1 1 0 1 0
2Vdc 0 1 0 0 0 1 1 1 0 0
−2Vdc 1 0 0 0 0 1 1 0 1 1
Vdc 1 1 0 0 0 0 0 0 0 1
−Vdc 0 0 0 0 0 0 0 1 1 0
0 0 1 0 1 1 0 0 0 1 0
5 Comparative Assessment Between Classical

and Modified CHB
In this paper both the classical cascaded H bridge and modified topology of multilevel
inverter have been compared on the basis of their advantages and limitations out of
which the modularity in structure and reduced switch count number in the modified
topology of MI for the same level of output voltage is a main consideration. The
overall performance for both topologies of multilevel inverter can be evaluated on
the basis of some desirable features that have to be always kept in mind such as their
modularity in structure, harmonic performance, switch count number, dc source
number and so on. In this paper all the comparisons are shown in Table 4 and also
the comparative graphs for voltage stress and number of switches are also shown in
Fig. 5.
Table 4 Comparative table between classical and modified CHB

Topology Output Switch DC source Modularity Configuration Voltage
voltage count required harmonic
levels (N) number content (%)
Classical 11 20 5 Possible Symmetrical or 11.8
CHB unsymmetrical
Modulated 11 10 5 Possible Symmetrical 18.7
CHB
262 H. Gupta et al.
Classical Classical
25 11- 25 11-
LEVEL LEVEL
CHB CHB
20 20
Number of Switches
Modulated Modulated
Voltage Stress
15 11-LEVEL 15 11-LEVEL
CHB CHB
10 10
5 5
0 0
Topology Topology
Fig. 5 Comparative graphs between classical 11-level CHB and modulated 11-level CHB
6 Simulation Result
To vindicate the effectiveness of modified topology of MI over the classical cascaded

H bridge, the Simulink results are shown in Figs. 6 and 7 for modified and classical
CHB respectively. The FFT analyses of these two topologies are also shown in
Figs. 8 and 9 which shows its voltage harmonic content. Both topologies of modulated
multilevel inverter and classical multilevel inverter are simulated for the same number
of voltage level that is eleven in which the switching frequency and load parameters
are also taken of same rating. Table 4 compares the modified and classical topologies
of multilevel inverters based on switch count number and essential parameters, in
addition to the simulation results.
11-level output voltage of proposed MMI

800
600
11-level ouput voltage
400
200
-200
-400
-600
-800
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
time
Fig. 6 Output voltage waveform of modulated 11-level CHB

11-level output of classical CMI 11-level ouput of classical CMI

800
600
400
200
-200
-400
-600
-800
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
time
Fig. 7 Output voltage waveform of classical 11-level CHB
Fundamental (50Hz) = 498.2 , THD= 11.87%

1
Mag (% of Fundamental)
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 100 200 300 400 500 600 700 800 900 1000
Frequency (Hz)
Fig. 8 THD spectra of classical 11-level CHB
Fundamental (50Hz) = 591 , THD= 18.75%

18
Mag (% of Fundamental)
16
14
12
10
8
6
4
2
0
0 100 200 300 400 500 600 700 800 900 1000
Frequency (Hz)
Fig. 9 THD spectra of modulated 11-level CHB

264 H. Gupta et al.
7 Conclusion
In this paper comparative assessment has been done between the modified topology
of multilevel inverter and classical topology of CHB on the basis of switch count
number, modularity in structure, harmonic performance for the same output voltage
levels. The modified multilevel inverter topology decreases the switch count that
is necessary to achieve output voltage of 11 levels thereby, the size and cost of
the equipment reduced. In this paper the multicarrier PWM technique are used for
classical topology of CHB and it is to be found that this technique is best suitable in
terms of simplicity and harmonic performance. The switched output for the changed
topology of the multilevel inverter is accomplished using a modulation strategy based
on the switching states listed in Table 3. The whole comparison has been shown in
Table 4 between two topologies. In future research, a range of more multilevel inverter
topologies can be developed for industrial high-power applications where the number
of switches, voltage level and stress can be given importance.
References
1. Gupta H, Yadav A, Maurya S (2016) Dynamic performance of cascade multilevel inverter

based STATCOM. In: 2016 IEEE 1st international conference on power electronics, intelligent
control and energy systems (ICPEICES), Delhi, India, pp 1–4. https://doi.org/10.1109/ICP
EICES.2016.7853479
2. Lai J-S, Peng FZ (1996) Multilevel converters-a new breed of power converters. IEEE Trans
Ind Appl 32(3):509–517. https://doi.org/10.1109/28.502161
3. Rodriguez J, Lai J-S, Peng FZ (2002) Multilevel inverters: a survey of topologies, controls,
and applications. IEEE Trans Ind Electron 49(4):724–738.https://doi.org/10.1109/TIE.2002.
801052
4. Dahidah MSA, Konstantinou G, Agelidis VG (2015) A review of multilevel selective harmonic
elimination PWM: formulations, solving algorithms, implementation and applications. IEEE
Trans Power Electron 30(8):4091–4106. https://doi.org/10.1109/TPEL.2014.2355226
5. McGrath BP, Holmes DG (2002) Multicarrier PWM strategies for multilevel inverters. IEEE
Trans Ind Electron 49(4):858–867. https://doi.org/10.1109/TIE.2002.801073
6. Perez MA, Bernet S, Rodriguez J, Kouro S, Lizana R (2015) Circuit topologies, modeling,
control schemes, and applications of modular multilevel converters. IEEE Trans Power Electron
7. Rohner S, Bernet S, Hiller M, Sommer R (2010) Modelling, simulation and analysis of a
Modular Multilevel Converter for medium voltage applications. In: 2010 IEEE international
conference on industrial technology, Via del Mar, Chile, pp 775–782. https://doi.org/10.1109/
ICIT.2010.5472634
8. Rao BN, Suresh Y, Panda AK, Naik BS, Jammala V (2020) Development of cascaded multilevel
inverter based active power filter with reduced transformers. CPSS Trans Power Electron Appl
5(2):147–157. https://doi.org/10.24295/CPSSTPEA.2020.00013
9. Dhanamjayulu C, Meikandasivam S (2018) Implementation and comparison of symmetric and
asymmetric multilevel inverters for dynamic loads. IEEE Access 6:738–746. https://doi.org/
10.1109/ACCESS.2017.2775203
10. Gupta H, Yadav A, Maurya S (2020) Multi carrier PWM for cascade topology of multilevel
inverter. In: 2020 international conference on recent trends on electronics, information, commu-
nication & technology (RTEICT), Bangalore, India, pp 328–332. https://doi.org/10.1109/RTE
ICT49044.2020.9315586
11. Hasan MM, Abu-Siada A, Dahidah MSA (2018) A three-phase symmetrical DC-link multilevel
inverter with reduced number of DC sources. IEEE Trans Power Electron 33(10):8331–8340.
12. Goel V, Kumar J, Gambhir J (2015) Different multilevel inverter topologies with reduced
number of devices. In: 2015 2nd international conference on recent advances in engineering &
computational sciences (RAECS), pp 1–6. https://doi.org/10.1109/RAECS.2015.7453375
13. Ebrahimi J, Babaei E, Gharehpetian GB (2012) A new multilevel converter topology with
reduced number of power electronic components. IEEE Trans Ind Electron 59(2):655–667.
https://doi.org/10.1109/TIE.2011.2151813
14. Zaid MM, Ro J (2019) Switch ladder modified H-Bridge multilevel inverter with novel pulse
width modulation technique. IEEE Access 7:102073–102086. https://doi.org/10.1109/ACC
ESS.2019.2930720
15. Lee SS, Sidorov M, Lim CS, Idris NRN, Heng YE (2018) Hybrid cascaded multilevel
inverter (HCMLI) with improved symmetrical 4-level submodule. IEEE Trans Power Electron
16. Cho Y, LaBella T, Lai J, Senesky MK (2014) A carrier-based neutral voltage modulation strategy
for multilevel cascaded inverters under unbalanced DC sources. IEEE Trans Ind Electron
61(2):625–636. https://doi.org/10.1109/TIE.2013.2254091
17. Napoles J, Leon JI, Franquelo LG, Portillo R, Aguirre MA (2009) Selective harmonic mitigation
technique for multilevel cascaded H-bridge converters. In: 2009 35th annual conference of ieee
industrial electronics, Porto, Portugal, pp 806–811. https://doi.org/10.1109/IECON.2009.541
5023
Low Speed Performance Improvement
of Dual VSI Fed Direct Torque
Controlled Five Phase Open-End
Winding Induction Motor
C. Venkata Subba Reddy and Swati Devabhaktuni
1 Introduction
With the great development of power electronic converters and modern Digital signal
processors, interest on Multi-phase machines tremendously increased in the begin-
ning of twenty-first century [1]. Multi-phase induction motors have many features
when compared with conventional three phase induction motor such as reduced per
phase current or per phase power without increase in per phase voltage, Higher
torque density, lower Torque pulsations, high fault tolerant capability for loss of
any one or more phases, Higher efficiency, reduced current ripples, reduced noise
which makes drive reliable, compact and smooth operation [1]. Due to these addi-
tional features, multi-phase induction machine can be used in ship propulsion, space
craft, and electric vehicles. Recently many researchers focused on implementation
innovative advanced control methods such as direct torque control, flux-orientation
control, predictive control for five leg voltage source inverter-fed five-phase induc-
tion motor, making drive performance faster and efficient [2, 3]. DTC of Induction
Motor drive is implemented in the stationary reference frame (α − β), and the
advantages of DTC compared with FOC are less sensitivity to motor parameter vari-
ation, simplicity, better dynamic performance, easy implementation, and no complex
coordinate transformation required makes preferred control technique in industrial
Drive applications [2, 3]. However the Traditional DTC scheme suffers from the
large torque ripples and variable switching frequency. DTC method implemented
with hysteresis controller based torque and flux comparators with a lookup table for
appropriate voltage vector selection of the inverter to control of stator flux and devel-
oped torque separately. Multi-phase inverter gives higher number voltage vectors,
C. V. S. Reddy (B) · S. Devabhaktuni

Department of Electrical Engineering, National Institute of Technology, Warangal, India
e-mail: cvsreddy@student.nitw.ac.in
S. Devabhaktuni
e-mail: swatikjm@nitw.ac.in
268 C. V. S. Reddy and S. Devabhaktuni
which will facilitate finer torque and flux control in Direct Torque Controlled Induc-
tion motor. Nowadays, many research works are in progress for improvements in
Direct Torque Control schemes in conventional Three Phase as well as multi-phase
drives.
Nowadays, PTC became popular and competitive with DTC for improving
dynamic performance. In PTC, the required switching voltage vector selection is
based on evaluation of cost function. However, it has some limitations such as
sensitive to model parameter change, more calculation time, require cost function
weighting factor tuning [2, 4, 5]. The available literature on various Direct Torque
Control schemes of multi-phase drives are as follows: In [6, 7], voltage space vector-
based DTC is proposed, which require additional PI controllers and carrier based,
maintains constant switching frequency; however, it does not guarantee stable oper-
ation under low speeds. In [8–12], Modified DTC schemes proposed to improve
performance of 2-level inverter fed five phase induction motor drive, where mainly
focused on xy components reduction, low speed performance improvement, and
common mode voltage reduction. Normal 2-level multi-leg inverter generates higher
THD voltages.
MLI topologies in multi-phase systems are implemented for improving the THD
of voltage waveforms. Furthermore, DTC approach can be applied for multilevel VSI
(MLI) fed multi-phase IM drives for precise flux and torque response in medium and
high power applications. Recently in [13–15], various DTC schemes of 3-level neutral
point clamped inverter (NPC) driven five phase Induction Motor are performed for
torque ripple reduction and low speed performance improvement, where it requires
clamping diodes and capacitor neutral point balancing issue is addressed. Also it is
not feasible for going to beyond three-level by NPC, cascade H-bridge, flying capac-
itor topologies due to hardware complexity and cost for multi-phase drives. Apart
from above investigated multi-level inverter topologies for multi-phase motor drive
applications, the dual VSI configuration fed open-end winding five-phase induc-
tion motor (FP-OEWIM) attains important features. They are simple configuration,
clamping diodes not required, and neutral point voltage balancing issues avoided.
Multilevel inversion is possible with the usage of two independent dc voltage sources
with different ratios in dual inverter configuration. Due to these features, dual-
inverter configuration fed FP-OEWIM attains important role in the field of multilevel
inverter topologies of electrical drives. Many researchers focused on dual inverter fed
three phase open end winding induction motor (TP-OEWIM), where TP-OEWIM
excited from two isolated DC sources. And very few researchers recently focused on
multi-phase Open end winding induction motor drives are as follows: In [16], FP-
OEWIM drive with dual inverters configuration with isolated dc link voltage sources
is discussed, where the SVM-DTC method is applied which suffers from above
stated issues same as DTC-SVM control of normal multi-phase machine. In [17], A
Modified DTC is implemented for OEWIM, which uses 20 adjacent and non-adjacent
voltage vectors for eliminating common mode voltage and suffers from higher torque
ripples; demagnetization under low speed and small xy components appears. In [18],
Hysteresis band based DTC implemented for OEWIM, which uses 10 virtual non-
adjacent voltage vectors and improves the DC bus utilization, xy components zero,
Low Speed Performance Improvement of Dual VSI … 269
CMV elimination and suffers from high torque ripple and demagnetization under
low speed region.
This paper concentrates on three-level dual voltage source Inverter fed FP-
OEWIM drive using the modified DTC scheme. The major objectives of this paper
are to suppress higher torque ripples under low speed operation, and avoiding flux
demagnetizing problem at zero/low speed due frequent selection of zero vectors in
dual-inverter fed FP-OEWIM drive. The conventional method of DTC generates
higher torque ripple under low speed operation in dual-inverter fed FP-OEWIM
drive due to selection of large voltage vectors. Besides this, flux demagnetization
occurs under zero speed condition for conventional method [17, 18]. All these above
stated problems in conventional Direct Torque Control of FP-OEWIM are overcome
in the proposed DTC technique with modified lookup table. This paper configured
as follows: In Sect. 2, Mathematical modelling of dual inverter fed FP-OEWIM;
in Sect. 3, theoretical analysis and method of implementation of Proposed DTC
method for five-phase OEWIM; in Sect. 4, comparative analysis of Proposed DTC
with Conventional DTC method along with simulation results; Sect. 5 discusses
about conclusions.
2 Mathematical Modelling of Dual VSI Fed Five Phase

OEW Induction Motor
The 3-level dual inverter configuration of five phase open end winding induction
motor (FP-OEWIM) can be modelled mathematically with 1:1 DC link voltage
(Vdc /2) by using individual pole voltages as shown below.
Each inverter pole voltages of dual inverter fed FP-OEWIM are expressed by (1)
and (2) as given below
⎛ ⎞ ⎛ ⎞
Va0 Sa+
⎜ V ⎟ ⎜ S+ ⎟
⎜ b0 ⎟ ⎜ b ⎟ V
⎜ ⎟ ⎜ ⎟ dc
⎜ Vc0 ⎟ = ⎜ Sc+ ⎟∗ (1)
⎜ ⎟ ⎜ ⎟ 2
⎝ Vd0 ⎠ ⎝ Sd+ ⎠
Ve0 Se+
⎛ ⎞ ⎛ ⎞
Va 0 Sa+
⎜ V ⎟ ⎜ S+ ⎟
⎜ b 0 ⎟ ⎜ b ⎟ V
⎜ ⎟ ⎜ ⎟ dc
⎜ Vc 0 ⎟ = ⎜ Sc+ ⎟∗ (2)
⎜ ⎟ ⎜ ⎟ 2
⎝ Vd 0 ⎠ ⎝ Sd+ ⎠
Ve 0 Se+
where Sa+ , Sb+ …. and Sa+ , Sb+ …. are upper switches of inverter 1 and 2 legs of dual
inverter.
The resultant pole voltage of dual inverter configuration exhibits three-level and
can be represented by (3) as shown below
⎛ ⎞ ⎛ ⎞
Vaa Va0 − Va 0
⎜ V ⎟ ⎜ V − V ⎟
⎜ bb ⎟ ⎜ b0 b0⎟
⎜ ⎟ ⎜ ⎟
⎜ Vcc ⎟ = ⎜ Vc0 − Vc 0 ⎟ (3)
⎜ ⎟ ⎜ ⎟
⎝ Vdd ⎠ ⎝ Vd0 − Vd 0 ⎠
Vee Ve0 − Ve 0
The common mode voltage (CMV) and phase voltage of dual inverter configura-
tion can be obtained from (1), (2), (3) and shown by (4) and (5) as below.
1
Vcm = ∗(Vaa + Vbb Vcc + Vdd + Vee ) (4)
5
⎛ ⎞ ⎛ ⎞
Vaa Vaa
⎜ V ⎟ ⎜ V ⎟
⎜ bb ⎟ ⎜ bb ⎟
⎜ ⎟ ⎜ ⎟
⎜ Vcc ⎟ = ⎜ Vcc ⎟ − (Vcm ) (5)
⎜ ⎟ ⎜ ⎟
⎝ Vdd ⎠ ⎝ Vdd ⎠
Vee Vee
Time varying mutual inductances in motor torque and voltage equations makes
induction motor dynamic behavior analysis complex. For simple analysis, five phase
induction motor is modelled as two-phase machine by transforming five phase
quantities to two phase quantities [19]
The block diagram representation of Direct Torque Controlled (DTC) dual inverter
fed FP-OEWIM is shown in Fig. 1, which contains 5-ph OEWIM, two 2-level
inverters, speed sensor, torque and flux estimation block, torque and flux hysteresis
controllers.
Fig. 1 Schematic diagram of DTC controlled dual inverter fed 5-ph open end winding induction
motor
The dynamic behavior of induction motor can be observed by dynamic modelling

of five phase induction which can be done by following equations under stationary
reference frame. Electrical quantities represented in space vector form.
Stator voltage equations in space vector form fixed reference frame are:
d
Vs = Rs Is + ϕs (6)
dt
Rotor voltage equations:
d
0 = Rs Ir + ϕr − ωr ϕr (7)
dt
By using above voltage equations stator flux linkages are found as follows:
ϕs = ∫(Vs − Is Rs )dt (8)
Flux linkages of stator conductors and rotor conductors can be written as
ϕs = L s i s + L m ir (9)
ϕr = L r ir + L m i s (10)
Electromagnetic torque developed in two axis machine stationary reference frame

can be written as
5P
Te = imag(ϕs ∗Is ) (11)
22
From machine mechanical dynamics, torque balancing equation can be written as
dωr
Te = Tl + j
dt
And

Te − Tl
ωr = dt (12)
J
where
p—no. of poles of machine
J—Moment of inertia
T l —Load torque
T e —developed Torque
ωr —Rotor mechanical speed.
Five phase induction can be modeled by using Eqs. (6)–(12) in stationary reference
frame.
3 Proposed DTC Operation for Dual Inverter

Configuration
The DTC schematic block diagram for 3-level dual VSI fed FP-OEWIM is shown
in Fig. 1. The speed of the motor is compared with set speed, and the speed error is
processed passed through speed PI controller, which gives the ref. torque command
(Tm*). The actual electromagnetic torque (Tm) and flux (s) are estimated from
motor voltages and currents as in Eqs. (8)–(11). Estimated torque compared with
commanded torque and estimated flux compared with reference flux (*s). Based
on Torque error, flux error and stator flux sector information appropriate switching
vector is selected from lookup table through five level torque and two level flux
hysteresis controllers. In DTC, Switching states directly influences the stator flux
and torque.
The effective voltage space vector Vs of dual inverter configuration can expressed
from voltage space vectors Vs1 , Vs2 of individual VSIs as in (13), which are repre-
sented in αβ plane in stationary reference frame. Null voltage vectors are obtained
during inv1 and inv2 have same switching states or all upper or lower switches of
either inverter ON simultaneously.
2 Vdc
Sa+ + Sb+ e− + Sc+ e− + Sd+ e− + Se+ e−
i2π i4π i6π i8π
Vs1 = 5 5 5 5
5 2
2 Vdc
Sa+ + Sb+ e− + Sc+ e− + Sd+ e− + Se+ e−
i2π i4π i6π i8π
Vs2 = 5 5 5 5
5 2
Vs = Vs1 − Vs2 (13)
The 5-leg 2-level VSI consists of fundamental plane (αβ) and harmonic x–y plane.
Fundamental plane consist torque producing components whereas x–y plane consists
of harmonic components which are not responsible for torque production. A 2-level
VSI consists of 32 switching vectors with 30 active voltage vectors and 2 zero voltage
vectors and categorized as large voltage vectors (10), medium voltage vectors (10)
and small vectors (10). The large voltage vectors in αβ-plane are appeared as to small
voltage vectors in xy harmonic plane or vice versa with opposite direction. Medium
voltage vectors in αβ plane mapped as medium vectors in xy harmonic plane.
For proposed dual inverter configuration, fundamental space vector plane αβ is
divided as 20 sectors, with 18° angle span. Dual inverter configuration generates
40 active space vectors and two null voltage vectors which are obtained from inv1
and inv2 large vector switching states which gives small xy components as shown in
Fig. 2 Switching voltage vector diagram of dual inverter in αβ plane
Fig. 2 and are divided into four groups as: Large vectors (20), Medium vectors (10),
small vectors (10) and Null voltage vectors (2) as follows.
Small voltage vector (10) is
Vs1 = V3inv1 − V4inv2 = 0.647Vdc ∠2π/5 − 0.647Vdc ∠(3π/5)

= 0.618Vdc ∠(0) (14)
Medium voltage vectors (10) is
Vm1 = V1inv1 − V9inv2 = 0.647Vdc ∠0 − 0.647Vdc ∠(9π/5)

= 0.757Vdc ∠(3π/10) (15)
Large voltage vector (10 + 10)
Vl1 = V1inv1 − V6inv2 = 0.647Vdc ∠0 − 0.647Vdc ∠(π )

= 1.29Vdc ∠0
(16)
Vl2 = V1inv1 − V7inv2 = 0.647Vdc ∠0 − 0.647Vdc ∠(6π/5)
= 1.23Vdc ∠(π/10)
Fig. 3 Torque, flux, speed response under high speed (100 rad/sec) and low speed (20 rad/sec) for
conventional method
Conventional DTC [17, 18] uses only 10 medium and 10 large vectors along with
null voltage vectors as in Fig. 3 for all possible speed as shown which causes higher
torque ripples and flux instability at lower/zero speeds under flux error +1 and torque
error 0. In conventional DTC, under flux error +1 and torque error 0, null voltage
vectors applied for long duration which causes flux drooping. But really it needs
increased stator flux under flux error +1 and consequently flux becomes unstable
especially under low speed. Whereas in case of high speed operation application
of null vectors are acceptable under flux error +1 and torque error 0 and flux is
maintained as constant due to short interval of null vectors.
In proposed lookup table based DTC, more no. of voltage vectors generated which
facilitates finer torque and flux variation. Use of small voltage vectors under low
speed or zero speed gives lower torque ripple. Modified lookup table uses large and
medium vectors for higher speeds along with null voltage vectors shown in Table 1
unshaded form and uses small vectors for small speeds as shown in Table 1. Flux
instability or flux drooping problem under low speeds overcome by replacing null
vectors with small vectors which gives higher flux component which really needed
and little influence on torque component under flux error +1 and torque error 0
which are shown in modified lookup Table 1 with shaded format. In sector 1, for
case d = +1 and dT = 0 condition small vector S2 selected which have large flux
component and little torque component. Similarly small voltage vectors selected for
other sectors under d = 1, dT = 0.
Table 1 Proposed DTC lookup table (unshaded-high speed, shaded-low speed)

Sector
Ψ
dΨ dT
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
2 L4 L5 L6 L7 L8 L9 L10 L11 L12 L13 L14 L15 L16 L17 L18 L19 L20 L1 L2 L3
1 S3 S3 S4 S4 S5 S5 S6 S6 S7 S7 S8 S8 S9 S9 S10 S10 S1 S1 S2 S2
M1 M2 M2 M3 M3 M4 M4 M5 M5 M6 M6 M7 M7 M8 M8 M9 M9 M10 M10 M1
1
S2 S2 S3 S3 S4 S4 S5 S5 S6 S6 S7 S7 S8 S8 S9 S9 S10 S10 S1 S1
0 Null
1
-1 M8 M9 M9 M10 M10 M1 M1 M2 M2 M3 M3 M4 M4 M5 M5 M6 M6 M7 M8 M8
- L18 L19 L20 L1 L2 L3 L4 L5 L6 L7 L8 L9 L10 L11 L12 L13 L14 L15 L16 L17
-2 S9 S9 S10 S10 S1 S1 S2 S2 S3 S3 S4 S4 S5 S5 S6 S6 S7 S7 S8 S8
L8 L9 L10 L11 L12 L13 L14 L15 L16 L17 L18 L19 L20 L1 L2 L3 L4 L5 L6 L7
2
1
Null
0
Null
-1
L14 L15 L16 L17 L18 L19 L20 L1 L2 L3 L4 L5 L6 L7 L8 L9 L10 L11 L12 L13
-2
The Dual inverter fed FP-OEWIM drive operation with proposed lookup table DTC
method is verified through MATLAB/Simulink software. Results obtained in the
proposed method are compared with conventional DTC approaches [17, 18]. All the
Direct Torque control methods are developed with same sampling frequency with
same torque and flux hysteresis bands for comparison purpose. The 5-level torque
and 2-level flux hysteresis bands are set at 0.2Nm and 0.02Wb for comparing both
proposed and conventional DTC methods. The 5-ph induction motor speeds below
20 rad/sec is treated as low speed and above that treated as high speed. The simulated
results for conventional method [17] and proposed method are shown in below figures
and discussed for comparison (Table 3).
4.1 Simulation Results
For the 5-ph induction motor set speeds of 100 rad/sec and 20 rad/sec, steady-state
speed, torque, and flux responses for conventional method [17] are given in Fig. 3
under low speed and high speed conditions. Similarly, speed, torque, flux responses
for the same reference speeds of 100 rad/sec and 20 rad/sec of a proposed DTC
schemes shown in Fig. 4. From Figs. 3 and 4, it can be concluded that proposed DTC
scheme produces appreciable torque ripple reduction under low speed with small
Table 2 Five-Ph induction

5-Ph OEW induction motor parameters
motor parameters
Stator resistance (Rs) 6.03
Stator inductance (Ls) 0.4893 H
Rotor resistance (Rr) 6.085
Rotor inductance (Lr) 0.4893 H
Mutual inductance (Lm) 0.45 H
No. of poles (P) 4
Rated speed 1440 rpm
Inertia constant (J) 0.0048
Inverter DC voltage (Vdc/2)
Fig. 4 Torque, flux, speed response under high speed (100 rad/sec) and low speed (20 rad/sec) for
proposed DTC method
decrement flux ripples due to selection of small voltage vectors when compared with
conventional DTC method.
Figures 5 and 6 show the switching pattern of Sa switch of inv1 and inv2 for
both conventional and proposed DTC, which concludes that proposed DTC scheme
gives lower switching frequency operation under low speeds when compared with
conventional method even though switching frequency almost same for higher speed
operation which facilitates greater reduction in switching losses under low speed
operation. Switching loss reduction occurs due to avoiding of frequent zero switching
under lower speed operation with replacement of small voltage vectors.
Figures 7 and 8 show the phase current of convention DTC and Proposed DTC
schemes under high speed and low speed conditions. It can observe that current
waveforms almost similar for both conventional and proposed methods under high
Fig. 5 Switching of Sa of inverter 1 and 2 under high speed and low speed for conventional DTC
method
Fig. 6 Switching of Sa of inverter 1 and 2 under high speed and low speed for proposed DTC
method
Fig. 7 Current response under high speed and low for conventional DTC method
speeds, but with little THD improvement of current under low speed operation of
proposed DTC.
Figure 9 shows the developed torque, stator flux, speed response of conventional
method of DTC under low/zero speed under no load condition. From Fig. 9 flux
instability occurs at zero speed i.e. unable to maintain rated flux due to frequent
Fig. 8 Current response under high speed and low for Proposed DTC method
Fig. 9 Torque, flux, speed response under low speeds (20 rad/sec, 0 rad/sec) for conventional DTC
method
selection of zero voltage vectors and after some time no active vectors switched on
and becomes zero flux.
Figure 10 shows the torque, speed and flux response under low speed operating
conditions. From Fig. 10, this can be stated that rated flux maintained i.e. demag-
netization problem eliminated under zero speed with no load due to selection small
voltage vectors in place of zero vectors. Torque ripples also less under low speed
operation due to selection small vectors.
Finally, from Table 3 it can be stated that the overall changes to conventional
lookup table in the proposed DTC scheme results in improvement of torque response,
switching frequency reduction, elimination of demagnetizing effect under low speed,
improvement of current THD under low speed operation.
Fig. 10 Torque, flux, speed response under low speeds (20 rad/sec, 0 rad/sec) for proposed DTC
method
Table 3 Results comparison table

Conv Proposed Conv Proposed Conv Proposed
method method method method method method
Torque ripple (p-p) Switching frequency Current %THD
High speeds 8% 6% 3.26 kHz 3.16 kHz 7.91 7.05
(100 rad/sec)
Low speeds 8% 3.5% 5 kHz 2.2 kHz 7.48 6.4
(20 rad/sec)
5 Conclusions
The proposed DTC scheme which uses 40 active vectors with modified lookup table
improves drive performance under low speed operation with dual inverter model.
Modified lookup tables for both high and low speeds gives following benefits under
low speed operation: torque ripple reduction under both high and low speeds partic-
ularly at low speeds, low switching frequency under low speed over conventional
method, stable flux operation possible under zero/low speeds under no load, less
current THD under low speed. Hence it is suggested that proposed DTC scheme
exhibits better performance results for dual inverter fed OEWIM under low speed
operations without degrading high speed performance. These benefits in proposed
DTC scheme verified with MATLAB/Simulink software.
References
1. Levi E, Bojoi R, Profumo F, Toliyat HA, Williamson S (2007) Multiphase induction motor
drives—a technology status review. IET Electr Power Appl 1(4):489–516
2. Wang F, Zhang Z, Mei X, Rodriguez J, Kennel R (2018) Advanced control strategies of induc-
tion machine: field oriented control, direct torque control and model predictive control.Energies
11.https://doi.org/10.3390/en11010120
3. Abu-Rub H, Iqbal A, Guzinski J (2021) High performance control of AC drives with
Matlab/Simulink models (pp 298–356)
4. Riveros J, Prieto Corvalán J, Barrero F, Toral SL, Levi E (2010) Predictive torque control for
five-phase induction motor drives. In: IECON proceedings (Industrial Electronics Conference),
pp 2467–2472.https://doi.org/10.1109/IECON.2010.5675341
5. Bhowate A, Aware MV, Sharma S (2020) Predictive torque control algorithm for a five-phase
induction motor drive for reduced torque ripple with switching frequency control. IEEE Trans
Power Electron 35(7):7282–7294. https://doi.org/10.1109/TPEL.2019.2954991
6. Lai Y-S, Chen J-H (2001) A new approach to direct torque control of induction motor drives for
constant inverter switching frequency and torque ripple reduction. IEEE Trans Energy Convers
16(3):220–227. https://doi.org/10.1109/60.937200
7. Mahanta U, Patnaik D, Panigrahi B, Panda A (2015) Dynamic modeling and simulation of
switching table based DTC of five phase induction motor. 1–6. https://doi.org/10.1109/EESCO.
2015.7253740
8. Khaldi BS, Abu-Rub H, Iqbal A, Kennel R, Mahmoudi MO, Boukhetala D (2011) Sensorless
direct torque control of five-phase induction motor drives. In: IECON 2011 - 37th annual
conference of the IEEE industrial electronics society, Melbourne, VIC, pp 3501–3506. https://
doi.org/10.1109/IECON.2011.6119875
9. Zheng L, Fletcher JE, Williams BW, He X (2011) A novel direct torque control scheme for a
sensorless five-phase induction motor drive. IEEE Trans Ind Electron 58(2):503–513. https://
doi.org/10.1109/TIE.2010.2047830
10. Gao L, Fletcher JE, Zheng L (2011) Low-speed control improvements for a two-level five-phase
inverter-fed induction machine using classic direct torque control. IEEE Trans Ind Electron
58(7):2744–2754. https://doi.org/10.1109/TIE.2010.2070775
11. Riveros JA, Durán MJ, Barrero F, Toral S (2012) Direct torque control for five-phase induction
motor drives with reduced common-mode voltage. In: IECON 2012 - 38th annual conference
on IEEE industrial electronics society, Montreal, QC, pp 3616–3621. https://doi.org/10.1109/
IECON.2012.6389317
12. Payami S, Behera RK (2017) An improved DTC technique for low-speed operation of a five-
phase induction motor. IEEE Trans Ind Electron 64(5):3513–3523. https://doi.org/10.1109/
TIE.2017.2652397
13. Tatte YN, Aware MV (2015) Torque ripple minimization in five-phase three-level inverter
fed direct torque control induction motor drive. In: 2015 17th European conference on power
electronics and applications (EPE’15 ECCE-Europe), Geneva, pp 1–6. https://doi.org/10.1109/
EPE.2015.7309400
14. Tatte YN, Aware MV (2017) Direct torque control of five-phase induction motor with common-
mode voltage and current harmonics reduction. IEEE Trans Power Electron 32(11):8644–8654.
15. Chikondra B, Muduli UR, Behera RK (2020) Performance comparison of five-phase three-
level NPC to five-phase two-level VSI. IEEE Trans Ind Appl 56(4):3767–3775. https://doi.org/
10.1109/TIA.2020.2988014
16. Karampuri R, Jain S, Somasekhar VT (2016) Phase displaced SVPWM technique for five-phase
open-end winding induction motor drive. In: 2016 IEEE students’ conference on electrical,
electronics and computer science (SCEECS), Bhopal, pp 1–6. https://doi.org/10.1109/SCE
ECS.2016.7509283
17. Mavila PC, Rajeevan PP (2019) A five level DTC scheme for dual inverter-fed five phase open-
end winding induction motor drives with single DC source. In: 2019 IEEE industry applications
society annual meeting, Baltimore, MD, USA, pp 1–6. https://doi.org/10.1109/IAS.2019.891
2445
18. Mavila PC, Rajeevan PP (2020) A new direct torque control scheme for five phase open-end
winding induction motor drives with reduced DC voltage requirement. In: 2020 IEEE inter-
national conference on power electronics, smart grid and renewable energy (PESGRE2020),
Cochin, India, p 16. https://doi.org/10.1109/PESGRE45664.2020.9070665
19. Levi E (2008) Multiphase electric machines for variable-speed applications. IEEE Trans Ind
Electron 55(5):1893–1909
Self-balanced Symmetric Source
Configuration of Nine Level Switched
Capacitor-Based Grid Connected
Inverter with LCL Filter
1 Introduction
In DC to AC power conversion applications, Multilevel Inverters (MLIs) play a very

important role to generate staircase waveform with imitation of sine wave [1]. As
compared with traditional two level inverters, MLIs give curtailed voltage stress,
less electromagnetic interference, higher efficiency, small size filters and less Total
Harmonic Distortion (THD) [2]. Mainly there are three conventional MLI topologies
are available in the literature, viz., Neutral Point Clamped (NPC) [3], Flying Capac-
itor (FC) [4] and Cascaded H-Bridge (CHB) [5]. In these, CHB MLI has higher
modularity but requires more number of isolated DC sources [6]. For generating
9-level output, the aforementioned topologies take more number of switches, gate
driver circuits and protection circuits. To curtail the voluminous of above conven-
tional topologies, Reduced Device Count (RDC) MLIs have been reported in [7]. In
RDC MLI topologies, one of the popular topologies is Multilevel DC Link Inverter
(MLDCL) topology and for generating nine level output, it takes 12 switches, 3 DC
sources and separate backend H-Bridge is required [8]. The nine level [9] MLI has
the disadvantages like control uncertainty and diversity of voltage factors have been
rapidly increasing because of high switching frequency devices. For generating nine
level output, the topology [10] requires 4 DC sources, ten switches and TSV/TBV
of 16.5Vdc. In [11], the topology comprises 2 DC sources, 12 switches, 3 capacitors
and TSV of 14Vdc have been reported. Similarly, the topologies [12, 13] take more
standing voltage for generating nine level output. Generally, TSV is the one of the
key parameters to analyze the cost requirement of devices.
To alleviate aforementioned problems, a new nine level symmetric source (1:1)
configuration of grid connected MLI topology has been introduced with less per
unit TSV. By taking asymmetric source configuration (1:2 and 1:3) that is binary and
trinary modes, it is possible to generate 13 level and 17 levels, respectively. Generally,
A. Bhanuchandar (B) · B. K. Murthy

284 A. Bhanuchandar and B. K. Murthy
for any MLI topology, the PWM technique has been placed a very important role for
making self-balancing of capacitors and to generate the required level output. For
1:1, 1:2 and 1:3 source configurations with RL load, the modified version of Nearest
Level Control (NLC) technique [14, 16] that is URCS has been applied for getting
required level output. In grid connection, a ULSPD-PWM scheme has been applied
for making successful operation of an inverter.
The rest of the work is summarized as follows: The operation and control schemes
of the proposed topology have been presented in Sect. 2. In Sect. 3, the comparative
study has been described. Section 4 gives the simulation results for both RL load and
grid connected cases. Finally, the conclusion has been presented in Sect. 5.
2 Operation of Proposed Topology and Control Schemes
A. Operation of Proposed Inverter Topology with LCL filter
The schematic of RSC inverter topology with LCL filter is depicted in Fig. 1. It
comprises 2 DC sources, 2 Switched capacitors and 10 unidirectional switches. For
generating 9 levels, 13 levels and 17 levels, the 1:1 (symmetric), 1:2 (binary) and 1:3
(trinary) source configurations have been considered respectively. And the operation
of respective switching arrangements for all configurations has been depicted in
Tables 1, 2 and 3, respectively. For 1:1, 1:2 and 1:3 source configurations, the voltage
peak values have been obtained as 2Vdc, 3Vdc and 4Vdc, respectively. In this paper,
for grid connection, the symmetric source configuration has been considered. The
S9
S1 S5
Vdc1 S2 S6
C1 +
A B + C2 Vdc2
S3
S7
S4 S8
S 10
I inv
LCL Filter
I grid
V grid
Fig. 1 Proposed RSC inverter topology with LCL filter
Self-balanced Symmetric Source Configuration … 285
Table 1 Switching arrangement-1:1 source configuration (9-level)

State S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 C1 C2
2VDC 1 1 0 0 1 1 0 0 0 1 – –
1.5VDC 0 1 0 1 1 1 0 0 0 1 D –
VDC 0 0 1 1 1 1 0 0 0 1 – –
0.5VDC 0 0 1 1 1 0 1 0 0 1 – C
0 0 0 1 1 0 0 1 1 0 1 – –

3VDC 1 1 0 0 1 1 0 0 0 1 – –
5VDC
2 1 0 1 0 1 1 0 0 0 1 C –
2VDC 0 0 1 1 1 1 0 0 0 1 – –
3VDC
2 1 0 1 0 0 1 0 1 0 1 C D
VDC 1 1 0 0 0 0 1 1 0 1 – –
VDC
2 1 0 1 0 0 0 1 1 0 1 C –
0 0 0 1 1 0 0 1 1 0 1 – –

4VDC 1 1 0 0 1 1 0 0 0 1 – –
7VDC
2 1 0 1 0 1 1 0 0 0 1 C –
3VDC 0 0 1 1 1 1 0 0 0 1 – –
5VDC
2 1 1 0 0 1 0 1 0 0 1 – C
2VDC 1 0 1 0 1 0 1 0 0 1 C C
3VDC
2 0 0 1 1 1 0 1 0 0 1 – C
VDC 1 1 0 0 0 0 1 1 0 1 – –
VDC
2 1 0 1 0 0 0 1 1 0 1 C –
0 0 0 1 1 0 0 1 1 0 1 – –
Table 4 gives the TSV/TBV calculation of all switches with 1:1 source configuration.
And Table 5 gives generalized equations for different source configurations.
Generally, for calculating SC values, the longest/maximum discharge period
method has been preferred. For better operation, the capacitance values have to
be taken little more than the actual calculated values. For finding switching and
conduction losses, the PLECS/PSIM software have been used in the literature and
according to datasheet values, it is possible to get easily the final efficiency of the
power converter. The proposed topology actually looks like three-level FC with two
Table 4 Per-unit TSV of the proposed topology-1:1

S. No. Switches Blocking voltage
1 S1 toS8 (0.5V DC )8
2 S9 andS10 (2V DC )2
Total standing voltage (TSV) in per unit 8V DC /2VDC = 4 p.u
Table 5 Generalized equations for different source configurations for the proposed topology
S. No. Description 1:1 source 1:2 source 1:3 source
configuration configuration configuration
U = Number of basic units/cells
1 Total number of levels 8U + 1 12U + 1 16U + 1
2 Total number of DC 2U 2U 2U
sources
legs and in between it consists additionally S9 and S10 switches. For this topology,
there is no need of bidirectional switches requirement for generating nine, thirteen
and seventeen levels.
After getting particular level output, the design of filter is very much important
for the grid connection. As compared with L, LC filters, LCL filter gives better
ripple attenuation. However, it suffers from resonance problem and finally the system
becomes unstable. To tackle these resonance problems, generally active damping
method has been preferred. Because of this method, no additional power losses
are present as compared with passive damping method. The values of inverter side
inductance, grid side inductance and filter capacitances have been calculated from
[15].
B. Proposed Control Schemes
(i) URCS-RL load case
For all source configurations with RL load, the URCS has been applied and its
strategy as depicted in Fig. 2. The proposed control scheme is the modified version
of NLC technique. The URCS provides less computation time as compared with
NLC scheme and SHE PWM technique.
Fig. 2 Unified rounding control scheme—URCS

Fig. 3 Proposed dq frame current control scheme-1:1 source configuration
In this scheme, modulating signal is given to rounding function and it generates

Aggregated Signal (ASG). The Switching Table Generation (STG) has been pene-
trated from Tables 1, 2 and 3 for the respective source configurations. Because of
truncation, there is no need of logic gate requirement for generating Gate Pulses
(GP). This control scheme is basically a fundamental switching frequency tech-
nique thereby switching losses greatly reduce and it also provides self-balancing of
capacitors. It is universally applicable for any MLI topology with valid switching
table.
(ii) dq Frame Current Control Strategy-Grid Connected Case
The control strategy for symmetric source configuration in grid connected case has
been depicted in Fig. 3. Firstly, sense the grid voltage (let, Vgrid = Vβ ) and through
proper PLL, the Vd and Vq components have been generated. Here, PLL has been
used to generate reference variable for grid synchronization. Similarly, the Id and
Iq components have been generated by sensing inverter current (let, Iinv = Iβ ). To
inject active and reactive power into the grid, Id(ref) and Iq(ref) have been used. From
decoupling control strategy, the modulating β-component signal has been generated.
By using Proper Gain (PG ) value, it is possible to generate desired modulating signal.
By taking absolute function, it is possible to curtail the number of high frequency
carriers present in Repeating Table Generation (RTG) block. After truncation of ASG
and STG, the required switching pulses/Gate Pulses (GP) have been generated. If by
taking Iq(ref) is equal to zero then it is possible to get UPF operation at grid side.
3 Comparative Analysis
Table 6 represents generalized comparison of symmetrical MLI topologies and Table

7 represents the comparison of nine level symmetrical inverters topologies. From this
Table 6 Comparison of conventional and proposed topologies

Components (L = number of voltage NPC FC CHB Proposed
levels)
5(L−1)
Number of switches 2(L − 1) 2(L − 1) 2(L − 1) 4
(L−1) (L−1) (L−1) (L−1)
Number of DC sources/DC bus 3 3 2 4
capacitors
(L−1)(L−2) (L−1)
FC 0 2 0 4
5(L−1)
Number of main diodes 2(L − 1) 2(L − 1) 2(L − 1) 4
Number of clamping diodes (L − 1)(L − 2) 0 0 0
TBV/TSV (L − 1)Vdc (L − 1)Vdc (L − 1)Vdc (L − 1)Vdc
Table.7 Comparison of nine level symmetrical inverter topologies

References DC sources Switches Capacitors Diodes TBV (*V DC )
NPC 01 16 00 56 08
FC 01 16 28 00 08
CHB 04 16 00 00 08
[9] 02 12 03 00 14
[10] 04 10 00 03 16.5
[11] 02 12 03 00 14
[12] 02 12 03 00 12
[6] 02 14 02 00 18
[13] 02 14 02 00 15
Proposed 02 10 02 00 08
table, it is concluded that the proposed topology has less number of switches and
less Total Blocking Voltage (TBV).
A. RL Load Case
Figures 4, 5 and 6 show inverter output voltage waveforms with 1:1, 1:2 and 1:3
source configurations, respectively. For generating 9, 13 and 17 levels, the URCS
has been applied. At t = 0.05 s, step changes in Modulation Indexes (MI) from 1
to 0.5 have been applied to all configurations. Table 8 shows simulation parameters
with RL load.
From all source configurations, the number of levels has been varied with respect
to change in MI values. Finally, from simulation results, it is concluded that the
applied URCS provides self-balancing of capacitors even if step changes in MIs.
Fig. 4 Inverter output voltage with 1:1 source configuration using URCS
Table 8 Simulation parameters-RL load case

S. No. Parameters 1:1 source 1:2 source 1:3 source
configuration configuration configuration
(9-level) (13-level) (17-level)
1 DC sources Vdc1 = Vdc2 = Vdc1 = 200 V; Vdc1 = 200 V;
200 V Vdc2 = 400 V Vdc2 = 600 V
2 Load-RL R = 100 , L = 30 mH
3 SC C1 = C2 = 3300μF
4 f0 50 Hz
5 Control scheme URCS/ULSFCS
Table 9 Simulation
S. No. Parameters Value
parameters-grid connected
case 1 Grid voltage 230 V(RMS)
2 Switching frequency 10 kHz
3 Grid frequency 50 Hz
4 Resonance frequency 1416 Hz
5 Rated power (S) 2 kVA
6 Vdc1 = Vdc2 200 V
7 Inverter side inductance 4 mH
8 Grid side inductance 4.3 mH
9 Filter capacitance 6 μF
10 Switched capacitors (C1 = C2) 4700 μF
B. Grid Connected Case-1:1 Source Configuration
The control strategy of Fig. 3 has been applied with the simulation parameters
mentioned in Table 9. For generating grid voltage 230 V (RMS), the peak value
of 400 V with nine level output has been considered and for that Vdc1 and Vdc2
values have been selected as 200 V. Figure 7 shows simulation results for UPF oper-
ation mode with step change in Id(ref) from 10 to 5 A at 0.05 s. From this, firstly,
the capacitors are self-balanced within permissible limits of voltage ripple then nine
level output is obtained with peak value of 400 V. Lastly, the grid voltage and grid
current maintain UPF operation even if step changes in Id(ref) . Generally, Id(ref) is used
to inject active power into the grid. Figures 8 and 9 show the harmonic spectrums of
nine level output voltage and grid current.
Fig. 7 Step changes in id(ref) with 1:1 source configuration: UPF operation mode
Fig. 8 Harmonic spectrum-nine level output voltage
5 Conclusion
With less number of switches, DC sources and capacitors, a new nine level
symmetric source configuration of SC-based MLI topology has been implemented.
This topology also provides binary and trinary source configurations for generating
more number of levels. For all source configurations, the URCS gives good results
to obtain desired output even if step changes in MI values. In this topology, the
capacitors are self-balanced without the need of auxiliary circuits and no backend
H-Bridge is required for generating negative voltage levels. A comparative study is
Fig. 9 Harmonic spectrum-grid current with id(ref) = 10 A
also done with other conventional topologies and finally it provides very less per unit
TSV. With dq frame current control strategy and ULSPD-PWM scheme, the UPF
operation has been obtained at grid side. The drawback of this topology is that it has
only unity voltage gain that means there is no boosting factor ability. This topology
is well suitable for photovoltaic and fuel cell-based grid-connected applications.
References
1. Rodriguez J, Bernet S, Wu B, Pontt JO, Kouro S (2007) Multilevel voltage-source-inverter

topologies for industrial medium-voltage drives. IEEE Trans Ind Electron 54(6):2930–2945
2. Abu-Rub H, Holtz J, Rodriguez J, Baoming G (2010) Medium-voltage multilevel converters—
state of the art, challenges, and requirements in industrial applications. IEEE Trans Ind Electron
57(8):2581–2596
3. Zhang L, Waite MJ, Chong B (2013) Three-phase four-leg flying capacitor multi-level inverter-
based active power filter for unbalanced current operation. IET Power Electron 6(1):153–163
Trans Ind Appl IA-17(5):518–523
5. Kersten A, Theliander O, Grunditz EA, Thiringer T, Bongiorno M (2019) Battery loss and
stress mitigation in a cascaded H-bridge multilevel inverter for vehicle traction applications by
filter capacitors. IEEE Trans Transp Electrific 5(3):659–671
6. Veenstra M, Rufer A (2005) Control of a hybrid asymmetric multilevel inverter for competitive
medium-voltage industrial drives. IEEE Trans Ind Appl 41(2):655–664
7. Vemuganti HP, Sreenivasarao D, Siva Kumar G, Sai Spandana A (2018) Reduced carrier PWM
scheme with unified logical expressions for reduced switch count multilevel inverters. IET
Power Electron 11(5):912–921. https://doi.org/10.1049/iet-pel.2017.0586
org/10.1109/TIA.2005.847306
9. Rajeevan PP, Sivakumar K, Gopakumar K, Patel C, Abu-Rub H (2013) A nine-level inverter
topology for medium-voltage induction motor drive with open-end stator winding. IEEE Trans
Ind Electron 60(9):3627–3636
10. Ali AIM, Sayed MA, Mohamed EEM, Azmy AM (2019) Advanced single-phase nine-level
converter for the integration of multi terminal DC supplies. IEEE J Emerg Sel Topics Power
Electron 7(3):1949–1958. https://doi.org/10.1109/JESTPE.2018.2868734
11. Li J, Bhattacharya S, Huang AQ (2011) A new nine-level active NPC (ANPC) converter for
grid connection of large wind turbines for distributed generation. IEEE Trans Power Electron
26(3):961–972
12. Barbosa P, Steimer P, Steinke J, Winkelnkemper M, Celanovic N (2005) Active-neutral-point-
clamped (ANPC) multilevel converter technology. In: Proceedings of European conference on
power electronics and applications, Dresden, Germany, p 10
13. Chaudhuri T, Barbosa P, Steimer P, Rufer A (2007) Cross-connected intermediate level
(CCIL) voltage source inverter. In: Proceedings IEEE power electronics specialists conference,
Orlando, FL, USA, pp 490–496
14. Rodriguez J et al (2009) Multilevel converters: an enabling technology for high-power
applications. Proc IEEE 97(11):1786–1817. https://doi.org/10.1109/JPROC.2009.2030235
15. Ali A, Shanmugham P, Somkun S (2017) Single-phase grid-connected voltage source converter
for LCL filter with grid-current feedback. In: 2017 international electrical engineering congress
(iEECON), Pattaya, Thailand, pp 1–6.https://doi.org/10.1109/IEECON.2017.8075720
and environment: towards clean energy technologies, pp 1–5. https://doi.org/10.1109/ICEPE5
0861.2021.9404491
Priority-Based Charging of Electric
Vehicles to Prevent Distribution
Transformer Overloading
Arjun Visakh, T. Sornavel, and M. P. Selvan
1 Introduction
Today, electric vehicles (EVs) are considered to be one of the most popular solutions
to the degrading air quality in cities across the world. The use of efficient electric
motors that produce zero tailpipe emissions makes EVs ideal for the urban commute.
As the popularity and penetration of EVs continue to grow, so does the challenge of
meeting their battery energy requirement. Meeting the chage requirement of a fleet
of EVs can overstress the power distribution system.
Allowing EVs to recharge as and when they are plugged in could lead to over-
loads, low efficiency, and voltage sags in the distribution system [1]. To overcome
these challenges and minimize the impact of EV charging, several charge scheduling
schemes have been developed in the past [2]. Some of these seek to prevent equip-
ment overloads by minimizing the contribution of EVs to the peak load [3]. Shifting
the charging of EVs from peak load to off-peak hours can lead to load leveling and
a higher load factor. This can be achieved by minimizing the variance of system
demand from a reference loading level [4]. A few authors have attempted to improve
the efficiency of power delivery by minimizing the power lost in distribution lines [5].
Research on improving the voltage profile using dynamic programming [6] has also
been undertaken. Dynamic pricing-based charge scheduling schemes that minimize
the charging cost borne by EV users [7] are also available in the literature. Thus,
optimization of EV charging may be motivated by a variety of objectives.
A situation in which several EV owners return home from work around the same
time and immediately plug in their EVs can result in the addition of a substantial
charging load on the system when the system demand is already quite high [8]. In
this scenario, as EV loads spread across the network, the total load on the system
will eventually exceed the transformer capacity. Avoiding prolonged overloads on
A. Visakh (B) · T. Sornavel · M. P. Selvan

National Institute of Technology, Tiruchirappalli, Tiruchirappalli, Tamil Nadu, India
e-mail: selvanmp@nitt.edu
296 A. Visakh et al.
transformers, an essential and costly equipment for power distribution, should be

given a top priority. By preventing overloads and reducing the risk of transformer
failure, the distribution network’s reliability can be improved considerably.
In this paper, the charging of EVs is scheduled with the objective of preventing
overloads on the distribution transformer. To prevent overloading, the EV loads
should be scheduled in a manner that reduces their concurrence with the peak load
hours of the system. One such EV management system that delays charging until
sufficient capacity becomes available was trialed in the UK [9]. During those intervals
when the transformer can accommodate some but not all of the EV loads, the EVs
are charged cyclically to distribute the available capacity evenly.
However, sharing the available energy equally among all the plugged-in EVs may
result in failure to meet the charging demand of some users, e.g., those planning long
trips or those plugged-in for a short duration. To avoid this and improve overall cus-
tomer satisfaction, this paper introduces a priority-based charging scheme. Priorities
are assigned to plugged-in EVs based on their initial SOC, desired SOC, arrival time,
and departure time. The priority values, being dependent on the amount of energy
required, and the charging time available reflect the vehicle user’s urgency of charge
requirement.
2 System Modeling
2.1 Network Topology
The test system on which the scheduling algorithm’s performance is assessed is based
on the IEEE 4-node test feeder. The system voltage is lowered from 24.9 kV to 400 V
to represent an Indian distribution system. Fifteen domestic customers are assumed
to be connected at bus 4, the terminal load bus (Fig. 1).
Each residence is supplied by a single-phase 230V connection rated at 3 kW.
The maximum load on the distribution transformer is thus 45 kW. Considering a
25% margin for future loads, the transformer should be capable of supplying a load
up to 56.25 kW. The nearest standard transformer rating is 63 kVA [10]. Thus, the
distribution transformer is rated at 11kV/400V and 63 kVA.
1 2 3 4
Household
Load
EV Load
11 kV 11 kV / 400 V
Distribution Distribution
System Transformer
Fig. 1 Test system

Priority-Based Charging of Electric Vehicles … 297
2.2 Base (Non-EV) Load on the System
The fifteen residential connections are uniformly distributed among the three phases,
i.e., five customers in each phase. Daily load profiles [11] that give the power demand
over 15-min intervals are randomly assigned to each consumer. A power factor of
0.90 lagging is assigned to all the residential loads. The resulting kVA load on the
transformer over a day can be seen in Fig. 2.
2.3 EV Charging Load
The level of EV penetration determines the number of consumers who own EVs. In
this paper, a 100% penetration level is considered, i.e., all the residences have EV
loads. The maximum power limit for EV chargers is taken as 3 kW [12]. The EVs
plugged into the network are modeled as constant power loads operating at unity
power factor for the load flow analysis [12]. All the EVs are assumed to have their
battery capacity equal to 20 kWh.
All the EVs are assumed to recharge their batteries overnight by slow AC charging.
Each EV’s energy demand is decided by the initial battery state of charge (SOC)
upon arrival and the desired SOC at departure. The random nature of EV mobility is
replicated by assigning a normal probability distribution to the EV arrival time. The
probability distribution for arrival time is defined with μ = 6 p.m. and σ = 2 h [13].
The initial SOC values upon arrival are uniformly distributed between 0.15 and 0.4
p.u. All the EVs are expected to be fully charged at departure, which is assumed to
be at 8 a.m. here.
R phase Y phase B phase Total

60
Load (in kVA)
40
20
0
12pm 3pm 6pm 9pm 12am 3am 6am 9am 12pm
Time of day
Fig. 2 Base load on the transformer

2.4 Load Flow Analysis
The transformer loading can be monitored throughout the day by performing load
flow analysis on the unbalanced three-phase network. In this work, OpenDSS soft-
ware was used to solve the test system’s load flow equations. The ‘time-series solu-
tion’ method was employed to run the load flow at the beginning of each 15-min
interval.
3 Priority-Based Charging
3.1 Motivation
Uncoordinated charging represents a situation in which the batteries start recharging

as soon as the EVs are plugged in. Most EV users plug in their cars for charging
upon returning home from work around 6 p.m. [13]. As this coincides with the
system’s peak load hours, the transformer is very likely to get overloaded. Such
an uncontrolled charging scenario was simulated with the parameters specified in
Sect. 2 using MATLAB and OpenDSS. Its impact on transformer loading is shown
in Fig. 3.
A considerable portion of EV charging overlaps with the peak base-load hours. As
a result, the transformer’s peak load is significantly higher than its rated capacity of
63 kVA. It is evident that during peak load hours, it is not feasible for the transformer
to support the charging of all EVs. Thus, there is a need to coordinate and schedule
the charging of EVs so as to stagger their aggregate load on the transformer.
One approach is to share the available transformer capacity cyclically among all
the plugged-in EVs. However, by giving equal preference to all EVs without giv-
ing adequate weightage to the amount of energy required and plug-in duration, the
urgency of charge requirement may get overlooked. Thus, in order to improve cus-
80
63 kVA
Load (in kVA)
60
40
20 Base Load
Uncontrolled EV charging
0
Time of day
Fig. 3 Impact of uncontrolled charging on transformer loading

tomer satisfaction without overloading the transformer, a priority-based scheduling

algorithm is required.
3.2 Methodology
t
Let T denote the set of all 15-min intervals in a day, starting at 12 noon. Let Nconn
denote the set of plugged-in EVs at time slot t (∀t ∈ T). Let N be the maximum
t
number of EVs that can be connected to the system without overloading the trans-
former at time slot t. The values of N t calculated over a day to avoid the transformer
loading from exceeding its rated capacity of 63 kVA are shown in Fig. 4. It is clear the
actual number of EVs connected to the system exceeds the acceptable value during
the peak load hours.
The value of N t is then compared with the number of EVs plugged-in at that
t
interval, i.e., the size or cardinality of the set Nconn . If the number of plugged-in EVs,
i.e., n(Nconn ) is lower than N , then the scheduling algorithm does not intervene,
t t
and the EVs may be charged normally. However, if n(Nconn t

) > N t , then each EV is
assigned a priority value based on its energy requirement and plug-in time. The EVs
t
in the set Nconn are then sorted based on their priority values, and the first N t EVs are
allowed to charge during that interval. The remaining [n(Nconn t
) − N t ] EVs have to
wait for their turn. At the next interval (t + 1), the priority-based sorting and charging
of EVs is repeated. Such a scheduling algorithm ensures that the transformer does
not get overloaded at any interval and prioritizes urgent charging requests.
3.3 Calculation of Priority
During each time slot t (∀t ∈ T), the priority values of all EVs in the set Nconnt
are
calculated to determine their charging status. The formula for calculating priority
comprises three parts. The first part is based on the battery’s current SOC, represented
15 Acceptable
Number of EVs
Actual
10
0
Time of day
Fig. 4 Number of EVs plugged-in to the distribution system

as S OCc . It is given by the formula
p1 = 1 − S OCc (1)
The second part is based on the percentage of charge completed and is given by
S OCc − S OCi
p2 = 1 − (2)
S OCd − S OCi
where the subscripts i, c, and d depict the initial, current, and desired values, respec-
tively. The third part is determined by the slack time. It is evaluated as

tav − tr eq
p3 = 1 − (3)
ttotal
where tav represents the available time left for charging, tr eq represents the time
required to meet the desired charge, and ttotal denotes the total plug-in time of the
EV. The final priority of each EV is then calculated as
p = p1 + p2 + p3 . (4)
4 Results and Analysis
In this section, the impact of uncoordinated charging on the distribution system is

compared with that of priority-based charging. The aggregate EV load on the system
for both cases is compared in Fig. 5. It can be seen that the bulk of the EV load has
been delayed from the peak load hours to the off-peak hours around midnight. The
impact of this staggering on transformer loading is shown in Fig. 6.
The transformer loading during each 15-min interval has now been reduced to a
value below its rated capacity. Furthermore, by charging the EVs according to their
priorities, every owner’s demand for a fully charged battery has been met, as evident
in Fig. 7. It can be observed that the SOC of all the EVs reaches 1 p.u. (full charge)
by 8 a.m., the hour of departure.
Under the prioritized-charging scenario, although the charging start time was
delayed for a few low-priority EVs, the energy requirement of all the vehicles was
met well before their departure time. As all the EVs are charged up to the SOC
desired by their users, customer satisfaction is guaranteed. The SOC profiles in both
scenarios have a constant slope as the charging rate is never altered from the nominal
value of 3 kW.
Load (in kW)

20
10
0
Time of day
(a) Uncontrolled EV charging
Load (in kW)
20
10
0
Time of day
(b) Prioritized EV charging
Fig. 5 Comparison of aggregate EV load on the system
80
63 kVA
Load (in kVA)
60
40
20
Base Load Uncontrolled Charging Prioritized Charging
0
Time of day
Fig. 6 Comparison of transformer loading
5 Conclusion
An EV charge scheduling system that prevents distribution transformers from getting

overloaded is presented in this paper. The EV user’s urgency to charge the battery is
also taken into account while scheduling the charging. The urgency is represented
by EV charging priority, formulated as a function of energy requirement and plug-in
duration. During peak load hours, the scheduling algorithm postpones the charging
of low-priority EVs. Thus, a portion of the EV load gets shifted from on-peak to
off-peak hours without defaulting on any high-priority charging deadline. Although
delayed from the time of plug-in, the charging of low-priority EVs is completed
before the departure time deadline. As a result, the aggregated load on the distribution
transformer does not exceed its capacity at any time, and the charge requirements of
all the EV users are satisfied.
SOC (in p.u.) 1
0.5
0
Time of day
(a) Uncontrolled EV charging
1
SOC (in p.u.)
0.5
0
Time of day
(b) Prioritized EV charging
Fig. 7 Comparison of SOC profile of EVs
References
1. García-Villalobos J, Zamora I, San Martín JI, Asensio FJ, Aperribay V (2014) Plug-in electric
vehicles in electric distribution networks: a review of smart charging approaches. Renew Sustain
Energy Rev 38:717–731
2. Solanke TU, Ramachandaramurthy VK, Yong JY, Pasupuleti J, Kasinathan P, Rajagopalan A
(2020) A review of strategic charging-discharging control of grid-connected electric vehicles.
J. Energy Storage 28:101193
3. Erdogan N, Erden F, Kisacikoglu M (2018) A fast and efficient coordinated vehicle-to-grid
discharging control scheme for peak shaving in power distribution system. J Mod Power Sys
Clean Energy 6(3):555–566
4. Kisacikoglu MC, Erden F, Erdogan N (2018) Distributed control of PEV charging based on
energy demand forecast. IEEE Trans Ind Inf 14(1):332–341
5. Suyono H, Rahman MT, Mokhlis H, Othman M, Illias HA, Mohamad H (2019) Optimal
scheduling of plug-in electric vehicle charging including time-of-use tariff to minimize cost
and system stress. Energies 12(8):1500
6. Dubey A, Santoso S (2015) Electric vehicle charging on residential distribution systems:
impacts and mitigations. IEEE Access 3:1871–1893
7. He Y, Venkatesh B, Guan L (2012) Optimal scheduling for charging and discharging of electric
vehicles. IEEE Trans Smart Grid 3(3):1095–1105
8. Deilami S, Masoum AS, Moses PS, Masoum MA (2011) Real-time coordination of plug-in
electric vehicle charging in smart grids to minimize power losses and improve voltage profile.
IEEE Trans Smart Grid 2(3):456–467
9. Saunders E, Butler T, Quiros-Tortos J, Ochoa LF, Hartshorn R (2015) Direct control of EV
charging on feeders with EV clusters. In: 23rd international conference on electricity distribu-
tion, pp 1–5
10. Power transformers—Part 1: General. Standard, International Electrotechnical Commission,
Geneva, CH (2000)
11. California ISO: Today’s outlook—demand trend (2021). http://www.caiso.com/
TodaysOutlook/Pages/default.aspx
12. Veldman E, Verzijlbergh RA (2014) Distribution grid impacts of smart electric vehicle charging
from different perspectives. IEEE Trans Smart Grid 6(1):333–342
13. Jin C, Tang J, Ghosh P (2013) Optimizing electric vehicle charging: a customer’s perspective.
IEEE Trans Vehicular Technol 62(7):2919–2927
Implementation of Three-Phase Hybrid
Energy System Integrated with UPQC
Shravan Kumar Yadav and Krishna Bihari Yadav
1 Introduction
Many power electronic-based systems are evolved in recent years, and when linked
to power system, such systems cause power quality issues and even decrease system
efficiency [1, 2]. The issue of drawing non-sinusoidal current and reactive electricity
from origin plagues power electronic devices, which are being most commonly used
during automotive, consumer, and residential areas. Some modules attached at the
same point of common coupling (PCC) suffer from voltage variations as a result of
this action [3]. The widespread usage of non-linear flows today has resulted in a slew
of power quality problems, including elevated current harmonics, voltage distortion,
and poor efficiency in the electrical grid [4, 5]. As a result, as a non-linear load
increases, discrete currents are injected into the device. This twisted supply voltage
and current triggers critical equipment to overload, as well as the breakdown of certain
safety features, transmission lines, and engines, as well as wire overloading. As a
result, it is much more important to install compensating equipment that compensate
for harmonic currents and voltages produced by uneven loads [6]. Active power
filters (APFs) have been investigated and established method of resolving this issue
in grid-connected renewable sources [7]. An addition of a green electricity supply to
the grid has a significant impact on the system efficiency. Harmonic frequencies are
pumped into the system mostly as a result of non-linear loads being associated, and
green technology necessitates the usage of reactive electricity upon its system. With
the use of non-linear forces, power requirements may impact utility and consumer
devices. It emphasizes the importance of power efficiency at different stages of
electricity use. Such power quality issues should be addressed in order to have stable
and high-quality power when combining renewable resources (RSs) [8–10]. To fix
S. K. Yadav (B) · K. B. Yadav

Department of EE, NIT, Jamshedpur, Jamshedpur 831014, Jharkhand, India
K. B. Yadav
e-mail: kbyadav.ee@nitjsr.ac.in
306 S. K. Yadav and K. B. Yadav
the significant electric power quality problems in indoor and outdoor, The hybrid
series active filter (The THSeAF) is built on the voltage source form converters.
This electricity quality compensator offers a cost-effective way to address existing
challenges while still incorporating solar energy to provide a long-term flow to the
load. To derive voltage and current harmonics and also voltage disrupts, the controller
uses a creative mixture of the p-q theory and Synchronous Reference Frame (SRF)
[11]. Results in the combining of voltage spikes with a short term voltage fluctuation
behind a dynamical load level can be defined as a distribution system which is mainly
affected by complicated voltage fluctuations. As a result, the dynamic voltage restorer
(DVR) integrated with battery would have sufficient storage to resolve large and
dense complicated voltage fluctuations exist in a supply chain with a PV device [12].
The DVR is a sequence filter which prevents sensitive loads from the electric power
quality issues including voltage swell, sag, harmonics and distortions. This indicates
that DVR will handle power quality problems at the output terminals. Harmonic has
become a significant power quality issue that pollutes the supply chain, allowing end-
user devices to refuse to operate due to voltage, current, or frequency fluctuations
[13].
The architecture and analysis study of a three-phase Hybrid energy system UPQC
is discussed in this article. To increase the variable output during load active current
extraction, a frequency-dependent d-q theory-based regulation is used. The below
are the key benefits of the proposed model:
• Coordination of renewable energy supply and improved power efficiency.
• Output voltage and output current improved at same time.
• Load current compensation has been improved.
The output of proposed devices is thoroughly examined using MATLAB-Simulink
tools.
The following is the general structure of the work: The literature work is depicted
in Sect. 2. Section 3 outlines the proposed control structure, while Sect. 4 defines
the device architecture and setup. In comparison, Sect. 5 addresses the observations,
and Sect. 6 takes the article to an end.
2 Literature Survey
Initially, that Shunt active power filter (SAPF) paradigm was touted as the ideal
method for reducing harmonic pollution [14], but its control algorithms can operate
harmonic removal, DC-link capacitance manipulation, current power, and command
signal strategies are both looked at and discussed. During non-ideal grid voltage envi-
ronments, an inverter of 3-level neutral-point clamped diode (NPC)-based filter offers
effective harmonic distortion stabilization and power factor correction [15]. The
recommended model is called the dual fundamental component extraction (DFCE)
method. The suggested process derives fundamental voltage and current elements
that are needed at the very same time in order to generate reference current and
Implementation of Three-Phase Hybrid Energy … 307
synchronization stages, accordingly. This literature includes a detailed analysis of

efficiency in [16] power systems, including DC and green energy systems. Technolo-
gies for measuring power condition and potential solutions to power quality problems
are thoroughly investigated. Then, we look at how customized energy devices like
D-STATCOM, UPQC, UPS, TVSS, and DVR can be used to mitigate these issues for
microgrid systems. STATCOM can be a viable option for renewable energy systems
due to its numerous benefits, while spinning reserve may improve power efficiency in
conventional systems. [16, 17] Introduces a 3F UPQC with efficient architecture. The
sequence compensator employs a basic DVR with PLL. Applying the d-q principle
to the controller portion of the shunt compensator, on the other side, improves the
efficiency of the UPQC. A series compensator based on a basic hysteresis controller
is often used. The suggested system’s key features are that (a) the voltage and current
are sustained concurrently, and (b) under a variety of harmonics, sags/swells, and
unstable situations, reliability is guaranteed. Fuel Cell Integrated UPQC (FCI-UPQC)
is used to improve the power quality in [18] 3F four-wire distribution grid. On the
shunt line, the planned FCI-UPQC with 4-leg converter, and on the sequence side, it
has 3-leg converter. Enhanced integration of renewable energy sources [19] into the
electricity network, as well as the spreading of uneven loads in the home and industry,
have exacerbated power quality problems and rendered coordination more difficult
than before. Extended renewable energy sources into the electricity network, as well
as the spreading of industrial and domestic critical loads, have exacerbated electric
power quality problems and rendered coordination more difficult than before. To
address this issue, a number of advanced PLLs with improved disturbance rejection
capabilities were developed by various researchers. Almost all of these PLLs may
be thought of traditional SRF-PLL with additional filters that can be added within or
before the SRF-PLL control loop. As the DG device is positioned between the grid
and either generic load, UPQC acts as a symmetric device. The DG device conducts
active power-line condition when delivering the energy provided by the PV into the
grid in [20, 21] grid-connected mode. To gain a complete function of the system
function and properly build the electrical power, a comprehensive power analysis
through the PV-UPQC is needed.
3 System Design and Configuration
Figure 1 depicts the HES-structure. UPQC’s three-phase framework is needed for the
HES-UPQC. Shunt and series compensators are attached to a standard DC-bus in the
HES-UPQC. At the load side, the shunt compensator is associated. The photovoltaic,
BESS, and wind energy systems are all linked to the UPQC through a standard dc
bus.
Via a dc-dc converter, the solar PV array, wind, and BESS are connected directly to
UPQC’s DC-link. The grid electric voltage sags/swells are compensated by the series
compensator when voltage is in regulation mode. The series and shunt compensators
Fig. 1 System configuration HES-UPQC
are connected to grid by integrating inductance. Through the series injection trans-
former, the voltage created by series compensator is injected into the grid. Harmonics
produced by the switching operation of converters are filtered with ripple filters.
Both a series and shunt compensator may be called a voltage source inverter (VSI)
with a balanced voltage source. Insulated gate bipolar transistors (IGBTs) with the
DC capacitors make up the circuit. The extreme dynamic energy required in DC
connection the voltage is a significant factor in VSI-based hybrid filter mitigation.
4 PV-UPQC Control
The series compensator and shunt compensator are the two primary modules of HES-
UPQC. Shunt compensator corrects issues with load control efficiency including
the load current harmonics and the reactive power. In the situation planned, the
shunt compensator in UPQC also acts as the power source for the solar PV collec-
tion. To derive the power of PV collection, the algorithm used is maximum power
point tracking (MPPT) by shunt compensator. Supplying sufficient voltage in phase
prevents the module from power quality issues.
4.1 Shunt Compensator Control
Through running the solar at its extreme energy point, shunt compensator draws the
extreme power from it. Figure 2 shows the overall layout of the SOGI-FLL dependent
control algorithm. The algorithm of MPPT is used to run the Photovoltaic array in
such a manner that the full electric power is provided. The technique SOGI-FLL
is used to remove the quadrature portion of basic load present. MPPT power and
compensator switching control are part of the control method. To produce reference
DC connection voltage, the MPPT control is used.
MPPT Control Approach
Full Power Point is discovered using the steepest descent method (MPP). The steep
descent path is used in the different phase size than traditional approach, which are the
key benefits which improves convergence significantly. P&O algorithm is known as
the gradient descent process, because it is optimization technique. MPP is calculated
as follows [22]:
(d PP V /d VDC )V =VDC (k)

VDC (k + 1) = VDC + (1)
kε
where k ε is the phase volume corrector, which defines how steep is each step
which is in the direction of gradient and the inverter’s DC connection voltage. (dp/dv)
represents the MPP variance in each phase. The dp/dv ratio is determined as follows:
Fig. 2 Architecture of shunt compensator control

d Ppv
= f (VDC (k), Ppv (k)) (2)
d VDC
Inverter Switching Control

The switching process involves calculating the magnitude of PCC voltages, computer
models, collecting process reference currents, a leakage element, and feed-forward
part of PV electricity.
Fundamental positive sequence extraction
Compensation would not be sufficient if residual voltages are used explicitly to
measure reference filter currents. To produce comparison filter currents, the basic
positive sequence components of three-phase voltages are collected.
Component Loss and Photovoltaic Feed Forward
In control system, the Photovoltaic (PV) feed-forward component can be used to
account for differences in solar generation and voltage magnitude. PV feed-forward
unit is calculated using PV power and PCC voltage [23].
2Ppv
f =
2
I pv (3)
3Vs
The loss component (I loss ) is obtained by Proportional Integral (PI) controller

with the feedback of a imbalance signal between V DCref and V DC . The loss factor is
assessed as follows:
k1
Iloss = (K p + )(VDCr e f − VDC ) (4)
s
Fundamental Load Current Estimation

Figure 3 demonstrates the arrangement of SOGI-FLL. SOGI-FLL contains two
blocks the updated SOGI and the frequency locked loop (FLL). The controller FLL
calculates grid operator frequency (iq ) by an in-phase component (id ) and a quadra-
ture element. The feature of the input electric current iL (with the offset DC of “A”
and magnitude Ipa ) is as follows [3]:

Iap ω A
I L (s) = 2 + (5)
s + ω 2 s
The transition parameters for I d and I q are shown in Fig. 3.

kω s
Id (s) = I L (s) (6)
s 2 + kω s + ω 2
Fig. 3 SOGI-FLL
ks 2
Iq (s) = I L (s) (7)
s 2 + kω s + ω 2
Basic load terms are derived from these I d and I q coefficients, as seen in Fig. 3.
The in-phase of fundamental load portion has no impact on control algorithm’s
output, in the case of the offset DC. The load current is orthogonal to the quadrature
portion (iq ). The FLL controller takes a quadrature portion (iq ) of the basic load
component and the error (between iL and id ), for grid frequency estimation. Further-
more, in steady-state operating conditions, FLL output is unaffected by the existence
of a DC offset in the load portion. As a consequence, the framework that adapts to
frequency of SOGI-FLL is efficient and accurate for estimating grid frequency. Zero
crossing detectors estimate the magnitude of three basic load currents processes (I ap ,
I bp , and I cp ). Load currents’ net load portion (I load ) is measured as follows:
Iap+ Ibp+ Icp

Iload = (8)
3
The feed-forward, load, and loss components are used to calculate the net weight
of the reference grid present. The cumulative weight present in the reference grid is
calculated as follows [23]:
Itot = Iload − I pv f 2 − Iloss (9)

Reference Current Control
i ar = t pa ∗ Itot (10)
i br = t pb ∗ Itot (11)
i cr = t pc ∗ Itot (12)
The error current is measured from the reference and the sensed grid electric
currents and provided to hysteresis current controller, which regulates the VSI
switching pulses.
4.2 Series Compensator Control
Compensation of in-phase, pre-sag, and energy-optimal compensation which are the

three management strategies of series compensator [19, 20] includes a comprehensive
overview of the multiple compensation mechanisms used for sequence compensator
power. The voltage is injected by the series compensator in the same step as the grid
voltage in this job, resulting in the series compensator injecting the least amount of
voltage necessary.
Figure 4 depicts the sequence compensator’s control system. The quantization of
the PCC voltage obtained by PLL is used to produce the zero axis in the dq-0 domain.
Fig. 4 Control architecture of series compensator

The voltage and load from the PCC are translated to the d-q-0 domain. At zero, a part
of the q-axis is held. Series compensator reference voltage is defined by discrepancy
between PCC voltage and the load reference voltage. The real voltages of series
compensator are determined by the discrepancy between the load voltage and the
PCC voltage. PI controllers use the disparity between reference and real voltage of
series compensator to produce acceptable reference signals. To produce sufficient
gating signals for inverter, such reference signals are translated in abc-domain and it
is passed via a PWM controller.
The HES-UPQC framework in the MATLAB-Simulink software is used to test its

effects and performance. The non-linear load used is made up of a 3F diode rectifier
bridge connected with the R-L load. The simulation was run with a 50e−6 s solver
phase scale. The machine is exposed to a number of complex situations, including
PCC voltage sags and swells.
Figure 5 depicts the complex output of PV-UPQC in the presence of PCC voltage
swells/sags. Irradiation level (G) is maintained at 1000 W/m2. DC-link voltage (Vdc),
PCC voltages (Vs), load voltages (VL), grid currents (is), load currents in each line
(iLa, iLb, iLc), and load currents (iL) are shown in Fig. 5.
6 Conclusion
The configuration and output of three-phase HES-UPQC have been examined

under grid voltage sags/swells. The system’s accuracy has been checked using
MATLAB/Simulink. It has been discovered that HES-UPQC mitigates the harmonics
caused by non-linear- loads. The use of control techniques has improved the effi-
ciency of d-q control, especially in load unbalanced conditions. It can be shown that
HES-UPQC, by combining distributed generation with power quality enhancement,
is a successful option for current delivery systems.
Fig. 5 Performance of HES-UPQC

References
1. De Yong D, Bhowmik S, Magnago F (2017) Optimized complex power quality classifier using
one vs. rest support vector machines. Energy Power Eng 09(10):568–587
2. Li D, Yang K, Zhu ZQ, Qin Y (2017) A novel series power quality controller with reduced
passive power filter. IEEE Trans Ind Electron 64(1):773–784
3. Modesto RA, da Silva SAO, de Oliveira Junior AA (2015) Power quality improvement using a
dual unified power quality conditioner/uninterruptible power supply in three-phase four-wire
systems. IET Power Electron 8(9):1595–1605
4. Swain SD, Ray PK, Mohanty KB (2017) Improvement of power quality using a robust hybrid
series active power filter. IEEE Trans Power Electron 32(5):3490–3498
5. Javadi A, Hamadi A, Woodward L, Al-Haddad K (2016) Experimental investigation on a hybrid
series active power compensator to improve power quality of typical households. IEEE Trans
Ind Electron 63(8):4849–4859
6. Li S, Qi W, Tan SC, Hui SR (2016) Integration of an active filter and a single-phase AC/DC
converter with reduced capacitance requirement and component count. IEEE Trans Power
Electron 31(6):4121–4137
7. Tareen WU, Mekhilef S, Seyedmahmoudian M, Horan B (2017) Active power filter (APF)
for mitigation of power quality issues in grid integration of wind and photovoltaic energy
conversion system. Renew Sustain Energy Rev 70:635–655
8. Kasa S, Ramanathan P, Ramasamy S, Kothari DP (2016) Effective grid interfaced renewable
sources with power quality improvement using dynamic active power filter. Int J Electr Power
Energy Syst 82:150–160
9. Eltigani D, Masri S (2015) Challenges of integrating renewable energy sources to smart grids:
a review. Renew Sustain Energy Rev 52:770–780
10. Mira MC, Zhang Z, Knott A, Andersen MA (2017) Analysis, design, modeling, and control
of an interleaved-boost full-bridge three-port converter for hybrid renewable energy systems.
11. Javadi A, Woodward L, Al-Haddad K (2018) Real-time implementation of a three-phase
THSeAF based on a VSC and a P+R controller to improve the power quality of weak distribution
systems. IEEE Trans Power Electron 33(3):2073–2082
12. Mansor MA, Othman MM, Musirin I, Noor SZM (2019) Dynamic voltage restorer (DVR) in a
complex voltage disturbance compensation. Int J Power Electron Drive Syst 10(4):2222–2230
13. Farooqi A, Othman MM, Abidin AF, Sulaiman SI, Radzi MAM (2019) Mitigation of power
quality problems using series active filter in a microgrid system. Int J Power Electron Drive
Syst 10(4):2245–2253
14. Hoon Y, Radzi MAM, Hassan MK, Mailah NF (2017) Control algorithms of shunt active power
filter for harmonics mitigation: a review. Energies 10(12)
15. Hoon Y, Radzi MAM, Hassan MK, Mailah NF (2018) Operation of three-level in-verter-
based shunt active power filter under non ideal grid voltage conditions with dual fundamental
component extraction. IEEE Trans Power Electron 33(9):7558–7570
16. Hossain E, Tur MR, Padmanaban S, Ay S, Khan I (2018) Analysis and mitigation of power
quality issues in distributed generation systems using custom power devices. IEEE Access
6(c):16816–16833
17. Rohouma W, Balog RS, Peerzada AA, Begovic MM (2020) D-STATCOM for harmonic miti-
gation in low voltage distribution network with high penetration of nonlinear loads, Renew
Energy 145:1449–1464
18. Hasan K, Othman MM, Rahman NFA, Hannan MA, Musirin I (2019) Significant implication
of unified power quality conditioner in power quality problems mitigation. Int J Power Electron
Drive Syst 10(4):2231
19. Sundarabalan CK, Puttagunta Y, Vignesh V (2019) Fuel cell integrated unified power quality
conditioner for voltage and current reparation in four-wire distribution grid. IET Smart Grid
2(1):60–68
20. Golestan S, Guerrero JM, Vasquez JC (2017) Three-phase PLLs: a review of recent advances.
21. Chen J et al (2021) Distributed control of multi-functional grid-tied inverters for power quality
improvement. IEEE Trans Circuits Syst I Regul Pap 68(2):918–928
22. Yadav SK, Raj S, Suman S, Madhuri M (2014) Design of low power Mppt with digital control
for thermo PV generator. IOSR J Electr Electron Eng 9:17–26
23. Devassy S, Singh B (2016) Design and performance analysis of three-phase solar PV integrated
UPQC. In: 2016 IEEE 6th international conference on power systems (ICPS), pp 1–6
Design of Permanent Magnet Brushless
DC Motor for Electric Vehicle Traction
Application
Sandeep Vuddanti, Sharankumar Shastri, and Surender Reddy Salkuti
1 Introduction
Nowadays many technologies have been developed to replace conventional internal

combustion (IC) engines, which are among the primary producers of greenhouse
gas emissions. Environmentally safer technologies are being researched and devel-
oped and the most prominent advancements have been in electric vehicles in the past
10–15 years. Electric vehicles (EVs) could play a major role in the development of
vehicles to replace IC engines, leading to a reduction of greenhouse gas emissions
[1]. The increased demand for electric vehicles brings in the need to improve their
operation to the maximum by employing proper parts like batteries, motors, trans-
mission, etc. Due to the everlasting need for higher efficiency, the utilities look to
implement compact motors with higher power density and efficiency. This brings in
a need to design high-efficiency motors specifically for this application, considering
design requirements and constraints. The introduction of newer materials, faster
electronic switches, and processors made it possible to develop and improve effi-
cient special electrical machines such as PM synchronous machines (PMSMs) [2].
PMSMs possess high torque density, high efficiency, less maintenance, and simplicity
of structure. These features make the PM synchronous motors suitable for traction
purposes, and they are also in the renewable energy sector, e.g., wind turbines [3].
S. Vuddanti
Department of Electrical Engineering, National Institute of Technology Andhra Pradesh (NITAP),
Tadepalligudem, India
S. Shastri
Department of Electrical Engineering, Indian Institute of Technology Delhi (IITD), New Delhi,
India
S. R. Salkuti (B)
Department of Railroad and Electrical Engineering, Woosong University (WSU), Daejeon,
Republic of Korea
e-mail: surender@wsu.ac.kr
318 S. Vuddanti et al.
PMSMs can be classified based on the wave shape of the back-electromotive force
[4]. They are further classified into permanent magnet brushless DC (PMBLDC)
motors and permanent magnet synchronous machines (PMSMs). PMSMs generate
a sinusoidal BEMF, while PMBLDC motors produce a trapezoidal one.
PMBLDC motors are synchronous motors that are fed DC input from an inverter
to produce AC to drive each phase of motor. These controllers help provide pulses
of current to motor windings, hence controlling the torque and speed of the motor.
PMBLDC motors have gained popularity since their inception in 1975, and are
being used in a wide number of applications due to factors like high efficiency
and simple structure [5]. Permanent magnet machines have recently been used for
more demanding applications, mainly due to the extensive use of low-cost power
electronics switches [6].
Reference [7] describes the applications, principles and topologies, materials,
electric circuit parameter computations, equivalent circuits, specifications and design
principles (above and below 100 kW), design optimization and performance analysis
for Induction motors, from which the stator design of the PM motor is mainly based
on. Reference [8] presents the methods to design permanent magnet motors. The
sizing procedure and main dimension calculations along with performance calcula-
tions of the designed motor are mentioned. Drive topologies for the permanent magnet
motors are also described. The rotor analysis, calculation of maximum magnet thick-
ness, width, etc., have been adapted from here. The magnetic flux accumulated in
this type of internal PM configuration is higher than a surface permanent magnet of
the same dimensions. Simulations are compared to the hardware of the motor.
An approach for the design of AC PM motors for EV traction is proposed in [9].
Reference [10] presents the modeling of a 10 kW BLDC Motor controller by using
the PSIM software. Reference [11] reviews the relevant literature on the vehicle
dynamics and analyzes the same for a three-wheeler auto-rickshaw, and its behavior
under varying drive cycle load. Extensive simulation is done on the auto-rickshaw
replacing its drivetrain with that of a parallel hybrid drivetrain. A detailed insight into
the design parameters and performance analysis for the permanent magnet motors is
one of the main references for the design formulae taken into account in Ref. [12].
The major contributions of this paper are as follows:
• A study on the current types of powertrains utilized in electric vehicles is presented
in this paper. Powertrain study includes the study on batteries and motor drives.
A brief insight about PMBLDC motors and their comparison with other motors
is also described.
• The dynamics of the vehicle based on a standard vehicle model are presented
based on the amount of power required for the motor.
• Dynamics of an electric vehicle over simulation is proposed to design and assess
the performance of a PMBLDC motor for the EV application.
• Calculation of motor dimensions using analytical calculations using references
is presented. The structure and the design procedure of the required PMBLDC
motor are addressed.
Design of Permanent Magnet Brushless DC Motor … 319
• Simulations of the analytically designed motor using ANSYS’s Maxwell package

are presented and compared with the simulation results obtained using RMXprt
(analytical) and Maxwell 2D (FEM-based). The design is optimized by varying
parameters.
• Compared the simulation of the assessed motor with FEM simulation to obtain a
model with the least error.
The organization of this paper is as follows. Section 2 presents the description
of brushless DC motors and the analysis of vehicle dynamics. Motor design using
analytical calculations is presented in Sect. 3. Section 4 describes the motor design
using ANSYS Maxwell software. Simulation results using RMxprt and ANSYS 2D
software are included in Sect. 5. Conclusion is drawn in the last section.
2 Brushless DC Motors and Analysis of Vehicle Dynamics
In this section, different motors used in EVs are compared and the focus is made on
BLDC motors. The efficiency and the reason for the choice of BLDC motors due to
their advantages are also covered. This section also estimates the power required by
the motor to run the EV at the speeds given in the Indian Driving Cycle and takes
into account all the forces acting on the vehicle [13].
2.1 Brushless DC Motors
Here the primary focus is on the application of BLDC motors in EVs, the following
part explains the construction and operating principle of brushless DC motors.
The advantages of using BLDCs include high operating speeds, low maintenance,
compact size, 90+% efficiency, no brush sparking, quick response, rotor made
of permanent magnets, less rotor inertia, linear speed/torque curve produces a
predictable speed regulation, brush inspection is eliminated, quieter operation, there-
fore, reduced electromagnetic interference, regenerative braking has good efficiency
[14]. As one can notice, the benefits of using BLDC motors are higher, especially
from its brushed counterparts and Induction motors, making it more suitable for EV
applications.
2.2 Analysis of Vehicle Dynamics
With all EVs, it is very important to predict the performance and range before it is
manufactured. Several types of models exist to predict vehicle performance through
simple calculations. Vehicle performance modeling includes the prediction of several
Fig. 1 Vehicle dynamics
parameters like acceleration, top speed, range, power, etc. It becomes essential to
necessitate the electric vehicle with enough performance to properly blend in with
general traffic. It is also essential to model the EV before manufacturing individual
parts to avoid over-powering or under-powering the system. This becomes a very
important method to cut the costs of the EV [15]. The performance parameter we
need for the motor is the ‘power required’, to predict what the power of the designed
motor should be. This helps to avoid costs in production and allows the motor to run
at rated power to give maximum efficiency. Figure 1 depicts the vehicle dynamics.
The model used here to evaluate performance is called the force model/tractive
effort model [16, 17]. The propelling force of vehicle forward, the tractive effort
should be able to accomplish greater than the aerodynamic drag, greater than the
rolling resistance, accelerate the vehicle, and provides the force needed to overcome
the vehicle’s weight component along the slope. It depends mainly on the frontal area
and how wind opposes this area [18]. The mathematical equation for the aerodynamic
drag is given by,
1
Fw = ρ A f Cd (V + Vw )2 (1)
2
where Vw is a component of the wind speed in the vehicle’s moving direction, which
has a positive sign when the wind is against the vehicle motion, and negative when
the wind is along the vehicle direction [19]. ρ is the air density, A f is the vehicle
frontal area, V is the vehicle speed, Cd is the aerodynamic drag coefficient. When the
tire is rolling, the tire causes the asymmetric distribution of ground reaction forces,
which leads to a resistance in the wheels opposing the rolling of the wheel. The
mathematical equation for the rolling resistance is given by [20],
Fr = μrr MV g (2)
where μrr is rolling resistance coefficient, g is the acceleration due to gravity and
MV is the mass of the vehicle. The typical value of μrr is 0.015 for a radial-ply tire
[21]. The force required by the vehicle when it is accelerating, i.e., when its velocity
is changing. The linear acceleration of the vehicle is given by,
Fa = MV a (3)
where a is the acceleration of the vehicle (m/s2 ). The grading resistance can be
expressed as [22],
Fg = MV g sin α (4)
where α is the road slope angle.
2.3 Net Tractive Effort and Power Required
Net tractive effort (FT E ) now becomes the sum of all the forces acting on vehicle
[23], and it is expressed as,
FT E = Fw + Fr + Fa + Fg (5)
where Fr is rolling resistance force, Fa is acceleration force, Fw is the aerodynamic

drag force, Fg is grading resistance force. The total power required to overcome
the net tractive effort is equal to the force times the velocity of the vehicle at each
recorded instant of time [24]. The average power can be calculated and the power of
the motor can be estimated.
P = FT E V (6)
where FT E (N) is the total tractive effort required to overcome the net resistance by
the vehicle, and V (m/s) is the velocity of the vehicle.
2.4 Vehicle (E-Rickshaw) Model and Specifications
In this section, the amount of power required by the motor to run the electric vehicle
at the speeds given is estimated in the Indian Driving Cycle and takes into account
all the forces acting on the vehicle. The reference for the rickshaw has been taken
from the Bajaj RE 2S model [25, 26] manufactured by Bajaj Auto Limited® , and it is
depicted in Fig. 2. The vehicle is equipped with a 145.5 cc single-cylinder (2-stroke)
Fig. 2 Bajaj RE 2S rickshaw as a reference model
engine. The transmission consists of a manual gearbox (4-speed) with reverse gear
and a wet-plate clutch. The curb weight of the vehicle is 284 kg and the maximum
weight (gross) is about 610 kg [27]. Different parameters of Bajaj RE 2S rickshaw
are presented in Table 1.
The analysis can be done through basic calculations, either by using software like
MATLAB or even Excel. The obtained aerodynamic drag force (Fw ) is 106.33 N ,
rolling resistance (Fr ) is 60.329 N , and grading resistance (Fg ) is 0 N. Driving cycles
are plots of the speed of the vehicle versus time, i.e., helps us visualize the speed
of the vehicle at every second. The driving cycles provide us with a more realistic
and practical test for vehicle testing than analyzing the slope or the distance covered
by the vehicle which could provide a wrong result during simulation [28, 29]. The
drive cycle adopted is called the modern Indian driving cycle which is developed by
Table 1 Different parameters

Component/Parameter Values
of Bajaj RE 2S rickshaw
Engine Single cylinder, 2 strokes, 145.5 cc
Tires (Front-rear, R) 4.00-8, 4PR
Dimensions (W × L × H) 1300 mm × 2625 mm × 1710 mm
Drag coefficient (Cd ) 0.45
Rolling resistance (μrr ) 0.01
Frontal area (A f ) 2 m2
Fig. 3 Modified Indian 60

driving cycle
50
Velocity (in kmph)

40
30
20
10
0
0 5 10 15 20 25
Time (in seconds)
ARAI, India to simulate the realistic driving models for the Indian roads. Figure 3
depicts the modified Indian Driving Cycle.
The values are then added in an Excel worksheet (or a MATLAB code or Simulink
model) [30]. The calculations were done and added to excel to calculate the solutions
at once. The final power values were obtained and the average was taken to be 1530 W
or 1.5 kW.
3 Motor Design Using Analytical Calculations
In this section, the design of a surface PMBLDC motor for e-rickshaw application
is discussed. Motor design is a rigorous process and needs many parameters to
be generally assumed due to their validity due to industry usage. Also, generally,
several iterations run to obtain the correct parameters and their desired values [31].
This continuous re-evaluation is what makes motor design rigorous and the whole
process can be semi-automated using scripting, motor simulations in software like
ANSYS Maxwell® , FEMM® , MotorSolve® , etc.
The quarter motor cross-section showing all the dimensions depicted in Fig. 4. In
this figure, Din is the stator inner diameter, Dout is the stator outer diameter, Dr is the
rotor outer diameter, hm is the magnet thickness, g is an air gap, bs2 is the slot width
top, bs1 is the slot width bottom, wt is the tooth width, hcs is the back iron length,
(hos + hw + hs ) is the slot depth [32].
Initial parameters were assumed and the calculations were carried out by using
the standard Refs. [33, 34]. The predicted results using analytical calculations for the
design of brushless DC machines are shaft torque (T), which is equal to 4.77 Nm;
the predicted power, which is equal to 1500 W; the predicted current, which is equal
to 6.57A; RMS armature current, which is equal to 5.36 A; the applied voltage to
the motor, which is equal to 120 V; the resistance obtained per phase, which is equal
to 0.25 ohms/phase; and the maximum back emf, which is expected to be 120 V.
These values can now be added to simulation results and checked for the validity of
Fig. 4 Quarter motor cross-section showing all the dimensions
the analytical method [35]. In this paper, the motor is designed using the ANSYS
Maxwell software [36, 37].
This section presents the obtained simulated results for the analytically designed
motor, using FEA software, ANSYS Maxwell. Its modules RMxprt and 2D are used
to simulate the motor and obtain more accurate outputs than the ones obtained from
analytical calculations.
4.1 Simulation Results Using RMxprt
The design sheet for RMxprt output for PMBLDC motor design is presented next.
The general data for motor design is: rated output power is 1.5 kW, rated voltage is
120 V. The stator data is as follows: the number of stator slots is 24, the outer diameter
of the stator is 120 mm, the inner diameter of the stator is 69 mm, type of stator slot
is 3. The stator slot having hs0 is 1 mm, hs1 is 2 mm, hs2 is 11.77 mm, bs0 is 2 mm,
bs1 is 4 mm, bs2 is 5 mm, rs is 1 mm, top tooth width is 5.8199 mm, bottom tooth
width is 7.90104 mm, skew width (number of slots) is 0, length of the stator core is
50 mm, stacking factor of the stator core is 0.95, and type of steel is steel_1010. The
slot insulation thickness is 0.5 mm, layer insulation thickness is 0 mm, end length
adjustment is 0 mm, the number of parallel branches is 1, the number of conductors
per slot is 16, the number of coils is 11, average coil pitch is 5, the number of wires
per conductor is 2, wire diameter is 0.912 mm, wire wrap thickness is 0 mm, slot
area is 65.5358 mm2 , net slot area is 38.6844 mm2 , limited slot fill factor is 75%,
stator slot fill factor is 68.8025%, and coil half-turn length is 125.28 mm.
The rotor data is as follows: minimum air gap is 0.75 mm, inner diameter is
30 mm, length of rotor is 50 mm, stacking factor of the iron core is 0.95, the material
is steel-1010, polar arc radius is 33.75 mm, mechanical pole embrace is 1, electrical
pole embrace is 0.968883, the maximum thickness of magnet is 2.5 mm, the width of
the magnet is 51.0509 mm, type of magnet is NdFe35, and there is no magnetic shaft.
Permanent magnet data is as follows: residual flux density is 1.23 T, coercive force is
890 kA/m, maximum energy density is 273.675 kJ/m3 , relative recoil permeability
is 1.09981, the demagnetized flux density is 7.50732e-005 T, recoil residual flux
density is 1.23 T, and recoil coercive force is 890 kA/m.
Full-load data is as follows: average input current is 13.4992 A, root mean square
armature current is 12.4873 A, armature thermal load is 211.426 A2 /mm3 , specific
electric loading is 22.1208 A/mm, armature current density is 9.55781 A/mm2 , fric-
tional and windage loss is 0 W, an iron-core loss is 0.00215665 W, armature copper
loss is 118.566 W, the total loss is 119.89 W, the output power is 1500.02 W, input
power is 1619.91 W, efficiency is 92.5989%, rated speed is 2737.67 rpm, rated
torque is 5.23223 Nm, locked-rotor torque is 35.4585 Nm, and locked-rotor current
is 236.321 A. The RMxprt modeled figure is depicted in Fig. 5.
When all the above-obtained parameters are added into the Ansys RMxprt simu-
lation module, the simulation results from RMxprt for 1500 W, 120 V, 24 slot BLDC
motor are presented next.
The efficiency versus speed curve is presented in Fig. 6. From this figure, it can
be noticed that the efficiency of the system at rated speed is quite high (95% +).
Figure 7 depicts the output torque versus speed curve. From this figure, it can be
noticed that the starting torque of the motor is quite high due to large starting currents
in BLDC (due to low BEMF), but at rated speed, the torque is about 5 Nm, which is
the desired value.
Cogging torque is caused due to the interactions between the stator and magnet
poles. This is more prominent at lower speeds due to the longer time of interaction
Fig. 5 RMxprt modeled

figure
ANSOFT
100.00 Curve Info
Efficiency
80.00
60.00
(%)
40.00
20.00
0.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 3500.00
n (rpm)
Fig. 6 Efficiency versus speed characteristics
between the stator and rotor. Cogging torque output in two teeth is depicted in Fig. 8.
The cogging torque here is quite high and can be efficiently reduced/optimized by
using fractional slot/pole ratio, skewing of rotor magnets, skewing of stator teeth,
and optimum rotor arc width.
Figure 9 depicts the air gap flux density. From this figure, it can be observed that
an air gap flux density of about 0.75 T, which is lesser than the 0.9 T assumed in the
calculations.
Figure 10 depicts the induced voltages at rated speed. Induced back emf wave-
forms, which are exactly equal to 99.6 V as estimated (ε = 0.83), after the losses are
considered.
Figure 11 depicts the winding currents at full load. The peak value of winding
currents at full load is about 18 A, which is normal.
ANSOFT
37.50 Curve Info
Output Torque
25.00
(N.m)
12.50
0.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 3500.00
n (rpm)
Fig. 7 Output torque versus speed
ANSOFT
0.75 Curve Info
Cogging Torque
0.50
0.25
(N.m)
0.00
-0.25
-0.50
-0.75
0.00 125.00 250.00 375.00
Electric Degree
Fig. 8 Cogging torque output
The winding voltages under load conditions are depicted in Fig. 12. The winding
voltages are not purely square waves, this waveform has to be enhanced by properly
adjusting the lead angle and the drive configuration in the Maxwell circuit editor.
A 2D design is created from the same RMxprt module and the outputs are
analyzed. Due to FEA, 2D and 3D analyses provide more accurate results due to
meshing.
ANSOFT
0.75 Curve Info
Air-Gap Flux
Density
0.50
0.25
(Tesla)
0.00
-0.25
-0.50
-0.75
0.00 125.00 250.00 375.00
Electric Degree
Fig. 9 Air gap flux density
ANSOFT
100.00 Curve Info
Phase Voltage ea
Line Voltage eab

75.00
50.00
25.00
(Volts)
0.00
-25.00
-50.00
-75.00
-100.00
0.00 125.00 250.00 375.00
Electric Degree
Fig. 10 Induced voltages at rated speed
4.2 Simulation Results Using ANSYS 2D
2D cross-section of the designed motor is depicted in Fig. 13. When the RMxprt
module is exported to 2D simulation, then the obtained results are presented next.
The torque output of the designed motor is depicted in Fig. 14. From this figure, it
can be observed that the torque output is about 5 Nm as estimated (average), but due
to high cogging torque, there is a high rippling that takes place. This is undesirable
72 torque and needs to be reduced. Torque ripple can be reduced by efficient driving
methods and skewing.
ANSOFT
20.00 Curve Info
Source Current
Phase Current ia
15.00
Phase Current ib
Phase Current ic
10.00
(Amperes)
5.00
0.00
-5.00
-10.00
-15.00
-20.00
0.00 125.00 250.00 375.00
Electric Degree
Fig. 11 Winding currents at full load
ANSOFT
125.00 Curve Info
Phase Voltage va
Line Voltage vab
62.50
(Volts)
0.00
-62.50
-125.00
0.00 125.00 250.00 375.00
Electric Degree
Fig. 12 Winding voltages at full load
Figure 15 depicts the winding currents under load. From this figure, it can be
observed that the currents at rated speed, under load, about 13 A average and 12 A
rms current.
Here, the simulated results are obtained for the analytically designed motor using
FEA software, ANSYS Maxwell. Its modules RMxprt and 2D are used to simulate
the motor and obtain more accurate outputs than the ones obtained from analytical
calculations.
Fig. 13. 2D cross-section of the designed motor
Torque Maxwell2DDesign2 ANSOFT

8.75 Curve Info
Moving1.Torque
Setup1 : Transient
Moving1.Torque [NewtonMeter]
6.75
4.75
2.75
0.75
-1.25
0.00 2.50 5.00 7.50 10.00 12.50 15.00 17.50 20.00
Time [ms]
Fig. 14 Torque output

Winding Currents Maxwell2DDesign2 ANSOFT

30.00 Curve Info
Current(PhaseA)
Setup1 : Transient
Current(PhaseB)
20.00 Setup1 : Transient
Current(PhaseC)
Setup1 : Transient
10.00
Y1 [A]
0.00
-10.00
-20.00
-30.00
0.00 2.50 5.00 7.50 10.00 12.50 15.00 17.50 20.00
Time [ms]
Fig. 15 Winding currents under load
5 Conclusions
This paper has discussed various types of electric vehicles, their powertrains, and the
importance of considering the brushless DC motors for traction applications. Based
on comparison with other motors, a permanent magnet brushless DC (PMBLDC)
motor is set to be designed for traction application, and the power is estimated using
a tractive force model in vehicle dynamics. Based on the required parameters, the
motor was designed using analytical calculations and the outputs were estimated. The
calculations are stopped at the point where the industry gets sufficient data to design
and fabricate a motor. The obtained parameters were input into design software like
ANSYS Maxwell and the analytical, and 2D parameters were obtained, which gave
us an idea of how some variables are varied, making analytical hand calculations
inaccurate at certain aspects on how they don’t consider a lot of factors such as
load. This again proves how simulation software can be more efficient to design a
motor. Using a parametric optimization toolbox in ANSYS to optimize the parameters
to extract better outputs and testing new magnetic arrangement configurations like
Halbach Array to enhance flux density in the air gap is the scope for future research
work. This future work can be done to enhance efficiency, reduce losses, and cogging
torque.
Acknowledgements This research work was funded by “Woosong University’s Academic

Research Funding—2022”.
References
1. Department of Heavy Industry, Ministry of Heavy Industries & Public Enterprises, Government
of India
2. Kumar BVR, Kumar KS (2016) Design of a new dual rotor radial flux Halbach array for an
electric vehicle. In: IEEE international conference on power electronics, drives and energy
systems, pp 1–5
3. Hsieh M, Lin I, Hsu Y, McMahon RA (2012) Design of brushless doubly-fed machines based
on magnetic circuit modeling. IEEE Trans Magn 48(11):3017–3020
4. Krishnan R (2001) Electric motor drives: modeling, analysis and control. Pearson
5. Lee KH, Cha HR, Kim YB (2016) Development of an interior permanent magnet motor through
rotor cooling for electric vehicles. Appl Therm Eng 95:348–356
6. Nguyen PH, Hoang E, Gabsi M (2013) Bi-criteria optimization design of an interior permanent
magnet synchronous machine for a hybrid electric vehicle application. Math Comput Simul
90:178–191
7. Boldea I, Nasar SA (2002) The induction machine handbook. CRC Press
8. Gieras JF (2010) Permanent magnet motor technology: design and applications, 3rd edn. CRC
Press
9. Goss J, Wrobel R, Mellor P, Staton D (2013) The design of AC permanent magnet motors for
electric vehicles: a design methodology. In: International electric machines & drives conference,
Chicago, IL, pp 871–878
10. Satria A, Farisi V, Rizqiawan A, Heryana N, Purwadi A, Haroen Y (2014) Synthesis modeling
of 10 kW BLDC controller for city electric car based on PSIM. In: 2nd IEEE conference on
power engineering and renewable energy, Bali, pp 251–256
11. Tatimapula S (2013) Study of a parallel electric hybrid three-wheeled motor taxi. Int J Comput
Appl 38–41
12. Hendershot JR, Miller TJE (2010) Design of brushless permanent-magnet motors. Motor
Design Books LLC
13. Zhu ZQ, Howe D (2007) Electrical machines and drives for electric, hybrid and fuel cell
vehicles. Proc IEEE 95:746–765
14. Raines GB (2009) Electric motor drives for battery, hybrid and fuel cell vehicles. In: Electric
vehicles: technology, research and development, Nova Science, New York, pp 1–40
15. Ehsani M, Gao Y, Longo S, Ebrahimi K (2018) Modern electric, hybrid electric, and fuel cell
vehicles, 3rd edn. CRC Press
16. Larminie J, Lowry J (2012) Electric vehicle technology explained, 2nd edn. Wiley
17. Buyukdegirmenci VT, Bazzi AM, Krein PT (2014) Evaluation of induction and permanent
magnet synchronous machines using drive-cycle energy and loss minimization in traction
applications. IEEE Trans Ind Appl 50:395–403
18. Cheng M, Sun L, Buja G, Song L (2015) Advanced electrical machines and machine-based
systems for electric and hybrid vehicles. Energies 8:9541–9564
19. Kreuawan S, Gillon F, Brochet P (2008) Comparative study of design approach for electric
machine in traction application. Int Rev Electr Eng 3(3):455–465
20. Wang J, Xia ZP, Howe D (2005) Three-phase modular permanent magnet brushless machines
for torque boosting on a downsized ICE vehicles. IEEE Trans Veh Technol 54:809–816
21. Gu W, Zhu X, Quan L, Du Y (2015) Design and optimization of permanent magnet brushless
machines for electric vehicle applications. Energies 8:13996–14008
22. Sandeep V, Shastri S (2019) Analysis and design of PMBLDC motor for three wheeler electric
vehicle application. In: E3S web of conferences, vol 87, pp 1–7
23. [Online] https://www.bajajauto.com/intracityvehicles/bajajre/our-range-compact.html. Last
accessed 26 Mar 2021
24. Sheikh MB, Manware PS (2016) Brushless DC motor design for electric traction system. J Res
2(4):18–22
25. Vuddanti S, Karknalli V, Salkuti SR (2021) Design and comparative analysis of three phase,
four phase and six phase switched reluctance motor topologies for electrical vehicle propulsion.
Bull Electr Eng Inf 10(3):1495–1504
26. Dalcali A (2018) Optimal design of high-performance interior PM motor for electric vehicle.
Int J Energy Eng Sci 3(2):26–35
27. Mahmoudi A, Rahim NA, Ping HW (2012) Axial-flux permanent-magnet motor design for
electric vehicle direct drive using sizing equation and finite element analysis. Prog Electromagn
Res 122:467–496
28. Bhatt P, Mehar H, Sahajwani M (2019) Electrical motors for electric vehicle—a compara-
tive study. In: Proceedings of recent advances in interdisciplinary trends in engineering &
applications, pp 1–10
29. Huang C, Lei F, Han X, Zhang Z (2019) Determination of modeling parameters for a brushless
DC motor that satisfies the power performance of an electric vehicle. Measure Control 52(7–
8):765–774
30. Sandeep V, Shastri S, Sardar A, Salkuti SR (2020) Modeling of battery pack sizing for electric
vehicles. Int J Power Electr Drive Syst 11(4):1987–1994
31. Niapour SAKHM, Garjan GHS, Shafiei M, Feyzi MR, Danyali S, Kouhshahi MB (2014)
Review of permanent-magnet brushless DC motor basic drives based on analysis and simulation
study. Int Rev Electr Eng 9(5)
32. Nugraha YU, Cahyadi A, Yuniarto MN, Sidharta I (2019) Design optimization for torque
density in brushless DC motor with IPM V-type using PSO Method. IOP Conf Ser Mater Sci
Eng 694:1–8
33. Salkuti SR (2019) Optimal operation of microgrid considering renewable energy sources,
electric vehicles and demand response. In: E3S web of conferences, vol 87, pp 1–6
34. Ludois DC (2017) Brushless and permanent magnet free wound field synchronous motors for
EV traction, Technical report
35. Guerra AL, Icaza LA, Torres L (2018) Brushless DC motor control with unknown and variable
torque load. IFAC-PapersOnLine 51(13):644–649
36. Wang W, Fu R, Fan Y (2018) Electromagnetic parameters matching of permanent magnet
synchronous motor for hybrid electric vehicles. IFAC-PapersOnLine 51(31):407–414
37. Ebadpour M, Bagher M, Sharifian B, Babaei E (2018) Modeling and synchronized control of
dual parallel brushless direct current motors with single inverter. Comput Electr Eng 70:229–
242
D-FACTS-Based Power Quality
Enhancement Using Power Balance
Theory
Vikash Anand and S. K. Mallik
1 Introduction
Electric users must be provided good quality signal, voltage/current, means great
promising power factors, viz., harmonizing with 3-F standards regular system,
voltage/current magnitude, harmonic contents, voltage steadiness grade, and system
frequency on stated significance. Every so often, users are delivered with deprived
power class which leads to harmless, consistent, and proficient equipment operation.
Whereas, reduced quality harms result since a range of complaints, for example,
voltage/current—sags/swells, glimmer, altered harmonic, disruptions and transitory
disturbances because of existence of prominent cause.
European Communities stated that sags (dips) are some decreased voltage below a
manipulator termed bounds—flanked in cycle and 2.55 s. Swells (surges) are similar
sags, apart from voltage go beyond a manipulator expressed higher edge. Even though
innumerable meanings of sag/swell relation toward breadth too period quiet over-
come, IEEE 1159-1995 commended run-through continuously observing electric
power fineness outlines “Sag (dip) as a decrease to between 0.1pu and 0.9pu in rms
voltage or current at the power frequency for durations of 0.5 cycle to 1 min and Swell
as an increase to between 1.1pu and 1.8pu in rms voltage or current at the power
frequency for durations from 0.5 cycle to 1 min” [1–3]. Research framework has
discussed that sags/swells may be diminished through various approaches. Usage of
tradition power electronics expedient alike FACTS device stays stared as best adept
technique [4–6].
V. Anand (B)
Department of Electrical & Electronics Engineering, SRM Institute of Science & Technology
NCR Campus, Ghaziabad, India
e-mail: vikasa@srmist.edu.in
S. K. Mallik
Department of Electrical Engineering, National Institute of Technology Patna, Patna, India
e-mail: skmallik@nitp.ac.in
336 V. Anand and S. K. Mallik
In late 1980s, Electric Power Research Institute (EPRI) familiarized Flexible AC

Transmission Systems (FACTS) by means of innovative methodology for resolving
problematic like planning, governing, and functioning power system. Along with
improving ability of existing as well as innovative lines, it compromises new possi-
bilities to control power with growing transmission volume, improving voltage
regulator, and enhancing system strength. At first in April 19, 1988, FACTS idea
was presented by N. G. Hingorani. Continually several FACTS devices such as
SVC, STATCOM, DVR, and Static Synchronous Series Compensator have been
announced. UPFC leads to best adjustable, proficient, and favorable kind of FACTS
device, and it was introduced in 1991. It functions to regulate concurrently all three
constraints, viz., voltage, impedance, and phase angle. Research recommends that
UPFC efficiently control drift in active/reactive power and progresses voltage contour
inside power system [7–10].
Well along, a new FACTS device, DPFC was familiarized to moderate
current/voltage aberration and enrich superiority of power. Its configuration is orig-
inated from organization of UPFC and comprises a shunt converter and several
little self-governing series converters. It works on principle of power theory of
non-sinusoidal components. Like UPFC, DPFC is capable to control concurrently,
viz., bus voltage, line impedance, and transmission angle. Because of having capa-
bility to remove vast dc link between shunt-series converter and establish real power
interchange through 3rd harmonic current, it can lead above UPFC [1, 11].
Underneath this framework, this paper validates DPFC operation in single
machine of power network comprising two parallel lines in SIMULINK environs.
With/without DPFC, performance of voltage/current is analyzed on strife taking
place at load side. It observes effervescent operation of system encompassing DPFC
control scheme centered d-q axis theory, and PI control is very supportive in forming
steady formal.
2 Distributed Power Flow Controller (DPFC)
DPFC has been consequent as of UPFC through eliminating dc linked capac-

itor amidst UPFCs series-shunt converters, through employing disseminated series
converters of numerous elements in succession along line conductor. Power give-and-
take way in DPFC is done on power model of non-sinusoidal modules. At altered
frequencies active power remains confiscated as of one another, and current/voltage
at one frequency haven’t been impacted on real power by another frequency. Self-
governing real power at altered frequencies offers chance where converter without
any power cause is able to produce real power at particular frequency and absorbs it
at a different frequency. At fundamental frequency shunt converter is able to capti-
vate real power from line and at harmonic frequency it can inject current into line.
Likewise, at fundamental frequency, series converter can insert voltage into line and
at the harmonic frequency it captivates active power from line [12]. Figure 1 depicts
DPFC positioned in power network of two-bus system.
D-FACTS-Based Power Quality Enhancement … 337
Vs Vr
R L
Vse1
AC
Ish1
Vsh1 Ish3 DC
AC Vsh,dc DC
Vsh3
Cse
DC Csh AC
Vsh1,ref Vsh3,ref
Shunt Series
Control Control
Ish1qref Ish3ref Vse1,ref
Central Control Unit
Fig. 1 Simple control circuit diagram of DPFC
Due to change in active power inside DPFC, third harmonic is nominated for the
reason that it remains zero-sequence harmonic that may be effortlessly stuck in -
winding of transformer. Supplementary, there is a need for a high-pass filter in order
to create closed-loop for harmonic current and chunk 1st harmonic current. High-pass
filter can remain substituted by cable linked between winding of transformer and
ground by using 3rd harmonic. Harmonic present can be routed to pounded through
this cable as depicted in Fig. 2 [13].
Shunt converter captivates generated power at fundamental frequency. 3rd
harmonic current remains stuck in Y- transformer and are inserted into neutral
of -Y transformer by yield point of shunt converter resulting in harmonic current
flow through transmission line. Series capacitors’ DC voltage is controlled by this
harmonic current.
Pse1 = –Pse3
Vs Vr
Vse1 Vse3
R L
I
Ish
Vsh1
Pse3 = –Psh3
Psh1= –Psh3
Vsh3
Fig. 2 Single line diagram of DPFC

abc / αβ Reference Voltage

Vse1qref (Series Converter)
αβ Vse1dref
Signal
dq Ish1qref
Reference Current
Ish3ref (Shunt Converter)
PLL
Fig. 3 Simple structure of DPFC central controller
2.1 DPFC Control Mechanism
The three governor schemes of DPFC are Central, Shunt, and Series Control. Each
has been discussed below [13, 14].
2.1.1 Central Control
Central controller produces locus pointers for shunt and series converters of DPFC
focusing on DPFC functions at level of power system, viz., short frequency power
oscillation damping, power flow control, and asymmetrical components toning
(Fig. 3).
As per scheme requirements, voltage-reference and reactive current signals for
series and shunt converters correspondingly stand given by central controller. All
locus pointers are produced by central controller stand by fundamental frequency.
2.1.2 Shunt Control
Shunt controller purposes imparting continual 3rd harmonic current in route for
providing real power for series converters. At fundamental frequency, 3rd harmonic
current remains sealed along with bus voltage.
PLL is secondhand to seizure bus voltage frequency in directive to generate a
simulated spin locus casing for 3rd harmonic constituent, and multiplies this one
yield phase signal by 3. Fundamental frequency control of shunt converter purposes
to impart manageable reactive current into line and to retain capacitor DC voltage
at persistent level. Two flowed controllers are existing to control constituents of 1st
frequency. Current control modulating shunt current at fundamental frequency is
inner control loop. Components q and d of shunt converters’ reference signal stays
got from central controller and DC control, respectively [13] (Fig. 4).
Vs PLL 1st frequency control loop
Ѳ1
from Ish1qref Inverse
Central control
Vsh1q D-Q Vsh1ref
Current Transforma on
VdcShunt Ish1dref Control Vsh1d
DC Control
-C-
Pulses to
+ PWM Generator converter
Vs Ѳ1
PLL 3
Ѳ3
Inverse
D-Q
from Vsh3q Transforma on
Central control Ish3ref Vsh3ref
Current
Control
Ishunt (transformer neutral) Vsh3d
3rd frequency control loop
Fig. 4 Simple structure of DPFC shunt controller
2.1.3 Series Control
Separately 1-F converter takes its particular series control over route. Series capacitor
voltages, line current, and series voltage reference are positioned controller inputs
in d-q frame [14] (Fig. 5).
In series control within DPFC, 3rd harmonic frequency control is prime control
loop. Vector control principle is adopted to control dc-voltage. For 1-F park transfor-
mation, 3rd harmonic current is designated as spin locus casing. It can stand straight
forwardly seized over phase-locked loop (PLL) in series-converter.
3 Performance Study of Demonstrated Results in MATLAB
Non-linear load linked system depicted in Fig. 1 is simulated in SIMULINK block

set environment through 3-F source. Load remains linked source via parallel lines
I and II of equal length. D-FACTS are integrated in line II independently. Simulink
block parameters are enumerated in Appendix. Simulation studies are carried out in
imperative to investigate working of DPFC with PI controller for compensation of
voltage sag, current swell, and drop in load voltage harmonics at 3-F fault near load
side.
Vse1qref. 1st frequency control loop
{
X
from
Central control +
Vse1dref.
X
1stvse,ref
+ Pulses
LPF PWM Generator to
(Low Pass Filter) Ѳ1 converter
1-ϕ
iline 3rdvse,ref
PLL Ѳ3
X
3rd Pass Filter
ref.V DC Control
Vse,dc,ref
3rd frequency control loop
Fig. 5 Simple outline of DPFC’s series controller
From simulated results, effect of voltage sag is observed through generating 3-F
fault in short of compensator. Time length in lieu of fault stands 0.06 s. (0.06–0.12 s.).
Feeder power determines sternness of let fall to 0.5-unit voltage as depicted in Fig. 6.
On employed DPFC, voltage sag mitigation is operative with pintsize transitory
spike in fault occurring-diminishing instant. Figure 7 depicts compensated voltage
sag waveform.
Due to 3-F fault in power arrangement, current swell is depicted for duration of
0.06 s in Fig. 8.
Without Compensation
Fig. 6 Load voltage sag in short of compensation

With DPFC Compensation
Fig. 7 Mitigated load voltage dip with DPFC
Fig. 8 Load current swell without compensation
It is noticeable point that with DPFC compensator line current swell is diminished
with spikes at beginning-ending of fault duration. It is depicted in Fig. 9. Mitigated
current swell is found in better quality with DPFC compensator.
Fig. 9 Mitigated load current swell with DPFC

Fig. 10 Injected voltage of series converters in fault line2 for duration of 0.06 s
Fig. 11 Load voltage’s THD Without Compensation

(total harmonic distortion)
deprived of compensation
At fault occurring instant, line2’s series converters inject voltage in fault line to
maintain the required bus voltage and stops after clearing the fault, which has been
depicted in Fig. 10.
Load voltage harmonics study, Fast Fourier Transform (FFT) is validated through
figures. Figure 11 illustrates load voltage harmonics short of compensator.
On unification D-FACTS compensator, even harmonics stay eradicated while odd
harmonics are abridged to adequate edge specified in THD of IEEE standards, i.e.,
below 5% [15]. It is shown in Fig. 12.
The respective THD ethics of load voltage are specified in Table 1.
From Table 1, load voltage THD is decreased to large extent on employment of
DPFC compensator. Therefore, simulation outcomes authenticate DPFC capability
to recover power excellence in better extent with wild vigorous reaction.
Fig. 12 Load voltages’ total With DPFC Compensation
harmonic distortion (THD)

using DPFC compensation
Table 1 Total harmonic

THD of load voltage Without compensation With DPFC
distortion (THD) of load
voltage Value in percentage 48.36% 1.67%
4 Conclusion
The results clearly show that at 3rd harmonic frequency, real power can be swapped
via line through shunt-series converters in DPFC. Controllable active-reactive power
may also stay filled by series converters at fundamental frequency. Redundancy of
series converter and non-requisite of the high-voltage segregation with low rating of
components clue to better reliability and lower cost of DPFC. Simulation results—
mitigated voltage sag, mitigated current swell and reduced harmonic distortion,
confirmations efficacy of DPFC in power quality enhancement. It tips to improve-
ment in stability and quality of power system encompassing DPFC with D-Q axis
mechanism.
Appendix
The investigation system in use aimed at simulation study of mounted DPFC in soli-
tary machine of power arrangement. This arrangement is comprised of two parallel
transmission line with load in MATLAB. Structure facts considered for validation in
simulation exist like this in Table 2 [2, 16].
Table 2 System figures in simulink

System components Parameter Ratings
Three-phase source Transmission voltage 220 kV
Power/Frequency 110 MW/60 Hz
X/R 3
Short circuit capacity 11 GW
Transmission line Resistance 1/100 pu/km
Inductive and capacitive Reactance 12/100 pu/km
Line length 100 km
Unsymmetrical fault Fault type ABC- G
Ground resistance 0.05
DPFC
Shunt converter 3-ϕ Nominal power 60 MVAR
DC linked capacitor 600 μF
Coupling transformer (shunt) Nominal power 100 MVA
Voltage rating 220 kV/11 kV
Series converter 1-ϕ Nominal power 6 MVAR
Voltage rating 6 kV
References
1. Omar R, Rahim NA, Sulariman M (2009) Modelling and simulation for voltage sag/swell
mitigation using dynamic voltage restore (DVR). J Theoret Appl Inf Technol 5(4):464–470
2. Barakati SM, Sadigh AK, Mokhtarpour E (2011) Voltage sag and swell compensation with DVR
based on asymmetrical cascade multicell converter. In: North American Power Symposium
(NAPS), pp 1–7
3. Jamshidi A, Barakati SM, Ghahderijani MM, Power quality improvement and mitigation case
study using distributed power flow controller. Case 978-1-4673-0158-9/12/$31.00. IEEE pp
464–468
4. Round SD, Yu Q, Norum LE, Underland TM (1996) “Performance of unified power flow
controller using a d-q control system”, conference publication No. 423 of the IEE sixth
international conference on AC and DC power transmission, London, pp 357–362, April–May
5. Edris A-A (1997) Proposed terms and definitions for flexible AC transmission system (FACTS).
IEEE Trans Power Delivery 12(4):1848–1853
6. Nita I, Patne R, Thakre KL (2008) Factor affecting characteristics of voltage sag due to fault
in the power system. Serbian J Electr Eng 5(1):171–182
7. Gyugui L, Schauder CD, Williams SL, Rietman TR, Torgerson DR, Edris A (1995) The unified
power flow controller: a new approach to power transmission control. IEEE Trans Power
Delivery 10(2):1085–1097
8. Papic I, Zunko P, Pouh D, Weinhold M (1997) Basic control of unified power flow controller.
IEEE Trans Power Syst 12(4):1734–1739
9. Padiyar KR, Kulkarni AM (1997) Control design and simulation of a unified power flow
controller. IEEE Trans Power Delivery 13(4):1346–1354
10. Hingorani NG, Gyugyi L (2000) Understanding FACTS, concepts and technology of flexible
AC transmission system. Institute of Electrical and Electronics Engineers
11. Singh P, Tiwari R (2017) Amalgam power flow controller: a novel flexible, reliable and cost
effective solution to control power flow. IEEE Trans Power Syst. https://doi.org/10.1109/
TPWRS.2017.2764954
12. Yuan Z, Sjoerd WHDH, Braham F, Dalibar C (2010) A FACTS device: distributed power flow
controller (DPFC). IEEE Trans Power Electron 25(10):1085–1097
13. Anand V, Mallik SK (2018) Power flow analysis and control of distributed FACTS devices in
power system. AEE 67(3):545–561
14. Anand V, Mallik SK (2019) Introducing supplementary control scheme to DPFC during series
converter failure. JEAS 14(7):2124–2134
15. http://www.schneider-electric.com.tw/documents/Event/2016_electrical_engineering_sem
inar/IEEE_STD_519_1992vs2014.pdf. Accessed Nov 2020
16. Dwivedi OP, Singh JG, Singh SN (2004) Simulation and analysis of unified power flow
controller using SIMULINK. In: National power systems conference, NPSC 2004, IIT, Madras,
Dec. 27–30, pp 1048–1054
17. Zhang R, Cardinal M, Szczesny P, Dame M (2002) A grid simulator with control of single-
phase power convertors in D-Q rotating frame. In: Power electronics specialists conference
(PESC), IEEE, Australia, pp 1431–1436
Design and Implementation of Electrical
Dynamic Braking System in Electric Car
Using Buck Converter
Walid Alqaisi and Claude Ziad El-Bayeh
1 Introduction
Conventional vehicles that run on fossil fuels contribute to the global warming
phenomenon, as a result, they have negative environmental impacts. The electric
vehicle is gaining popularity as a promising alternative.
As a conventional system, the hydraulic car braking drives the static brake pad
against the brake drum. A friction force, which retards the vehicle movement, is
produced between them. Recently, Electrical Brake (EB) systems in electric vehicles
(EVs) have drawn the attention of researchers for their advantages in real applica-
tions. The development of power converters nowadays plays a big part in improving
the controllability of EBs. EB systems are eco-friendly because they don’t contain
copper, mercury, lead, cadmium, and chromium, unlike hydraulic brake pads which
are harmful materials. Fine particles these metals flake off on our roadways, then
flow into our rivers and lakes when it rains [1]. The EB system is able to provide a
low-cost large braking force using only a relatively small electrical energy. Pulse-
Width Modulation (PWM) provides control and regulation of the total output DC
voltage. Compared to traditional hydraulic braking components, EBs have a compact
design and better braking performance. Furthermore, the torque developed from an
EB is faster than a hydraulic braking system by 10–100 times [2].
An EB is an electro-mechanical system that requires a precise control system in
order to develop a safe and reliable braking force. In [3], a gain-scheduled nonlinear
PI controller is designed to regulate a tire slip in an EB system. In [4], monitoring the
behavior of the rear wheels is utilized to estimate tire slips in an antilock brake system
(ABS). In [5], an adaptive EB which is changing with driving and road conditions
is designed. The three-phase AC output of the inverter is used to drive and stop
W. Alqaisi (B)
School of Engineering Technology, College of the North Atlantic - Qatar, Doha, Qatar
C. Z. El-Bayeh
Canada Excellence Research Chairs Team, Concordia University, Montreal, Canada
348 W. Alqaisi and C. Z. El-Bayeh
the motor [6]. Although the good performance of the mentioned system, simpler
approaches can be used with similar results.
In [7], the braking of an induction motor is analyzed for different values of elec-
tronics parts. A Braking scheme for a Six-phase induction motor is used in [8]. A
braking approach for squirrel-cage induction motors is demonstrated using a three-
phase rectifier [9]. A Regeneration during deceleration is designed for an electric
vehicle [10]. Generative energy storage using bidirectional power conversion is used
in [11], where the converter’s output is fixed and it is not a variable, which makes
the selection of fast or smooth braking not possible for an operator to choose.
In electric vehicle braking operation, induction motor braking is mainly of three
types [12]: Regenerative braking, Plugging Braking, and Dynamic braking. Regen-
erative braking of an induction motor is possible if the motor’s speed is higher than
the synchronous speed. The first method is suitable for other applications, while the
second method generates a very high current. It is found that DC dynamic braking is
the most suitable method for electrical braking. Figure 1 shows the structure of the
overall system.
The main highlighted points in this paper are:
– The proposed system uses a buck controller for its simplicity.
– An induction motor braking is selected because it produces less heat than other
braking types.
– Motor braking voltage is variable as desired and controlled by a PID controller.
– It is not required to know the exact dynamic model and the application can be
applied to other satiations and applications.
– The entire system’s advantages are high precision, low cost, and high reliability.
Fig. 1 Electric braking system

Design and Implementation of Electrical Dynamic Braking System … 349
This paper is presented as follows. Section 2 demonstrates the buck converter

design. Section 3 clarifies control design formulation. Section 4 describes the calcu-
lation of braking design and braking travel time. Section 5 explains the electric
braking system. Section 6 presents Matlab/Simulink simulation and Sect. 7 concludes
the paper.
2 Proposed Design of the DC to DC Buck Converter
The structure of the used converter is shown in Fig. 2. The designed converter is
built to provide stable regulated output. Voltage value, PWM width, and frequency
are the controller parameters, while the speed of the vehicle and the time to stop are
the braking parameters. The Buck converter is a step-down converter consisting of a
low-pass LC filter. The capacitor is in parallel with the load. The inductance prevents
high current spikes. The free-wheeling diode is added to absorb the inductor’s current
momentum. Correct design of the components allows high-frequency switching.
The operation of the circuit involves open and closed conditions of the insulated-
gate bipolar transistor (IGBT). During the charging period, the capacitor voltage
moves from the minimum to the maximum. The function of the capacitor is to
regulate the output voltage ripples, noting that large capacitance causes spikes in the
source current when the transistor opens. Capacitor selection is given as [13]
L I L2 ,max
C= (1)
(Vout + Vout,over shoot )2 − Vout
2
where L is the inductor value, Vout is the output voltage, Vout,over shoot is the peak
overshoot voltage allowed on the output, I L2 ,max is the peak inductor current.
The inductance prevents a high current spike in addition to providing current at
the off-state of the IGBT. Inductor current never reaches zero, it is ranging from Imin
Fig. 2 System block diagram

to Imax . Peak to peak current is I L = Imax − Imin and I Avg is the average current.
The value of the capacitance is selected as [14]
(Vin − Vout )Vout

L= (2)
Vin Fsw I L
where Vin is the input voltage, Vout is the output voltage, Fsw is the switching
frequency (Hz), L is the inductance in Henry (H).
3 Control Design
Circuit analysis is based on the two switching states of the circuit, on-state and
off-state as shown in Fig. 3.
During the on-state condition, Fig. 3a, the circuit is analyzed as follows:
di L
L = vg (t) − vc (t) (3)
dt
In the same way, for the capacitor current
dvc
C = il (t) − i o (t) (4)
dt
Then we can write
dvc il (t) i o (t)
= − (5)
dt C C
dvc il (t) vo (t)
= − (6)
dt C RC
Equations (3) and (6) can be written in state space form as
Fig. 3 On-state a and off-state b of the DC power converter

−1
dil
0 −1
il (t)
dt = L
−1 −1 + L vg (7)
dvc
dt C RC
vc (t) 0
Then it can be written in the state space form as
ẋ = Aon + Bon u (8)
where
−1
−1
il (t) 0
x= , Aon = −1 L
−1 , Bon = L , u = vg
vc (t) C RC
0
During the off-state condition, Fig. 3b, the transistor is open, and the load current
is supplied by the stored energy in the inductor, with the circuit completed by the
diode. The circuit can be analyzed as follows:
During this time, inductor voltage is
vl (t) = −vc (t) (9)
dil −vc (t)

= (10)
dt L
In the same way, for capacitor current
i c (t) = il (t) − i o (t) (11)
Then, we can write
dvc il (t) vo (t)

= − (12)
dt C RC
Equations (10) and (12) can be written in the state space form as follows:

dil
0 −1
il (t) 0
dt = L
−1 −1 + vg (13)
dvc
dt C RC
vc (t) 0
Then, it can be written in the state space form as given below
ẋ = Ao f f x + Bo f f u (14)
where
−1

il (t) 0 0
x= , Ao f f = −1 L
−1 , Bo f f = , u = vg
vc (t) C RC
0
The converter is controlled by the pulse width modulator (PWM) signal. The
converter on-time is in the time interval ton = T dn and the converter off-time to f f =
T (1 − dn ). The state space form is given as follows:

ẋ = Aon d(t) + Ao f f (1 − dn ) x + Bon d(t) + Bo f f (1 − dn ) u (15)
The widely used controller, PID, is utilized where the tracking error is fed to the
PID controller which computes the proportional, derivative, and integral error signal.
The output of the PID controller u(t) applied to the plant is given as

de(t)
u(t) = K p e(t) + K i e(t)dt + K d (16)
dt
where K p , K i , and K d are the integral, proportional, and derivative gains, respec-
tively.
4 Calculation of Braking Travel Time
The function of the proposed system is to provide the three-phase winding with a
stationary magnetic field φ to set up in the stator bore. This magnetic field causes
a voltage charge to be induced in the cage-rotor bars as long as the motor rotor is
moving. When motor speed is reduced, induced voltage frequency, as well as the
inductive impedance, falls. Braking torque is approximately equal to the rated motor
torque on average. Braking current I DC is nearly 2–2.5 times of the rated motor
current I N , i.e.
I DC = 2.5I N (17)
Higher I DC values produce harder dynamic braking as in Fig. 4, where I DC1 >
I DC2 > I DC3 , while softer braking occurs for lower current values. Lowering the
controlled voltage is required if harsh braking is experienced during practical imple-
mentation. It is complicated to calculate braking time because of the nonlinearity
of involved parameters such as the changes of the mechanical load, inertia, motor
r.p.m, rated current, and pulses. Velocity to current curve is shown in Fig. 4. The
kinetic energy of the rotating parts is equal to the dissipated rotor losses, and the
braking torque is proportional to the square of the dc braking current. The electrical
resistance of motor windings is very low. Thus, a small injected voltage creates a
high current. Such current generates a solid braking effect on motor bars and the
rotor. Figure 5 shows that the reverse torque falls to zero as speed reduces to stand
still. Braking torque is shown in Fig. 5 [15], where M is the instantaneous braking
torque, Mav is the average value of the braking torque, n is motor speed, n sy is the
synchronous speed. As velocity dropped down, the motor ohmic resistance turned
Fig. 4 Current—r.p.m curve

for a fixed load
Fig. 5 Current—r.p.m curve

for a fixed load
to be more dominant which will strengthen the braking force. This braking torque
falls away and finishes when there is no further movement.
5 Electric Brake Systems
Electric braking (EB) systems are environment-friendly because they do not use
hydraulic parts or any mechanical disks. Brake pads in cars contain copper, mercury,
lead, cadmium, and chromium which are harmful materials that flow into our rivers
and lakes when it rains [1]. An EB is able to provide a safe and large braking force. To
generate a proper braking force, induction motors are used due to their efficiency and
durability. In addition, induction motors have high efficiency-to-size characteristics
when compared to other types of motors which must be larger for the same efficiency.
They are connected to each wheel and controlled by a PID closed-loop controller
where the desired braking force is proportional to the output of the converter. The
Fig. 6 The block diagram of the system
value of the setpoint is proportional to the vehicle speed in order to have smooth
braking as the speed is being reduced.
The structure requires velocity continuous measurement. The block diagram of
the control system is shown in Fig. 6, while the overall system is shown in Fig. 1.
6 Simulation
The simulation was performed using Matlab/Simulink in three parts built as shown
in Figs. 7 and 10. A variable desired voltage is expected to enter the controller, thus
the buck converter in Fig. 7 was tested for a voltage changing from ten to zero volts
as in Fig. 8 and for a fixed voltage as in Fig. 9. The desired voltage in the system is
a function of the tire’s r.p.m.
The second part was simulated by using an induction motor model provided by
the Simscape tool in Matlab Simulink as in Fig. 10. Testing was performed on three
time sequences, at first, the motor is driven to the nominal speed ωn , then the power
is disconnected letting the speed decrease gradually. At the third stage, DC braking
is applied at VDC = 14, 10, 6.3 V, as a result, different baking forces are obtained
which led to a complete stop at different times as in Fig. 11. The first stage ends at
t1 = 0.5 second, the second stage ends at t2 = 1 second, while the third stage ends
when the speed reached zero at t3 .
The stator current is an important factor to consider, a higher value of stator current
for a long time can damage the windings. Figures 12, 13, 14 show the stator currents
for VDC = 14, 10, 6.3, respectively. Braking currents peak at t2 = 1, but they remain
within the acceptable range. Note that I1 , I2 , and I3 are the stator winding currents.
Fig. 7 The buck DC converter in simulink
Fig. 8 Testing the converter for a variable desired voltage

Fig. 9 Testing the converter for a fixed desired voltage
Fig. 10 Induction motor braking simulink model
The Proposed Scheme

Simulation results show braking travel time achieved at t = [1 − 3.5] seconds, which
means that 2.5 s to come to a complete stop (Fig. 15). The buck converter controls
the voltage as in Fig. 16, where a high shock happens when braking starts at t = 1
second, and then it returns to stability quickly. Then the controller reduces the voltage
Fig. 11 Testing was performed on three time sequences
Fig. 12 Induction motor winding currents for V DC = 14

Fig. 13 Induction motor winding currents for V DC = 10
Fig. 14 Induction motor winding currents for V DC = 6.3
to zero in order to prevent unnecessary stator current. The stator currents are shown
in Fig. 17, it can be seen that the condition in Fig. 16 is met where the braking current
remains in the acceptable limit.
Fig. 15 Braking travel time
Fig. 16 Buck converter regulated voltage
7 Conclusion
In this paper, we proposed a design of closed-loop control of pulse width modulated

DC Buck converter with a PID controller for an electrical braking system in an elec-
tric car. This system is the future development of braking moving vehicles for its
Fig. 17 Stator current in the three windings
advantages, i.e., the designed system provides a strong braking force and it is environ-
mentally friendly, unlike conventional systems. On the other hand, the strong brake
force allowed controlling the stop time. The system was analyzed mathematically
and simulated by Matlab/Simulink in order to prove its efficiency.
References
1. The hidden pollutant in our brake pads|Environment Canterbury. https://www.ecan.govt.nz/

get-involved/news-and-events/2018/the-hidden-pollutant-in-our-brake-pads/. Accessed 31 Jan
2021
2. Solyom S, Rantzer A (2014) Synthesis of a model-based tire slip controller. Taylor Fr
41(6):475–499. https://doi.org/10.1080/004231105123313868
3. Zhang L, Li L, Lin C, Wang C, Qi B, Song J (2014) Coaxial-coupling traction control for a
four-wheel-independent-drive electric vehicle on a complex road. Proc Inst Mech Eng Part D
J Automob Eng 228(12):1398–1414
4. Choi SB (2008) Antilock brake system with a continuous wheel slip control to maximize the
braking performance and the ride quality. IEEE Trans Control Syst Technol 16(5):996–1003
5. Dousti M, Baslamısli SC, Onder ET, Solmaz S (2015) Design of a multiple-model switching
controller for ABS braking dynamics. Trans Inst Meas Control 37(5):582–595
6. Awari P, Sawarkar P, Agarwal R, Khergade A, Bodkhe S (2017) Speed control and electrical
braking of axial flux BLDC motor. In: 2017 6th international conference on computer applica-
tions in electrical engineering—recent advances, CERA 2017, Apr. 2018, vol. 2018-January,
pp 297–302. https://doi.org/10.1109/CERA.2017.8343344
7. Singh R, Umashankar S, Vijaykumar D, Kothari DP (2013) Dynamic braking of induction
motor-Analysis of conventional methods and an efficient multistage braking model. In: 2013
international conference on energy efficient technologies for sustainability, pp 197–206
8. Duran MJ, Gonzalez-Prieto I, Barrero F, Mengoni M, Zarri L, Levi E (2015) A simple braking
method for six-phase induction motor drives with diode front-end rectifier. In: IECON 2015–
41st annual conference of the IEEE industrial electronics society, pp 1542–1547
9. Rata M, Rata G (2016) Study solution of induction motor dynamic braking. In: 2016
international conference on development and application systems (DAS), pp 33–37
10. Mahapatra AA, Gopalakrishna S (2014) Regenerative braking in induction motor drives in
applications to Electric Vehicles. In: 2014 IEEE students’ conference on electrical, electronics
and computer science, pp 1–5
11. Agrawal SK, Kumar V, Alam A, Thakura P (2014) Regenerative braking for induction motor
drive. In: 2014 6th IEEE power India international conference (PIICON), pp 1–6
12. Induction motor braking regenerative plugging dynamic braking of induction
motor|Electrical4U. https://www.electrical4u.com/induction-motor-braking/. Accessed 11
Mar 2021
13. Maimone G (2010) Selecting l and c components in the power stage of the mc34700 switching
regulators. Free. Semicond. Appl. Note AN4067
14. Mondal A (2014) Digital PID controller design for DC-DC buck converter
15. Gear T, Specialist M, Starting, braking and positioning with three-phase cage induction motors
A Reduced Device Count of Self
Balancing Five-Level Switched
Capacitor-Based Grid-Connected
Inverter
1 Introduction
In Renewable Energy Source (RES) applications, the Multilevel Inverters (MLIs)

are very popular and produce staircase waveform with imitation of sinusoidal wave
[1]. Conventionally, three types of MLIs have been reported, viz., Neutral Point
Clamped (NPC), Flying Capacitor (FC), and Cascaded H-Bridge (CHB). Generally,
for any MLI, the number of switch count, input DC sources, gate drivers, discrete
diodes, capacitors, and Total Blocking Voltage (TBV/TSV) are important to decide
the size of the inverter and requirement of filter size in grid connection. For generating
five-level output, the NPC inverter takes more number of diodes, more number of
capacitors, and voltage gain becomes unity. Additionally, to balance the capacitors,
separate auxiliary circuits or complex control schemes are required [2]. The number
of requirements is more as compared with NPC, and in this, for better operation, the
capacitors must be precharged [3]. As compared with NPC and FC, CHB MLI has
higher modularity but requires more number of isolated DC sources [4]. The TBV
is more in [5] for generating five-level output with a single DC source. Generally,
in most of MLIs, less TSV has been preferred for reducing the cost. The topology
[6] provides five-level output with a single DC source and single capacitor, but TBV
becomes more. For generating five-level output with a single DC source, it takes TSV
of 11Vdc [7]. The topologies [8, 9] produce TBV of 12Vdc with the same number of
component count for generating five-level output and a separate backend H-Bridge,
which is required in [9] for generating negative voltage levels. The topology [10]
requires two DC sources with TSV of 10Vdc for generating five-level output, but the
drawback of this topology is unity voltage gain, which means there is no boosting
ability. For generating five-level output, the topologies [11, 12] require two capacitors
and a single DC source. Here, both topologies give TSV of 14Vdc. The topology
[13] gives less TBV, but requires two DC sources and there is no boosting ability. To

alleviate the aforementioned problems, a new RDC five-level output inverter topology
has been introduced in this paper and it provides self-balancing of capacitors and
boosting ability with a single DC source. Conventionally, there are different types of
modulation techniques [14, 15] that are available and, in this paper, a new ULSFCS
has been reported in the RL load case by taking reference of Nearest Level Control
(NLC) technique. For higher ripple attenuation purpose, the LCL filter has been
integrated with the grid [16, 17].
The rest of the paperwork is categorized as follows. The operation and control
schemes of the proposed topology have been presented in Sect. 2. In Sect. 3, the
comparative study has been described. The Sect. 4 gives the simulation results for
both RL load and grid-connected cases. Finally, the conclusion has been presented
in Sect. 5.
2.1 Operation of Proposed Inverter Topology with LCL Filter
The schematic of RDC inverter topology with LCL filter is depicted in Fig. 1. It
comprises eight unidirectional switches, one bidirectional switch and one SC with
a single DC source supply. Here, the switches (S1, S2), (S3, S4), (S6, S7), and (S8,
S9) are in complementary mode and the bidirectional switch S5 is connected in
common emitter configuration mode for reducing the gate driver circuit by one as
compared with common collector configuration mode. The operation of the proposed
topology is depicted in Table 1. Generally, the capacitors are charged in parallel and
discharged in series, then finally with proper switching action, the capacitor has
been self-balanced. From Table 1, it is concluded that this topology provides voltage
boosting of two with respect to a single DC source. Generally, the voltage stress of
each switch is indexed with TBV/TSV. In this topology, all switches face an equal
Fig. 1 Proposed five-level

RDC inverter topology with
LCL filter S1 S3 S 6 S8
VDC A B +
C
S2 S 4 S7 S9
Iinv S5
LCL Filter
Igrid
Vgrid
A Reduced Device Count of Self Balancing Five-Level … 365
Table 1 Switching arrangement—five-level SC inverter

Switches 0Vdc 1Vdc 2Vdc −1Vdc −2Vdc
S1 0 1 1 0 0
S2 1 0 0 1 1
S3 1 1 0 1 1
S4 0 0 1 0 0
S5 1 1 0 1 0
S6 1 1 1 1 0
S7 0 0 0 0 1
S8 0 0 0 1 1
S9 1 1 1 0 0
Table 2 Standing voltage of

S. No. Switches Standing voltage
all switches
1 S1 Vdc
2 S2 Vdc
3 S3 Vdc
4 S4 Vdc
5 S5 Vdc + Vdc
6 S6 Vdc
7 S7 Vdc
8 S8 Vdc
9 S9 Vdc
Total standing voltage in per unit 10 Vdc /2 Vdc = 5 pu
amount of voltage stress Vdc individually, then per unit value of TBV becomes five
and it is depicted in Table 2. Based on the maximum discharging period (MDP)
method, it is possible to get proper switched capacitance value within permissible
limits of voltage ripple. However, in this paper, little more capacitance value has
been considered for making successful operation of the grid.
As compared with L and LC filters; the LCL filter provides higher ripple attenua-
tion, but it suffers from resonance problem and finally the system becomes unstable.
To alleviate the resonance problem, the active damping method has been preferred,
and in this method, there are no additional power losses as compared with the passive
damping method. The PLECS/PSIM software are very helpful for finding switching,
conduction losses of any MLI topology. Lastly, the advantages of proposed topology
is that it has less TBV, boosting ability and self-balanced capacitor with single dc
source supply. This DC source supply can be considered PV with boost converter or
fuel cell with boost converter with proper MPPT algorithms or suppose if PV panels
are more then it is possible to integrate with single stage grid connected systems also.
2.2 Proposed Control Schemes
(i) ULSFCS-RL load case
The new ULSFCS has been applied to the proposed topology and its strategy as
depicted in Fig. 2. The proposed control scheme is the modified version of the NLC
technique and it is also called as Unified Rounding Control Scheme (URCS). The
URCS provides less computation time as compared with conventional NLC and SHE
PWM techniques.
In this scheme, modulating signal is given to rounding function and it generates an
aggregated signal (AS). The Switching Table Generation (STG) is taken from Table
1. Because of truncation, there is no need for a logic gate requirement for generating
gate pulses (GP). This control scheme is basically a fundamental switching frequency
technique, where switching losses greatly reduce and it also provides self-balancing
of capacitors. It is universally applicable for any MLI topology with a valid switching
table.
(ii) dq frame current control strategy-Grid Connected Case
The control strategy for five-level SC inverter in grid-connected case has been
depicted in Fig. 3. Firstly, sense the grid voltage (let, Vgrid = Vβ ) and through proper
Fig. 2 Unified rounding control scheme-URCS/ULSFCS
Fig. 3 Proposed dq frame current control scheme

Table 3 Comparative analysis with different conventional five-level inverter topologies

S. No. Description [5] [6] [7] [8] [9] [10] [11] [12] [13] P
1 NLevels 5 5 5 5 5 5 5 5 5 5
2 NSwitches 12 6 6 8 6 8 8 8 10 10
3 NDrivers 12 6 6 8 5 8 8 8 9 9
4 NDiodes 12 8 7 8 6 10 8 10 10 10
5 NSources 1 1 1 2 2 1 2 2 1 1
6 NCapacitors 2 1 1 0 0 2 0 0 2 1
7 TSV 20 12 11 12 10 14 12 10 14 10
PLL, the Vd and Vq components have been generated. Here, PLL has been used to
generate reference variable for grid synchronization. Similarly, the Id and Iq compo-
nents have been generated by sensing inverter current (let, Iinv = Iβ ). To inject active
and reactive power into the grid, Id(ref) and Iq(ref) have been used. From the decou-
pling control strategy, the modulating β-component signal has been generated. By
using Proper Gain (PG ) value, it is possible to generate desired modulating signal.
By taking absolute function, it is possible to curtail the number of high-frequency
carriers present in the Repeating Table Generation (RTG) block. After truncation of
ASG and STG, the required switching pulses/gate pulses (GP) have been generated.
By taking Iq(ref) equal to zero, it is possible to get UPF operation on the grid side.
3 Comparative Study
Table 3 shows the comparative analysis with different conventional five-level inverter
topologies. From this table, it is concluded that the proposed topology gives less TBV
with a single DC source supply, thereby decreasing the cost requirement of switches..
Generally, switch requirement reduces means; respective protection circuit, gate
driver circuit, and heat sink also reduces thereby decreasing the size of the inverter.
4.1 RL Load Case
Table 4 shows simulation parameters-RL load case. In this, for getting peak value
of voltage 400 V, the input DC supply Vdc = 200 V has been considered. The
ULSFCS has been applied to the proposed topology with step change (at t = 0.05 s)
in Modulation Index (MI) values 1 and 0.5. Figure 4 shows simulation results for
five-level SC inverter with RL load. From this, the capacitor is self-balanced at 200
Table 4 Simulation
parameters—RL load case
1 Vdc 200 V
2 Voltage gain 2
3 TSV (pu) 5
4 R 100
5 L 30 mH
6 SC 3300 μF
7 MI 1 and 0.5
Fig. 4 Simulation results for five-level SC inverter with RL load: ULSFCS/URCS
volts even if the step changes in MIs. With MI = 1 and 0.5, the inverter generates
five-level and three-level outputs, respectively.
4.2 Grid-Connected Case
Table 5 shows the simulation parameters of the grid-connected case. The LCL filter
values have been calculated from [16, 17]. Figure 5 shows simulation results for step
changes in Id(ref) from 20 to 10 A: UPF operation mode. From Fig. 5, the capacitor
is self-balanced at 200 V and generates five-level output then grid voltage and grid
current are maintains UPF operation even step changes in Id(ref). By the proposed
control strategy of Fig. 3, the grid current exactly reaches the peak value according
to Id(ref). Generally, Id(ref) is used to inject active power into the grid. Figures 6 and 7
shows harmonic spectrums of inverter output voltage and grid current with respect
to Id(ref) = 20A.
Table 5 Simulation
parameters—grid-connected
case 1 Grid voltage 230 V (RMS)
2 Switching Frequency 10 kHz
5 Rated power(S) 2 kVA
6 Vdc 200 V
7 Inverter side inductance 3.98 mH
8 Grid side inductance 4 mH
9 Filter capacitance 5.95 μF
10 Switched capacitor 4700 μF
Fig. 5 Step changes in Id(ref) . UPF operation mode
5 Conclusion
The proposed topology provides less per unit TSV and voltage boosting ability with
less number of devices. In this, the capacitor is self-balanced without the need for
auxiliary circuits and separate complex control schemes. The voltage stress of each
switch does not exceed the input DC supply. A comparative study is also done with
other conventional five-level inverter topologies. With RL load and grid-connected
cases, the proposed control schemes give good results in the output. Even if step
changes in direct axis reference current values, the grid voltage and grid current still
maintain UPF operation. This topology is well suitable for photovoltaic and fuel
cell-based grid-connected applications.
Fig. 6 Harmonic spectrum—five-level output voltage
Fig. 7 Harmonic spectrum—grid current with Id(ref) = 20A
References
1. Rodriguez J, Lai J-S, Peng FZ (2002) Multilevel inverters: a survey of topologies, controls,
and applications. IEEE Trans Ind Electron 49(4):724–738. https://doi.org/10.1109/TIE.2002.
801052
Trans Ind Appl IA-17(5):518–523. https://doi.org/10.1109/TIA.1981.4503992
3. Shukla A, Ghosh A, Joshi A (2007) Capacitor voltage balancing schemes in flying capacitor
multilevel inverters. In: 2007 IEEE power electronics specialists conference, Orlando, FL,
USA, pp 2367–2372. https://doi.org/10.1109/PESC.2007.4342381

4. Peng FZ, Lai J-S, McKeever J, VanCoevering J (1995) A multilevel voltage-source inverter
with separate DC sources for static VAr generation. In: IAS ‘95. Conference record of the 1995
IEEE industry applications conference thirtieth IAS annual meeting, Orlando, FL, USA, vol
3, pp. 2541–2548. https://doi.org/10.1109/IAS.1995.530626
5. Cheng C, He L (2016) Flying-capacitor-clamped five-level inverter based on switched-capacitor
topology. In: 2016 IEEE energy conversion congress and exposition (ECCE), Milwaukee, WI,
USA, pp 1–5. https://doi.org/10.1109/ECCE.2016.785512
6. Ye Y, Cheng KWE, Liu J, Ding K (2014) A step-up switched-capacitor multilevel inverter
with self-voltage balancing. IEEE Trans Industr Electron 61(12):6672–6680. https://doi.org/
10.1109/TIE.2014.2314052
7. Barzegarkhoo R, Kojabadi HM, Zamiry E, Vosoughi N, Chang L (2016) Generalized structure
for a single phase switched-capacitor multilevel inverter using a new multiple DC link producer
with reduced number of switches. IEEE Trans Power Electron 31(8):5604–5617. https://doi.
org/10.1109/TPEL.2015.2492555
8. Ruiz-Caballero DA, Ramos-Astudillo RM, Mussa SA, Heldwein ML (2010) Symmetrical
hybrid multilevel DC–AC converters with reduced number of insulated DC supplies. IEEE
Trans Industr Electron 57(7):2307–2314. https://doi.org/10.1109/TIE.2009.2036636
org/10.1109/TIA.2005.847306
10. Hari Priya V, Sreenivasarao D, Siva Kumar G (2017) Improved pulse-width modulation scheme
for T-type multilevel inverter. IET Power Electron 10(8):968–976. https://doi.org/10.1049/iet-
pel.2016.0729
11. Niu D et al (2019) A novel switched-capacitor five-level T-type inverter. In: 2019 2nd interna-
tional conference on smart grid and renewable energy (SGRE), Doha, Qatar, pp 1–6. https://
doi.org/10.1109/SGRE46976.2019.9021066
12. Ranjan A, Gupta KK, Kumar L, Jain S (2012) A switched-capacitors based multilevel boost
inverter with single input source. In: 2012 IEEE 5th India international conference on power
electronics (IICPE), Delhi, India, pp 1–6. https://doi.org/10.1109/IICPE.2012.6450373
13. Sivakumar K (2015) A fault-tolerant single-phase five-level inverter for grid-independent PV
systems. IEEE Trans Ind Electron 62(12):7569–7577. https://doi.org/10.1109/TIE.2015.245
5523
0861.2021.9404491
16. Ali A, Shanmugham P, Somkun S (2017) Single-phase grid-connected voltage source converter
(iEECON), Pattaya, Thailand, pp 1–6. https://doi.org/10.1109/IEECON.2017.8075720
17. Zhang N, Tang H, Yao C (2014) A systematic method for designing a PR controller and active
damping of the LCL filter for single-phase grid-connected PV inverters. Energies 7:3934–3954.
https://doi.org/10.3390/en7063934
Pitch Angle Control Using Fuzzy Logic
for DFIG-Based WECS
Kirti Kaharwar and Bhavnesh Kumar
1 Introduction
Electrical energy has a vital crucial role in our everyday life. The world’s progress
is greatly dependent upon this energy as the industrialization process and economic
developments are closely associated with it. However, with decreasing fossil fuels
and increasing environmental issues like global warming and air pollution, the use
of renewable and cleaner energy generation processes is growing. Thus, there is an
acute need for green energy resources like solar, wind, hydro, tidal, and geothermal
[1, 2].
Wind and solar energy harnessing systems are more attractive among all the
available renewable energy sources as they are free, abundant, and available all over
the globe. Among the two, wind energy is mostly popular because of its availability
at night also. However, wind patterns show many deviations from the regular pattern
leading to output power fluctuations. Hence, output power leveling is very much
required in these systems [3].
For maintaining constant output power in WECS (Wind energy conversion
system), an effective control mechanism is required to regulate the pitch angle.
Conventionally, PI controllers are used for leveling output power by twisting the
blades by a certain angle in case the wind speed is higher than the rated flows [4].
Some researchers have also explored the system’s performance with PID controllers
[5]. Although, simple to implement, these controllers are impotent in handling nonlin-
earities effectively. In [6], the H-infinity controller was proposed to provide reliable
performance to the system even if there are deviations in wind and turbine parame-
ters. However, it was a quite complex strategy to implement. The technique Sliding
mode control (SMC) was also investigated to control pitch angle [7]. Prominent
K. Kaharwar (B) · B. Kumar

Instrumentation and Control Engineering Department, Netaji Subhas University of Technology,
New Delhi, India
374 K. Kaharwar and B. Kumar
results were presented for effectively handling all the nonlinearities and uncertain-
ties present in the system. But the disadvantage of this method is that it requires wind
data and the chattering effect is also there. To counter these drawbacks, IMC-PID
(Internal Model Control-PID)-based controllers were implemented to regulate the
pitch angle [8]. It provides the finest performance only at some specific operating
conditions, but the demerit is that it is unable to account for all operating condi-
tions, which change according to wind speed. For regulating and maintaining the
extracted output power constant, a neural network-based SMC was designed [9].
This method was able to cater to all the uncertainties present in the aerodynamic
model and continuous function was used instead of sign function to reduce the chat-
tering phenomenon. Control of pitch angle using General predictive control (GPC)
has also been introduced [10]. In this method, the error is directly proportional to GPC
(General predictive control) law, so if there is a large error, then the controller will
become unstable. Some authors also proposed Model Predictive Control (MPC).
In [11], MPC (Model Predictive Control) provides an optimized solution despite
considering all uncertainties and system constraints. It can predict future behavior
using the system’s model. But the problem with MPC (Model Predictive Control) is
to solve quadratic programming online. Therefore, an intelligent controller named
fuzzy logic is designed that can handle nonlinear functions of the system [12]. Math-
ematical data of wind turbines is not required and it is simple to implement and
very effective than the above controllers. There are several pieces of kinds of litera-
ture about this fuzzy controller. The controller Linear Active Disturbance Rejection
Control (LADRC) was proposed in [13]. LADRC can estimate the external as well
as internal disturbances and can compensate for them in real time. Its purpose is to
modulate the output power and the speed of the generator and to overcome safety
issues. In [14, 15], the fuzzy logic controller was introduced for MPPT (Maximum
power point tracking). In [16], a fuzzy logic-based pitch angle controller was designed
in which the pitch angle was dependent on wind speed and the error voltage. In [17],
the input to a fuzzy controller is wind speed, which will make the system costlier
because of the anemometer required to compute the wind speed. In [18], a pitch
angle controller based on fuzzy logic is implemented for the stabilization of the
power system. In [19], mechanical power variations are given as input to the FLC
(fuzzy logic controller) and if a sudden variation occurs in the wind speed, then
the controller was not effective and fluctuations are observed and MPPT (Maximum
power point Tracking) was also not there as compared to conventional PI controllers.
Some researchers integrated the neural network, fuzzy logic, and PID and proposed
a new scheme called Adaptive Fuzzy PID [20]. This method was designed to control
the disturbances that can harm the system’s efficiency. This method is capable of
self-tuning the parameters under various conditions.
Hence, this paper outlines the leveling of output power, which is done by regulating
the pitch angle using the fuzzy logic controller. The inputs applied to the controller
are Error (change in power), variation in error, rotor speed, and the pitch angle, which
is the output of the controller [21]. This approach is applied to the nine MW wind
farm and can sustain constant power output for speeds exceeding the rated wind
speeds.
Pitch Angle Control Using Fuzzy Logic for DFIG-Based WECS 375
Section 2 outlines the DFIG modeling and wind turbine modeling. Section 3
represents the implementation of leveling of output power, i.e., control of pitch
angle with the help of a fuzzy-based controller, and Results and simulation studies
are depicted in Sect. 4. The conclusion is drawn in Sect. 5 and references follow
thereafter.
2 System Modeling and Description
2.1 A DFIG Model
Figure 1 illustrates the schematic of WECS (Wind energy conversion system) under
investigation with pitch angle control mechanism. The complete system comprises a
wind turbine, gear train, shaft, doubly fed induction generator (DFIG), three-phase
transformer, transmission line, grid, converters of Rotor, and Grid side connected
through a DC link capacitor to regulate DC voltage. The DFIG modeling can be
explained through the following equations.
The voltages equations in a d-q reference frame are as follows:
dds
vds = rs i ds + − ωs qs (1)
dt
dqs
vqs = rs i qs + + ωs ds (2)
dt
Turbine
3 Phase Transformer
Vw Ps, Qs
Pm Tem
Gear Train DFIG Grid
Wind
flow
Pr, Qr
Pr, Qr
Converter of Converter of
ß Rotor side Grid side
Pm
Wind Turbine Controller
Pitch Angle Pitch Controller
Pref
Fig. 1 Description of DFIG-based wind turbine model

ddr
vdr = rr i dr + − ωr qr (3)
dt
dqr
vqr = rr i qr + + ωr dr (4)
dt
where rs and rr is defined as the stator and rotor windings resistance, respectively,
ωs and ωr is defined as the rotational speed of stator and rotor, respectively.
The flux linkages are given by the below equations
ds = L s i ds + Mi dr (5)
qs = L s i qs + Mi qr (6)
dr = L s i dr + Mi ds (7)
qr = L s i qr + Mi qr (8)
where L s is defined as the stator inductance and L r is defined as the rotor inductance
and M is mutual inductance.
By using the above Eqs. 1–8, the DFIG (Doubly fed Induction Generator) model
can be expressed as
dds rs rs
= − i ds + ωs qs + Mi dr + vds (9)
dt Ls Ls
dds rs rs
= − i qs − ωs ds + Mi qr + vqs (10)
dt Ls Ls
di dr rr M dds M
= − i dr + ωr σ i qr − + ωr qs + vdr (11)
dt σ L s dt Ls
di qr rr M dqs M
= − i qr − ωr σ i dr − − ωr ds + vqr (12)
dt σ L s dt Ls
2
where σ = L r − ML s is known as the dispersion factor.
The electromagnetic torque expression for the DFIG (Doubly fed Induction
Generator) system is described below
3 M
Tem = P ds i qr − qs i dr (13)
2 Ls
The useful and reactive power are defined below and the power losses are
considered negligible:
3
Ps = vds i ds + vqs i qs (14)
2
3
Qs = vqs i ds − vds i qs (15)
2
2.2 Wind Turbine Modeling
The aerodynamic power Pm is expressed as
1
Pm = ρ AC p (λ, β)vw3 (16)
2
And wind turbine torque is expressed as
Pm
Tm = (17)
ωr
1 v2
Tm = ρπ R 3 C p (λ, β) w (18)
2 λ
where ρ is the density of air in kg/m3 , A = πr 2 denotes turbine blade area in m2

and R denotes blade radius in m and ωr represents the generator rotational speed in
rad/sec.
The expression for Cp (power coefficient) is obtained by the relation given below,
which is derived by turbine characteristics

a2 a
− 5
C p (λ, β) = a1 − a3 β − a4 e λi + a6 λ (19)
λi
1 1 1
= − 3 (20)
λi λ + 0.08β β +1
C p is the coefficient of power, and from the Eqs. (19) and (20), it can be stated
that C p is a function of both blade tip speed ratio λ and the pitch angle of the blade
β. The coefficients a1 to a6 are defined as: a1 = 0.5176, a2 = 116, a3 = 0.4, a4 = 5,
a5 = 21, and a6 = 0.0068. Figure 2 depicts the deviation of C p (power coefficient)
with respect to λ (tip speed ratio) for distinct values of the pitch angle β and also
given that Cpmax = 0.5 is obtained at β = 0 degrees and for λ = 10.
0.6
β=0°
0.5
β=2°
0.4 Cpmax
β=4°
β=6°
0.3
Cp
0.2 β=8°
β=10°
0.1
β=14°
β=12°
0
0 5 10 15 20
Lambda
Fig. 2 Characteristics of Cp (λ, β)
For each pitch angle β, there exists a single λopt , which gives the coefficient Cp
for the maximal value of power.
The tip speed ratio (TSR) is expressed as
ωr R
λ=
vw
At any given wind speed vw , there will be the optimal rotational speed ωr opt at
which power will be maximum and from this optimal rotational speed, the optimal
tip speed ratio can be calculated, which then gives maximum Cp (power coefficient).
The optimal value of tip speed ratio can be calculated by knowing the optimized
rotational speed
ωr opt R
λopt =
vw
The wind turbine operation can be studied into four regions (Fig. 3). The region
where the wind speed is smaller than the cut-in speed (the speed at which power
generation starts) is denoted by region 1 and the power generation is zero in this
region. Region 2 corresponds to the region between the cut-in and rated wind speed
at which power generation starts growing. And for the winds having speeds greater
than rated speed and below cut out speed is region 3 where power is to be sustained
at the rated value. Above cut-out wind speed, the turbine is taken out of operation to
safeguard the generator parts, and the generated power is zero in this region. Pitch
control action will be taken in region 3 to sustain the output power.
Prated
9 Region 1 Region 2 Region 3 Region 4
Power(MW)
0 6(Vcut-in) 12(Vrated) 30(Vcut-out)

wind speed (Vw) m/s
Fig. 3 Mechanical power versus wind speed curve
3 Proposed Control Strategy: FLC (Fuzzy Logic

Controller)
Figure 4 represents the schematic of the proposed FLC. It comprises a fuzzifier,

knowledge rule base, inference engine, and a defuzzifier. The role of the fuzzifier
is to convert crisp inputs into fuzzifier inputs. The knowledge rule base defines the
relationship between the input and the output membership functions. The inference
engine is the core of the fuzzy logic controller, which can simulate the fuzzified
inputs with the help of a knowledge rule base developed by humans to accomplish
the desired control action. Defuzzifier converts the fuzzified quantities obtained from
the inference engine into crisp quantities.
Mamdani type of inference technique is used and the defuzzification method used
is center of gravity (COG). Three inputs are applied to the fuzzy controller and the
controller as shown in Fig. 5 produces one output.
Inference
Defuzzifier Crisp
Crisp Inputs Fuzzifier Engine
(COG) Outputs
(Mamdani)
Knowledge Rule
Base
Fig. 4 Block diagram of FLC

Pg ΔP
+
- β (Pitch angle)
Δ(ΔP) Fuzzy Logic Rate
d/dt Saturator
Pref
Controller Limiter
ωr
Fig. 5 Proposed fuzzy logic control strategy
Three inputs applied are:

(a) Change in power (P)
(b) Variation in the change in power ((P))
(c) Rotor speed (ωr).
One output, i.e., Pitch angle (β) is there.
P = Pm − Pr e f
δ( P) = P(i) − P(i − 1)
In fuzzy, linguistic variables are defined and are expressed as small (S), Medium
(M), Big (B), Large Negative (LN), Small Negative (SN), Zero (Z), Small Positive
(PS), Large Positive (LP). The fuzzy rules are formulated as given in the table. These
rules can be interpreted as if the rotor speed.
( r ) is B (Big) and P is SP (Small Positive) and δ (P) is LP (Large Positive)
then pitch angle is LP (Large Positive). The fuzzy rule base, inputs, and output’s
membership functions are given in Table 1 and Fig. 6, respectively.
Table 1 Rule Base for FLC

r S M B
P LN SN Z SP LP LN SN Z SP LP LN SN Z SP LP
δ(P) LN LN LN LN LN N SN SN SN P P SP SP P P LP
SN LN LN LN N N SN SN Z P P SP SP P P LP
Z LN LN N SN SN SN Z Z P P SP P P LP LP
SP LN N N SN SN SN Z Z P SP P P LP LP LP
LP N N SN N SN Z Z SP SP SP P P LP LP LP
a) Error(change in output power) b) change in error
c) Rotor Speed d) Pitch Angle
Fig. 6 FLC membership functions
4 Simulation Results and Analysis
To examine the potency of the WECS on the presented FLC, simulation is performed
on a DFIG-based wind turbine having a capacity of 9 MW. The simulation is
performed on both the controllers, i.e., the PI controller and the fuzzy logic controller.
In the PI controller, the values of Kp and Ki were varied until the desired tuned
response is obtained [22] and were found to be 50 and 50, respectively.
First, the results of both the controllers are compared for a particular wind speed
of 15 m/s as given in Fig. 7. In Fig. 7a, the output power is maintained constant,
i.e., 9 MW, but oscillations are present in PI controllers as compared to the FLC.
There is a sudden overshoot initially for PI controller, whereas in FLC, overshoot is
minimal and the constant power and pitch angle is maintained constant for particular
wind speed (Fig. 7b). Similarly, for the coefficient of power in Fig. 7c, respectively,
oscillations are sustained for the PI controller, whereas in FLC oscillations die out
after a few seconds. In Figs. 7d, e, the FLC is maintaining tip speed ratio and rotor
(a) (b) Pitch Angle
(c) (d)
(e)
Fig. 7 Responses for PI controller and FLC at a particular wind speed of 15 m/s: a generated output
power, b pitch angle, c coefficient of power, d tip speed ratio, e rotor speed
speed constant for particular wind speed, whereas the controller PI is unable to
perform so. These parameters are continuously decreasing. Therefore, overall FLC
performs better rather than the PI controller.
For realizing the effective implementation of the given FLC, the system has been
examined on the wind profile illustrated in Fig. 8.
Responses recorded for case-II are shown in Fig. 9. In Fig. 9b, FLC is maintaining
the power constant and reducing the oscillations also but the controller PI is unable to
perform well and sustain oscillations. In Fig. 9c, FLC is maintaining the fixed pitch
angle for fixed wind speed, instantly responding to step-change in wind speed, but
PI is incompetent to do so. There are so many variations in the PI controller for pitch
angle as well as for all the other parameters like Cp, lambda (tip speed ratio), and
Fig. 8 Wind profile for

case-II
rotor speed. Overall FLC is a very effective controller in maintaining the constant
output power and regulating other parameters.
5 Conclusion
A 9 MW wind farm is examined for its operation on wind speed higher than rated. The
absolute system model is created in MATLAB/Simulink environment for examina-
tion. The control strategy based on fuzzy logic is described and developed to manage
the pitch angle. It is evident from the results that the system with FLC outperforms
the PI controller in maintaining the constant output power for wind speeds exceeding
the rated wind speeds. It also helps in reducing the oscillations that occur in output
power in comparison to the PI controller. FLC (Fuzzy logic controller) can also regu-
late other parameters like rotor speed, coefficient of power, tip speed ratio, and pitch
angle. Overall, FLC outperforms the PI controller in regulating the power output and
other parameters and gives a stable response to all the deviations, which occur in
wind speed for a particular time.
(c) (d)
Fig. 9 Responses for PI controller and FLC at a constant variation in wind speed: a generated
output power, b pitch angle, c coefficient of power, d tip speed ratio, e rotor speed
References
1. Committee on Environmental Impacts of Wind Energy Projects, “Environmental Impacts of

Wind Energy Projects”, Washington, The National Academies Press (2007)
2. Report of New and Renewable Energy Authority. http://www.nrea.gov.eg. Last access 20 Oct
2014
3. Mi Y, Bao X, Jiang E, Deng W, Li J, Ren L, Wang P (2014) The pitch angle control of squirrel-
cage induction generator wind power generation system using sliding mode control. In: 16th
European conference on power electronics and applications, Lappeenranta, pp 1–10
4. Hansen H, Larsen L (2005) Control design for a pitch-regulated, variable speed wind turbine.
Riso, National Laboratory Roskilde, Denmark
5. Vidal Y, Acho L, Luo N, Zapateiro M, Pozo F (2012) Power control design for variable-speed
wind turbines. Energies 5:3033–3050
6. Gaur P, Pathak D, Kumar B, Chauhan YK (2016) PI and fuzzy logic controller based tip speed
ratio control for smoothening of output power fluctuation in a wind energy conversion system.
In: 2016 7th India international conference on power electronics (IICPE), Patiala, India, pp
1–6. https://doi.org/10.1109/IICPE.2016.8079480
7. Shahgholian G, Karimi H, Mahmoodian H (2012) Design a power system stabilizer based on
fuzzy sliding mode control theory. Int Rev Model Simul. 5(5):2191–2196
8. Naik KA, Gupta CP (2016) Improved oscillatory behavior of a grid-connected wind farm using
IMC-PID based pitch angle controller. In: 2016 IEEE 6th International Conference on Power
Systems (ICPS), New Delhi, pp 1–6
9. Dastres H, Mohammadi A Shamekhi M (2020) A neural network based adaptive sliding mode
controller for pitch angle control of a wind turbine. In: 2020 11th power electronics, drive
systems, and technologies conference (PEDSTC), pp 1–6. https://doi.org/10.1109/PEDSTC
49159.2020.9088373
10. Senjyu T, Sakamoto R, Urasaki N, Funabashi T, Fujita H, Sekine (2006) Output power leveling
of wind turbine generator for all operating regions by pitch angle control. IEEE Trans Energy
Convers 21(2):467–475
11. Spencer MD, Stol KA, Unsworth CP, Carter JE, Norris SE (2013) Model predictive control of
a wind turbine using short-term wind field predictions. Wind Energy 16(3):417–434
12. Chowdhury M, Hosseinzadeh N, Shen W (2012) Smoothing wind power fluctuations by fuzzy
logic pitch angle controller. Renew Energy 38(1):224–233
13. Ullah F, Ali S, Ying D, Saeed A (2019) Linear active disturbance rejection control approach
base pitch angle control of variable speed wind turbine. In: 2019 IEEE 2nd international
conference on electronics technology (ICET), pp 614–618. https://doi.org/10.1109/ELTECH.
2019.8839443
14. Chen Z, Gomez SA, McCormick M (2000) A fuzzy logic controlled power electronic system
for variable speed wind energy conversion systems. In: Eighth international conference on
power electronics and variable speed drives, Publication No. 475
15. Abo-Khalil AG, Lee D, Seok J (2004) Variable speed wind power generation system based on
fuzzy logic control for maximum output power tracking. In: 35th annual IEEE power electronics
specialists conference, Achen, Germany
16. Demata JV, Dimazana JNM, Fausto RS, Magwili GV, Chua EE, Pacis MC (2019) Utilization
of fuzzy logic controller for pitch control of a small scale wind energy conversion system
(WECS). In: 2019 IEEE 11th international conference on humanoid, nanotechnology, infor-
mation technology, communication and control, environment, and management (HNICEM),
pp 1–6. https://doi.org/10.1109/HNICEM48295.2019.9073448
17. Zhang J, Cheng M, Chen Z, Fu X (2008) Pitch angle control for variable speed wind turbines.
In: Proceedings of the 3rd international conference on DRPT, pp 2691–2696
18. Jauch C, Cronin T, Sørensen P, Jensen BB (2007) A fuzzy logic pitch angle controller for power
system stabilization. Wind Energy 19–30
19. Macedo AVA, Mota WS (2012) Wind turbine pitch angle control using fuzzy logic. In:
Proceeding of the IEEE/TDC-LA, Montevideo, Uruguay, Sep, pp 1–6
20. Ali MMM, Youssef A, Abdel-Gaber G, Ali AS (2018) Adaptive fuzzy-PID based pitch angle
control of wind turbine. In: 2018 twentieth international Middle East power systems conference
(MEPCON), pp 1110–1114. https://doi.org/10.1109/MEPCON.2018.8635229
21. Hosseini E, Shahgholian G (2017) Output power leveling for DFIG wind turbine system using
intelligent pitch angle control. Automatika 58(4):363–374. https://doi.org/10.1080/00051144.
2018.1455017
22. Reynoso-Meza G, Sanchis J, Herrero JM, Ramos C (2012) Evolutionary auto-tuning algorithm
for PID controllers. IFAC Proc 45(3):631–636. https://doi.org/10.3182/20120328-3-IT-3014.
00107
Class-E Power Amplifier-Based Wireless
Power Transfer System
Jay Prakash Narayan, Anamika Das, and Ananyo Bhattacharya
1 Introduction
Wireless power transmission (WPT) is a means of transmitting electrical power with

help of transmitter and receiver over a large distance in the absence of conducting
wires or cables. There are three techniques to transfer the power wirelessly, magnetic
couplings, capacitive coupling, and electromagnetic radiation. Through the electro-
magnetic wave, power can be transferred over a very large distance, but it has lower
efficiency and it has greater detrimental effect on the surroundings in which we are
living. Due to loose coupling, the magnetic field energy and coupling coefficient
quickly attenuate with increasing distance in electromagnetic induction. Due to this
problem, the power can transferred only for few centimeters or sometimes even in
millimeters by using the electromagnetic induction. In the electromagnetic radiation
mode, the transfer efficiency declines as the wavelength of power transmission rises.
Magnetic coupling method combines the benefits of both of the above technologies.
It can transfer power over a large distance almost 2–4 times of the diameter of the
resonating coils. In the recent, year the demand of the wireless Power Transfer (WPT)
is increased tremendously. The WPT system finds its application in electric vehicles,
charging cables, transportation security, medical implants, pacemaker, etc.
J. P. Narayan (B) · A. Das · A. Bhattacharya

Department of Electrical Engineering, National institute of technology, Jamshedpur, Jharkhand,
India
A. Das
e-mail: 2019rsee002@nitjsr.ac.in
A. Bhattacharya
e-mail: ananyo.ee@nitjsr.ac.in
388 J. P. Narayan et al.
2 System Model
Figure 1 illustrates the WPT System model block diagram. The system consists of
an AC source which is used as an input source, with a diode rectifier rectifying
the input voltage. Using class E power amplifiers, the high voltage is transformed to
high frequency AC. After that, the transmitter is supplied with the oscillating signals.
Power is transmitted wirelessly by maintaining sufficient resonant coupling between
the transmitter and receiver stages.
The proposed scheme of WPT system is illustrated in Fig. 2. At first, the diode
rectifier converts the input AC voltage to DC.
The DC inputs from the rectifier is collected by the class-E amplifier and trans-
forms it to high frequency AC. In switched mode resonant converter, the class E
amplification is very effective. Generally, transistor is used as a switch in the power
amplifier, which have high efficiency due to low power losses in the switch.
The Class-E amplifier is made up of a parallel RF choke (L 2 ) and a Resonant
circuit. The output coming from the class-E amplifier is directly interconnected with
tank circuit, which are made up off transmitting coils and capacitor C2 as shown in
Ac input Class E power

Recfier Transmier setup
source Amplifier
Dc load Recfier stage Receiver setup
Fig. 1 Schematic diagram of (WPT)
L1 L3
D1 L2
D3 D1 D3
C1 C3
260 V ac
Co
S1
D4 D2 Lt Lr C4 Ro
C2
D2
D4
Transmitter Receiver
Side Side
Fig. 2 Circuit configuration of proposed WPT

Class-E Power Amplifier-Based Wireless Power Transfer System 389
Fig. 2. The receiver is made up of the transmitting coils and a tank circuit consisting
of capacitor C4 , as well as a full wave rectifier that transforms the AC into DC. A
filter Co eliminates harmonics before being fed to the load R0. Power is transmitted
between the transmitter and receiver at the resonant condition.
3 Design of Parameters of WPT System
The assumption made in this WPT system are as follows:

(1) It is assumed that the component used in the WPT system is ideal.
(2) The switch resistance during the On state, as well as their parasitic capacitance,
are not taken into account.
(3) The drop of voltage across the diodes are not taken into account.
(4) The value of capacitor is considered to be very large.
According to the requirements of power, the value of load can be calculated by
the given equation

(vin )2 0 · 45 0 · 40
RL = 0 · 576 1 − − (1)
P0 QL Q 2L
where Q L is a constant which is taken by the designer, generally Q L value is between

0.5 to 2. Here Q L value is taken approximately 0.90.
Then, using the equation below, you can determine the value of shunt capacitance
(C1 and C2 ) that is connected through the switch.

1 0.91 1 0.6
C1 = 2 + 1+ − 2 + (2)
2π f o R L π4 + 1 π2 QL QL (2π f o )2 L 1

1 1 1.014 0·2
C2 = + + 1+ − 2 (3)
2π f o R L Q L − 0.104 Q L − 1.78 2π f o L 1
The value of L 1 is randomly selected, which is taken as 1.465 µh and L 2 value is

calculated as;

RL
L2 = QL (4)
2f o
4 Simulation Analysis
The proposed WPT model was simulated with the help of MATLAB/SIMULINK.
The theoretical model has been checked with a 260 V input voltage (Fig. 3).
The output of the system is calculated across the load Ro . The various parameters
used in this WPT system to get the maximum efficiency are shown in Table 1.
The rectifier stage’s output voltage is determined to be 234.2 V DC, which is then
fed to the class-E power amplifier which converts it into high frequency AC and a
300 V output is obtained. Figure 6 displays the waveform of the WPT system output
voltage at the receiving end. Because of the inclusion of an output capacitor, which
acts as a filter, the system’s output voltage becomes stable across time (Fig. 4).
As seen in Fig. 5, the output current of the designed wireless transmission system
flows through the secondary side. The current flowing at the output of proposed WPT
model is 60 mA.
Fig. 3 Implementation of proposed WPT system using MATLAB\SIMULINK
Table 1 Parameters values

Parameter Value
Input voltage 260 V
L1 1.465 µh
L2 70 µh
L3 1.07 mh
C1 187 nf
C2 187 nf
C3 165 nf
C4 65.8 pf
C0 0.5 mf
R0 500
Fig. 4 Waveform of WPT system’s output voltage
The WPT system has an output power of approximately 183.36 W. which is shown
in the Fig. 6 (Table 2).
5 Conclusion
This paper describes a model of WPT system for power transfer in the medium
range. The above system is suitable for the appliances whose output power is within
the 300W. The model takes a 260 v AC input, which is then converted into high
frequency AC. The above proposed WPT models give an output of 300 V across the
output resistor. Generally, two coil WPT systems have efficiency within the range of
45–60%. In this WPT system, an efficiency of 59.4% is observed which indicates a
highly efficient system. The WPT system with full wave bridge rectifier outperforms
the WPT system without full wave rectifier in terms of voltage, power, efficiency,
and other factors.
Fig. 5 Waveform of WPT system’s output current
Fig. 6 Output Power waveform of WPT system

Table 2 Below is a comparison of the outputs of a wireless power transmission system with and
without a full wave bridge rectifier
WPT model Input source Input Output Output Output Ƞ in
voltage (V) voltage (V) current Power (W) percentage
(Amp)
Without full DC source 12 45 0.25 28 54.6
wave rectifier
With full AC source 260 300 0.60 183.36 59.4
wave
rectifier
References
1. Kursk A, Karalis A, Moffatt R, Joannopoulos JD, Fisher P, Soljacic M (2007) Wireless power
transfer via strongly coupled magnetic resonances. Science 317(5834):83–86
2. Hu AP (2009) Wireless/contactless power supply: inductively coupled resonant converter
solutions. Saarbrucken, Germany: VDM
3. Lee S-H, Lorenz RD (2011) Development and validation of model for 95%-efficiency 220-W
wireless power transfer over a 30 cm air gap. IEEE Trans Appl Ind 47(6):2495–2504
4. Pinuela M, Yates DC, Lucyszyn S, Mitcheson PD (2013) Maximizing DC-to-load efficiency
for inductive power transfer. IEEE Trans Power Electron 28(5):2437–2447
5. Sample A, Waters B, Wisdom S, Smith J (2013) Enabling seamless wireless power delivery in
dynamic environments. Proc IEEE 101(6):1343–1358
6. Florian C, Mastri F, Paganelli R, Masotti D, Costanzo A (2014) Theoretical and numerical
design of a wireless power transmission link withgan-based transmitter and adaptive receiver.
IEEE Trans Microw Theory Tech 62(4):931–946
7. Ahn D, Hong S (2014) Wireless power transmission with self-regulated output voltage for
biomedical implant. IEEE Trans Ind Electron 61(5):2225–2235
8. Aldhaher S, Luk P-K, Bati A, Whidborne J (2014) Wireless power transfer using Class E
inverter with saturable DC-feed inductor. IEEE Trans Ind Appl 50(4):2710–2718
9. Hui S, Zhong W, Lee C (2014) A critical review of recent progress inmid-range wireless power
transfer. IEEE Trans Power Electron 29(9):4500–4511
10. Prasanth V, Bauer P (2014) Distributed ipt systems for dynamic powering: misalignment
analysis. IEEE Trans Ind Electron 61(11):6013–6021
11. Aldhaher S, Luk P-K, Whidborne JF (2014) Electronic tuning of misaligned coils in wireless
power transfer systems. IEEE Trans Power Electron 29(11):5975–5982
12. Huang S-J, Lee T-S, Huang T-H (2014) Inductive power transfer systems for PT-based ozone-
driven circuit with flexible capacity operation and frequency-tracking mechanism. IEEE Trans
Ind Electron 61(12):6691–6699
13. Raval P, Kacprzak D, Hu A (2015) Multiphase inductive power transfer box based on a rotating
magnetic field. IEEE Trans Ind Electron 62(2):795–802
14. Ahn D, Hong S (2015) Wireless power transfer resonance coupling amplification by load-
modulation switching controller. IEEE Trans Ind Electron 62(2):898–909
15. Zhong W, Zhang C, Liu X, Hui S (2015) A methodology for makinga three-coil wireless power
transfer system more energy efficient than atwo-coil counterpart for extended transfer distance.
16. Fu M, Zhang T, Ma C, Zhu X (2015) Efficiency and optimal loads analysis for multiple-receiver
wireless power transfer systems. IEEE Trans Microw Theory Tech 63(3):801–812
17. Fu M, Tang Z, Liu M, Ma C, Zhu X (2015) Full-bridge rectifier input reactance compensation in
megahertz wireless power transfer systems. In: Presented at the IEEE PELS workshop emerging
technologies, wireless power, Daejeon, Korea
18. Ahn D, Mercier P (2016) Wireless power transfer with concurrent 200kHz and 6.78 MHz
operation in a single transmitter device. IEEE Trans Power Electron 31(7):5018–5029
Enhancement of Power Factor
and DC-Link Voltage Stabilization Using
Multiple Techniques of HPFC
Satya Venkata Kishore Pulavarthi, Jami Rajesh, Nakka Jayaram,

and Sukanta Halder
1 Introduction
Rectifiers are basically two types. The first one is line commutated rectifiers which use
diodes and thyristors. The second one is Power Factor Correction (PFC) rectifiers
which use MOSFET and IGBT. So, they need force commutation. PFC rectifiers
are also classified as non-regenerative and regenerative PWM rectifiers. Boost and
Vienna rectifiers [1] are non-regenerative PWM rectifiers in which the power flow
is only in one direction from AC source to DC load. Voltage source rectifier and
current source rectifier are regenerative PWM rectifiers in which the power flow is
bidirectional [2]. Regenerative PWM voltage source rectifier is called High-Power
Factor Converters (HPFCs).
Multi-pulse rectifiers [3, 4] are used to reduce the harmonics in the source side
but the use of phase-shifting transformers restricted the use of these multi-pulse
rectifiers.
One can use passive filters to reduce the harmonics. Different types of commonly
used passive filters [5] are bandpass, high-pass, double bandpass, and composite
filters. The main limitation of the passive filters is that they are not suitable for
changing system conditions. The size of the filter and the tuned frequency can’t be
changed so easily.
Next active filters [6] came with power electronic equipment to reduce the
harmonics seen by the source. The disadvantages of active filters include their large
size, rating, control sophistication, and cost.
HPFCs are introduced to overcome these drawbacks and meet the above standards
[7]. This lead HPFCs to be an essential segment of any AC–DC conversion system
due to their well-regulated DC output, reduced size, and higher efficiency. There are
S. V. K. Pulavarthi (B) · J. Rajesh · N. Jayaram · S. Halder

National Institute of Technology Andhra Pradesh, Tadepalligudem, India
e-mail: kishorepulavarthi.sclr@nitandhra.ac.in
396 S. V. K. Pulavarthi et al.
variants of HPFCs, which are widely inherited for their characteristics such as sinu-
soidal input currents, negligible THD of input currents, good supply pf, controlled
DC output voltage, fewer ripples in the DC output voltage, low switching stresses,
and low Electro-Magnetic Interference (EMI) emissions.
Predictive reasoning-based adaptive hysteresis band control for the three-phase
PWM rectifier is presented in [8]. But the authors have not shown the basic character-
istics of PWM rectifier such as unity power factor and DC-link voltage stabilization.
Model predictive control based on two vectors [9] is applied to the PWM rectifier
to control the power flow. A novel low complexity model predictive direct power
control (LC-MPDPC) for the PWM rectifier is presented in [10] to realize the active
and reactive power control. This control has the advantage of eliminating the AC
voltage sensor but the control involves complexity. A power predictive control for
a three-phase PWM rectifier is proposed in [11]. This method has the advantage
of eliminating the PI controller. A vector control method without an input voltage
sensor for the PWM rectifier was proposed in [12]. The steady-state model of the
filter inductor and the output of the current controller estimate the input voltage infor-
mation required by the PLL. Authors in [13] proposed a new backstepping control
system for the Three-phase PWM rectifier to control the active and reactive powers
instantaneously. This method gave better results compared to the PI controller and
PR (proportional resonant) controller.
In this paper, different control techniques used for three-phase HPFC are discussed
and simulated [14, 15].
The remaining paper is structured as follows. Section 2 discusses the operation
of the HPFC. Control schemes are discussed in Sect. 3. Simulation results of the
HPFC controlled by the above schemes are presented in Sects. 4 and 5 represents
the conclusion.
2 High-Power Factor Converter
Figure 1 shows the circuit diagram of the three-phase HPFC. A feedback control
loop is used in it to make the output voltage/DC-link voltage constant at the desired
value. The DC-link voltage (Vo ) is measured and compared with the desired reference
voltage of Voref . The resultant error signal is routed through the control block, which
turns on and off the converter’s switching devices. HPFC is a bidirectional power flow
converter, the power flow between the load and the source can be regulated based
on the DC-link requirements. Controlling the DC-link voltage allows the current
flow at the DC link to be reversed, allowing power reversal to be regulated [6]. The
waveforms and phasor diagrams of the HPFC in both rectifier and regenerative mode
are shown in Fig. 2.
The current flowing to the load (Io) is positive in the rectifier mode of operation,
and the capacitor (CD) is discharged via the DC load. In this case, the control circuit
requests more power from the AC supply due to the error signal. For producing the
necessary PWM signals for the switching devices, the control circuit draws power
Enhancement of Power Factor and DC-Link Voltage Stabilization … 397
idc Io
V s Is Ls
Vmod Vo DC
CD
Load
error
Control circuit Voref
Fig. 1 High-power factor converter
Fig. 2 Rectifier (top) and inverter (bottom) modes of operation
from the supply. The capacitor voltage is restored as more current flows from the AC
to the DC side.
Io becomes negative in the inverter or regenerative mode of operation, and the
capacitor CD is overcharged. The control circuit is commanded by the error signal
to discharge the capacitor, allowing power to flow to the AC mains.
Both the active and reactive powers can be regulated by the PWM. Thus, power
factor correction can be done by using this type of rectifiers. The current waveforms
at the AC side can also be maintained almost sinusoidal, thus reducing the harmonic
contamination to the mains supply.
3 Control Techniques
Three control schemes for the control of three-phase High-Power Factor Converter
are discussed. They are (a) Current control technique, (b) Vector control technique,
and (c) Voltage control technique.
3.1 Current Control Technique
The generalized control block diagram is shown in Fig. 3. The output voltage (Vo ) is
compared with the desired reference voltage (Voref ) and the error signal is obtained
as given in Eq. (1).
Error, e = Voref − Vo (1)
This error passed through the PI controller will give the current reference template
I which is shown in Eq. (2) as follows.
t
I = Kp e + KI edt (2)
0
Here, Kp and KI are the proportional gain and integral gain, respectively. This
current template I is multiplied with a unit template Vxu (a sinusoidal signal with the
same phase and frequency of the AC source voltage) to obtain the AC input current
reference Ixref given by Eq. (3).
Ixref = I * Vxu (3)
This Ixref is compared with the actual line current (Ix ) and the error is again passed
through the P controller to get modulating signal(m) given in Eq. (4), which has the
desired phase and frequency.
m = Kpi ∗ (Ixref − Ix ) (4)
I Ixref SPWM
+ PI * + P PULSES
Voref - - Generator
Vo Vxu Ix
Fig. 3 Block diagram of current control technique

where Kpi is the proportional gain used to control the converter current. To produce
the necessary pulses to the converter switches, this modulating signal is compared
to the triangular carrier.
3.2 Vector Control Technique
Another method to control the HPFC is in d-q vector space. The control block diagram
for the vector control technique is shown in Fig. 4.
The input source currents (ia , ib , and ic ) can be represented by using the vector
notation such as is = id + jiq , which is defined by Eq. (5).
⎡ i ⎤
id 2 cos θ sin θ 1− 1
− 1 a
⎣ ib ⎦
= √2 √2 (5)
iq 3 − sin θ cos θ 0 2 − 2
3 3
ic
The actual DC-link voltage/output voltage (Vo ) is compared with the desired refer-
ence voltage (Voref ) and the error is passed through the PI (proportional + integral)
controller, which gives the value of idref and, the iqref is set to zero in order to main-
tain the pf nearly unity. These references are compared with the input currents which
are in d-q coordinates given in Eq. (5). Typically, two proportional controllers are
used to pass these current errors to generate the values of Vmod_d and V mod_q.
By applying the complete inverse transformation given by Eq. (6), the modulation
signals in the abc reference frame can be generated. The SPWM technique can be
used to generate gate drive pulses for the regulated power switches in the HPFC.
⎡ ⎤ ⎡ ⎤
Va mod 1 √0
⎣ Vb mod ⎦ = ⎢ ⎥ cos θ − sin θ V mod _d
⎣ − 21 2√3 ⎦ (6)
sin θ cos θ V mod _q
Vc mod 1
2
− 23
idref Vmod_d
Voref + PI + P
- -
PWM
id Gene- PULSES
Vo
rator
iqref Vmod_q
+ P
-
iq
Fig. 4 Block diagram of vector control technique

3.3 Voltage Control Technique
Voltage Vxmod is generated using the differential equations that govern the HPFC
circuit. The following differential Eq. (7) can be derived from Fig. 1
dix
Vx (t) = Ls + Rix + Vx mod (t) (7)
dt
Assume that, Vx (t) = Vm sin(ωt + φ), and the solution for Eq. (7) to make the
converter work at constant power factor should be of the form given in Eq. (8)
ix (t) = I(t) sin(ωt + φ) (8)
Solving the Eqs. (7), (8), and Vx (t) will yield the expression Vx mod shown in
Eq. (9), which will make the converter work at a fixed power factor.

dI
Vx mod = Xs I sin φ + Vm − RI − Ls cos φ sin ωt
dt
(9)
dI
− Xs I cos φ + RI + Ls − Vm sin φ cos ωt
dt
For unity power factor operation,cos φ = 1 and sin φ = 0.

With these values, the above Eq. (10) can be reduced as Eq. (6) shown below.

dI
Vx mod = Vm − RI − Ls sin ωt − Xs I cos ωt (10)
dt
The block diagram of this technique implemented for unity pf is shown in Fig. 5.
The advantage of this technique is that no input current sensing is needed. Here, sin
ωt and cos ωt are obtained from the AC source.
VmSinwt
Sinwt
Voref I -R-L (d/dt)
+
- PI s
+
Vamod PWM
Vo Vbmod Gene- PULSES
-XS
rator
Coswt Vcmod
Fig. 5 Block diagram of voltage control technique for one leg of a converter
The three-phase HPFC is simulated in MATLAB/SIMULINK with all the control

techniques described above. The parameters taken for the simulation are shown in
Table 1.
4.1 Current Control Technique
The circuit of three-phase HPFC with the current control technique is implemented
in MATLAB/SIMULINK. Figure 6 shows that the DC-link voltage is tracking the
reference value of 600 V. The voltage and current waveforms at the AC side under
steady state which are in phase with each other are shown in Fig. 7. The harmonic
spectrum of the input current wave is shown in Fig. 8 and THD is found to be 1.22%
which is less than the IEEE519 standards [16].
Another 25 load is connected from time 0.1 to 0.2 s, therefore current increases
during this time. We can observe the same thing in Fig. 9. To work the circuit in
inverter mode, a voltage source more than the reference voltage is connected across
the DC bus from 0.3 to 0.4 s so that the current and voltages are in opposition to
each other that can be observed in Fig. 10. From time 0.5 to 0.6 s, the source voltage
Table 1 Parameters used for

Parameter Value
simulation √
AC source peak value 230 (2)
Source inductance 4.5 mH
Load resistance 25
Load side capacitance 3300 µF
Vref 600 V
Sampling time 10 µs
Fig. 6 DC-link voltage

Fig. 7 Voltage and current

waveforms at the AC side
Vs
Is
Fig. 8 AC side current

harmonics
Fig. 9 Waveforms during

load change
Vs Is
is reduced by 20%, therefore the source current increases to maintain the DC-link
voltage back to its reference voltage, and both voltage and current waveforms still
maintain the near unity power factor. The corresponding waveforms are shown in
Fig. 11. Figure 12 represents the DC-link voltage which is regulated by maintaining
the reference voltage at all load conditions and disturbances.
Fig. 10 Waveforms during

inverter (Regenerative) mode
Vs
Is

waveforms during source
perturbation Vs
Is

during disturbances
4.2 Vector Control Technique
The HPFC circuit with the vector control technique is also implemented in
MATLAB/SIMULINK. The DC-link voltage is maintained constant which is shown
in Fig. 13. The voltage and current waveforms at the AC side are shown in Fig. 14

waveforms
Vs
Is
and the harmonic spectrum of the input current wave is shown in Fig. 15 and the
THD is observed to be 2.14%.

harmonics
4.3 Voltage Control Technique
The Three-phase HPFC is also simulated using the voltage control method described
above. Figure 16 shows the DC bus voltage waveform which is made constant at
600 V. AC side voltage and current waveforms shown in Fig. 17 are in phase with
each other. The harmonic spectrum of the AC side current is shown in Fig. 18. and
the THD is observed to be 2.58%. Table 2 shows the comparison of THD for the
above three methods.

waveforms Vs
Is

harmonics
Table 2 Comparison of AC
Method %THD
side current THD
Current control 1.22
Vector control 2.14
Voltage control 2.58
5 Conclusion
Three methods to control the HPFC are discussed. The circuits are simulated in
MATLAB/SIMULINK with parameters given in the above Table 1. The DC-link
voltage is maintained constant during disturbances also and at the same time, the
power factor at the AC side maintained nearly unity. THD is compared and tabulated
for all the control techniques. The results show that the current-control technique
gave the better THD compared with the other two techniques. This HPFC can be
used in the integration of renewable energy sources to the grid.
References
1. Uddin MN, Abedin AH, Bashar KL, Islam S, Choudhury MA (2017) Three phase one switch
modular-boost/vienna power factor corrected (PFC) rectifier. In: 2017 IEEE international
conference on industrial technology (ICIT), pp 230–235
2. Rodriguez JR, Dixon JW, Espinoza JR, Pontt J, Lezana P (2005) PWM regenerative rectifiers:
state of the art. IEEE Trans Ind Electron 52(1):5–22
3. Bin Wu (2006) High power converters and ac drives. IEEE PRESS. A John Wiley & Sons Inc.,
Publication
4. Kocman S, Kolar V, Trung Vo T (2010) Elimination of harmonics using multi-pulse rectifiers.
In: Proceedings of 14th international conference on harmonics and quality of power—ICHQP
5. Das JC (2004) Passive filters-potentialities and limitations. IEEE Trans Ind Appl 40(1)
6. Singh B, Haddad KA, Chandra A (1999) A review of active filters for power quality
improvement. IEEE Trans Ind Electron 46:960–971
7. Rashid MH (2001) Power electronics handbook. Academic Press. ISBN 0-12-581650-2
8. Begag S, Belhaouchet N, Rahmani L (2009) Three phase PWM rectifier with constant switching
frequency. J Electr Syst 01:7–12
9. Zhang Y, Peng Y, Yang H (2016) Performance improvement of two-vectors-based model
predictive control of PWM rectifier. IEEE Trans Power Electron 31(8):6016–6030
10. Li H, Lin M, Yin M, Ai J, Le W (2019) Three-vector based low complexity model predictive
direct power control strategy for PWM rectifier without voltage sensors. IEEE J Emerg Select
Top Power Electron 7(1):240–251
11. Cho Y, Lee K-B (2016) Virtual flux based predictive direct power control of three-phase PWM
rectifiers, with fast dynamic response. IEEE Trans Power Electron 31(4):3348–3359
12. Upamanyu K, Ameta C, Narayanan G (2020) Simplified input voltage sensorless vector control
for PWM rectifiers. IEEE Trans Ind Appl 56(4):4051–4060
13. Wai R, Yang Y, Wang Y (2018) Design of backstepping control for direct power control of
three-phase PWM rectifier. In: 2018 3rd international conference on intelligent green building
and smart grid (IGBSG), pp 1–4
14. Jungreis A (2012) High power factor isolated buck-type power factor correction converter. U.S.
Patent 8,102,678, issued January 24
15. Vaideeswaran V, Sankar N (2018) Control techniques of three phase PWM rectifier. Int J Eng
Adv Technol (IJEAT) 8(2S). ISSN: 2249-8958
16. EEE recommended practices and requirements for harmonics control in electric power systems.
IEEE Std 519 (1992)
Design and Modeling of 3.3 kW
GaN-Based HPFC Converter
for Onboard EV Charger
Piyush Kumar, Amit Singh Tandon, Naveen Yalla, and A. V. J. S. Praneeth
1 Introduction
Power factor correction (PFC) converters are broadly used in EV applications for
unity power factor [1]. Due to the use of rectifier and passive elements (inductor
and capacitors), the power factor of the converter is less than desired. Hence to
improve PF we use the PFC method. In the conventional PFC converter because of
the bridge rectifier diode, there is high conduction loss which reduces the efficiency
of the overall circuit. Researchers have been done research on bridgeless PFC boost
converters to increase the overall efficiency. HPFC converter has low conduction
loss, but when it operates in the CCM there is high reverse recovery loss of the
MOSFET body diode [2, 3]. So, GaN HEMT is used in converter with less switching
loss, less conduction loss and zero reverse recovery charge which overall increases
its efficiency [4, 5]. GaN HEMT switches offer significantly better performance than
MOSFETs. GaN reduces the physical size and improves the power density. High
power density is desirable, especially for EV applications.
In this paper, we discuss two switching operations: high-frequency switching
scheme and reduced switching control scheme. In a high-frequency switching
scheme, all the four switches are operating at high frequency, but in this switching,
we have high dv/dt and high distortion. The above problem is overcome by a reduced
frequency switching scheme. In this, two switches (GaN switches) are operating at
high frequency, whereas the other two switches (MOSFETs switches) are operating
at power frequency. The reduced switching reduces the dv/dt stress on the switches
and it also reduces the distortion which overall reduces the THD of input current.
Voltage control and current control controller are used in the converter circuit. A
small-signal averaging model is done and the transfer function will be determined
P. Kumar · A. S. Tandon (B) · N. Yalla

Department of EEE, NIT Tiruchirappalli, Tiruchirappalli, India
A. V. J. S. Praneeth
R&D Electrical, Borg Warner Inc-Luxembourg, Luxembourg, Germany
410 P. Kumar et al.
using the state-space averaging method. Inductor and capacitor values are calculated
based on the circuit parameter of the HPFC converter. HPFC converter-based GaN
HEMT is analyzed and the results are discussed in this paper.
2 HPFC Configuration
Figure 1a presents the onboard EV charger and Fig. 1b shows the PFC topology
which is our area of research. In Fig. 1b we have two switches S1 and S2 as GaN
HEMT which we are operating at high frequency and the switches S3 and S4 are Si
switches that operate at power frequency.
2.1 Operating Modes
Interval 1: Positive line cycle inductor charging mode

Using Fig. 2a, S2 is the principal switch during the positive half line cycle (line >
neutral), and S1 is driven by a complementary PWM signal. The boost DC/DC stage
is made up of SS21 and L 1 . Half-bridge leg S4 is turned on during this positive half-
cycle, when S3 is still dormant. Current flows from L 1 → S2 → S4 and then back
to N when the primary switch S2 is switched on. During this time, the inductor L 1
charges current and the output capacitor C L discharges to keep the output voltage in
control.
Interval 2: Positive line cycle inductor discharging mode
In Fig. 2b, while still in the positive half line cycle, when S2 is switched off, S1 is
turned on, and current flows via S1 and back to N through S4 throughout the time (1
− D). The inductor L 1 is discharging current to the output during this time when the
output capacitor, C L , is charging voltage.
(a) (b)
Fig. 1 a Onboard charger, b HPFC topology

Design and Modeling of 3.3 kW GaN-Based HPFC Converter for Onboard EV Charger 411
(a) Positive AC half cycle: t = D (b) Positive AC half cycle: t =1- D
(c) Negative AC half cycle: t =D (d) Negative AC half cycle: t =1- D
Fig. 2 Current flow in the circuit
As S4 is performing all of the time in positive half line cycle operation (intervals 1
and 2), the DC bus ground Vdc is tied to N potential. For the duration of the positive
half cycle, intervals 1 and 2 loop repeatedly.
Interval 3: Negative line cycle inductor charging mode
Using Fig. 2c as an example: The operation of the negative half line cycle (neutral
> line) is analogous to that of the positive half line cycle (neutral > line), with the
exception that the top and bottom switches are switched. S1 is now the main switch,
S2 is free-wheeling, and S3 is activated, leaving S4 dormant. Current flows from L 1
→ S3 → S1 and then back to L when the primary switch S1 is switched on. During
this time, the inductor L 1 charges current and the output capacitor C L discharges to
keep the output voltage controlling.
Interval 4: Negative line cycle inductor discharging mode
In Fig. 2d, while still in the negative half line cycle, when S1 is switched off, S2
is turned on, and current flows via S3 and back to L through S2 throughout the
timeframe (1 − D). The inductor, L1, is discharging current to the output during this
time, when the output capacitor, CL, is charging voltage.
Because S3 is performing all of the time during negative half line cycle operation
(intervals 3 and 4), the DC bus VDC+ is tied to N potential. For the duration of the
negative half cycle, intervals 3 and 4 loop repeatedly.
412 P. Kumar et al.
2.2 Modeling, Analysis and Design
HPFC converter has two GaN switches (S1 , S2 ) and two MOSFETs (S3 , S4 ), as shown
in Fig. 2 [6, 7]. S4 is switched on during the positive half-cycle, S2 is the active switch
and S1 is the synchronous switch. The supply charges the inductor when S2 is turned
on, and when S1 is turned on the inductor discharges the energy to the capacitor. S3
and S1 are the active switches in the negative half cycle, and S2 is the synchronous
switch.
From Fig. 2a in which active switch is ON in positive cycle and inductor is charged
by the AC supply, we have state space as given below:
diL

1
0 0 iL
dt
dvc = −1 + L Vac (1)
dt
0 RCvc 0

iL
VO = 0 1 (2)
vc
The coefficient of state-space matrix when the active switch is ON is as follows:

1
0 0
A1 = −1 B1 = L (3)
0 RC 0

C1 = 0 1 (4)
From Fig. 2b in which active switch is OFF in positive cycle and inductor is
discharging the energy into dc-link capacitor, we have state-space as given below:
diL
−1
1
0 iL
dt
dvc = L
1 −1 + L Vac (5)
dt C RC
vc 0

iL
VO = 0 1 (6)
vc
The coefficient for the state-space matrix when the synchronous switch is ON is
as follows:
−1 1
0 L
A2 = 1 −1 B2 = L (7)
C RC
0

C2 = 0 1 (8)
After applying the disturbance on X,U and D,

Ẋ + ẋˆ = D + d̂ A1 + 1 − D − d̂ A2 X+x̂

+ D + d̂ B1 + 1 − D − d̂ B2 (U + u̇)
where X, U and D are the steady-state value of the state vector, the input voltage
vector and the duty cycle, respectively. x, u, d are the disturbance of the state vector,
the input voltage vector and the duty cycle, respectively.
Ẋ = [D A1 + (1 − D)A2 ] X + [DB1 + (1 − D) B2 ] U (9)
ẋˆ = Ax̂ + Bû+[(A1 − A2 )X + (B1 − B2 )U]d̂ (10)
In order to obtain the power level transfer functions should be changed into S-
domain form
1 VC
I˙L 0 −(1−D) iL vac
= L + L −ILL (11)
V˙C (1−D)
C
−1
RC v c 0 C d

IL 10 iL
= (12)
VO 01 vc
By using the state-space we derive the gain:
iL (1 − D)IL +(sC+ R1 )VO

Gid = = at vac = 0 (13)
d LCs2 + RL s + (1 − D)2
vc (1 − D)VO − sLIL
Gvd = = at iL = 0 (14)
d LCs2 + RL s + (1 − D)2
(a) Inductor
We have to choose inductor size so that the ripple in the input side is of specified
limit [8].
Vo D(1 − D) 400 ∗ 0.5 ∗ (1 − 0.5)

L≥ = = 0.5 mH
f I 500 ∗ 1000 ∗ 0.02 ∗ 20.28
Since here we are taking the ripple to be 2% of the input current, the frequency
of the switch is to be 500 kHz. As higher the switching frequency the inductor size
is reduced by that proportion. The higher the inductance the less will be the ripple.
(b) Output capacitor
The output capacitor was sized to meet both the hold-up time or the AC line voltage
period of 20 ms, and the low-frequency voltage ripple requirements. The capacitor
value is selected to have the larger value among the two equations given below:
414 P. Kumar et al.
2 ∗ Po ∗ thold 2 ∗ 3300 ∗ 20 ∗ 10−a

C ≥ = = 2171 µF
Vo − Vo,min
2 2 4002 − 3402
Po 3300
C ≥ = = 2626 µF
2π ∗ fline ∗ V*Vo 2π ∗ 50 Hz*10*400
We will choose the higher value of capacitance, C = 2626 mF.
3 Control Implementation
Since CCM operation was chosen for this design, average current mode control will
be used as it is the most convenient method for achieving a CCM operation PFC
[9]. Power factor is defined as the ratio of the actual and apparent power of an AC
input, as noted in the literature review. PF can be defined as the product of current
distortion and phase shift, assuming that the input voltage is a perfect sine wave. The
PFC control loop functions are:
• Control the output voltage: Making the output voltage equal to the reference
value we will implement the voltage control loop. In Fig. 3, v is measured feedback
voltage and v∗ is our reference voltage. The error value is going to the PI controller,
and the output of the PI controller is shown in Fig. 8.
• Control the input current: For current to be sinusoidal and maintain the same
phase as the input voltage, we implement the current control loop. Here in Fig. 4, i
is the measured feedback current which is compared with i∗ which is our reference
current. The error is going to the PI controller and the output of the PI controller
is shown in Fig. 9. For generating the pulse we use the repeating sequence and
relation operator (Fig. 5).
In control scheme we have the following:
Fig. 3 Voltage control

scheme
Fig. 4 Current control scheme

Fig. 5 Power circuit simulation diagram
• Reduced switching control scheme: When two switches are GaN switches, the
other two are MOS switches operating at power frequency (Figs. 6 and 7).
• High switching frequency control scheme: All four switches are high-frequency
operating switches.
Voltage control is also called the outer control loop. With the help of voltage
control loop, we are keeping the output voltage equal to 400 V.
Fig. 6 Control circuit simulation diagram for reduced switching control
Fig. 7 Control circuit simulation diagram for high switching frequency control
416 P. Kumar et al.
Fig. 8 Output of voltage control loop
Fig. 9 Output of current control loop
Current control is also the inner control loop. With the help of inner loop control,
we are keeping the input current which is also equal to the inductor current in phase
with the input voltage (Figs. 8 and 9).
Output of control loop
4.1 Component Selection for the Power Circuit
Parameter Value
Input voltage (RMS) 85–265 V
Output voltage 400 V
Output power steady state 3.3 kW
(continued)
Fig. 10 Input voltage and current for the reduced switching control
(continued)
Parameter Value
Output voltage 100 Hz ripple 10 V peak to peak
Power factor >0.98
Current THD 5% (max)
Reduced Switching Frequency control Scheme

Figure 10 shows input voltage and input current which are in phase with each other,
i.e., at unity power factor. Figure 11 shows the output voltage and Fig. 12 shows the
pole voltage for the reduced switching frequency.
Fig. 11 Output voltage for the reduced switching control
Fig. 12 Pole voltage for the reduced switching control

418 P. Kumar et al.
High switching Frequency Control Scheme

Figure 13 shows that input current and input voltage are in phase, i.e., at unity PF.
Figure 14 shows the output voltage and Fig. 15 shows the pole voltage for the high
switching frequency scheme.
THD & Input PF
Figures 16 and 17 show the THD for input current of the high switching frequency
and reduced switching frequency. Here we show the THD value for two powers at
500 W and at 3.3 kW which are given below.
Figure 18 shows the THD versus power graph for high switching frequency and
reduced switching frequency. As observed, the THD value starts decreasing as power
increases. For high-frequency switching THD at 3.3 kW is 3.08% and for reduced
switching frequency it is 2.27%.
Fig. 13 Input voltage and current for high switching frequency control
Fig. 14 Output voltage for high switching frequency control
Fig. 15 Pole voltage for high switching frequency control

Fig. 16 Current THD for high-frequency switching for 1 and 3.3 kW, respectively
Fig. 17 Current THD for reduced frequency switching for 1 and 3.3 kW, respectively
Fig. 18 Current THD versus power for high-frequency switching and reduced frequency switching,
respectively
Figure 19 shows the input power factor for the high and reduced frequency
switching scheme for the HPFC converter. As observed, the input PF increases as
we start increasing the power from 500 W to 3300 kW.
5 Conclusion
Design and modeling of 3.3 kW GaN-based HPFC converter for onboard EV charger
is investigated with the software platform. A small-signal analysis of high PFC has
420 P. Kumar et al.
Fig. 19 PF versus power for high-frequency switching and reduced frequency switching, respec-
tively
been carried out. State-space averaging is used to derive the transfer function of the
PI controller for the voltage and current loop controller. Two PWM control strategies
are implemented to get the desired output. Power quality aspects like input current,
THD, and output voltage ripple are investigated in both cases. It can be observed that
both the methods are providing nearly sinusoidal input currents.
References
1. Mohanty PR, Panda AK, Das D (2016) An active PFC boost converter topology for power factor
correction. Ann IEEE Ind Conf (INDICON)
2. Babu KSH, Holde R, Singh BH (2018) Power factor correction using boost converter operating
in CCM for front-end AC to DC conversion. ICSESP
3. Turchi J, Dalal D, Wang P, Jenck L (2014) Power Factor Correction (PFC) handbook. ON
Semiconduct
4. GaN Systems, Appl. Note GN001 (2018) Application guide-design with GaN enhancement
mode HEMT
5. GaN Systems, Appl. Note GS665BTP-REF rev 170905 (2017) High efficiency CCM bridgeless
totem pole PFC design using GaN E-HEMT
6. Erickson RW, Maksimovic D (2001) Fundamentals of power electronics, 2nd edn. Kluwer
Academic Publishers, Norwell
7. Mohan N, Undeland TM, Robbins WP (2003) Power electronics: converters, application and
design 3rd edn. John Willey & Sons, New York
8. Johnson B, Instrument T (2018) Application Report SLUA896. Power factor correction design
for on-board chargers in electric vehicles
9. NXP Semiconductors, Appl. Note DRM174 (2016) Totem-pole bridgeless PFC design using
MC56F82748
A Novel Thirteen-Level Two-Fold Gain
Inverter Topology with Reduced Voltage
Stress
M. S. H. Naidu, C. Sadanala, and S. Patnaik
1 Introduction
Currently, multilevel inverters are playing a very crucial role in developing high
power applications like HVDC, traction and propulsion systems, renewable energy
sources, electric vehicles, AC drives, etc. [1]. They can withstand higher voltages,
operate at high switching frequencies by avoiding higher voltage stresses and produce
the stepped sinusoidal wave with low distortions. Also, they can reduce filter size at
the output. However, the conventional three-level inverters are not suitable to meet
the above demands, so the inverters that can produce multi levels are proposed in the
literature [2]. Mainly there are three different configurations for multilevel converters:
(1) Neutral point clamped MLI (NPC-MLI), (2) flying capacitor MLI (FC-MLI),
and (3) cascaded H bridge MLI (CH-MLI). Besides the advantages of these above
converters, there are some drawbacks also mentioned in the literature. In the case of
NPC-MLI and FC-MLI, there is a problem of voltage balancing in the capacitors. If
the number of levels increases at the output then the requirement of the capacitors also
increases in these topologies. In CH-MLI, a large number of DC sources and more
number of switches are required which increases the cost of the inverter. So there are a
wide variety of topologies available in the literature to encounter the aforementioned
problems. In [3], asymmetrical MLI topologies were proposed with less number of
switches but multiple DC sources are used. In [4], two sources are used but more
number of components are used. In switched capacitor multilevel inverters [5] only
one source is used but total standing voltage and maximum blocking voltages are
very high, so they are not appropriate for high-voltage applications. In FC-MLIs,
TSV and MBV are low but cannot boost the voltage up to the requirements. By
considering all the above problems, this paper proposed a hybrid 13-level inverter
topology that uses the reduced count of components with low voltage stress on the
M. S. H. Naidu (B) · C. Sadanala · S. Patnaik

Department of Electrical Engineering, NIT Raipur, Raipur 492010, India
e-mail: naidusaiharsha@gmail.com
422 M. S. H. Naidu et al.
switches. It has a gain of 2 and MBV is within Vdc. Moreover, the self-balancing
of capacitor voltages will take place without the use of any additional circuitry or
control methods.
2 Presented Topology
The presented topology consists of 13 switches (s1–s13), 3 capacitors (C1–C3), and

1 dc input source, as shown in Fig. 1. C1 and C2 are the flying capacitors which will
have steady-state voltages in the ratio 1:2 and the other one is switched capacitor
used to bring out the intermediate voltage levels. The different voltage levels which
are generated by using switching configurations can be seen in Table 1 and the
corresponding current paths for positive levels can be understood from Fig. 2.
2.1 Natural Balancing of Capacitors
As per the switching table the voltage levels ±1/3, ±2/3, ±4/3, ±5/3 Vdc, can
be produced in three different states. This shows us that these levels have extra
switching states which will help in balancing the capacitors in a natural way. The
current second balance is used for balancing the capacitor voltage, i.e., the capacitors
will reach the steady-state if and only if they maintain the current second balance. The
current second balance means the amount of energy stored by the capacitor should be
released in one complete switching cycle so that each of the extra switching states for
the specified voltage levels in the table should be given to the converter in each cycle
for maintaining the current second balance of the capacitors; therefore the flying
capacitors are required to be managed in three cycles with respect to fundamental of
modulating (reference) wave to maintain the required steady-state voltages. Hence,
the time period of the fundamental modulating wave is equal to three times of time
S1 S3 S5 S7 S10 S12
C1 C2 100V S9 C3
S2 S4 S6 S8 S13
S11
LOAD
Fig. 1 Circuit diagram for the presented 13-level inverter

Table 1 Switching states for the presented 13-level inverter
O/P voltage level Capacitor states (Vs, Vc1, Vc2, Vc3) S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13
2 Vdc Vs + 0 + 0 + C3 1 0 1 0 1 0 0 1 1 1 0 0 1
2
1 3 Vdc Vs − C1 − C2 + C3 1 0 0 1 1 0 0 1 1 1 0 0 1
C3 + C2 1 0 1 0 0 1 0 1 1 1 0 0 1
C3 + Vs − V1 0 1 1 0 1 0 0 1 1 1 0 0 1
1
1 3 Vdc Vs + 0 − C2 + C3 0 1 0 1 1 0 0 1 1 1 0 0 1
C3 + C2 − C1 0 1 1 0 0 1 0 1 1 1 0 0 1
C3 + C1 1 0 0 1 0 1 0 1 1 1 0 0 1
1 Vdc Vs 1 0 1 0 1 0 1 1 0 1 1 0 1
2
3 Vdc Vs + C1 − C2 + 0 1 0 0 1 1 0 0 1 0 0 1 0 1
C2 1 0 1 0 0 1 0 1 0 0 1 0 1
Vs − C1 0 1 1 0 1 0 0 1 0 0 1 0 1
1
3 Vdc Vs − C2 + 0 0 1 0 1 1 0 0 1 0 0 1 0 1
C2 − C1 0 1 1 0 0 1 0 1 0 0 1 0 1
A Novel Thirteen-Level Two-Fold Gain Inverter Topology …
C1 1 0 0 1 0 1 0 1 0 0 1 0 1
0 1 0 1 0 1 0 1 0 0 1 0 1 0
− 13 Vdc −Vs + 0 + C2 + 0 1 0 1 0 0 1 1 0 0 1 0 1 0
−C2 + C1 1 0 0 1 1 0 1 0 0 1 0 1 0
−C1 0 1 1 0 1 0 1 0 0 1 0 1 0
− 23 Vdc −Vs − C1 + C2 + 0 0 1 1 0 0 1 1 0 0 1 0 1 0
−C2 0 1 0 1 1 0 1 0 0 1 0 1 0
−Vs + C1 1 0 0 1 0 1 1 0 0 1 0 1 0
(continued)
423
Table 1 (continued)
424
O/P voltage level Capacitor states (Vs, Vc1, Vc2, Vc3) S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13
−1 Vdc −Vs 0 1 0 1 0 1 1 0 0 1 0 1 0
1
−1 3 Vdc −Vs + 0 + C2 − C3 1 0 1 0 0 0 1 0 1 0 1 1 0
−C3 − C2 + C1 1 0 1 0 1 0 1 0 1 0 1 1 0
−c3 − c1 0 1 1 0 1 0 1 0 1 0 1 1 0
2
−1 3 Vdc −Vs + C1 + C2 − C3 0 1 1 0 0 1 1 0 1 0 1 1 0
−C3 − C2 1 0 0 1 1 0 1 0 1 0 1 1 0
−C3 − Vs + C1 0 1 1 0 1 0 1 0 1 0 1 1 0
−2Vdc −Vs + 0 + 0 − C3 0 1 0 1 0 1 1 0 1 0 1 1 0
M. S. H. Naidu et al.
A Novel Thirteen-Level Two-Fold Gain Inverter Topology … 425
(a)
(b)
(c)
(d)
Fig. 2 Switching states of the proposed model

(e)
(f)
Fig. 2 (continued)
period of the capacitor voltages. Here the positive value of the capacitor indicates
discharging and the negative value indicates charging. The value of voltages across
the capacitors can be found by using the time-domain analysis [6].
2.2 Analysis of Voltage Levels
The output voltage levels 0, ±Vdc, ±2 Vdc are the same for all three cycles. Out of
those 0, +Vdc, +2 Vdc levels are explained below:
Level 0: Fig. 2a shows the circuit diagram for o/p voltage level 0. It is produced by
switching all upper switches or all lower switches.
Level 1: Fig. 2b shows the circuit diagram for o/p voltage level 1 Vdc. It is produced
by using switching states S5, S3, S1, S13, S11, S8, S7, S10, and the voltage source
provides load current. All the capacitors are disconnected.
Level 2: Fig. 2c explains the circuit diagram for o/p voltage level 2 Vdc. It is produced
by using switching states S5, S3, S1, S13, S11, S8, S7, S10, and the input voltage
source and switching capacitor C3 supply the load current. Flying capacitors are
disconnected.
All the three equivalent circuits (Fig. 2d–f) of the inverter will use different
switching configurations to generate the same voltage level. These three different
configurations are applied to the converter in three different cycles for obtaining
natural balancing of the capacitors [6]. One-third level in three cycles is explained
below.
Cycle 1: Fig. 2d shows the circuit diagram for o/p voltage level 1/3 Vdc. It is produced
by using switching states S5, S4, S2, S13, S11, S8, and the input voltage source and
capacitor C2 contribute to the output voltage level.
Cycle 2: Fig. 2e explains the circuit diagram for o/p voltage level 1/3 Vdc. It is
produced by using switching states S3, S2, S13, S11, S8, and S6. Here the flying
capacitors C1 and C2 contribute to the output voltage levels.
Cycle 3: Fig. 2f shows the circuit diagram for o/p voltage level 1/3 Vdc. It is produced
by using switching states S1, S13, S11, S8, S6, and S4. Here capacitor C1 contributes
to the output voltage level.
2.3 Parameters Comparison
In this module the other recent topologies are compared with the presented topology
in terms of the number of active/passive components, TSV, MBV and CF in Table
2. This topology uses single DC source to generate 13 levels and also gives the low
value of TSV and cost function (CF) compared to other recent topologies. In [5, 7]
the authors use few number of switches but TSV is very high. All the topologies
in the table can generate the 13 levels. The CF will evaluate the total cost of the
converter [8] which can be expressed as

N sw + N gd + N d + N cap + αTSV(p,u) ∗ Nsources
CF = (1)
N level
where α is the weight coefficient.
The presented model is validated in the simulation environment of MATLAB for R

and RL load. A single voltage source of 100 V DC is used at the input side and the
output voltage has the maximum value of 200 V which is due to the input voltage
source and switched capacitor voltage. Flying capacitors are used for producing the
intermediate voltage levels in the waveform. All the capacitors are well balanced
at their desired values to produce the 13 levels at the output. The simulation circuit
parameters are shown in Table 3. The total harmonic distortion of R load output
voltage is found to be 9.48% without using any harmonic filter at the load side.
The output voltages and currents from R and RL load and the voltages across the
428
Table 2 Comparative analysis of 13-level inverters

Paper Sources Gatedrivers Capacitors Diodes Switches Gain TSV MBV Cost function
α = 0.5 α = 1.5
[9] 1 3 4 4 10 6 33 1 2.88 5.42
[10] 1 13 3 2 13 6 32 3 3.61 6.076
[8] 1 12 4 0 12 1.5 6 1 2.38 2.84
[11] 1 14 3 1 14 6 5.50 3 2.67 3.09
[4] 2 9 2 14 14 6 32 6 8.46 9.84
[12] 1 29 5 5 29 6 4.83 1 13.38 19.68
[7] 4 6 0 0 10 1 32 5 5.41 5.78
[5] 1 7 5 0 10 6 59 1 3.96 4.80
[3] 3 8 0 0 8 2 4 5 8.5 7.807
[13] 1 19 5 0 19 6 39 3 4.15 5.07
Proposed 1 9 3 0 13 2 9 1 2.26 2.96
M. S. H. Naidu et al.
Table 3 Parameters of the

Input voltage 100 V
presented circuit
Peak voltage at the output 200 V
Capacitor values 1200 uF
capacitors are also taken from the simulation. THD analysis is also done for R load
output voltage (Fig. 3d).
4 Conclusions
In this paper, a novel 13-level two-fold gain inverter topology is presented with
a reduced number of components which aim to reduce the voltage stress on the
circuit components. The presented model comprises 13 switches, 3 capacitors and
1 DC input source to produce 13-level output voltage having a gain of 2. Different
operating modes and their switching states of the circuit are explained elaborately for
the presented model. Capacitor voltage balance will take place in open loop without
using voltage or current sensing circuits. The comparison analysis was made between
the other recent topologies and the presented topology, and it has been found that the
proposed one has the lowest cost function. The voltage and current waveforms for R
and RL load are taken from the simulation and the THD analysis is done for R load
output voltage.
Fig. 3 Output voltages, currents and capacitor voltages for a RL load b RL load between 0 and
0.1 s c R load d THD analysis for R load output voltage
Fig. 3 (continued)
References
1. Rodriguez J, Lai J-S, Peng FZ (2002) Multilevel inverters: a survey of topologies, controls, and
applications. IEEE Trans Industr Electron 49(4):724–738. https://doi.org/10.1109/TIE.2002.
801052
2. Kouro S et al (2010) Recent advances and industrial applications of multilevel converters. IEEE
Trans Industr Electron 57(8):2553–2580. https://doi.org/10.1109/TIE.2010.2049719
3. Siddique MD et al (2019) Low switching frequency based asymmetrical multilevel inverter
topology with reduced switch count. IEEE Access 7:86374–86383. https://doi.org/10.1109/
ACCESS.2019.2925277
4. Samadaei E, Kaviani M, Bertilsson K (2019) A 13-levels module (K-type) with two DC sources
for multilevel inverters. IEEE Trans Industr Electron 66(7):5186–5196. https://doi.org/10.1109/
TIE.2018.2868325
5. Ye Y, Cheng KWE, Liu J, Ding K (2014) A step-up switched-capacitor multilevel inverter

with self-voltage balancing. IEEE Trans Industr Electron 61(12):6672–6680. https://doi.org/
10.1109/TIE.2014.2314052
6. Vasu R, Chattopadhyay SK, Chakraborty C (2020) Seven-level packed U-cell (PUC) converter
with natural balancing of capacitor voltages. IEEE Trans Ind Appl 56(5):5234–5244. https://
doi.org/10.1109/TIA.2020.3008397.
7. Samadaei E, Gholamian SA, Sheikholeslami A, Adabi J (2016) An envelope type (E-
Type) module: asymmetric multilevel inverters with reduced components. IEEE Trans Industr
Electron 63(11):7148–7156. https://doi.org/10.1109/TIE.2016.2520913
8. Zeng J, Lin W, Cen D, Liu J (2020) Novel K-type multilevel inverter with reduced components
and self-balance. IEEE J Emerg Select Topi Power Electron 8(4):4343–4354. https://doi.org/
10.1109/JESTPE.2019.2939562
9. Ye Y, Chen S, Wang X, Cheng K-WE (2021) Self-balanced 13-level inverter based on switched
capacitor and hybrid PWM algorithm. IEEE Trans Industr Electron 68(6):4827–4837. https://
doi.org/10.1109/TIE.2020.2989716
10. Sandeep N (2021) A 13-level switched-capacitor-based boosting inverter. IEEE Trans Circuits
Syst II Express Briefs 68(3):998–1002. https://doi.org/10.1109/TCSII.2020.3017338
11. Kim K-M, Han J-K, Moon G-W (2021) A high step-up switched-capacitor 13-level inverter
org/10.1109/TPEL.2020.3012282
12. Taghvaie A, Adabi J, Rezanejad M (2018) A self-balanced step-up multilevel inverter based
on switched-capacitor structure. IEEE Trans Power Electron 33(1):199–209. https://doi.org/
10.1109/TPEL.2017.2669377
13. Hinago Y, Koizumi H (2012) A switched-capacitor inverter using series/parallel conversion
with inductive load. IEEE Trans Industr Electron 59(2):878–887. https://doi.org/10.1109/TIE.
2011.2158768
PV-Array-Integrated UPQC for Power
Quality Enhancement and PV Power
Injection
Harsh D. Patel, Priyank Gandhi, and Pranav Darji
1 Introduction
Harmonics, unbalanced current, reactive power, voltage dip and swell result in losses,
less utilization grid, poor efficiency, tripping of motors, etc. and cause electrical
pollution in the distribution system. Also, pollution of the environment is a major
concern, therefore the world is shifting toward green energy generation through
PV-based solar rooftops in households and commercial buildings [1]. PV array is an
intermittent source of energy and connecting a large number of such sources in a weak
distribution system may result in a dip and swell in the voltage, consequently resulting
in instability of the grid [2]. To improve the power quality and to transmit the green
energy, two power electronic-based devices are required. This paper aims to address
the function of power quality enhancement and green energy transfer in a distribution
system using single power electronic-based device. In order to overcome the power
quality issues and integrate clean energy source in the grid, a multifunctional PV-
integrated UPQC is discussed in the paper.
A single-stage system is used to deal with power quality issues and it is also
performing the function of PV power injection in the three-phase distribution system
[3]. In [4], a single-phase system for power quality improvement on the load side
and PV power injection is discussed. But the problem with integrating PV system
with the shunt filter is that it is not able to achieve unity instantaneous power factor
and regulate the voltage of PCC at the same time. In order to fulfill both the purpose
that is unity power factor and voltage regulation, UPQC can be used. PV array
integration in UPQC can be done to improve the power quality problem and green
energy generation. In [5, 6], PV-UPQC has been reported with d-q-o theory, in which
single-stage PV-UPQC has been discussed. This has several advantages such as grid
H. D. Patel (B) · P. Gandhi · P. Darji

Sardar Vallabhbhai National Institute of Technology, Gujarat, Surat 395007, India
P. Darji
e-mail: pra-nav@eed.svnit.ac.in
434 H. D. Patel et al.
and load-side power quality improvement, protecting the critical load from a grid-
side problem and improving the fault ride-through capacity of the converter. Design
and configuration of various distributed flexible AC transmission system (D-FACTS)
devices are discussed in [7].
A case study of distribution static compensator (DSTATCOM) having second-
order Butterworth low-pass filter to filter out dc component from d-axis and q-axis
load current is mentioned in [8]. In [9, 10] moving average filter-based UPQC is
implemented to enhance the performance of SRF control, particularly in nonlinear
and unbalanced load condition. So in this paper, PV-UPQC having a moving average
filter in DSTATCOM control system for extracting dc component is used. In [11, 12]
configuration, structure and control of dynamic voltage restorer (DVR) to mitigate
supply-side power quality issues is given. In [13, 14] various MPPT algorithms for
bringing out maximum power from the PV array are discussed. In [15] performance
analysis of PV-UPQC for sensitive load is carried out and it is learned that oscil-
lation in induction motor due to dip and swell in the supply voltage is reduced. A
generalized cascaded delay signal cancellation for fundamental frequency positive
sequence voltage or current generation for PV-UPQC is discussed and performance
anatomization is carried out in [16]. A battery energy storage-based PV-UPQC that
can also supply the real power to the grid during cut-off of the supply power with
power quality enhancement is discussed in [17].
To control PV-integrated UPQC there are several control strategies for the gener-
ation of reference current/voltage and they can be classified in time-domain control
strategies and frequency-domain control strategies. Time-domain control strategies
consist of unit template generation, instantaneous reactive power theory, SRF theory,
etc. Frequency-domain control strategies such are Fourier series theory, discrete
Fourier transform theory, etc. [7]. SRF theory is used in this paper to control
DSTATCOM and DVR. In the case of unbalanced loading, the d-axis load current
contains a double harmonic part. To remove the double harmonic part, a low-pass
filter having a very low corner frequency is needed. Since the dynamic response of
a low-pass filter during the extraction of an active load component is poor, a moving
average filter has been utilized to strain the d-axis load current [5].
In this paper a two-stage PV-UPQC is considered. The first stage is of the boost
converter and the second stage of UPQC. The key benefits of the system are as
follows:
1. Voltage and current quality improvement.
2. Power quality improvement with PV power penetration.
3. Improvement of dynamic performance during load active current extraction
using moving average filter [5].
4. Stable performance under dynamic conditions such as dip, swell, insolation
change and different loading condition.
The performance of the system is investigated for different conditions using
MATLAB/Simulink.
PV-Array-Integrated UPQC for Power Quality Enhancement … 435
2 Configuration and Design of the System
The configuration of PV-array-integrated UPQC using MPPT circuit is exhibited in

Fig. 1. MPPT circuit consists of a boost converter and MPPT algorithm. PV-UPQC
consists of DSTATCOM and DVR which share a common dc-link capacitor and the
PV system is connected to the dc-link capacitor using an MPPT circuit. DSTATCOM
is connected to the load side and DVR is affixed to the source side. A transformer
connected in series with the supply is used to feed the voltage generated by the DVR
in order to reduce dip and swell from the supply. A series RC filter (ripple filter) is
used to reduce the harmonics of the DVR and DSTATCOM.
2.1 Design of PV-UPQC
PV-UPQC design entails determining the correct PV array rating, dc-link capacitor
value, interconnecting inductor value for DVR and DSTATCOM, series transformer
rating and dc-link voltage. DSTATCOM is designed to provide active power from
PV array, supply reactive power of load and mitigate load harmonics. The DVR’s
design allows it to minimize up to 30% dip and swell in the voltage [12]. PV array
Fig. 1 Layout of PV-UPQC

rating is such that it will supply power to the load and feed the remaining power to
the grid. Values of PI controller, PV array specification, ripple filter values, etc. are
given in the Appendix.
Voltage magnitude of dc-link: As given in [7], the magnitude of dc-link voltage
depends upon the phase voltage of the supply and depth of the modulation. The
supply phase voltage (Vph ) is 239.6 V and the depth of the modulation (m) is set to
1. The magnitude of the dc-link voltage must be at least twice as high as the peak
phase voltage [7].
√
2 2V ph
Vdc =
m
= 677.69 V (1)
The dc-link voltage turns out to be 677.69 V and the value is taken approximately
700 V.
Value of the dc-link capacitor: The voltage magnitude of the dc-link capacitor and
the power supplied by DSTATCOM is used to calculate the value of the dc-link
capacitor [7]. The value of the capacitor can be found by using the energy balance
equation given as follows:
3kaV ph Ish t
Cdc = 2
0.5 ∗ Vdc − Vdc1 2
= 9.3 mF (2)
where average dc voltage (Vdc ) is 700 V, lowest dc voltage (Vdc1 ) is 677.7 V, phase
voltage (Vph ) is 239.60 V, the time required to attain steady-state value (t) is 30
ms, phase current of the DSTATCOM (Ish ) is 44.17 A, the factor for considering a
variation in the energy during dynamics (K) is 0.1 and the overloading factor (a) is
1.5.
Interconnecting inductor for DSTATCOM: The value of the interconnecting
inductor for DSTATCOM depends upon the switching frequency, dc-link voltage,
maximum overload and ripple in the current supplied by the DSTATCOM. The value
of the interconnecting inductor for the DSTATCOM can be estimated from the below
equation [5]. The value of the inductor is chosen as 1 mH.
√
3mVdc
Lf =
12a f sh Ir c, pp
= 800 μH (3)
where the depth of modulation index (m) is 1, switching frequency (fsh ) is 10 kHz,
ripple in the inductor current of DSTATCOM (Irc,pp ) is 10.52 A and the value of
maximum overload (a) is 1.2 pu.
Series feeding transformer for DVR: PV-integrated UPQC given in this paper can
compensate 30% dip and swell in the voltage appearing at load which is 71.88 V. So
the voltage required to be injected is 71.88 V; hence the series feeding transformer
turns ratio required can be given as
Vvsc 239.6
K se = =
Vse 71.88
= 3.33 (4)
In order to get the KVA rating of the series feeding transformer, the primary side
current of the transformer should be known which is the same as the supply current.
With a 30% dip condition it will be 46 A, so the KVA rating is given as.
SSe = 3V Ise,sag = 3 ∗ 72 ∗ 46 = 10 KVA (5)
Interconnecting inductor for DVR: The value of the interconnecting inductor for
DVR depends on the switching frequency, dc capacitor voltage, depth of modulation
and inductor ripple current of DVR [5]. The value of the interconnecting inductor
for the DVR can be estimated from the following equation:
√
3mVdc K se
Lr =
12a f se Ir
= 3.6 mH (6)
where the depth of modulation index (m) is 1, switching frequency (fse ) is 10 kHz ,
ripple in the inductor current of DVR (Ir ) is 6.9 A, the value of maximum overload
in pu (a) is 1.2 and series transformer turns ratio (Kse ) is 3.
3 Control of PV-UPQC
PV-integrated UPQC consists of DSTATCOM, DVR along with PV array and MPPT
circuit to transfer the maximum active power from the PV array. DSTATCOM is
connected to the load side and it is used to mitigate power quality problems appearing
due to loads, such as unbalanced current, reactive power and harmonics. DSTATCOM
also performs the function of transferring real power from the PV array to the grid.
DVR is used to mitigate supply-side power quality problems such as dip and swell
in the supply voltage.
Fig. 2 Control strategy for DSTATCOM
3.1 Control of DSTATCOM
SRF-based control strategy having a moving average filter for DSTATCOM is

depicted in Fig. 2. In this control strategy the load current is changed to d-q-o domain
from a-b-c domain. PLL is synchronized with PCC voltage. The dc component is
extracted from d-axis by using a moving average filter to generate reference compen-
sating current. The dc voltage control loop is provided to control the dc-link capacitor
voltage. A low-pass filter having a corner frequency of 10 Hz can be used in place of
the moving average filter, but the low-pass filter has a poor dynamic response than
the moving average filter [9]. Since the smallest order harmonic in the d-axis load
current is having a double harmonic part, the window length (Tw ) of the moving
average filter is held at 10 ms. The moving average filter’s transfer function is given
as
1 − e−Tw s
M AF(s) = (7)
Tw s
3.2 Control of DVR
Different control strategies for DVR are discussed in [10, 12] such as in-phase
compensation, energy optimal compensation and pre-sag compensation. The control
structure for DVR is shown in Fig. 3. To requite for dip and swell in source voltage,
Fig. 3 Control strategy for DVR
DVR injects voltage that is in phase or out of phase with the source voltage. Load and
source voltages are converted to d-q-o domain to obtain actual compensating voltage.
Reference compensating voltage and actual compensating voltage are compared to
produce an error signal. This error signal is then converted back to a-b-c domain
and given to pulse width modulation (PWM) voltage control to generate a switching
signal. The reference value of d-axis load voltage (VLd*) is taken as reference load
voltage peak, while the reference value of q-axis load voltage (VLq*) is taken as
zero [10].
3.3 MPPT Circuit
In order to integrate the PV system with UPQC and to transfer the maximum power
available with the PV array to the load and source, the MPPT circuit is used. MPPT
circuit consists of a boost converter and MPPT algorithm. Perturb and observe (P&O)
strategy is utilized in the MPPT algorithm. To be able to transfer the maximal
power from PV array to the UPQC, boost converter does the function of impedance
matching such that effective impedance across boost converter become equivalent
to the impedance of PV array [13]. Impedance matching is done by controlling the
duty cycle for the boost converter and it is shown in (Fig. 4). The equation relating
the duty cycle and impedance for the boost converter is given as follows:
Fig. 4 Block diagram of MPPT
Ri = (1 − D)2 Rl (8)
where Rl is load impedance and Ri is input impedance of the boost converter.

The duty cycle for the boost converter is produced via sensing voltage and current
of the PV array which is then given to perturb and observe algorithm which will
produce suitable duty cycle for the boost converter. Flowchart of the P&O is depicted
in Fig. 5.
4 PV-UPQC Simulation Analysis
To investigate the dynamic and steady-state behavior, PV-UPQC as shown in Fig. 1

is considered and its response is analyzed using MATLAB/Simulink. Nonlinear load
(diode bridge rectifier) and unbalanced load are taken. The step size considered for
the simulation is 1e-5 s. The dynamic conditions such as change of solar radiation,
dip in voltage, swell in voltage and load variation are considered. Details of the
different parameters are provided in the Appendix.
4.1 Performance Evaluation of PV-UPQC During

the steady-State Condition
Figures 6 and 8 are of the system without connecting PV-UPQC and it is clear that
the PCC was unbalanced current and contains harmonics. Also, the instantaneous
power factor was not unity. After connecting PV-UPQC in the system it is observed in
Figs. 7 and 9 that instantaneous power factor has turned out to be unity, harmonics are
Fig. 5 Flowchart for the perturb and observe algorithm
Fig. 6 PCC voltage and source current before connecting PV-UPQC
reduced, unbalanced PCC current has become balance and power is being supplied
by PV-UPQC to source and load. As active power is being supplied from PV-UPQC
to the source, the current at PCC is 180° out of phase with PCC voltage (Fig. 10).
The dc-link voltage settles at 705 V. THD in the source current before applying
the PV-UPQC was 21.61%, which is then reduced to 3.28%.
Fig. 7 PCC voltage and source current after connecting PV-UPQC
Fig. 8 Load current for nonlinear and unbalanced load
Fig. 9 Source current under steady-state condition
4.2 Change of Solar Radiation
Solar radiation received by the PV array is changed from 1 kW/m2 to 0.2 between 0.4
and 0.6 s, as shown in Fig. 11 and it is observed in Fig. 12 that MPPT is efficiently
tracking the change in the solar radiation. As seen in Figs. 13 and 14, the current
supplied by the source increases while the current supplied by the DSTATCOM
decreases due to changes in radiation. Because of the change in the power delivered
Fig. 10 DC-link voltage under steady-state condition
Fig. 11 Change of solar radiation
Fig. 12 Ideal and extracted maximum power from PV array
by the PV array, now DSTATCOM is providing power only to the load. Change in
radiation has resulted in a small dip in the capacitor voltage, and it is reduced from
705 to 701 V. After checking a few different data points in Fig. 12, the efficiency of
the MPPT circuit obtained is around 97–98%, that is 97–98% of maximum power
available with PV array is being extracted by MPPT (Fig. 15).
Fig. 13 Current supplied from the source under change in radiation
Fig. 14 Current supplied by DSTATCOM under change in radiation
4.3 Dip and Swell in Supply Voltage
The change of supply voltage from 1 pu to 0.8 pu is made between 0.3 and 0.4 s to
create a dip in the voltage, and the change of supply voltage from 1 to 1.2 pu is made
between 0.4 and 0.5 s to create a swell in the supply voltage as depicted in Fig. 16.
Fig. 15 DC-link voltage under change of radiation

Fig. 16 Source voltage during dip and swell
Fig. 17 Load voltage during dip and swell
It is observed in Fig. 17 that the effect of voltage dip and swell is overcome in the
load voltage by using PV-UPQC. DVR will supply in-phase voltage to overcome the
dip in the supply voltage and DVR will supply 180° out of phase voltage in order
to overcome the swell in the voltage as depicted in Fig. 18. Small changes in the
capacitor voltage are noticed during the dip and swell condition in the source voltage
(Fig. 19).
Fig. 18 Voltage supplied by DVR during dip and swell

Fig. 19 Voltage of dc-link capacitor during dip and swell
4.4 Dynamic Load Change
Changes in load are made to observe the system’s dynamic response. First, a nonlinear
load is connected in parallel with the balanced load, then the unbalanced load is
introduced in parallel with the existing load at 0.4 s and it is disconnected from the
system at 0.6 s. Changes in DSTATCOM current and dc-link voltage caused by a
change in load settle in 0.08 s. With the increase in load there is no change in the
voltage at the load side due to PV-UPQC and it is maintained at 240 V (Figs. 19, 20,
21, 22).
The voltage across a dc-link capacitor is depicted in Fig. 23, and a negligible
change in dc-link capacitor voltage is observed due to change in loading condition.
Fig. 20 Load current during change in load

Fig. 21 Source current during change in load
Fig. 22 Current supplied by DSTATCOM during change in load
Fig. 23 DC-link voltage under change in load
5 Conclusion
The performance of PV-UPQC is examined under steady-state and diverse circum-

stances, and the designing of various parameters is done. Simulation studies of
PV-UPQC are carried out in MATLAB/Simulink. From the simulation study, it is
perceived that PV-UPQC is able to supply reactive power demand of load and hence
achieves unity instantaneous power factor at PCC. It is able to reduce THD in load
current from 21.61 to 3.28%. It has made the unbalanced source current balanced.
The simulation study revealed that PV-UPQC is capable of reducing power quality
issues and supplying real power in diverse circumstances. By the use of an MPPT
circuit, PV-UPQC is able to deliver 97–98% of power available with a PV array.
It is observed from the simulation study that it is performing well to lessen power
quality issues and PV-UPQC appears to be a viable option for injecting green energy
through UPQC.
Appendix
Simulation parameter
PCC line voltage is 415 V, diode bridge rectifier load having P = 5000 W Q = 500
Var, unbalance load having P1 = 1000 W, P2 = 2000 W, P3 = 1500 W, Q1 = 400
Var, Q2 = 350 Var, Q3 = 600 Var, Cdc = 9.3 mF, switching frequency for DVR is 10
kHz, RC filter with R = 10 and C = 10 µF, PI controller in DSTATCOM having
Kp = 1.5 and Ki = 0.1 and PI controller in DVR having Kp = 8 and Ki = 1200, Ls =
0.1 mH, Rs = 0.02 . The PV array model used is having three parallel connected
module and 23 series connected module Voc = 864 V, Isc = 62.65 A; Vmpp = 701 V;
Impp = 58.94 A; Ppv = 41.35 KW.
References
1. Mountain B, Szuster P (2015) Solar, solar everywhere: opportunities and challenges for
Australia’s rooftop PV systems. IEEE Power Energy Mag 13(4):53–60. https://doi.org/10.
1109/MPE.2015.2416113
2. Yang BYY, Enjeti P, Blaabjerg F, Wang H (2015) Wide-scale adoption of photovol-taic energy.
IEEE Ind Appl Mag 21–31
3. Singh B, Jain C, Goel S (2014) ILST control algorithm of single-stage dual purpose grid
connected solar PV system. IEEE Trans Power Electron 29(10):5347–5357
4. Wu T-F, Nien H-S, Shen C-L, Chen T-M (2005) A single-phase inverter system for PV power
injection and active power filtering with nonlinear inductor consideration. IEEE Trans Ind Appl
41(4):1075–1083
5. Dhagate DR, Dhamaal SS, Thakre MP (2020) Design and performance evaluation of 3-phase
photovoltaic integrated UPQC. In: 2020 international conference on power, energy, control and
transmission systems (ICPECTS), pp 1–7
6. Devassy S, Singh B (2017) Design and performance analysis of three-phase solar PV integrated
UPQC. IEEE Trans Ind Appl 54(1):73–81
7. Singh B, Chandra A, Haddad KA (2015) Power quality: problems and mitigation techniques.
Wiley, Hoboken, NJ, USA
8. Padiyar KR (2007) FACTS controllers in power transmission and distribution
9. Raja A, Santhosh KNSK (2019) Mitigation of current and voltage harmonics using MAF based
UPQC, pp 2436–2441
10. Dash SK, Ray PK (2021) A new PV-open-UPQC configuration for voltage sensitive loads
utilizing novel adaptive controllers. IEEE Trans Ind Inf 17(1):421–429
11. Hossain MI, Rahaman I, Rahman MN, Hasan MF, Hasan MM, Sarker RC (2020) Voltage sag
compensation in distribution system using dynamic voltage restorer. In: 2020 2nd international
conference on advanced information and communication technology (ICAICT), pp 492–497
12. Jayaprakash P, Singh B, Kothari D, Chandra A, Al-Haddad K (2014) Control of reduced-
rating dynamic voltage restorer with a battery energy storage system. IEEE Trans Ind Appl
50(2):1295–1303
13. Solanki CS (2015) Solar photovoltaics: fundamentals, technologies and applications. Phi
learning pvt. Ltd.
14. Singh R, Yadav R, Varshney L, Sharma S (2021) Analysis and comparison of PV array MPPT
techniques to increase output power. Int Conf Adv Comput Innov Technol Eng (ICACITE)
2021:168–172
15. Dheeban SS, Muthu Selvan NB (2020) PV integrated UPQC for sensitive load. In: 2020 inter-
national conference on emerging trends in information technology and engineering (ic-ETITE),
pp 1–7
16. Kumar K, Ravi K (2019) New performance upgrade grid system with PV Integrated UPQC-S
control. In: 2019 international conference on computation of power, energy, information and
communication (ICCPEIC), pp 1–7
17. Mansor MA, Hasan K, Othman MM, Noor SZBM, Musirin I (2020) Construction and perfor-
mance investigation of three-phase solar PV and battery energy storage system integrated
UPQC. IEEE Access 8:103511–103538
A Fuzzy-Based Buck-Boost Photovoltaic
Inverter for Voltage Stabilization During
Mismatched Environmental Conditions
Varun Bhardwaj and Prerna Gaur
Nomenclature
L1 , L2 Energy storage inductors (mH)

Lg Filter inductor (mH)
Vpv Photovoltaic array (PVA) voltage (V)
Ipv Photovoltaic array (PVA) current (A)
Cpv Photovoltaic array (PVA) to ground parasitic capacitor (uF)
fs Switching frequency (Hz)
Co1 , Co2 Voltage ripple reduction capacitors (uF)
V co1 , V co2 Rated voltage capacity of the capacitors (V)
Pco1 , Pco2 Rated power capacity of the capacitors (W)
fg Grid frequency (Hz)
Vg Grid voltage (V)
R Rectified version of unity sinusoidal function
Vm Amplitude of the grid voltage (V)
Vdc DC-link voltage (V)
1 Introduction
Considering the photovoltaic array as the optimal renewable source of power gener-
ation as compared to other renewable sources, it is used in most grid systems [1].
But there is a greater drawback for the photovoltaic array source as the power output
is fluctuating from the photovoltaic array modules because the solar insolation is
V. Bhardwaj (B) · P. Gaur

Instrumentation and Control Engineering Department, Netaji Subhas University of Technology,
New Delhi, India
452 V. Bhardwaj and P. Gaur
not constant. An array is formed with different PV panels connected in parallel

and series combinations which generate variable power and voltage output. In [2],
module integrated (MI) inverters are integrated with the ac grid parallelly, and a
flyback transformer is used for boosting the voltage. But this configuration results
in lower efficiency. In [3], the PV system is integrated with the grid through trans-
formerless H6-type inverter configuration, which is capable of eliminating problems
such as common-mode voltage and leakage current associated with the non-isolated
system. But to achieve desired voltage at the inverter input terminals, the count of
series-connected PV modules increases, due to which the output power from this
configuration is not constant during partial shading conditions.
So, various MPPT (maximum power point tracking) techniques were designed for
producing maximum output power and boosting voltage during mismatched envi-
ronmental conditions [4-6]. Some researchers used a power electronic equalizer [7]
or a DC–DC converter [7–10] for extracting maximum power and boosting voltage.
But there are limitations to these schemes. These schemes require a large number
of components; hence the cost and complexity of the system increase. The model
presented in [8] uses a DC–DC converter after every PV module to produce maximum
power output, but this structure suffers from low efficiency because of the increased
converter stages in this. In [11], all the PV modules are split up into two subarrays
to reduce the number of overall intermediate converter stages and then MPPT is
employed in each subarray. This scheme reduces the number of converter stages, but
still the efficiency is low. In conventional topology, the photovoltaic array variable
output is stabilized by a DC–DC booster converter [12], followed by an individual
inverter that feeds power into the grid. Due to this two-stage operation, there are many
conversion losses and losses related to switching included in the topology reducing
the system’s efficiency. To avoid these losses the photovoltaic array modules are
connected to a new single-stage buck-boost inverter which is tied to the grid directly
without any transformer connected between the grid and photovoltaic array inverter.
Because of this single-stage inverter, the conversion losses and switching losses
are reduced, thus improving the system’s efficiency [13]. In [14], low-voltage PV
modules are integrated with a power-optimized-based converter and voltage source
inverter to deliver stable voltage to the grid. These converters are used to enhance
the DC voltage and also to produce the rated power by using a control scheme called
fuzzy logic in partial shading conditions. The cost and the complexity of this archi-
tecture increase because each PV module requires a DC–DC converter to boost the
DC voltage. In [15] the authors proposed a methodology that uses a boost converter
and five-level neutral point clamped inverter to produce stable DC voltage at the grid
in mismatched environmental conditions. But there is a high number of switches
requirement in the inverter stage which makes it complex and difficult to use.
In this paper, the introduced inverter is connected in series with two photovoltaic
array modules at the input. The two photovoltaic modules are operated at different
sun insolation conditions which generate power and voltages at different levels.
The buck-boost inverter topology needs to stabilize the output voltage for different
operating photovoltaic array modules and convert DC voltage to single-phase AC
voltage for grid power injection. The introduced novel dual buck-boost single-phase
A Fuzzy-Based Buck-Boost Photovoltaic Inverter for Voltage Stabilization … 453
Fig. 1 Introduced novel DBBI topology
inverter (DBBI) topology is presented in Fig. 1 [16]. With the proposed topology
in Fig. 1, the number of devices connected can be reduced with a less complicated
control structure and improved efficiency.
Organization: This paper includes the working principle of the introduced DBBI
(dual buck-boost single-phase inverter) topology in Sect. 2, and control structure
modeling for the introduced inverter including the FIS module in Sect. 3. In Sect. 4
simulation results explanation for different operating conditions of PV modules are
presented. In Sect. 5 conclusion of the paper along with references are presented.
2 Working Principle DBBI
Figure 1 shows the circuit in which converter 1 has two controlled switches S1 and S3
with body diodes D1 and D3 . These diodes are freewheeling diodes connected anti-
parallel to the switches to dissipate the residual energy [17] in the switch during the
OFF state to avoid peak voltage generation during switching. Similar to converter 1,
converter 2 also includes switches S2 and S4 connected with anti-parallel diodes D2
and D4 . The two converters include individual energy storage inductors L1 and L2 ,
respectively, which makes the converter operate in buck mode or boost mode. The
inputs of the converters are connected with filter capacitors Cf1 and Cf2 in parallel
to the photovoltaic array modules which reduce the input voltage ripple [18]. At
the outputs of the converters two separate voltage ripple reduction capacitors, Co1
and Co2 are connected for DC voltage stabilization. The circuit is also included
with two freewheeling diodes Df1 and Df2 which operate during the OFF state of
all the switches S1 , S2 , S3 and S4 . The passive element calculation for the circuit is
represented as.
V pv1
L1 = (1)
4I L1 f s
V pv2
L2 = (2)
4I L2 f s
In the above equation, Vpv1 and Vpv2 have rated photovoltaic array module volt-
ages, I L1 and I L2 are the allowable current ripple for the inductors L1 and L2 ,
which is considered to be 15% of the rated peak current, and f s is the switching
frequency of the switches. The values of Co1 and Co2 are taken from the expression
given in [19].
x Pco1
Co1 = 2
(3)
2π f gVco1
x Pco2
Co2 = 2
(4)
2π f gVco2
Here, Vco1 and Vco2 are the rated voltage capacity of the capacitors [19], Pco1 and
Pco2 are the rated power capacity of the capacitors, f g is the grid frequency, and x is
the multiplication factor considered to be 2.5%.
With the above passive element calculations, the circuit is modeled and operated
as per the requirement. During buck mode [20] of operation of DBBI (dual buck-
boost single-phase inverter), only the switches S1 and S2 of both the converters are
operated with switches S3 and S4 being completely in OFF state. In buck mode, there
are four states of operation, where two are positive conduction states and two are
negative conduction states.
(a) Buck charge positive conduction (mode 1):
In this mode, S1 and S2 switches are turned ON charging the L1 and L2 inductors.
In a positive conduction state the inverter switches S5 and S8 are maintained ON
conducting current [4] in a positive direction on the grid side. In this mode, the filter
diodes Df3 and Df4 are in forward-biased conditions. The current conduction path
for mode 1 is shown in Fig. 2 [16].
(b) Buck discharge positive conduction (mode 2)
In this mode, the switches S1 , S2 are turned OFF and S5, S8 switches are maintained
ON. The charge stored in inductors L1 and L2 is discharged to the inverter with
positive conduction [21] to the grid. The freewheeling diodes Df1 , Df2 , Df3 , and Df4
are in forward-biased condition in this mode. The current conduction path for this
mode can be seen in Fig. 3 [16].
Similar to positive conduction modes, the negative conduction modes operate in
the same way but the switching of S6 and S7 switches are different. The switches S1
and S2 are operated at high switching frequency to store energy in inductors L1 and
L2 . The current conduction for the negative conduction mode is shown in Fig. 4 [16].
Fig. 2 Buck charge positive conduction mode (mode 1)
Fig. 3 Buck discharge positive conduction mode (mode 2)
Fig. 4 Buck mode negative conduction modes (mode 3 and mode 4)

(c) Boost charge positive conduction (mode 5)

In this mode of operation, the switches S1 and S2 are completely maintained ON
with S3 and S4 turned ON. Similar to mode 1 the switches S5 and S8 of the inverter
are turned ON for positive condition to the grid. During this mode of operation,
the inductors L1 and L2 are charged [8] by the PV module voltages. The current
conduction for this mode is shown in Fig. 5 [16].
In this mode, all the diodes are in reverse-biased condition and the output
capacitors Co1 and Co2 provide voltage at the output terminals.
(d) Boost discharge positive conduction (mode 6)
During this mode, the switches S1 and S2 are still maintained ON but the switches
S3 and S4 are turned OFF. In this mode, the freewheeling diodes Df3 and Df4 are in
forward-biased condition and the charged inductors L1 and L2 are in series with the
photovoltaic array modules [8] adding up the voltage. Therefore, the output voltage
is very high making it operate in boost mode. The current conduction for this mode
6 is shown in Fig. 6 [16].
Fig. 5 Boost charge positive

conduction mode (mode 5)
Fig. 6 Boost discharge

positive conduction mode
(mode 6)
Fig. 7 Boost mode negative conduction modes
After this mode the inverter needs to conduct in a negative direction; therefore,
the switching of the switches S1 , S2 , S3 , and S4 will be the same but the inverter
switches S6 and S7 are maintained ON [22]. The boost mode negative conduction
operations are shown in Fig. 7 [16].
3 Control Structure
For the introduced DBBI (dual buck-boost single-phase inverter) topology, the
inverter has to operate in synchronization to the grid voltage. The DC-link voltage of
the buck-boost converter needs to be stable even during shaded conditions. For this,
an MPPT algorithm [22] is used for the operation of switches S1 , S2 , S3 , and S4 . In the
MPPT algorithm, an incremental conductance is adopted into the control structure
for the DC voltage stabilization. The complete control structure of the introduced
DBBI is shown in Fig. 8.
The MPPT algorithm takes feedback from the photovoltaic array voltage (VpvN )
and current (IpvN ), where N is the number of photovoltaic array modules. The
Fig. 8 Introduced control

structure
maximum estimated voltage is compared to photovoltaic array voltage VpvN and

the error is fed to the PI controller for power output (PpvN ) estimation of the photo-
voltaic array. The pulses for the inverter switches are generated by PLL [23] with
feedback taken from grid voltage. During the positive voltage of the grid, switches
S5 and S8 are turned ON, and during the negative voltage of the grid, switches S6
and S7 are turned ON. The operating condition of the buck-boost converter depends
on the photovoltaic array voltage (VpvN ).
If VpvN ≥ VcoN ; buck mode ON;
If VpvN < VcoN ; boost mode ON;
VcoN is represented as.

Ppv1
Vco1 = · Vm · R (5)
Ppv1 + Ppv2

Ppv2
Vco2 = · Vm · R (6)
Ppv1 + Ppv2
Here ‘R’ is the rectified version of the unity sinusoidal function, generated from
unity sinusoidal function X (Fig. 8), synchronized with grid voltage and ‘Vm ’ is the
amplitude of the grid voltage at the output. In any mode inductor current estimation
is taken as reference (i L Nr e f ) for duty ratio generation of the switches S1 , S2 , S3 , and
S4 . The reference inductor current for a buck converter is represented as
vcoN
i L Nr e f = (7)
R pcoN
The reference inductor current for boost converter is represented as.
vcoN
2
i L Nr e f = (8)
R pcoN ∗V pv N
Here R pcoN is represented as

2
vcoN
√ m
2
R pcoN = (9)
Ppv N
The duty ratio from both the PI controllers [23] is compared with the high-
frequency sawtooth waveform-generating pulse A and pulse B. During buck mode,
pulse A is fed to switches S1 , S2 , and switches S3 , S4 are in OFF condition. During
boost mode, pulse B is fed to switches S3 , S4 , and switches S1 , S2 are in ON condition.
The conventional PI controller for the duty ratio generation of buck and boost
modes of operation is replaced with a FIS controller [24] with two input variables,
i.e., error (E) and change in error (CE), and one output variable (duty ratio D). Each
variable has seven membership functions [25] with output arranged as per the rule
base represented in Table 1.
The modeling of membership functions of the input and output variables is shown
in Fig. 9.
Table 1 Rule base

E CE
NB NM NS ZE PS PM PB
NB NB NB NB NM NS NS ZE
NM NB NM NM NM NS ZE PS
NS NB NM NS NS ZE PS PM
ZE NB NM NS ZE PS PM PB
PS NM NS ZE PS PS PM PB
PM NS ZE PS PM PM PM PB
PB ZE PS PS PM PB PB PB
a) Error b) change in error
c) output
Fig. 9 Inputs (a, b) and output (c) membership functions

The results comparison with PI and FIS controller are represented in Sect. 4 with
the simulation of the introduced circuit with different sun insolation conditions on
PV1 and PV2 .
4 Simulation Results Analysis
Figure 10 describes circuit modeling of DBBI (dual buck-boost single-phase inverter)

with interconnection to grid through inductor filter Lg for reduction of harmonics in
the current injection from the photovoltaic array module. The above test system is
modeled with the below parameters given as per Table 2 [16].
The values of L1 L2 , Co1 , and Co2 are calculated as per Eqs. (1–4) substituted
with given PV rating parameters. The simulation is run for 0.1 s with sun insolation
represented for PV1 is 1000 W/mt2 and with sun insolation represented for PV2 is
700 W /mt2 depicts that it is due to the presence of dust particles on PV2 array or a
Fig. 10 Modeling of introduced DBBI connected to the grid
Table 2 Parameter values

Passive element parameter Value
Grid Vg and fg 230 Vrms 50 Hz
L1 , L 2 , Lg 0.6, 0.6, 0.4 mH
Co1 , Co2 5, 5 uF
Cpv1 , Cpv2 0.1, 0.1 uF
S1–S4 fs 50 kHz
S5–S8 fs 50 Hz
Vmp (voltage of photovoltaic array at 116 V
maximum power)
Vdc link voltage 400 V
cloud is intervened between the surface of the PV2 module and Sun. Figure 11 is a
single-phase AC voltage waveform at PCC (point of common coupling) where the
photovoltaic array module and AC grid are connected.
Figures 12 and 13 show the DC-link voltage with PI and FIS controller of buck-
boost converter, respectively, and Fig. 14 shows the DC-link voltage comparison
between PI and FIS controller of the buck-boost converter.
Fig. 11 AC voltage output of inverter
Fig. 12 DC-link voltage with PI controller

Fig. 13 DC-link voltage with FIS controller
Fig. 14 DC-link voltage comparison between PI and FIS controller
Figure 15 draws the comparison of active power injection from PV array-

connected DBBI (dual buck-boost single-phase inverter) for PI and FIS controllers
under mismatched environmental conditions. Along with DC-link voltage compar-
ison, the power injection comparison, the THD of PCC (point of common coupling)
Fig. 15 Active power injection comparison between PI and FIS controller
voltage with PI, and the fuzzy controller is also studied using the FFT analysis
toolbox available in the powergui tool. Figures 16 and 17 describe the THDs of the
PCC voltages with PI and FIS controllers, respectively.
Fig. 16 THD of PCC voltage with PI controller

Fig. 17 THD of PCC voltage with FIS controller
5 Conclusion
As per the comparison of the results in Sect. 4, it can be concluded that even during
different sun insolation for the photovoltaic array modules representing mismatched
environmental conditions, the DC-link voltage is more stable at 400 V for the FIS
controller module. The ripple in DC-link voltage with FIS is very less than obtained
from the PI controller module. The power injected by DBBI with PI controller is
recorded at 1.5 kW, whereas for the FIS controller the DBBI injects 3 kW which is
much higher. Therefore, the DBBI module with different sun insolation conditions
of PV modules controlled by FIS is more stable and efficient than a conventional PI
controller. As the DC-link voltage is stable for the FIS controller, the THD of the
PCC voltage is also less for the same. The recorded THD for the PI controller is
3.62% and for the FIS controller is 2.67%. All the results are plotted with time as a
reference using the powergui toolbox of MATLAB software.
References
1. Foster R, Ghassemi M, Cota A (2010) Solar energy: renewable energy and the environment.
CRC Press Taylor and Francis Group, Inc
2. Umuhoza J, Zhang Y, Liu Y, Moquin J, Farnell C, Mantooth HA (2015) Interleaved flyback
based micro-inverter for residential photovoltaic application in remote areas. In: IEEE 16th
workshop control and modeling for power electronics (COMPEL) Vancouver BC 1–6
3. Ji B, Wang J, Zhao J (2013) High-efficiency single-phase transformerless PV H6 inverter with
hybrid modulation method. IEEE Trans Industr Electron 60(5):2104–2115
4. Bidram A, Davoudi A, Balog RS (2012) Control and circuit techniques to mitigate partial
shading effects in photovoltaic arrays. IEEE J Photovoltaics 2(4):532–546
5. Kaushika ND , Gautam NK (2003) Energy yield simulations of interconnected solar PV arrays.

IEEE Trans Energy Conver 18(1):127–134
6. Patel H, Agarwal V (2008) Maximum power point tracking scheme for PV systems operating
under partially shaded conditions. IEEE Trans Industr Electron 55(4)1689–1698
7. Villa LFL , Ho T-P, Crebier J-C, Raison B (2013) A power electronics equalizer application
for partially shaded photovoltaic modules. IEEE Trans Industr Electron 60(3)1179–1190
8. Olalla C, Deline C, Clement D, Levron Y, Rodriguez M, Maksimovic D (2015) Performance
of power-limited differential power processing architectures in mismatched PV systems. IEEE
Trans Power Electron 30(2):618–631
9. Femia N, Lisi G, Petrone G, Spagnuolo G, Vitelli M (2008) Distributed maximum power
point tracking of photovoltaic arrays: novel approach and system analysis. IEEE Trans Industr
Electron 55(7):2610–2621
10. Alharbi BM, Alhomim MA, McCann RA (2020) An efficient high voltage gain using two-stage
cascaded interleaved boost converter for solar PV system with MPPT technique. In: 2020 IEEE
power & energy society innovative smart grid technologies conference (ISGT), pp 1–4
11. Debnath D, Chatterjee K (2016) Maximising power yield in a transformerless single-phase
grid connected inverter servicing two separate photovoltaic panels. IET Renew Power Gener
10(8):1087–1095
12. Wai R-J, Wang W-H, Lin C-Y (2008) High-performance stand-alone Photovoltaic generation
system. IEEE Trans Industr Electron 55(1):240–250
13. Ankit SK, Sahoo SS, Yanine FF (2018) Review and comparative study of single-stage inverters
for a PV system. Renew Sustain Energy Rev 91:962–986
14. Pardhi PK, Sharma SK (2020) Power optimized architecture based converter and its control
for single phase grid-tied solar photovoltaic system. IEEE international conference on power
electronics, smart grid and renewable energy (PESGRE2020)
15. Abdelkrim T, Bouarroudj N, Lakhdari A, Benlahbib B, Borni A, Benamrane K (2019)
Design novel fuzzy logic controller of photovoltaic conversion cascade based five levels
inverter for stand-alone applications. In: International conference on environment and electrical
engineering (EEEIC)
16. Dutta S, Chatterjee K (2018) A buck & boost based grid connected PV inverter maximizing
power yield from two PV arrays in mismatched environmental conditions. IEEE Trans Ind
Electron 65(7):5561–5571
17. Faranda R, Leva S, Maugeri V (2018) MPPT techniques for PV systems: energetic and
cost comparison. In: 2008 IEEE power and energy society general meeting—conversion and
delivery of electrical energy in the 21st century, pp 1–6
18. Liu F, Duan S, Liu F, Liu B, Kang Y (2008) A variable step size INC MPPT method for PV
systems. IEEE Trans Industr Electron 55(7):2622–2628
19. Wu W, He Y, Blaabjerg F (2012) An LLCL power filter for single phase grid-tied inverter.
20. Araujo SV, Zacharias P, Mallwitz R (2010) Highly efficient single-phase transformerless
inverters for grid-connected photovoltaic systems. IEEE Trans Industr Electron 57(9):3118–
3128
21. Quesada GV, Gispert FG, Lopez RP, Lumbreras MR, Roca AC (2009) Electrical PV array
reconfiguration strategy for energy extraction improvement in grid-connected PV systems.
22. Patrao I, Garcera G, Figueres E, Gonzalez-Medina R (2014) Grid tie inverter topology with
maximum power extraction from two photovoltaic arrays. IET Renew Power Gener 8(6):638–
648
23. Sinha A, Jana KC, Das MK (2018) An inclusive review on different multi-level inverter topolo-
gies, their modulation and control strategies for a grid connected photovoltaic system. Sol
Energy 170:633–657
24. Won CY, Kim DH, Kim SC, Kim WS, Kim HS (2009) A new maximum power point tracker
of photovoltaic arrays using fuzzy controller. In: Proceedings of the 25th annual IEEE power
electronics specialists conference, pp 396403
25. Islam MS, Nepal MP, Skitmore M, Attarzadeh M (2017) Current research trends and application
areas of fuzzy and hybrid methods to the risk assessment of construction projects. Adv Eng
Inform 33:112–131
Modified Hysteresis Current Control
Implementation for Three-Phase
Grid-Connected Inverter
Mayank Arora and C. Vyjayanthi
1 Introduction
The distributed generation (DG) system is gradually becoming a new development

path in the global energy sector. The majority of DG systems utilize advanced tech-
nologies such as solar PV systems and wind turbines with an inverter connected to
deliver power to the grid.
PV system is cheaper, cleaner, and is provided free from nature. However, it is
widely used either in stand-alone or grid-connected systems. The main application
of stand-alone [1] is household [2], water pumping [3], electric vehicle charging
[4], military applications [5], and street lighting [6], whereas the grid connection is
mainly made in the hybrid system.
In a hybrid system like PV grid integration, the power from the PV module is
directly fed to the utility grid without using energy storage devices such as a battery.
It reduces the overall system investment by reducing the maintenance cost and less
space for installing equipment [7]. The integration of the grid requires a controller that
is efficient in operation. However, the performance of the inverter controller largely
determines the power quality of an inverter connected to the DG system. The most
widely used control method for grid-connected inverters is pulse width modulation
(PWM). Compared to voltage PWM converter operating in an open loop, the current
control PWM technique provides various advantages like a fast response, accurate
dc-link voltage stabilization, and protection features [8].
For providing the quick current control with stability, accuracy, and ease of imple-
mentation, the hysteresis current control outperforms other current control techniques
like sinusoidal PWM. Also, the error in the current can be reduced by changing the
bandwidth of the signal, which controls the switching frequency of the system. As
the switching frequency increases, the current waveform is improved. But there are
limitations to increasing the switching frequency as it increases the switching loss and
M. Arora (B) · C. Vyjayanthi

National Institute of Technology Goa, Veling, India
468 M. Arora and C. Vyjayanthi
problems related to EMI [9]. Hence, this interface system combines the three-phase
inverter with the pulse generation using hysteresis current control and power opti-
mizer operating at MPPT, capable of extracting maximal power from the PV system
[10, 11]. Therefore, this paper implements a hysteresis current controller with PI
for pulse generation of the three-phase inverter while maintaining the constant dc
voltage.
This paper is categorized as basic elements involved in grid integration in Sect. 2,
and the proposed methodology is presented in Sect. 3. Calculation, results, and
discussion are performed in Sects. 4 and 5. And, the conclusion is bestowed in
Sect. 6.
2 Basic Elements of Solar PV Grid Integration
The solar PV with the grid integration comprises various elements like PV module
of desired rating, power optimizer for increasing or decreasing the voltage as per
our applications, inverters, filters, and the utility grid as shown in Fig. 1. There is a
unique design in which a battery is also employed at the PCC to enhance the system
reliability, but it makes the cost and complexity go high [12].
2.1 PV Module
The solar PV system consists of various cells joined together in series or parallel
to form a panel, and multiple panels joined to form an array, and when the number
of arrays is combined, it creates a module [13]. However, for high voltage and high
power production, various modules are connected and feed power to the grid while
maintaining the constant voltage at the dc bus.
Figure 2 shows the equivalent circuit of the PV cell [14].
Fig. 1 Elements of solar PV grid integration

Modified Hysteresis Current Control Implementation for Three-Phase … 469
Fig. 2 Mathematical model

of solar cell
Fig. 3 P–V and I–V curve of solar module
In contrast, Fig. 3 shows the PV module’s voltage, current, and power characteris-
tics. Table 1 manifests the panel rating, which is utilized to design hysteresis current
control for the implementation of three-phase grid-connected inverter topology.
Table 1 Panel rating for grid

S. no Parameters Value
integration
1 Voc 36.3 V
2 Isc 7.84 A
3 Cells per module 60
4 Parallel string 47
5 Series connected module 10
6 Maximum power 213.15 W
Fig. 4 Three-phase VSI
2.2 Power Optimizer
Various types of power optimizers are available that can be used according to our
desired applications. In this three-phase grid connection, the rating of the solar PV
system is 363 V at the input and 600 V at the output. So, the converter which can be
utilized must be boost, buck-boost, SEPIC, or zeta converter [12].
2.3 Inverter
It is the most critical part of the grid-connected system, which has to be controlled by
switching pulses. Switching pulses are generated by using various methods which are
available in the literature. In this paper, the hysteresis control technique is employed
due to its simplicity and cost-effectiveness [15]. The inverter’s primary function is
to convert DC from PV array to AC and feed that AC power to the utility grid by
maintaining acceptable power quality levels. In this, a three-phase voltage source
inverter is utilized for integrating with the grid as shown in Fig. 4. It comprises six
switches which are controlled by the pulses given by the hysteresis controller.
2.4 Grid
The grid is connected at the end of the system. A few criteria should be matched
before interconnecting the grid with solar PV systems, like the voltage, frequency,
and phase sequence matching [16]. If these criteria are ignored, it injects harmonics
into the system, which decrease the stability of the interconnected system.
3 Proposed Methodology
The solar PV grid-connected system comprises a solar PV module that produces dc

voltage and current under varying solar irradiation. It is connected with the power
optimizer operating with the maximum power point tracking technique. This power
Fig. 5 Model of solar PV with grid integration
optimizer (SEPIC converter) is further connected to a three-phase inverter, which

converts DC power into AC to supply power to the connected grid [17], as shown in
Fig. 5.
The elements and its operation which are involved in the pulse generation using
hysteresis current control are discussed below:
3.1 MPPT
The solar module converts the solar energy from the sun to electrical energy. The
solar radiation throughout the day is not constant and varies with time, making our
system deviate from MPPT and operate with low efficiency. However, to trace that
maximum powerpoint, various MPPT techniques are used, and classified into three,
i.e., classical, optimization, and intelligent approach [15]. For this hysteresis control,
the P&O algorithm is implemented because of its ease of operation and simplicity
[9].
The flowchart of the P&O methodology is shown in Fig. 6.
It senses the current and voltage from the solar PV panel and calculates the change
in power and voltage change. If the change in power is greater than 0, check for the
change in voltage; if it is greater than 0, increase the Vref and if it is lesser than
0, decrease the Vref. Similarly, if the change in power is lesser than 0, check for a
change in voltage; if this change is more than zero, decrease the Vref; otherwise,
increase the Vref, where Vref represents the array voltage.
Fig. 6 Perturb and observe

MPPT algorithm
3.2 Hysteresis Current Controller
The hysteresis current controller is used to generate pulses for the inverter connected
between PV and grid system. It senses the voltage of point of coupling (PCC) and
compares it with the voltage that needs to be maintained constant, i.e., 600 V for this
system, as shown in Fig. 7.
The error is then fed to the PI controller, which converts the voltage into current
quantity. This current is then converted from dq to abc reference frame and compared
with the actual grid current, which is further passed through the hysteresis band for
the generation of pulses. And these pulses from T 1 to T 6 are fed to a three-phase
inverter for maintaining the constant voltage at the PCC.
The modified current control technique uses the PI controller with the hysteresis
band to reduce the error at the steady-state. The hysteresis controller gives varying
switching frequencies when operated alone, which causes noise in the system. Hence,
Fig.7 Hysteresis controller design

Fig. 8 Hysteresis band

pulses
the PI controller is introduced by changing the reference frame that reduces hysteresis
controller bandwidth [8].
Figure 8 indicates +h and –h as the limits of the hysteresis band and δ as the slope.
In phase a of hysteresis band, as the slope is positive, the controller makes switch
T 2 to turn on and switch T 1 to turn off, and when the slope is negative, it makes T 1
as on and T 2 as off, which is then fed to VSI. Similarly, the same case follows for
the phases b and c.
In this case, when the voltage at the dc bus is measured and compared with
the constant voltage, it produces an error signal. This error, when given to the PI
controller, produces the current. And this current is compared with the grid current
which gives error as output. This error is given by the following equations:
When the switch is operating, the dc bus voltage can be written as
di output
V PCC = L filter +V grid (1)
dt
The error fed to the hysteresis band is
δ = i output − i ref (2)
As the output current is the phase a current, so substituting Eq. (2) in Eq. (1) we
get
d
V PCC = L filter (Iref + δ)+V grid (3)
dt
dδ
V PCC − Vgrid = L filter (4)
dt
As considering dynamic nature, the change in error is
dδ V PCC − Vgrid
= (5)
dt L filter
The above equation shows the change in error concerning time oscillating between
the band for the production of pulses for VSI.
4 Calculations
4.1 Designing of SEPIC Converter
SEPIC converter is designed by considering the following parameters [12]:

Vo = 600 V(PCC voltage), Vs = 363 V (solar PV voltage), P = 100 KW, fsw =
5 kHz
Po 100 ∗ 103
Input Current of Inverter Iin = = = 275.48 A
VS 363
Po 100 ∗ 103
Output Current of Inverter Io = = = 166.67 A
Vo 600
VO
Duty Cycle of Converter D = = 0.6230 = 62.30%
VS +V o
Average inductor current and change in inductor current is given by [12]
I L2 = IO = 166.67 A
Vo ∗ Io 100 ∗ 103
I L1 = = = 275.48 A
VS 363
Assuming change in inductor current = 40%
I L1 = 0.4 ∗ 275.48 = 110.191 A
I L2 = 0.4 ∗ 166.67 = 66.668 A
Vs ∗ D
Ripple in inductor current is given by i L =
f ∗L
Vs ∗ D 363 ∗ 0.6230
L1 = = = 410.4 mH
f ∗ i L1 5000 ∗ 110.192
Vs ∗ D 363 ∗ 0.6230
L2 = = = 678 mH
f ∗ i L2 5000 ∗ 66.668
Table 2 Parameters of
SEPIC converter
1 Inductor 1 410.4 mH
2 Inductor 2 678 mH
3 Capacitors 1.735 mF
Now assuming the capacitor ripple voltage to be 2% of output voltage (Table 2).
D
C1 = C2 = = 1.735 mF
R ∗ (V O /VO ) ∗ f
4.2 Filter Design
Switching frequency = 10 kHz, Fres = Fsw/10 = 1000 Hz.

Calculating for single phase [18]: For 100 KW, 230 V(P–P), 50 Hz
3
100∗10
0.05 ∗ S
C= = 0.05 ∗ 3
= 100.28 mF
V2 ∗ 2 ∗ π ∗ f 230 ∗ 2 ∗ π ∗ 50
2
Grid current in a single phase of the three-phase system is (Table 3)
100KVA/3
Ig = = 144.92 A
230
Ig(sw) = 0.3% of Ig = 0.003 * 144.92 = 0.434 A.
Vi(sw) = 0.9 * Vg = 0.9 * 230 = 207 V
1
L= = 76.68mH
I g sw
(w SW ∗ V i sw
∗ (1 − (wsw/wres )2 ))
76.68
L1 = L2 = = 38.34mH(MinimumValue)
2
0.2 ∗ Vgrid 0.2 ∗ 230
L max = = = 1mH
2 ∗ π ∗ 50 ∗ I 2 ∗ π ∗ 50 ∗ 144.92
1mH
L1 = L2 = = 500 μH(MaximumValue)
2
Table 3 Parameters of filter

design
1 Inductor 1 500 mH
2 Inductor 2 500 mH
3 Capacitor 100.28 mF
Fig. 9 Grid voltage
5.1 Simulation at Standard Testing Condition
At STC, i.e., 1000 W/m2 and 25 °C grid voltage for a three-phase grid connection
as shown in Fig. 9 is around 320 V. And the grid current is around 200 A, as shown
in Fig. 10.
Figure 11 shows the PCC voltage waveform controlled by the hysteresis controller
by maintaining the constant 600 V supply at the dc bus. It operates in a transient
state while starting time from 0 to 0.035 s, and after 0.035, it becomes stable at 600
V, the so-called steady-state operation.
Figure 12 shows the active power waveform. Initially, it is in a transient state, and
when the PCC voltage reaches the stable region at 0.035 s, active power becomes
100 KW, which is the rating of our grid-connected PV system.
The filter design and fast-acting hysteresis controller effectively reduce the
harmonics of grid-side current. Figure 13 shows that the total harmonics distortion
is 1.09%, and the magnitude of the fundamental component is 193.5.
Fig. 10 Grid current
Fig. 11 Voltage at PCC
Fig. 12 Active power at the grid

Fig. 13 THD of grid current
6 Conclusion
In this paper, the hysteresis current control technique for the pulse generation of the
three-phase inverter is modified, which is simple and efficient in operation as it takes
lesser time to track and reach the steady state.
The simulated results prove that the proposed system of current control is efficient
in operation while maintaining the MPPT of the power optimizer at all conditions. All
the work is carried out in the MATLAB Simulink model, which shows the dynamic
behavior and performance of the grid-connected system with the calculation of filter
design and the power optimizer, which is chosen to be SEPIC because it provides
lesser THD compared to the ZETA converter. Overall, this hysteresis current control
of pulse generation is simple, robust, and efficient in operation.
References
1. Al-Waeli AH, Kazem HA, Chaichan M (2016) Review and design of a stand-alone PV system
performance. Int J Comput Appl Sci 1–6
2. Dhiman B, Kumar T, Rituraj G, Bhalla K, Chakrabarti D (2019) Study of small scale
photovoltaic applications in rural Indian household context. J Phys Conf Ser 1343
3. Patil SS, Zende RM (2017) Solar-powered water pumping system. In: 2017 third international
conference on sensing, signal processing and security (ICSSS), Chennai, pp 186–190
4. Porselvi T, Nisha J, Thendral K, Vasantha DP (2019) Solar Pv fed electric vehicle charging
system with hybrid energy storage system. In: 2019 international conference on computation
of power 2019 energy, information and communication (ICCPEIC), Melmaruvathur, India, pp
202–206
5. Jenkins P, Walters R (2017) Photovoltaic technology for Navy and Marine corps applications.
In: 2017 IEEE 60th international midwest symposium on circuits and systems (MWSCAS).
Boston, MA, USA, pp 958–961
6. Shivaleelavathi BG, Vinay ME, and Sucheeth V (2018) Solar based smart street lighting system.
In: 2018 international conference on electrical, electronics, communication, computer, and
optimization techniques (ICEECCOT), Msyuru, India, pp 1479–1483
7. Li Y, Chi Y, Wang X, Tian X, Jianqing J (2019) Practices and challenge on planning with large-
scale renewable energy grid integration. In: 2019 IEEE 3rd conference on energy internet and
energy system integration (EI2), Changsha, China, pp 118–121
8. Jena S, Babu BC, Naik AK, Mishra G (2011) Performance improvement of single-phase grid—
connected PWM inverter using PI with hysteresis current controller. In: 2011 international
conference on energy, automation and signal, pp 1–5
9. Rajamohamed S, Jeyanthy A, Devaraj D (2018) Hysteresis-based voltage and current control
techniques for grid connected solar photovoltaic systems: comparative study. Int J Electr
Comput Eng (IJECE), pp 2671–2681
10. Rajalakshmi B, Soumya U, Kumar AG (2017) Vehicle to grid bidirectional energy transfer: grid
synchronization using hysteresis current control. In: 2017 international conference on circuit,
power and computing technologies (ICCPCT), Kollam, India, pp 1–6
11. Sabegh MRZ, Bingham CM (2019) Impact of hysteresis control and internal thermal mass on
the energy efficiency of IoT-controlled domestic refrigerators. In: 2019 IEEE 7th international
conference on smart energy grid engineering (SEGE), Oshawa, Canada, pp 103–107
12. Pachauri RK, Chauhan YK (2016) Modeling and simulation analysis of PV fed Cuk, Sepic, Zeta
and Luo DC–DC converter. In: 2016 IEEE 1st international conference on power electronics,
intelligent control and energy systems (ICPEICES), pp 1–6
13. Hashimoto J, Ustun TS, Suzuki M, Sugahara S, Hasegawa M, Otani K (2021) Advanced grid
integration test platform for increased distributed renewable energy penetration in smart grids.
IEEE Access 9:34040–34053
14. Mohammed AY, Mohammed FI, Ibrahim MY (2017) Grid connected photovoltaic system.
In: 2017 international conference on communication, control, computing and electronics
engineering (ICCCCEE), pp 1–5
15. Bollipo RB, Mikkili S, Bonthagorla PK (2020) Critical review on PV MPPT techniques:
classical, intelligent and optimization. IET Renew Power Gener 1433–1452
16. Roselyn Dr, Preetha J, Sen D, Lal P, Purkayastha N, Nithya C (2020) Development of hysteresis
current controller for power quality enhancement in grid connected Pv system. Int J Electr Eng
Technol 11(4):8–21
17. Mohammed AY, Mohammed FI, Ibrahim MY (2017) Grid connected photovoltaic system.
In: 2017 international conference on communication, control, computing and electronics
engineering (ICCCCEE), Khartoum, Sudan, pp 1–5
18. Jo J, Liu Z, Cha H (2019) A new design method of LCL filter for single phase grid connected
power converter. In: 2019 international symposium on electrical and electronics engineering
(ISEE), pp 189–193
A Quasi-Direct Buck-Boost Fuzzy
Controlled Rectifier for Rapid Charging
of Electric Vehicle Battery
Rohan Jain and M. S. Manisha
1 Introduction
With increase in pollution due to emission of gases from internal combustion vehicles
the environment is deteriorating causing climatic disasters. Continuing the same is
also causing global warming where the temperature of earth’s surface is increasing
day by day. To reduce these impacts on environment and to save earth from these
climatic disasters, zero emission methods of transportation need to be adopted by us.
This can be only achieved by electric vehicles with battery as the source of power
for drive train [1]. But utilizing battery, replacing internal combustion is a greater
challenge as the battery needs to be charged regularly after discharge because of drive
train. There are many ongoing researches running for better charging capabilities of
the heavy battery. New technologies [1] are introduced to charge the battery faster
at rapid rate with high charging current injection to the battery. In conventional
topologies [2] the charging time takes around eight to ten hours, which is a major
issue for economy non-commercial vehicles. This charging time needs to be reduced
to around 1–2 h for charging the battery full from lowest state of charge (10%).
To solve this issue of slow battery charging a novel circuit topology Quasi-direct
buck-boost (QDBB) is introduced in this paper with three-phase AC input. The circuit
uses a six IGBT switches controlled rectifier [3] for producing stable DC link voltage.
The constant DC voltage is taken as input to the three-legged buck-boost converter
which controls the output current fed to the battery. In the three-legged buck-boost
converter [4], each leg comprises of two IGBT switches represented as buck and boost
switch. The ON and OFF time of these switches determine the charging current of
the battery [5]. The proposed QDBB topology is shown in Fig. 1.
As seen in the above proposed circuit the controlled rectifier is connected to three-
phase AC supply with LCL filter in between for mitigation of harmonics generated by
the rectifier due to high-frequency switching [6]. The rectifier DC voltage Vdc is fed
R. Jain (B) · M. S. Manisha

Netaji Subhas University of Technology, Delhi, India
482 R. Jain and M. S. Manisha
Fig. 1 Proposed QDBB topology with input three-phase supply
to three-legged buck-boost converter as mentioned previously. For ripple reduction

in the DC voltages Cf and Co filter capacitances are connected at the outputs of both
the converters, respectively. The battery is charged with voltage (uo) controlled by
the feedback loop control system.
This paper is included with the working principle of QDBB converter in section
II followed by section III with controlled design of the proposed topology and FIS
modeling. Section IV has the simulation results of the proposed system providing
graphical comparative analysis of QDBB converter with PI and FIS controllers.
2 Working Principle of QDBB Converter
As mentioned previously the QDBB topology includes three-phase controlled recti-

fier operated by sinusoidal PWM technique. The controlled rectifier has six IGBT
power electronic switches connected in three leg format fed from three-phase source
terminals through LCL filter [6]. The complete circuit is divided as upper region
switches and lower region switches. The positive region switches include Sap, Sbp
and Scpand negative region switches include San, Sbn and Scn. The positive region
switches and the negative region switches operate alternatively with NOT gate signals
connected to the switches. Each IGBT switch is connected with antiparallel diode
for energy dissipation in the power IGBT switch during OFF condition to avoid
overshoot of voltage at the output [7]. As per the input voltage the switches are
operated to generate DC voltage [8]. For AB phase voltage switches Sap Sbn San
Sbp are operated. During positive voltage Vab, switches Sap Sbn are turned ON,
and during negative voltage Vba, switches Sbp San are turned ON. As similar to the
above conduction for BC positive voltage (Vbc), switches SbpScn are turned ON
and during negative voltage (Vcb), switches ScpSbn are turned ON. For AC posi-
tive voltage (Vac), switches Sap Scn are turned ON and for negative voltage (Vca),
switches San Scp are turned ON. The rectification for B phase can be seen in Fig. 2.
In the above circuit the idc is the DC current generated at the output terminals
of the controlled rectifier [9]. The filter capacitance Cf is connected at the output
A Quasi-Direct Buck-Boost Fuzzy Controlled Rectifier … 483
Fig. 2 Rectifier operation for phase B
Fig. 3 Current conduction path for phase B positive rectification
terminals to reduce ripple in the DC voltage. The current conduction path for the
phase B rectification is shown in Fig. 3.
The pulse generation for the controlled rectifier is explained in detail in Sect. 3. At
the output terminals of three-phase rectifier buck-boost converter [10] is connected
with three legs connected with individual inductors connected [11, 12] to each leg
for energy storage (Fig. 4).
Each inductor Lb has charging currents io1 , io2 and io3 accumulated to io. The
complete current is given as
i o = i o1 + i o2 + i o3 (1)
As seen in the circuit topology above the switches S1p S2p S3p are the buck
switches [13] and switches S1n S2n S3n are the boost switches. The ON time and OFF
time of these switches are controlled by voltage feedback PI controller further updated
with FIS controller. The control structure of rectifier and buck-boost converter is
discussed in Sect. 3.
Fig. 4 Three-legged Buck-Boost converter
3 Control Design
The controlled rectifier is operated with sinusoidal PWM technique [14] where refer-
ence sin waveform is compared to high-frequency triangular wave for generation of
pulses. The control structure of the module is shown in Fig. 5.
A two component current is produced with current feedback from three-phase
source (ia, cib, cic, c). The Clarks transformation of the current is given as
⎡ i ⎤
−1 −1 a,c
iα 2 1 ⎣ i b,c ⎦
= √2 √
2
3 − 3 (2)
iβ 3 0 2 2 i c,c
The measured two iα iβ components are compared to reference components i α∗ i β∗ .

The reference components is given as
Fig. 5 Controller of controlled rectifier

i α∗ = G ∗ .Vα (3)
i β∗ = G ∗ .Vβ (4)
Here, Vα Vβ are the two voltage components of the three-phase source voltages
u a u b u c [15] generated using Eq. 2. The multiplication factor G* is given as
Ki
G ∗ = (K p + )(u ∗o − u o ) (5)
s
Here, K p is the proportional gain, K i is integral gain of PI controller fed from

comparison of output DC voltage (u o ) of the three-legged converter with reference
DC voltage value (u* o ).
The reference two component duty ratios for the generation of pulse for the rectifier
is given as
dα = Vα + Cα (i α∗ − i α ) (6)
dβ = Vβ + Cβ (i β∗ − i β ) (7)
The two component reference duty ratios are converted to three signal format
given as
⎡ ⎤ ⎡ ⎤
u a∗
2
3 √
0
3 ⎢
⎣ u ∗ ⎦ = ⎣ −1 3 ⎥ dα
b 3 √
3 ⎦ (8)
2 −1 dβ
u ∗c 3
− 3
3
The above reference voltage signals are compared to high frequency triangular
waveform [14] for generation of pulses for Sap Sbp and Scp . The lower switches San Sbn
and Scn get the complimentary signals through NOT gates.
The three-legged buck-boost converter control structure is shown in Fig. 6.
∗
As per Eq. 5 the reference inductor discharge current (i o,n ) is given as
∗
i o,n = G ∗ .u ∗o (9)
Here, n is the number of the leg of the buck-boost converter. The above reference
inductor current is compared to measured inductor current (i o,n ). The compared error
current is fed to PI controller [15] generating required duty ratio change for the buck
and boost switches. The final duty ratio is given as
m Bn = Dn + m buck (10)
Fig. 6 Controller of buck-boost converter
The above resultant duty ratio is compared to phase shifted high-frequency trian-
gular waveforms for generation of pulses for S1p S2p S3p and the NOT gate signals
of the same are fed to S1n S2n S3n.
The PI controller of the DC output voltage comparison [8] as in Eq. 5 is replaced
with FIS control structure with seven membership functions two input variables
and one output variable with similar membership functions. The input variables are
modeled with ‘gauss’ type membership functions [16] and the output variable is
modeled with ‘triangular’ type membership functions. The FIS modeling is shown
in Fig. 7.
In the above figure ZE is zero region, PS is Positive Small region, PM is Positive
medium region, PB is positive big region, NS is negative small region, NM is negative
medium region, NB is negative big region. The first input variable is error (E) [17] set
with a range −400 to 400, and second variable change in error (CE) is set with a range
−1 to 1. The output variable is set with a range of -4 to 4. As per the membership
functions the rule [17] is given as (Table 1).
With the above If-and-If–then rule base the PI controller is replaced with FIS
module and the results are compared in the next section with graphical representation
plotted with respect to time.
4 Simulation Results and Analysis
With all the above modules a simulation model is developed using MATLAB
Simulink software as shown in Fig. 8.
In the above modeling the input is a three-phase source connected to a step down
transformer. The secondary of the transformer is connected to LCL filter for reduction
of harmonics as the controlled rectifier operates at high switching frequency injects
harmonics into the grid. The simulation is run for 1 s and the possible results are
recorded for different operating conditions (Fig. 9).
Fig. 7 FIS modeling
The above is the FIS controller modeling with two input variables (E and CE) and
one output variable (G*). The reference inductor current is generated by multiplica-
tion of G* value with the reference output voltage Vo*. The G* value is specifically
limited in a range using saturation block. As per Fig. 6 the reference inductor current
io*n is compared to measured current ion and from the error, duty ratio for the
switches is generated using PI controllers. For the given reference value below are
the three-phase source input voltages and currents for the controlled rectifier (Fig. 10).
Table 1 A 49 rule base

E CE NB NM NS ZE PS PM PB
NB NB NB NB NM NS NS ZE
NM NB NM NM NM NS ZE PS
NS NB NM NS NS ZE PS PM
ZE NB NM NS ZE PS PM PB
PS NM NS ZE PS PS PM PB
PM NS ZE PS PM PM PM PB
PB ZE PS PS PM PB PB PB
Fig. 8 Simulation modeling of the proposed system
Fig. 9 FIS modeling in controller
As observed the three-phase voltages and currents have very less harmonics in
them reduced by the LCL filter connected between the source and the controlled
rectifier. Below is the DC link voltage [18] of 380 V generated by the controlled
rectifier maintained constant throughout the simulation time of 1 s (Fig. 11).
For the same simulation time the battery charging characteristics like battery
voltage, battery current and State of charge (SOC) of the battery are recorded and
battery charging currents are compared with PI and FIS controllers [19] (Figs. 12,
13, 14).
In the above figure the charging current comparison is taken and it is observed
that the charging current of the battery is higher for FIS controller. Therefore the
converter charges the battery faster with FIS control structure reducing the charging
time (Fig. 15).
As seen above the power factor of input with FIS is near to unity whereas with PI
controller the power factor of the three-phase source is very low.
Fig. 10 Three-phase input voltages and currents
Fig. 11 DC output voltage

after controlled rectifier
5 Conclusion
A successful simulation implementation of quasi-direct buck-boost converter is done

with different controllers improving the performance of the converter. As per the
results compared the DC voltage is more stable and generated with reduced ripple
Fig. 12 Battery charging characteristics with PI controller
Fig. 13 Battery charging characteristics with fuzzy logic controller

Fig. 14 Battery charging current comparison with PI and FIS controllers
Fig. 15 Power factor comparison with PI and FIS controller

content charging the battery with three-legged buck-boost converter. The generation
of better reference inductor signal by the FIS controller makes the battery to charge
faster as compared to traditional PI controller. Even the power factor is maintained
near to unity with fuzzy controller whereas the power factor with PI controller is low
at 0.1. Hence the performance of the converter is improved with FIS controller and
the battery is charged rapidly reducing the charging time.
References
1. Mohan N, Undeland T, Robbins W (2002) Power electronics: converters applications and

design, 3rd edn. Wiley, NewYork
2. Singh B, Singh BN, Chandra A, Al-Hadad K, Pandey A, Kothari D (2003) Are view of single-
phase improved power quality AC-DC converters. IEEE Trans Ind Electron 50(5):962–981
3. Prasad AR, Ziogas P, Manias S (1991) Anactive power factor correction technique for three-
phase diode rectifiers. IEEE Trans Power Electron 6(1):83–92
4. Stihiand O, Ooi B (1998) A single-phase controlled-current PW Mrectifier. IEEE Trans Power
Electron 3(4):453–459
5. Ooi BT, Dixon JW, Kulkarni AB, Nishimoto M (1998) An integrated AC drive system using a
controlled-current PWM rectifier/inverter link. IEEE Trans Power Electron 3(1):64–71
6. Dixonand JW, Ooi BT (1988) Indirect current control of a unity power factor sin usoidal current
boost type three-phase rectifier. IEEE Trans Ind Electron 35(4):508–515
7. Pahlevani M, Jain PK (2020) Soft-switching power electronics technology for electrical vehicle:
a technology review. IEEE J Emerg Selected Top Indus Electron 1
8. Kuperman A, Levy U, Goren J, Zafransky A, Savernin A (2013) Battery charger for electic
vehicle traction battery switch station. IEEE Trans Ind Electron
9. Gomez JC, Morcos MM (2003) Impact of ev battery chargers on the power quality of
distribution systems. IEEE Power Delivery
10. Chan CC, Chau KT (1997) An overview of power electronics in electric vehicles. IEEE Trans
Ind Electron
11. Chen G, Deng Y, Dong J, Hu Y, Jiang L, He X (2017) Integrated multiple-output synchronous
buck converter for electric vehicle power supply. IEEE Trans Vehicular Technol 66(7):5752–
5761
12. Ray O, Josyula A, Mishra S, Joshi A (2015) Integrated dual-output converter. IEE Trans Ind
Electron 62(1):371–382
13. Vasiladiotis M, Rufer A (2015) A modular multiport power electronic transformer with inte-
grated split battery energy storage for versatile ultrafast ev charging stations. IEEE Trans Ind
Electron
14. Menzi D, Bortis D, Kolar JW (2018) Three-phase two-phase-clamped boost-buck unity power
factor rectifier employing novel variable dc link voltage input current control. In: Proceedings
of 2nd IEEE International power electronics and application conference and exposition (PEAC)
15. Soeiro TB, Bauer P (2019) Three-phase unidirectional quasi-single-stage delta-switch rectifier
+ dc-dc buck converter. In: Proceedings of 39th annual conference of the industrial electronics
and society (IECON)
16. Sugeno M, Nguyen HT, Prasad NR (1999) Fuzzy modeling and control: selected works of M.
Sugeno. CRC Press, Boca Raton
17. Passino KM, Yurkovich S (1998) Fuzzy control. Addison Wesley Longman, Menlo Park, CA
18. Wu H, Zhang J, Xing Y (2015) A family of multi-port buck-boost converters based on DC-
link-inductors (DLIs). IEEE Trans Power Electron 30(2):735–746
19. Zhai L, Hu G, Lv M, Zhang T, Hou R (2020) Comparison of two design methods of EMI filter
for high voltage power supply in DC–DC converter of electric vehicle, vol. 8, pp 66564–66577
Comparative Analysis of Controlling
Methods for Doubly Fed Induction
Generator Based Wind Energy System
Aftab Ahmed Ansari and Giribabu Dyanamina
1 Introduction
Despite the effects of the COVID-19 recession, the wind industry will continue to
advance rapidly and can contribute significantly to future economic crises. The latest
market outlook released by Global wind energy council (GWEC) market intelligence
predicts that 771.3 GW of wind infrastructure will be set up in 2020 considering the
impact of COVID-19, a 6% reduction from pre-COVID Q1 2020 forecasts. The
forecast for offshore wind stations in 2020 has increased by 5% from pre-COVID
predictions to 6.5 GW, owing to the construction rush in China. From over 348
GW planned to be constructed between 2020 and 2024, the wind farm industry is
on target to accomplish record growth in the next five years, bringing global total
wind capacity to nearly 1000 GW by the end of 2024, a 54% increase from 2019.
In the face of COVID-19, the wind industry’s continued growth, despite the fact that
other economic sectors, such as coal and oil, have seen varying prices and enormous
demand drops, demonstrates its role in building sustainable and reliable economies
[1].
The increasing popularity of wind farms in the electrical network has been intrinsi-
cally tied to the wind farm research and development in frequent years. Wind turbine
electrical systems have become increasingly crucial in the interface between the
mechanical systems of wind farms and grids to which wind turbines are connected
[2]. Usually, induction machines or synchronous and permanent magnet generators
are chosen to transform mechanical power into electrical power. The squirrel-cage
induction generator is popularly used due to its simple design, high weight and low
cost compared to synchronous generators. Squirrel cage induction generator (SCIG),
a device of constant speed wind turbine generators (WTG), which has variety of
disadvantages, poor voltage regulation, wind intensity variations reflected in the grid
and lower wind energy transformation efficiency. By using variable speed WTGs,
A. A. Ansari (B) · G. Dyanamina

Maulana Azad National Institute of Technology, Bhopal, India
494 A. A. Ansari and G. Dyanamina
although most of the drawbacks may be eliminated. With recent advancements in

modern power electronics, the DFIG-based wind power plant has attracted great
attention.
The DFIG system is an effective system where the power electronic integration
manages rotor currents so that the variable speed needed for optimum power capture
during fluctuating winds is achieved. DFIG provides the advantages of speed control
at lesser cost and lower power loss because the power electronic switches handle
only the power of the rotor, usually less than 25% of the total output power. The use
of power converters enables the circuit of rotor to provide bidirectional power flow
and increases the generator’s speed range. By performing MPPT relatively higher
range of speed (±30%), robustness against power system disruption, and improved
dynamic efficiency are advantages compared with the other turbines [3–5]. These
characteristics have made DFIG WECS one of the wind industry’s leading advanced
technologies today with a share of the market of around 50%.
WECS interconnected to the system is equipped with various sophisticated control
methodologies. Among all these techniques field oriented control (FOC) is the widely
used approach in DFIG-based WTs because of its some feature like consistent
and accurate steady-state performance, lesser converter frequency but it also has
some disadvantages such as its dependency on system parameter variation [6] so to
overcome this drawback several nonlinear control strategies, such as direct torque
control/direct power control (DTC/DPC), have been introduced. The rapid advance-
ment has placed much importance on direct-control techniques for AC machinery
due to its simplicity and high dynamic performances. Takashi pioneered the use of
direct torque control technique in 1985 [7]. It has outstanding performance char-
acteristics and solid strength with regard to changes in the machine parameters.
Its principle is based on controlling the machine’s developed torque directly by
choosing the appropriate voltage vector using torque and flux information from a
table known as a switching table. Hysteresis controllers are used in traditional DTC
methods for induction motors to mitigate errors. The usage of hysteresis controllers
causes conventional DTC to have certain drawback which includes varying/changing
switching frequency, troubles during starting and medium speed operations, unnec-
essarily high torque ripple. Much advancement in direct-control strategies such as
DTC with space-vector modulation was introduced to counteract these disadvantages
(DTC-SVM) [8]. Furthermore, the use of smart control methods has been imple-
mented to all places in various industries; the DFIG control system has successfully
used the artificial neural system, the sliding-mode control (SMC), the fuzzy logic
control (FLC).
In the same way as DTC, another control methodology named direct power
control (DPC) is formulated relying on the relatively similar principles as DTC.
This approach does not demand any complicated transformations and computation
complexity in comparison to the vector control approach. The real and reactive power
is controlled by the selection of the most suitable voltage vectors from the switching
table directly and without modulation. In this way DPC has a lower dynamic response
and sensitivity to system parameters than traditional vector control. Ripple in high
power in steady-state operation is the primary disadvantage of the DPC mechanism
Comparative Analysis of Controlling Methods … 495
[9]. All three of these strategies have distinct characteristics and are widely used.
These control methods are reviewed in this article, and their performance at different
operating conditions is analyzed and compared using simulation.
This paper is divided into the following sections: Sect. 2 of this paper discusses
DFIG’s mathematical modeling, whereas Sect. 3 of this paper reviews DFIG’s opera-
tional control system strategies. Section 4 summarizes the results obtained. Section 5
compares the various operational techniques and finally, the conclusion is Sect. 6.
2 DFIG Mathematical Modeling
The energy transfer is achieved by implementing 2 subsystems: a mechanical power

extraction device known as a turbine and an electrical power conversion machine
known as a generator. DFIG is an example of a machine with a wound rotor. In the
dq reference frame, the space-vector electrical equations for DFIG are written as
[10]:
døsd
Vsd = Rs Isd + − ωs øsq (1)
dt
døsq
Vsq = Rs Isq + + ωs øsd (2)
dt
dør d
Vrd = Rr Ird + − ωre ørq (3)
dt
dørq
Vrq = Rr Irq + + ωre ørd (4)
dt
where,
Ird is the d-axis orientation current of the rotor
Irq is the q-axis orientation current of the rotor
Isd is the d-axis orientation current of the stator
Isq is the q-axis orientation current of the stator
ωs is the synchronous electrical speed
ωre is the electrical speed of rotor
Rs and Rr are the static and rotating motor windings resistances.
The stator and rotor flux are found to be:
øsd = L s Isd + M Ird (5)

øsq = L s Isq + M Irq (6)
ørd = L r Ird + M Isd (7)
ørq = L r Ird + M Isd (8)
where the stator flux is øsd and øsq toward the direction of the d and q axes, respectively.
Rotor flux ørd , ørq is direction of the axis d and q.
Ls and Lr are, respectively, leakage inductances.
M is the mutual inductance.
The real and reactive powers of the stator are described as:
Ps = Vsd Isd + Vsq Isq (9)
Q s = Vsq Isd + Vsd Isq (10)
For DFIG, the torque equation can be represented as:
pM
Tem = Ird øsq − Irq øsd (11)
Ls
This DFIG electrical model is a widely known model for designing and
determining the control objectives of a DFIG wind energy conversion system.
3 DFIG Control Methods
Wind power system research primarily focuses on issues such as cost reduction,
improved power quality and simpler system operation to address the unpredictable
nature of wind flow [11]. In WECS, control really plays a crucial role in addressing a
constantly changing as well as nonlinear characteristic of wind. Different techniques
of controlling to resolve this problem have been reviewed in the literature.
3.1 Field-Oriented Control (FOC)
DFIG control systems usually use field-oriented control (FOC) because it controls the
speed of the motor quite effectively and also it’s lower building cost. The approach
uses independent torque and flux control. In general, there is current-oriented control
Wr
Iqs
Vqs Vαs
Speed Synchronous d-q Voltage
Space Vector
Regulator frame current to Source
PWM
Regulator
Vds α-β inverter(VSI)
λr Ids Vβs
Flux Sinθ Cosθ
Regulator
λs
Vab
Estimation of
Flux and Speed Vbc
Ias
Ws
Ibs
Fig. 1 Field-oriented control
and stator-flux-oriented control. For maximum wind power extraction, the stator-flux-
oriented control is extensively used in the DFIG by regulating the power generated
by DFIG and it uses the q-axis component of current to control the real power and the
reactive power control is offered by the d-axis component of current. A schematic
field-oriented control block diagram is shown in Fig. 1. The current of stator in the
FOC method is controlled by altering it into a rotary dq reference framework. The real
and reactive powers are controlled by RSC, while the DC-link potential difference
is controlled by GSC [12].
The sensorless FOC is addressed in [13] on the basis of a model reference observer
and its performance will be compared to the open-loop estimator strategy. It is
suggested in [14] that a WECS based on DFIG with a DC-link battery energy storage
system maintains constant grid power with a stator-flux-oriented control technique.
The vector field control allows the decomposition of real and reactive powers on the
side of stator. As a consequence, it is possible to gain high-performance control of
power. A simulation result for RST and PI controller integrated with FOC for DFIG-
based WT is presented in [15] and it also describes two types of control based on
the calculation of the angle of flux space vector. FOC can be applied as a direct-field
oriented control (DFOC) where the machine’s flux model is employed and Indirect-
field oriented control (IFOC), which uses the calculated slip frequency and rotor flux
to determine the angle of the flux vector. A comparative analysis between the DFOC
regulation and the IFOC regulation of the DFIG incorporated in the wind farm is
illustrated in [16]. Sensorless positioning FOC control of DFIG requires the use of
at least four to six current sensors [17]. Proposes a new rotor positioning and current
estimation approach that allow the controlling operation of DFIG only with three
current sensors. Eliminating a single current sensor increases hardware reliability
by reducing the number of hardware elements that must be used, thus minimizing
complexity of the system and cost of the system.
SMC (Sliding-Mode Control) has become extremely popular recently. The
machine states are transmitted to the sliding board. Once achieved, SMC maintains
the system state in the near vicinity [18]. Numerous papers explored FOC utilizing
SMC for DFIG, such as [19], which presents a DFIG-based SMC for wind farm.
In [20], sliding-mode control of second order is employed for the tracking of stator
reactive power and electro-mechanical torque desired value is examined. In [21], a
Sliding mode block is used to monitor and control the RSC and GSC for a DFIG.
In traditional sliding-mode control, the major downside is that the system is respon-
sive to parametric variation; moreover, Chattering can disrupt the performance of the
controlled system and create a dangerous high-frequency vibration [22]. However, to
actually achieve increased performance of FOC, an accurate understanding of DFIG
parameters is necessary; these parameters cannot in practice be exactly known [23].
3.2 Direct Torque Control (DTC)
A further potential alternative to the FOC control strategy is DTC. A schematic

direct torque control block diagram is shown in Fig. 2. It uses a lookup table and
hysteresis control to regulate the stator flux and electro-magnetic torque. This is
basically a dynamic current limiting method. It restricts to some degree the use of
crowbar protection during voltage dips [24]. The plural current loops are omitted in
the conventional DTC (CDTC) method, and the control pulses are selected straight-
forwardly via lookup table to control the torque and rotor flux combination. Limi-
tations of the CDTC include high torque pulsation, variable frequency of switching,
torque, and flux control problems at low speeds. These adverse effects are reduced
in many DTC research papers by the use of the SVM scheme [25, 26] however,
Torque Ref.
+
Pulse Generation
(Look-up Table)
Vector Selection
Torque Hysteresis
controller
Flux Ref.
+
-
Flux Hysteresis
controller iabcr
DFIG
Sector ω re
Flux
Torque Flux and Sector ω re
Estimation
Torque iabcs
Fig. 2 Direct Torque control

control robustness is scarified. In general, the DTC scheme is preferred because of

lower calculation requirements and improving system dynamics. The DTC method
implemented for the dc voltage control in [27] is effective in reducing harmonics
and is easy to execute. An improved DTC procedure for a wind system DFIG that
enhances both the rotor flux position’s division accuracy and the number of basic
voltage vectors is presented in [28].
In [29] a new robust DTC technique is presented depending on the sliding second-
order mode and on the modulation of space vectors of a doubly fed induction gener-
ator installed WT system. The DTC technique suggested would reduce the flux,
torque, and current ripples depending on the second-order SMC. It also decreases
the frequency fluctuations in the control of the induction-machine. To control these
surfaces, two separate sliders are used, for example, flux and torque sliding surfaces.
Rotor flux control directly does not necessarily ensure reactive power control, as
the relationship between them is uncertain. Therefore, a random selection of a flux
reference does not, therefore, guarantee good power factor control. An updated DTC
method is therefore suggested in [28], to simultaneously handle the stator power
factor when regulating the generator’s electro-magnetic torque and rotor-flux. The
idea presented in this article presented a plan to control the rotor field flux of a wind
power converter. The controls are supported by a rotor vector for controlling the rotor
flux as well as electro-magnetic torque whereas the power factor at the terminals of
stator is kept under control. Direct rotor flux control does not inherently guarantee
control of reactive power as the correlation between the reactive power is not evident.
The random choice of flux reference, therefore, does not ensure a reasonable control
of the power factor. This fuzzy direct torque control scheme is investigated in [30]
for a DFIG-based wind turbine. It has become more popular recently because of
the high performance it provides over traditional FOC, especially regarding electric
machines. In [31] a novel fuzzy DTC system of DFIG was developed to improve
DTC performance and high torque characteristics can be attained. This model uses
the concept of the fuzzy logic control system and improves the system performance
by reducing the ripples of electro-magnetic torque and optimizing the total harmonic
distortion (THD) in the electro-magnetic current.
3.3 Direct Power Control (DPC)
DTC performance gets poorer as speed decreases and also at the time of start-up.
Direct power control (DPC) was considered as a solution to avoid this problem with
the DTC. The background of the DPC is drawn mainly from the DTC. Figure 3
illustrates the DPC block diagram. Direct Power Control is a major beneficial link
in regulating the DC-link voltage and adjusting the system reactive power. In DPC,
we also often take advantage of the geometrical connection between rotor and stator
fluxes and the fact that we can regulate the flux of rotor through the RSC in DPC.
In DPC rather than controlling torque and rotor flux directly, we control the impact
these magnitudes have on stator’s real and reactive power.
The issues related to wind turbines based on DFIG were addressed in [32]. The
electro-magnetic torque oscillations could be removed under unbalanced conditions
using a simplistic DPC technique, followed with a real and reactive power reference
approach, which avoid mechanical stress in wind turbines and exchanges sinusoidal
currents with the grid.
Conventional DPC includes methods of switching frequencies of a component,
by using an optimal table for the switching transitions. This is an easy and effective
strategy which ensures good outcome in applying. But there are several problems with
the conventional DPC, which make the wind power generation system based on DFIG
hard to use. For instance, the voltages generated by the rotor through a discrete and
fixed switching table could not reach an accurate and reactive power control during
maximum wind energy extraction. In addition, the vibration of the electro-magnetic
torque due to the conventional DPC is much larger than that in the vector control at
the same frequency of sampling. Furthermore, the traditional DPC moreover exac-
erbates the AC filter structure because the switching frequency is variable in DPC.
Many redesigned or improved DPC methods have been observed in [33, 34] in order
to overcome these limitations. In [39] article addressed three types of DPC strategies
with theoretical study and a comprehensive simulation investigation for wind power
generation systems based DFIG. A comparative analysis is provided in [36] between
the traditional DPC based on the standard position of flux of rotor and the DPC based
on the stator flux position for the wind turbine oriented DFIG. Specifically, in [37]
the author presented a novel DPC for DFIG with MPPT strategy for wind farm appli-
cation which is based on hybrid artificial intelligence with fixed switching frequency.
A fractional-order sliding-mode control (FOSMC) based DPC Scheme for DFIG-
based WECS is proposed in [38]. In order to construct sliding surfaces, the principle
of Caputo derivatives is implemented and used. Additionally, a comparative anal-
ysis is performed using simulations of FOSMC-based DPC and SMC-based DPC in
four different operational scenarios a steady-state environment, variable wind speed,
and variable grid-voltage condition. [40] presents a robust and simple enhanced new
voltage modulated-DPC control scheme for the DFIG system’s BTB converter. The
proposed model is a simple feed forward and feedback arrangement that eliminates
the use of PLL or Park transformations, thereby reducing computational stress and
simplifying the architecture and global exponential stabilization is ensured by the
suggested method.
This paper is a comparative analysis of three of the most successful DFIG control
methodologies, namely: field-oriented control (FOC), direct torque control (DTC),
and direct power control (DPC). The simulation results are analyzed and descriptive
and analytical output comparisons are brought out. In Table 2, the parameters used
are mentioned. The DFIG performance for wind speed rises to 12 m/s from 10 m/s
in 1.5 s and the current waveform of the stator for FOC, DTC and DPC is shown
in Fig. 4. The stator current waveform has more ripples in the field-oriented control
compared to direct-control techniques.
The WECS performance for different parameters such as speed, torque, active
and reactive power during starting transient is presented in Fig. 5. The speed of rotor
for field-oriented control at the start of the generator has a higher peak overshoot.
This high peak overshoot can be minimized by using the DPC controller instead of
the FOC controller. The rotor speed varies in all situations depending on the wind
velocity, but in the case of FOC the settling requires a long time. Introducing DTC in
DFIG-based WECS also tends to improve torque ripples by improving their current
waveform.
The behavior of WECS in three control strategies in the course of wind speed
changes by a step as shown in Fig. 6. Compared to the two other approaches, the
performance of the direct power control is better. The torque and speed of rotor are
Ps, Qs Grid
Calculation
Hysteresis/ Flux
N Control
Stator Flux
Estimation
Is Vs
Speed
Optical Switching
RSC GSC
Table
PWM
Pulses
Fig. 3 Direct power control
(a) (b)
(c) (d)
Fig. 4 a Variation of Wind Speed, Waveform of the stator current for b FOC control of DFIG c
DTC control of DFIG d DPC control of DFIG
adjusted and can be easily settled down for the DPC controller during the gradual
shift in wind velocity.
(a) (b)
(c) (d)
Fig. 5 DFIG’s transient performance for three control methods a speed, b active power, c Torque
and d reactive power
(a) (b)
(c) (d)
Fig. 6 DFIG performance for three control methods a speed b active power, c torque and d reactive
power throughout step change in wind velocity at 1.5 s
5 Comparison of Operational Control Strategies for DFIG

Based WECS
In last sections, we summarize the comprehensive review of the control strategies

and show the simulation results of the reviewed control strategies under transient
conditions and at the time of step change in wind velocity. Now, we present a set of
descriptive and analytical comparisons that show the advantages and disadvantages
of the control method in the DFIG-based WECS application. FOC is indeed the most
complicated control strategy due to its advanced applications of reference frame
transformations and model-based structure. The DPC, on the other side, emerges out
as the least complicated control mechanism with full machine-model independence.
In comparison to the FOC requirements, the sampling times related to direct must be
shorter and the sensor-bandwidth must be greater: this empowers direct strategies to
focus particularly on controlled quantities in order to keep them inside the specified
hysteresis bands.
The FOC execution incurs lower instrumentation constraints, although struc-
turally complex. With regard to computational tasks within MATLAB Simula-
tions, FOC seems to be very resource-consuming due to the multiple reference-
transformations and model-based nature whereas at the other side the direct tech-
niques involve simplified data handling, mostly using hysteresis comparators and
lookup tables so overall implementation complexity of FOC is more. The dynamic
reaction of FOC is overcome by both direct-control technologies, although it is fast
and performing. A number of references in relation to the operation under FOC of
the DFIG [37, 38] concludes that the variability of the reference-currents must be
limited in an aim to maintain the control system’s stability. The sampling times of
up to 200 μs seemed to be possible for VC method before any significant nega-
tive impact could be reported on the control. As regards direct methods, the quality
or strength of current waveforms at control time exceeding 50 μs was detrimental as
the controlled signals were likely to leave the hysteresis belts. Minimizing sampling
time doesn’t really guarantee better direct-control efficiency: if the sensor-bandwidth
is exceeded and the observed signal is not exact, reducing sampling time has negative
effects on direct-control methods. We defined robustness as a measurement of the
control technique’s ability to achieve success in the case of structure disturbances and
unsafe conditions along with modeling errors, parameter variations, noise measure-
ment and so on. Section 4 simulation results show that FOC and DPC are fairly robust
control systems with different capabilities: Under parameter deviation FOC becomes
weaker and DPC is not very successful under noise measurement. The MPPT wind
turbine operation mode calls for mechanical parameters to be measured or estimated.
Only DPC is insufficient for MPPT, as stated in the control designs for the control
methods under review. However, by introducing a position-encoder to DPC control
system, this weakness can be quickly overcome. Table 1. is the final comparable
overview depending on different results mentioned in this paper.
Table 1 Comparison of Operational Control Strategies for DFIG Based WECS

Performance criteria FOC DTC DPC
Computational The most DTC is the moderate DPC steps out as
complexity complicated control complex control simplest control method
technique is FOC technique
Machine model FOC refers to a DTC is the strategy DPC is
dependency technique that is that marginally parameter-independent
more dependent on depend on machine
system parameters parameter
Sample time constraint Low High High
Transitory response Low transitory Medium transitory High transitory
performance performance performance
Settling time Medium Low Low
Harmonic content Low Medium Medium
Instrumentation FOC Instrument High High
constraint constraint is
Moderate
Overall implementation FOC is in fact the DTC is moderately DPC needs only a few
complexity most difficult complex strategy to signals to be sensed so
method to implement less complex strategy to
implement implement
Robustness High Medium High
MPPT stability High High Low
Table 2 DFIG ratings and

Variables Values implemented
parameters
Power 2 * 106 (VA)
Nominal voltage (L-L) 690 (V)
Frequency (Grid) 50 (cycle/second)
Resistance of stator 0.03513 (pu)
Inductance of stator 0.04586 (pu)
Resistance of rotor 0.03488 (pu)
Inductance of rotor 0.04586 (pu)
Mutual inductance 1.352 (pu)
Pairs of pole 2
Factor of friction 0 (pu)
6 Conclusion
All three comparative control methods have demonstrated remarkable performance

in both the transient and the step change of wind. The simulations and review yield
valuable insights into the benefits and drawbacks of each control scheme. FOC seems
to be the most complicated control strategy due to its lengthy use of frame transfor-
mations of reference and model-based structure. On the other hand, DPC manages
to stand out as the least complicated control mechanism with full machine method
independence. Under parameter deviation VC becomes weaker and DPC is not very
successful under noise measurement. The FOC execution incurs lower instrumenta-
tion constraints, although structurally complex. With regard to computational tasks
within MATLAB Simulations, FOC seems to be very resource-consuming due to the
multiple reference-transformations and model-based nature. The direct techniques,
on the other side, require simplified data handling, mostly using hysteresis compara-
tors and lookup tables so overall implementation complexity of FOC is more. The
simulation results and comparison show that the DPC method, as outlined in this
work, performed better than the other two strategies.
References
1. Global Wind Energy Council, Global Wind Energy Council (2021) [Online]. Available: https://
gwec.net/. Accessed 29 Jan 2021
2. Yaramasu V, Wu B, Sen PC, Kouro S, Narimani M (2015) High-power wind energy conversion
systems: State-of-the-art and emerging technologies. In: Proceedings of the IEEE, vol 103, no.
5, pp 740–788
3. Bhowmik S, Spee R, Enslin JHR (1999) Performance optimization for doubly fed wind power
generation systems. IEEE Trans Indus Appl 35(4):949–958
4. Ekanayake J, Jenkins N (2004) Comparison of the response of doubly fed and fixed-speed
induction generator wind turbines to changes in network frequency. IEEE Trans Energy Conver
19(4):800–802
5. Flannery PS, Venkataramanan G (2008) A fault tolerant doubly fed induction generator wind
turbine using a parallel grid side rectifier and series grid side converter. IEEE Trans Power
Electron 23(3):1126–1135
6. Aydin E, Polat A, Ergene LT (2016) Vector control of DFIG in wind power applications. In:
2016 IEEE international conference on renewable energy research and applications (ICRERA),
Birmingham, UK, pp 478–483. https://doi.org/10.1109/ICRERA.2016.7884383
7. Takahashi, Noguchi T (1986) A new quick-response and high efficiency control strategy of an
induction motor. IEEE Trans Ind Appl IA 22:820–827
8. Lai Y, Chen J (2001) A new approach to direct torque control of induction motor drives for
constant inverter switching frequency and torque ripple reduction. IEEE Trans Energy Convers
16:220–227
9. Nian H, Song Y (2014) Direct power control of doubly fed induction generator under distorted
grid voltage. IEEE Trans Power Electron 29(2):894–905. https://doi.org/10.1109/TPEL.2013.
2258943
10. Golnary F, Moradi H (2019) Dynamic modelling and design of various robust sliding mode
controls for the wind turbine with estimation of wind speed. Appl Math Model 65:566–585
11. Jha D (2017) A comprehensive review on wind energy systems for electric power generation:
Current situation and improved technologies to realize future development. Int J Renew Energy
Res (IJRER) 7(4):1786–1805
12. Karad S, Thakur R (2019) Recent trends of control strategies for doubly fed induction generator
based wind turbine systems: a comparative review. Arch Comput Methods Eng 28(1):15–29
13. Esmaeeli MR, Kianinezhad R, Razzaz M (2012) Field oriented control of DFIG in wind energy
conversion systems. J Basic Appl Sci Res 2(11):11486–11493
14. Rahul Charles CM, Vinod V, Jacob A (2015) Field oriented control of doubly fed induction
generator in wind power system. In: IEEE international conference on computational intelli-
gence and computing research (ICCIC), Madurai, pp 1–5. https://doi.org/10.1109/ICCIC.2015.
7435760
15. Poitiers F, Bouaouiche T, Machmoum M (2009) Advanced control of a doubly-fed induction
generator for wind energy conversion. Electric Power Syst Res 79(7):1085–1096. https://doi.
org/10.1016/j.epsr.2009.01.007
16. Ihedrane Y, El Bekkali C, Bossoufi B (2017) Direct and indirect field oriented control of
DFIG-generators for wind turbines variable-speed. In: 14th international multi-conference on
systems, signals & devices (SSD), Marrakech, pp 27–32. https://doi.org/10.1109/SSD.2017.
8166915
17. Prasad RM, Mulla MA (2019) A novel position-sensorless algorithm for field-oriented control
of DFIG with reduced current sensors. IEEE Trans Sustain Energy 10(3):1098–1108. https://
doi.org/10.1109/TSTE.2018.2860993
18. Tanvir AA, Merabet A, Beguenane R (2015) Real-time control of active and reactive power
for doubly fed induction generator (DFIG)-based wind energy conversion system. Energies
8:10389–10408
19. Barambones O, Cortajarena JA, Alkorta P, Gonzalez de Durana JM (2014) Real-time sliding
mode control for a wind energy system based on a doubly fed induction generator. Energies
7:6412–6433
20. Beltran B, Benbouzid M, Tarek AA (2012) Second-order sliding mode control of a doubly fed
induction generator driven wind turbine. IEEE Trans Energy Convers 27:261–269
21. Ruiz-Cruz R, Loukianov AG, Sanchez EN (2013) Real-time sliding mode control for a doubly
fed induction generator. In: Decision and control (CDC) IEEE 52nd annual conference on
2013, pp 6786–679
22. Patnaik RK, Dash PK, Mahapatra K (2016) Adaptive terminal sliding mode power control of
DFIG based wind energy conversion system for stability enhancement. Int Trans Electr Energy
Syst 26:750–782
23. Utkin V, Guldner J, Shi J (2009) Sliding mode control in electro-mechanical systems, 2nd edn.
CRC Press. https://doi.org/10.1201/9781420065619
24. Naveen G, Sarvesh PKS, Krishna BR (2013) DTC control strategy for doubly fed induction
machine. Int J Eng Adv Technol 3(1)
25. Chen SZ, Cheung NC, Wong KC, Wu J (2010) Integral sliding-mode direct torque control
of doubly-fed induction generators under unbalanced grid voltage. IEEE T Energy Conver
25:356–368
26. Zhu X, Liu S, Wang Y (2014) Second-order sliding-mode control of DFIG-based wind turbines.
In: IEEE 2014 3rd renewable power generation conference, Naples, Italy. IEEE, New York,
NY, USA, pp 1–6
27. Gundavarapu A, Misra H, Jain AK (2017) Direct torque control scheme for DC voltage regula-
tion of the standalone DFIG-DC system. IEEE Trans Indus Electron 64(5):3502–3512. https://
doi.org/10.1109/TIE.2016.2644623
28. Li Y et al (2016) An improved DTC controller for DFIG-based wind generation system. In:
IEEE 8th international power electronics and motion control conference (IPEMC-ECCE Asia),
Hefei, pp 1423–1426. https://doi.org/10.1109/IPEMC.2016.7512499
29. Boudjema Z et al (2017) A novel direct torque control using second order continuous sliding
mode of a doubly fed induction generator for a wind energy conversion system. Turk J Electr
Eng Comput Sci 25:965–975
30. Ayrir W, Ourahou M, Hassouni BE, Haddi A (2020) Direct torque control improvement of a
variable speed DFIG based on a fuzzy inference system. Math Comput Simul 167:308–324
31. El Ouanjli N, Motahhir S, Derouich A, Abdelaziz E, Ali C, Mohammed T (2019) Improved
DTC strategy of doubly fed induction motor using fuzzy logic controller. 5:271–279. https://
doi.org/10.1016/j.egyr.2019.02.001
32. Abad G, Rodríguez MÁ, Iwanski G, Poza J (2010) Direct power control of doubly-fed-
induction-generator-based wind turbines under unbalanced grid voltage. IEEE Trans Power
Electron 25(2):442–452. https://doi.org/10.1109/TPEL.2009.2027438
33. Escobar G, Stankovic AM, Carraso JM et al (2003) Analysis and design of direct power
control (DPC) for a three phase synchronous rectifier via output regulation subspaces. IEEE
Trans Power Electron 18(3):8232830
34. Xin W, Yue CD, Yu ZC (2005) A torque ripple minimization scheme in direct torque control
of induction motors. Tech Autom Appl 24(8):72–74
35. Xiong L, Wang J, Mi X, Khan MW (2018) Fractional order sliding mode based direct power
control of grid-connected DFIG. IEEE Trans Power Syst 33(3):3087–3096. https://doi.org/10.
1109/TPWRS.2017.2761815
36. Zhi D, Xu L (2007) Direct power control of DFIG with constant switching frequency and
improved transient performance. IEEE Trans Energy Convers 22(1):110–118
37. Fayssal A, Chaiba A (2016) A novel direct power control for grid-connected doubly fed induc-
tion generator based on hybrid artificial intelligent control with space vector modulation. Revue
Roumaine des Sciences Techniques Serie Electrotechnique et Energetique 61:263–268
38. Petersson A, Thiringer T, Harnefors L, Petru T (2005) Modeling and experimental verification
of grid interaction of a DFIG wind turbine. IEEE Trans Energy Convers 20(4):878–886. https://
doi.org/10.1109/TEC.2005.853750
39. Heller M, Schumacher W (1997) Stability analysis of doubly-fed induction machines in stator
flux reference frame. In: Proceedings of the 7th european conference on power electronics and
applications (EPE’97), vol 2, pp 707–710
40. Gao S, Zhao H, Gui Y, Zhou D, Blaabjerg F (2021) An improved direct power control for
doubly fed induction generator. IEEE Trans Power Electron 36(4):4672–4685. https://doi.org/
10.1109/TPEL.2020.3024620
Adaptive Pitch Control of Wind Turbine
Generator to Counter Transient
Contingencies in a Multi Machine System
Hailiya Ahsan, Abdul Waheed Kumar, and M. D. Mufti
1 Introduction
With every passing year, as wind power gains significant impetus, it is considered
as a mainstream solution to all major energy crisis problems. A number of factors
collectively come to its rescue in disguise, viz., hike in fuel costs, fuel crisis, environ-
mental degradation and the urge towards a green Earth. The increasing wind power
capacity, however, burdens the existing power systems with an instability threat. In
reality, the very intermittent nature of wind is primarily the inherent cause of facilitat-
ing staggering power outputs [1]. A highly variable wind speed results in an uneven
consumption of reactive VAR’s by the reactive power compensation devices like
capacitor banks. This digresses the power factor from a unity power factor operating
template [2]. It is due to this reason, that the wind turbine generator (WTG) generates
fluctuating voltage at its point of common coupling with the grid and thus hampers
power quality [3].
Pitch controlled wind turbines are technically encouraged to stabilize the WTG’s
power outputs, when the wind speed faces excursions [4]. Inefficient pitch control
technique determining the turbine blade position can lead to consequential stress
on the blades, leading to blade wear and tear and reduced life [3, 5]. In [6], the
authors have presented a pitch neuro-control architecture based on neuro-estimators
of the effective wind to control the pitch angle. The current wind and the forecasting
wind have been included in the controller. A pitch angle control scheme for rapid
power control of a permanent magnet synchronous generator based wind power plant
has been shown in [7]. The effectiveness of variable pitch sliding mode controller
based on disturbance observer to improve dynamic performance and robustness to
wind disturbances has been demonstrated in [8]. In [9], the authors have validated
H. Ahsan (B) · A. W. Kumar · M. D. Mufti

Department of Electrical Engineering, National Institute of Technology, Hazratbal Srinagar,
J&K 190006, India
e-mail: waheed_02phd17@nitsri.ac.in
510 H. Ahsan et al.
the higher performance nature of Linear Active Disturbance Rejection Control over
conventional Proportional-Derivative (PD) control method to improve wind turbine
dynamics in a varying wind profile.
In this paper, a novel adaptive pitch control scheme of WTG’s is presented with
the aim of smoothing the power output as well as the terminal voltage of the machine.
The said adaptive control technique is a self tuning scheme, which targets the reduc-
tion/mitigation of the driving point variable to zero, which in this case happens to
be the network frequency. A trained recursive least squares algorithm is adopted
for obtaining the model parameters, at each sampling instant. An optimized control
signal is issued by the APC strategy by minimizing a quadratic index serving as the
control law [10]. This control signal reflects the blade pitch angle and effectively
reduces the driving point frequency error to zero.
The main highlights of this paper are:
1) A self tuned adaptive control scheme determines the pitch angle of a WTG,
thereby imparting comprehensive stability to a multi machine system.
2) Detailed mathematical modeling of the WTG is delineated .
3) Modified Newton Raphson load flow technique is outlined in depth.
4) Stability evaluation performed in light of turbulent contingencies (wind as well
as 3-φ faults) [11, 12].
2 Wind Turbine Generator Fed Multi Machine System
The arrested mechanical power by a variable speed wind turbine can be put forth as
[13],
Pwtur b = 0.5ρπ R 2 vw 3 C p (λ, β) (1)
where, Pwtur b is the arrested mechanical power from the wind, ρ is the density of
air in kgm−3 , R is the radius of the turbine blade in m, vw is the speed of the wind in
ms−1 ; and C p is the function of the tip speed ratio (λ) and the blade pitch angle (β),
also known as the power coefficient [14].
The wind turbine power coefficient is mathematically put as (See Fig. 1),
1
Cp = (λ − 0.022β 2 − 5.6)e−0.17λ (2)
2
The lumped one mass model torque equation can be defined as [15],
1
ω̇ = (Pwtur b /ω B − Te ) (3)
2Hwtg
In this paper, a squirrel cage induction generator is adopted as the wind turbine
generator due to its cost effectiveness and rugged construction [16]. The wind driven
induction machine’s electrical torque modelling can be done as,
Adaptive Pitch Control of Wind Turbine Generator … 511
Fig. 1 Relation between

C p − λ and varying pitch
angles (β)
Table 1 Induction generator data

Parameters Symbol Values
Rotor time constant, sec To 0.8871
Stator resistance, p.u. r swtg 0.0017
Turbine-generator inertia H wtg 0.42
constant, p.u.
Induction machine reactance, X 5.0167
p.u.
Transient reactance, p.u. X’ 0.0333
X-X’, p.u. b0 4.9834
Tewtg = E dwtg Idwtg + E qwtg Iqwtg (4)
The induction machine has been modelled as an induction motor in the system,
running in generating mode, claiming negative slip as well as torque in operation.
All induction generator details are given in Table 1 [17].
The stator algebraic equations pertaining to the induction generator at wtg bus
are:

Vdwtg E dwtg rswtg −X wtg Idwtg
= + (5)
Vqwtg E qwtg X wtg rswtg Iqwtg
The differential set of equations describing the rotor winding dynamics can be
put as,
E d˙wtg 1 −E dwtg 0 b0 Idwtg
= −
E q˙wtg To −E qwtg b0 0 Iqwtg
(6)
s E qwtg
+ ωo
−s E dwtg
Steady state analysis using a slightly modified Newton Raphson load flow is
performed on the system which yields the pre-fault (or pre-disturbance) state values
512 H. Ahsan et al.
Table 2 Steady state analysis: modified load flow results

Bus no. Voltage Angle Generation Load
magnitude degrees
MW MVAR MW MVAR
1 1.040 0.000 0.000 0.000 21.66 41.60
2 1.025 10.791 0.000 0.000 163.0 9.800
3 1.025 6.157 0.000 0.000 85.00 −7.800
4 1.017 −0.676 0.000 0.000 50.00 −21.50
5 0.989 −2.495 125.0 50.00 0.000 0.000
6 1.006 −2.190 90.00 30.00 0.000 0.000
7 1.024 5.220 0.000 0.000 0.000 0.000
8 1.014 2.214 100.0 35.00 0.000 0.000
9 1.031 3.454 0.000 0.000 0.000 0.000
Induction machine initial slip: −0.0074
of network variables like generator injection currents, bus voltages and injected real
and reactive powers at all the buses.
On setting the derivative terms in the induction machine modelling equations
to zero, and solving for the real and reactive power contributions of the induction
machine, a fourth order equation for machine slip computation is obtained. Mathe-
matically, it can be jotted down as
S1 s 4 + S2 s 3 + S3 s 2 + S4 s + S5 = 0 (7)
where,
S1 = a 4 rs4 3 + X − rswtg Vwtg 2 /Pwtg
2

S2 = a 3 b0 2rswtg − Vwtg 2 /Pwtg

S3 = a 2 b0 2 − 2rswtg Vwtg 2 /Pwtg + 2rs4 2 + 2b0 X + 2X
2

S4 = ab0 2rswtg − Vwtg 2 /Pwtg
S5 = rswtg 2 − Vwtg 2 /Pwtg + 2b0 X + X + b0 2

2
and
b0 = X − X
a = ω0 T0
The power contribution can also be inferred from Eq. 4 and can be put in real and
reactive terminologies as,
ω ref
Ki 1
s Tm
1/Ts 1 1/Tc 1 1/T5 1
s s s
ω meas 1/R
1/Tc
1/Ts 1/T5
Fig. 2 Governor dynamics’ modeling with automatic generation control
β1 |Vwtg |2
Pwtg = (8)
(β1 2 + β2 2 )
β2 |Vwtg |2
Q wtg = (9)
(β1 2 + β2 2 )
and simplification terms can be put forth as,
sab0
β1 = rswtg +
1 + s2a2
b0
β2 = X +
1 + s2a2
The load flow initiation is done using a flat voltage at the induction machine
connected bus. On solving the slip equation, i.e., Eq. (7), the maximum negative
value of the slip is taken as the initial slip for the operation of the induction machine.
Equation (9) is solved for a flat voltage, initial slip and initial real power status of
the machine.
The obtained reactive power is fed to the load flow algorithm, so as to yield a nearly
perfect approximation for the reactive power drawn by the induction generator from
the capacitor banks or grid. This is the modified load flow algorithm implementation
with hybrid topologies of synchronous as well as induction machines. Thenceforth,
the regular load flow procedure is followed.
The load flow results of the present system under consideration have been men-
tioned in Table 2. The initial conditions of the four machines are mentioned in Tables 3
and 4.
The mathematical modeling of the synchronous machines of the multi machine
system is given in detail in [18, 19]. All reactances in terms of line, capacitor banks,
transformers and so on are translated into an admittance format and the load com-
putation at h nodes is done using [15, 20],
Yh = (Ph − j Q h )/|Vh |2 (10)
Automatic generation control with governor transient dynamics has been taken
into account, and the modeling scheme can be shown as in Fig. 2 [21]. The parameters
514 H. Ahsan et al.
Table 3 Initial conditions: synchronous generators

Parameters SG1 SG2 SG3
I SG0 0.2083 − 0.4007i 1.5800 + 0.2039i 0.8163 + 0.1645i
δ0 1.0719 61.9350 54.4525
I d0 0.4045 1.2983 0.5686
I q0 0.2008 0.9233 0.6084
E d0 ’ 0.0000 0.6164 0.6132
E q0 ’ 1.0644 0.7986 0.7850
E f d0 1.0989 1.8060 1.4282
Table 4 Initial conditions: WTG

Parameters Values
E D40 1.0109
E Q40 0.0050
I D40 0.4998
I Q40 0.2053
s0 −0.0074
Table 5 Governor control parameters

Parameters SG1 SG2 SG3
R 0.05 0.05 0.05
ki 0.2362 0.5020 0.2618
Tc 0.40 0.40 0.40
T5 5 5 5
Ts 0.1 0.1 0.1
T3 0 0 0
T4 0 0 0
T max
m 1.6 3.2 1.7
T min
m 0 0 0
of the speed governor control for the three areas of synchronous generators have
been given in Table 5. The automatic voltage control has been incorporated using the
IEEE-Type I exciter, the details of which can be found in [17].
Fig. 3 a Conventional pitch 100 /s

control: PI controlled
ω PI 90
β
technique, b adaptive pitch β 1
control technique Control
1+Td s
0
Rate Limiter
ω ref
(a)
Adaptive
ω β
Pitch
Control
ω ref
(b)
3 Adaptive Pitch Control
An adaptive control system may be defined as one that adjusts automatically on-line
the parameters of its controller, so as to maintain satisfactory performance when the
parameters of the plant(controlled system) vary with time [22].
Problem Formulation
A linear discrete time model of the plant for controller design is required and is
assumed to be of the form,
N N
z(k) = m i z(k − 1) + n i p(k − 1) (11)
i=1 i=1
where, z(k) is the deviation of the controlled variable from its nominal value and p(k)
is the control signal at kth sampling instant. In Eq. (11), mi and ni are the model
parameters for all values of i, which are estimated on-line at each sampling instant.
Quadratic Indexed-Control Law
For devising a control law, the following performance index is defined, where
516 H. Ahsan et al.
W = [ẑ(k + 1)]2 + d[ẑ(k + 1) − z(k)]2

(12)
+ c[ p(k) − B ∗ p(k − 1)]2
The second term can however be approximated easily as,
d [z(k + 1) − z(k)]
{z(k + 1)} =
dt T
Mathematical conveniences like derivative terms are introduced to impart fast
correcting actions and early damping. Similarly, integral correcting action is incor-
porated with binary constant (B∗ ) to facilitate inclusion/exclusion of past control
signal, as needed. d and c are weighing constants, attached to the controller.
Optimal control signal p(k) is obtained by first replacing z(k + 1) in Eq. (12) by
z(k+1) obtained from Eq. (11) and then finally setting
∂W
=0
∂ pk
The optimal control signal is yielded as,
[g1 dz(k) + cB ∗ p(k − 1) − f (k)]

p(k) = (13)
[c + g1 2 (1 + d)]
where,
N N

f (k) = g1 (1 + d) m i z(k − i + 1) + n i p(k − i + 1)
i=1 i=1
Figure 3 delineates the difference in the implementation of the conventional pitch

control and the proposed APC technqiue.
The system output is assumed to be the digression in network frequency [23].
While regulating this driving point parameter, the pitch angle of the WTG is altered.
This is because, the system frequency gets affected when either the wind speed
anticipates turbulence or the network entertains some fault scenario.
APC in a Gist
The implementation of the adaptive control scheme requires the following steps in
each sampling period:
(1) Sample the system output.
(2) Estimate the model parameters.
(3) Compute and apply the derived control signal.
In order to assess/illustrate the application of the proposed technique in the West-

ern System Coordinating Council, a simulation was performed. Figure 4 depicts the
fault location as implemented in simulation studies. The fault is sustained for 95
milliseconds. The wind generator meets with a staggering wind profile, as shown in
Fig. 5.
It can be seen from Fig. 6, that the system parameters converge to their final values
in a few milliseconds. This implies the attainment of a steady state, when the adaptive
pitch control scheme is in action.
Figure 7 illustrates the WTG responses, when the system is perturbed with a 3-φ
short circuit [25]. The deviations in power, voltage as well as pitch angle of the
induction generator are appreciably lesser with APC equipped WTG than with a
conventionally employed PI controlled WTG. The peak overshoot is reduced by
over 12.2% in wind power output with APC driven generator. Similarly, the terminal
Fig. 4 Wind turbine generator fed Western System Coordinating Council [24]
Fig. 5 Wind perturbation

518 H. Ahsan et al.
Fig. 6 Evolution of model

parameters online
Fig. 7 Simulation
results-Wind turbine
generator responses: a power
scene, b voltage scene, c
pitch angle response
WTG voltage reflects considerable smoothing and flat profile, seconds after the fault
is aborted. Significant improvement in pitch angle regulation is noticed, where the
blade is turned at lower angles as the nature of wind and fault inception changes.
Clearly, the APC based blade does not furnish any pitch angle spikes. This means, the
blade does not require instantaneous movement. On the contrary, the PI controlled
pitch angle shows irregular responses, which might be lethal for the turbine blade.
Some inferences drawn from the simulation study results are:
(1) Even though PI controllers are categorized as one of the simplest controllers,
which can be tuned to suit a plethora of operating conditions, but adaptive con-
trollers have a number of key advantages. Firstly, there is no requirement of a
transfer function in adaptive control implementation. Secondly, these controllers
adapt/self-tune to the changing operating conditions on-line. The variations is
system operating states does not allow the PI controllers to modulate their control
signal [26, 27]. However, this ability is found in adaptive control stategies.
(2) APC driven WTG robustly handles the turbulent wind and network fault contin-
gency [28].
(3) Comprehensive power and voltage smoothing is witnessed.
(4) Lastly, the blade pitch angle is effectively regulated without any spikes in pitch
angle.
5 Conclusion
The development and application of an Adaptive Pitch Control scheme has been
proposed for a simple squirrel cage induction generator, which serves as a wind
turbine generator in a multi machine power system. Detailed mathematical modeling
and load flow implementation is presented. A self tuned adaptive control strategy,
which is based on a quadratic index qualified control law, determines the pitch angle
of the WTG. Comprehensive stability is accorded to the wind generator, by smoothing
the power and voltage profiles, in addition to regulation of the turbine blade angular
movement. A fault is simulated in the system model, with a varying wind in the
background, and a comparative analysis is carried out between an APC based WTG
and a PI based WTG. Simulation studies uphold the potency of the proposed scheme.
In future, the aforementioned APC based turbine blade control can also be extended
to doubly fed induction generator driven systems, under wind turbines. Also, a drive
train model of the wind turbine can also be considered with minimum inertia and
damping.
520 H. Ahsan et al.
References
1. Farzin H, Fotuhi-Firuzabad M, Moeini-Aghtaie M (2017) A stochastic multi-objective frame-

work for optimal scheduling of energy storage systems in microgrids. IEEE Trans Smart Grid
8(1):117–127
2. Ghofrani M, Arabali A, Etezadi-Amoli M, Fadali M (2013) Energy storage application for
performance enhancement of wind integration. IEEE Trans Power Syst 28(4):4803–4811
3. Hasanien HM, Muyeen S (2012) Design optimization of controller parameters used in vari-
able speed wind energy conversion system by genetic algorithms. IEEE Trans Sustain Energy
3(2):200–208
4. Zhang S, Mishra Y, Shahidehpour M (2016) Fuzzy-logic based frequency controller for wind
farms augmented with energy storage systems. IEEE Trans Power Syst 31(2):1595–1603
5. Milano F, Dörfler F, Hug G, Hill DJ, Verbič G (2018) Foundations and challenges of low-inertia
systems. In: Power systems computation conference (PSCC). IEEE, pp 1–25
6. Sierra-García JE, Santos M (2021) Improving wind turbine pitch control by effective wind
neuro-estimators. IEEE Access 9:10 413–10 425
7. Thapa KB, Jayasawal K (2020) Pitch control scheme for rapid active power control of a PMSG-
based wind power plant. IEEE Trans Indus Appl 56(6):6756–6766
8. Wan-Jun H, Shi-Yuan H, Yi-Shi J, Xin-Jing C, Song-Qing C, Yan-Hui Q, Zhe L (2018) Variable
pitch sliding mode control of wind turbine based on disturbance observer. In: Chinese control
and decision conference (CCDC), pp 5221–5225
9. De T, Rashid A, Ying D, Sheng LH (2020) Pitch angle control of modern variable speed variable
pitch wind turbine based on linear active disturbance rejection control approach. In: IEEE/IAS
industrial and commercial power system Asia (I CPS Asia), pp 1382–1387
10. Arabali A, Ghofrani M, Etezadi-Amoli M, Fadali MS, Baghzouz Y (2013) Genetic-algorithm-
based optimization approach for energy management. IEEE Trans Power Delivery 28(1):162–
170
11. Chaiyatham T, Ngamroo I (2017) Improvement of power system transient stability by PV farm
with fuzzy gain scheduling of PID controller. IEEE Syst J 11(3):1684–1691
12. Ahsan H, Mufti M-UD (2020) Sweeping power system stabilization with a parametric fuzzy
predictive control of a generalised energy storage device. IET Gener Transmission Distrib
13. Ahsan H, Mufti M-D (2019) Modelling and stability investigations of an aggregate wind farm–
fed multi-machine system. Wind Eng 0309524X19862759
14. Muyeen S, Takahashi R, Ali MH, Murata T, Tamura J (2008) Transient stability augmentation
of power system including wind farms by using ecs. IEEE Trans Power Syst 23(3):1179–1187
15. Wang L, Truong D-N (2013) Stability enhancement of DFIG-based offshore wind farm fed to
a multi-machine system using a Statcom. IEEE Trans Power Syst 28(3):2882–2889
16. Ahsan H, Mufti M-D (2019) Distributed storage approach versus singular storage approach: a
dynamic stability evaluation. Int J Power Energy Syst 39(1)
17. Sauer PW, Pai MA, Chow JH (2017) Power system dynamics and stability: with synchrophasor
measurement and power system toolbox. Wiley
18. Ahsan H, Salam Z, Lone SA et al (2017) Modeling and simulation of an energy storage based
multi-machine power system for transient stability study. In: Energy Conversion (CENCON),
IEEE Conference on 2017. IEEE, pp 78–83
19. Ahsan H, Mufti M-D (2019) Dynamic performance improvement of a hybrid multimachine
system using a flywheel energy storage system. Wind Eng 0309524X19849853
20. Ahsan H, Mufti MD (2017) Modeling and simulation of a superconducting magnetic energy
storage based multi-machine power system for transient stability study. In: 2017 6th interna-
tional conference on computer applications in electrical engineering-recent advances (CERA).
IEEE, pp 347–352
21. Hosseini S, Etemadi A (2008) Adaptive neuro-fuzzy inference system based automatic gener-
ation control. Electric Power Syst Res 78(7):1230–1239
22. Lim C (1989) A self-tuning stabiliser for excitation or governor control of power systems.
IEEE Trans Energy Convers 4(2):152–159
23. Ahsan H, Mufti MUD (2020) Systematic development and application of a fuzzy logic equipped
generic energy storage system for dynamic stability reinforcement. Int J Energy Res
24. Ahsan H, Mufti MD (2020) Comprehensive power system stability improvement with ROCOF
controlled SMES. Electric Power Components Syst 48(1–2):162–173
25. Ghosh S, Kamalasadan S (2017) An energy function-based optimal control strategy for out-
put stabilization of integrated DFIG-flywheel energy storage system. IEEE Trans Smart Grid
8(4):1922–1931
26. Silva-Saravia H, Pulgar-Painemal H, Mauricio JM (2016) Flywheel energy storage model, con-
trol and location for improving stability: the Chilean case. IEEE Trans Power Syst 32(4):3111–
3119
27. Lee D-J, Wang L (2008) Small-signal stability analysis of an autonomous hybrid renewable
energy power generation/energy storage system part i: time-domain simulations. IEEE Trans
Energy Convers 23(1):311–320
28. Taj TA, Hasanien HM, Alolah AI, Muyeen SM (2015) Transient stability enhancement of a
grid-connected wind farm using an adaptive neuro-fuzzy controlled-flywheel energy storage
system. IET Renew Power Gener 9(7):792–800
Detection of Broken Rotor Bar Fault
in an Induction Motor Employing Motor
Current Signature Analysis
Alok Verma, Pratul Arvind , Somnath Sarangi, Jayendra Kumar,

and Anumeha
1 Introduction
The induction motor plays a necessary part in any industrial plant. An early detection
of faults in the motor will be very helpful to avoid costly breakdowns in industrial
plant.
According to IEEE broken rotor bar fault is one among the supreme faults in
an induction motor [1, 2]. Uncontrolled vibrations, greater thermal stress or torque
fluctuation as well as poor starting performances are some of the most common faults
in a broken rotor bar. For the detection of a broken rotor bar fault, various methods
were already used. The diagnosis of the rotor faults in an induction motor can be
detected using various sensor signals in which vibration monitoring is a well-known
approach [3, 4].
Mechanical systems often exhibit non-linear behaviours, because of vibrations in
parameters like friction and damping. Therefore, the normal vibration signals may be
changed to a complex and non-linear form [5]. There may have some limitations in
A. Verma (B)
Advanced Remanufacturing and Technology Centre, Agency for Science, Technology and
Research (A*Star), Singapore, Singapore
e-mail: alok_kumar_verma@artc.a-star.edu.sg
P. Arvind
ADGITM, New Delhi, India
S. Sarangi
Indian Institute of Technology, Patna, India
e-mail: somsara@iitp.ac.in
J. Kumar
National Institute of Technology, Jamshedpur, India
Anumeha
Government Women’s Polytechnic Jamshedpur, Jamshedpur, India
524 A. Verma et al.
often applied techniques that include time and frequency domain techniques. Conse-
quently, the techniques for non-linear dynamic parameter estimation are needed that
can give a useful option to retrieve the fault-related parameters hidden in the complex
as well as non-linear samples [5]. Diverse experiments have already been performed
for parameter extraction and fault diagnosis on-linear dynamic data set [6]. However,
some literatures used approximate entropy (ApEn) as a fault diagnosis tool for non-
linear signals. ApEn has also been described [5, 6] in the fields of the physiological
signal and vibration signal processing of the rotating machine. Nevertheless, ApEn
showed very close similarity with the time series and self-matching which makes
ApEn majorly dependent on the length of time series [7]. Later, a new method for
entropy called sample entropy (SampEn) which excludes, self-matches was offered
by Richman and Moorman [8] to overcome the limitations of ApEn, and has brought
a lot of consideration in this field of work.
In this paper, faults of the broken rotor bars were investigated by employing rotor
vibration as diagnostic media. To tackle the non-linear behaviour existing in the rotor
bar using vibration, multiscale entropy (MSE) was used which describes the unifor-
mity in the diagnostic information. Despite that, the support vector machine (SVM)
was used for the categorization and diagnosis of fault in broken rotor bars. New
entropy named multiscale entropy was discussed for the measurement of complexity
in the recent literature [9]. Various researchers applied this latest method for making
a distinction between healthy and congestive hearts failure of a person [9, 10]. Few
works have been done where multiscale entropy (MSE) has been employed in the area
of fault diagnosis of the machine and that was only for the vibration signals. Zhang
et al. [11] has introduced SVM based intelligent fault diagnosis method by extracting
multiple time scale of entropy (MSE) from vibration signal. An approach to differ-
entiate the aligned motor with a misaligned one using vibration was performed by
Lin et al. [12]. MSE of vibration signal was used for bearing fault diagnosis using
MSE to tackle the nonlinearity that existed with the bearing fault by Zhang et al.
[13]. In the later stage, adaptive neuro-fuzzy inference system (ANFIS) was used to
diagnose the bearing faults from the system. Several research [14–17] have also been
performed to diagnose the machine failures using machine learning techniques.
In this paper, a new approach is proposed to detect the faults in the broken rotor
bar by the authors, to tackle the nonlinearity which exists in the vibration of a broken
rotor. For the extraction of information from vibration, a new approach of multiscale
entropy analysis is being processed that differentiates the fault of broken rotor bars
from the healthy rotor bar.
2 Multiscale Entropy
The MSE algorithm is employed in the present work that depends on the SampEn for
variety of scales. A statistical method for noisy time series identified as approximate
entropy (ApEn) was introduced by Pincus [7] in 1991. This method was used to
measure the regularity of time series data. In the presented study, instead of revising
Detection of Broken Rotor Bar Fault in an Induction Motor … 525
Fig. 1 SampEn with respect to the scale factor for white and 1/f noises time series
the previously used regularity measure ApEn statistical, the MSE algorithm is used
for various scales of the similar process. As for refinement of ApEn, SampEn is
employed and it measures the regularity in series data. To develop the entropy algo-
rithms, two important parameters such as m and r need to be selected properly;
where two m sequences are measured with the tolerance r. SampEn may increase as
‘r’ decreases; therefore, the determination of these two parameters is important. From
the previous works of literature, the presented work selected r = 0.15 × standard
deviation of the time series data, and ‘m’ = 2.
A new algorithm called MSE was proposed by Costa et al. [10], in which SampEn
was calculated for multiple scale called scale factor. MSE of the time series X 1 ,…,
X N can be calculated as
1 jτ
y (τ
j =
)
Xi (1)
τ i=( j−1)τ +1
where τ is known as scale factor, and j varies from 1 to Nτ . In case of τ = 1, the output
of MSE time series is the original series. When τ increases, the length of time series
decreases and when the scale factor increases, SampEn of the white noise decreases
as shown in Fig. 1. On the other hand, SampEn of 1f noise (pink) remains stable with
the rise of scale factor.
3 Experimental Arrangement
As shown in Fig. 2, an experiment is carried on the system. The whole systems

consist of three- induction motor, different sensors, data acquisition system,
computer/controller, and display as shown in Fig. 2. For the experiment as shown in
526 A. Verma et al.
Fig. 2 Experimental test rig set-up: a overall experimental setup, b 3- induction motor and current
sensor used, c speed controller/vector drive, d proximity sensor, e accelerometer
Fig. 2a three phases, 0.75 HP inductions motor without any load was used. Figure 2b
shows the data acquisition system. Figure 3 shows the 3- induction motor with
broken rotor bar used in the experiments.
An important sensor that is employed in this experiment is the current sensor.
Current sensor was clamped on the stator winding of the induction motor as depicted
Fig. 3 3- induction motor

with broken rotor bar used in
the experiments
Table 1 Specifications for experimental components

Sensors/machine Manufacturer Sensitivity/specification
Induction motor Marathon Electric (healthy and broken 0.75 HP, 50 Hz, 2850 RPM
rotor bars)
Current probe Fluke 100 mV/A
Triaxial Accelerometer Bruel and Kjaer 9.812 mV/(m/s2 )
in Fig. 2d. The overall specifications for the components used in the experiments are
represented in Table 1.
3.1 Experimental Procedure
The experimental setup as depicted in Fig. 2 is used to conduct the experiments. For
experiments, we have acquired four sets of real-time current signals. The first set of
the data was acquired with the healthy motor (N1) running at different speeds from
700 to 1600 rpm with increment of 30 rpm. The second set of the current data was
acquired from the motor with broken rotor bar faults (B1) and running at different
speeds (700–1600 rpm) by the interchange of the healthy motor to the motor with
broken rotor bars. For the collection of the third set of healthy data (denoted by N2),
the motors were again interchanged from faulty motor to healthy motor and then
data was stored. Finally, the fourth set of the current data was acquired by changing
the motor from healthy (N2) to the again motor with broken rotor bar (B2). Total
31 samples were collected for each data set and each trial data has 15,364 signals
values which was collected in 3 secs at 5.12 S/s. The control factors and their levels
are shown in Table 2.
Wavelet transform is a mostly used method to remove noise from the signals,
Although, wavelet transform is used as per the types of the signals. Figure 4 represents
the MSE curves (SampEn with respect to scale factor) that is calculated from the
current signals with the denoised signals from healthy motor and faulty motor with
broken rotor bars, all rotating at 1600 rpm. The idea to use the SampEn to distinguish
healthy signals to faulty signals is given by this observation. Diagnosis of fault
in the broken rotor bars of an induction motor depends on the method of current
measurement which is presented as follows.
The final features were calculated by differentiating the MSE features and the
denoised current signals for the detection purpose. A classification algorithm like
SVM has used these calculated values as an input to detect motor fault. MSE-related
Table 2 Levels and control

Control factors Level
factors set for the experiments
A Motor condition Normal and faulty
B Rotational speed 700–1600 rpm
528 A. Verma et al.
Fig. 4 Sample Entropy at a various Scale factor using motor current from healthy motor and motor
with broken rotor bar fault at 1600 rpm before and after denoising
features are discussed in result section which was based on the performance of the
classifiers. The Daubechies wavelet transform from the current signals was used to
perform the denoising and the algorithm was developed using MATLAB with the
parameters settings as (tptr = “rigrsure”; n = “2”; wav = “db4”).
The explained method for broken rotor bar fault detection underlies the concept of
statistical analysis. The proposed method employs the use of multiscale entropy in
order to tackle the non-linear behaviour that is present in the motor vibration. Also,
the regularity in the diagnostic information is explained. Further, the work presented
in the paper describes an approach for analysis of multiscale features which easily
differentiate the healthy and faulty samples of stator current of the experimental
induction motor.
The approaches used for the statistical analysis have been deliberated above. The
experimental approach performed the tests in the range of speed between 700 and
1600 rpm with an increase of 30 rpm. Table 2 demonstrates the parameters along
with their levels used for the experiment. Further, 03 trials of current samples were
performed. The three different tests are further elaborated. For the first trial, 62 current
samples of N1 U B1 set were considered. For the second test, 62 current samples of
N2 U B2 set were also considered. In both the above cases, 31 samples were used
as training and the rest 31 samples were used for the purpose of testing. The third
experiment considered the current samples received from N1 U B1 for the purpose
Table 3 Precision value of fault predicted for the training set employing Test 1–3 of current signal
(i.e. test-1 = N1 U B1, test-2 = N2 U B2 and test-3 = N3 U B3) that comprises of in total 62 current
samples, training set = 31 samples
Test Training Cost Gamma(γ ) Cross-validation Support Time
data function accuracy (%) vector (secs)
(C) approx
1 31 10,000 0.001 91.8 661 2
0000
2 31 10,000 0.001 90.4 448 2
0000
3 31 10,000 0.002 90.1 701 3
0000
Table 4 Precision value of fault predicted for the testing set employing for Test 1–3 of current
signal (i.e. test-1 = N1 U B1, test-2 = N2 U B2 and test-3 = N3 U B3) that comprises of total 62
samples, testing set = 31 samples
Test Testing samples Accuracy Time (sec)
1 31 94.68% 2
2 31 94.1% 2
3 31 93.3% 3
of training while the current samples of N2 U B2 were used as testing. Support

vector machine using multiscale entropy-related features were used to determine the
performance as already discussed.
The inputs considered for the classifier are speed and healthy/faulty current and the
output is the detected fault using current. Test-1 resulted in the maximum exactness
and performed well for all the three tests. In order to get the optimal accuracy, the
classifier has been trained and tested for numerous values. Tables 3 and 4 yield the
best value of fault detection obtained by current.
Results assert that the tested current accuracy is about 90.1–94.68% for all the
cases from experiments which is quite good. The presented results clearly reveal
that the developed SVM model can be trained and tested to predict the broken rotor
bar fault using MSE features of current signals at an early stage from failure with
reasonable accuracy.
5 Conclusion
A comprehensive information regarding the response of current arising due to the

broken rotor bar is presented in this work by using statistical analysis. This exper-
imental investigation is a contribution towards the improvement in the rotor fault
530 A. Verma et al.
diagnosis by using the multiscale entropy. Following conclusions may derive based
on the investigation:
(i) This paper is an important experimental contribution for the analysis of rotor
fault on the current signal by using MSE signal.
(ii) Effective modelling of the current using SVM during the rotor fault detection
was shown in the results obtained by experiments. The predicted results are
within 95% which is promising.
References
1. Bonnet AH (1988) Analysis of rotor failures in squirrel cage induction machines. IEEE Trans
Ind Appl 24:1124–1130
2. Bonnet AH, Soukup GC (1992) Cause and analysis of stator and rotor failures in three-phase
squirrel-cage induction motors. IEEE Trans Ind Appl 28:921–937
3. Nandi S, Toliyat HA (1999) Fault diagnosis of electrical machine—a review. In: International
conference on electric machines and drives, Seattle, WA, pp 219–221
4. Thomson WT, Fenger M (2001) Current signature analysis to detect induction motor faults.
IEEE Ind Appl Mag 7:26–34
5. Yan RQ, Gao RX (2007) Approximate entropy as a diagnostic tool for machine health
monitoring. Mech Syst Signal Process 21:824–839
6. Yan RQ, Gao RX (1995) Machine health diagnosis based on approximate entropy. In:
Instrumentation and measurement technology conference, Italy, pp 2054–2059
7. Pincus SM (1991) Approximate entropy as a measure of system complexity. PNAS 88:2297–
2301
8. Richman JS, Moorman JR (2000) Physiological time-series analysis using approximate entropy
and sample entropy. Am J Physiol H 278:2039–2049
9. Haitham MA, Alan VS (2002) Use of sample entropy approach to study heart rate variability
in obstructive sleep apnea syndrome. IEEE Trans Bio Eng 50:1900–1904
10. Costa M, Goldberger AL, Peng CK (2002) Multiscale entropy analysis of complex physiolog-
ictime series. Phys Res Lett 89:68–102
11. Zhang L, Xiong G, Liu H, Zou H, Guo W (2009) An intelligent fault diagnosis method based
on multiscale entropy and SVMs. Lect Notes Comput Sci 5553:724–732
12. Lin JL, Liu JYC, Li CW, Tsai LF, Chung HY (2010) Motor shaft misalignment detection using
multiscale entropy with wavelet denoising. Expert Syst Appl 37:7200–7204
13. Zhang L, Xiong G, Liu H, Zou H, Guo W (2010) Bearing fault diagnosis using multi-scale
entropy and adaptive neuro-fuzzy inference. Expert Syst Appl 37:6077–6085
14. Verma A, Sarangi S, Kolekar M (2016) Misalignment faults detection in an induction motor
based on multi-scale entropy and artificial neural network. Electr Pow Compo Sys 44:916–927
15. Verma A, Sarangi S (2015) Fault diagnosis of broken rotor bars in induction motor using multi-
scale entropy and backpropagation neural network. In: Intelligent computing and applications,
vol 343. Springer, New Delhi, pp 393–204
16. Verma A, Sarangi S, Kolekar M (2014) Stator winding fault prediction of induction motors
using multiscale entropy and grey fuzzy optimization methods. Comput Electr Eng 2246–2258
17. Arvind P, Maheshwari RP (2012) A wavelet packet transform approach for locating faults in
distribution system. In: 2012 IEEE symposium on computers & Informatics (ISCI), Penang,
Malaysia, pp 113–118
Performance Analysis
of Photovoltaic-Grid Connected System
for Electric Vehicle Charging
Manoj Kumar Sharma, Karanbir Singh, and Satish Kansal
1 Introduction
The transportation sector has seen substantial growth in recent years owing to ever
increasing population and improving economic conditions. This has led to increased
use of internal combustion engines (ICE) vehicles. The ICE vehicles depend on
fossil fuels for their operation that adds up to global air pollution substantially.
The increasing air pollution and greenhouse effect, caused by the release of CO2
from the ICE vehicles as well as limited availability of fossil fuels in the future,
has necessitated shifting to an alternate clean approach. Electric Vehicles (EVs)
being a clean source of transportation are an effective alternative. With features like
immediate torque, silent ride and premium performance along with lower fuel and
maintenance costs, the acceptance of EVs is on the rise. According to a report, the
market share of electric vehicles will increase from 1% in 2015 to 9% in 2025 [1]. As
per another forecast, EV market share will rise from 1.1 million worldwide in 2017
to 11 million worldwide by 2025 surging to 20 million by 2030 [2]. Although EVs
provide us with a lot of benefits over the ICE vehicles, there are still some economical
and technical issues like the stability of the power grid with regard to the usage of
high-power chargers, battery management, such as thermal ratings, safety and cell
balancing, which need to be addressed. Operating the power plants at their maximum
generation range raises many issues regarding their capabilities especially the aging
factor and increase in pollution levels by operating coal-fired plants for longer hours.
Another issue of charging time for the battery of EVs is a primary concern, as the
vehicle needs to absorb power from the grid [3]. The various types of charging
M. K. Sharma (B) · K. Singh

Department of Electrical and Electronics Engineering, UIET, Panjab University, Chandigarh, India
e-mail: mks_uiet@pu.ac.in
S. Kansal
Department of Electrical Engineering, BHSB Institute of Engineering and Technology, Sangrur,
Punjab, India
532 M. K. Sharma et al.
Fig. 1 Percentage penetration forecast of electric vehicles
methods have been compared in [4] to select the most efficient charging method.
Ali et al. have modelled hybrid energy storage system using EV batteries and super
capacitors for enhancing lifetime of the EV batteries and compared the performance
for different optimization strategies [5]. A conceptual architecture and an assessment
framework were proposed to explore integration scenarios of EVs and renewable
energy generation in distribution networks [6]. The combination of PV energy and
EVs in uncontrolled charging and smart charging strategies has been studied [7],
as was a two-stage framework for the economic operation of an EV parking deck
with renewable energy generation [8]. Most of the researches on charging of EVs and
renewable energy generation have focused on design, operation and optimal charging
strategies [9–13] (Fig. 1).
The issue of electric vehicle charging and power supply from grid and renewable
energy sources is a topic of research worldwide. The integration of PV with the EV
charging system has been on the rise due to rapid growth of EVs and concern over
the effects of greenhouse gasses. An overview of different charging approaches of
EVs is discussed and a PV-grid charging method incorporating a battery manage-
ment system is elaborated by Bhatti et al. [14, 15]. Khan et al. [16] have presented
the charging technologies of EVs, their sustainable development and characteristics.
They also forecast the current status, economic assessment, power market operation
and control and safety aspects of charging EV system. An electric vehicle charging
system incorporating energy storage system (ESS) for power quality improvement
and providing fast charging is discussed [17, 18]. A Battery Energy Storage Station
(BESS)-based hybrid power station to tackle the disadvantageous unstable power
output from PV and wind power generations is elaborated and SOC-based control
strategy has been adopted for smoothing the output fluctuation [19]. Mortaz and
Valenzuela have modeled a micro grid consisting of renewable energy resources
connected within a grid and using EV batteries in the parking facility for energy
storage [20]. A rule-based energy management system to provide uninterrupted
Performance Analysis of Photovoltaic-Grid Connected … 533
constant price charging is proposed. The model incorporates a PV-grid system along
with an energy storage unit. A daytime charging system using a PV, an ESU unit and
a micro grid with the help of heuristic rule-based strategies is proposed [21, 22]. The
model has helped to reduce burden on the micro grid as well as provided efficient
cost saving.
A PV-grid system for EV charging using particle swarm optimization PSO algo-
rithm to find optimum size of PV modules and ESU’s need has been modeled [23].
The proposed system allows charging EVs at a fixed energy rate without facing
economic losses. However, not much has been discussed about the power quality
issues related to electric vehicle charging on a PV-Grid system.
In an electric power system, a harmonic is a voltage or current at a multiple of the
fundamental frequency of the system, produced by the action of non-linear loads such
as rectifiers, discharge lighting or saturated magnetic devices. The Total Harmonic
Distortions (THD) occurring in a system can be calculated using Eq. (1).

V22 + V32 + V42 + · · · + Vn2
Total Harmonic Distortion, THD = (1)
V1
Harmonic frequencies in the power grid are a frequent cause of power quality
problems. Harmonics in power systems result in increased heating in the equipment
and conductors, misfiring in variable speed drives, and torque pulsations in motors.
Thus it becomes of utmost importance to tackle this problem. The effect of harmonic
distortions on the distribution transformers has been analyzed in [24]. To tackle these
power quality issues coordinated charging has been proposed in [25] to minimize
the power losses and to maximize the main grid load factor. As charging of electric
vehicles has led to main power quality issues, thus, in this work an analysis of
performance of PV-Grid system for EV charging is done.
In the present paper, a PV-grid system for electric vehicle charging has been
designed for four charging stations having eight vehicles each. The battery specifica-
tions of EVs considered are same as that of being used in real-life batteries of EVs.
The analysis is done for effect of different internal resistances and different SoCs of
EVs, total harmonic distortion (THD) and scheduling of PV-grid system. Section 2
of the paper describes the proposed PV-grid model for EV Charging. The description
of specified model and its working is given in Sect. 3. The simulations, performance
analysis and results are discussed in Sect. 4. Sections 5 and 6 of the paper describe
the economic aspect and conclusion, respectively.
2 Proposed PV-Grid Connected Model for EV Charging
One of the most viable and abundant renewable energy sources is solar energy. This
solar energy can be harnessed and used for day-to-day power needs. This harnessed
energy can be used for the purpose of electric vehicle charging as well. So rather than
depending on the AC grid solely for charging of electric vehicles, the PV module can
be scheduled in a way so that maximum energy can be accessed from it for charging
during its operation and the rest of the charging needs can be fulfilled from the AC
grid. This decreases the overall stress on the grid due to electric vehicle charging.
For this purpose, the model has been designed with four charging stations, each
serving eight electric vehicles at a time. The power generated from the PV module
will be continuously measured and the shift between PV and grid will take place on
the basis of comparison between these measured values and a predefined threshold
value of 10 kW.
2.1 Design Parameters
The proposed PV-Grid model is designed in MATLAB/Simulink. The various

components along with their ratings, used for designing of the proposed PV-Grid
system are given in Table 1:
Table 1 Components and

S. no Component Ratings
their ratings used in the
PV-Grid system 1 AC grid 3-phase ac source 20 MVA,
13.8 kV, 50 Hz
3-phase 20 MVA,
transformer 13.8/230 kV,
50 Hz
3-phase active load 5 MW
3-phase reactive 500 kVA
load
2 PV module PV module rating 236 kW, 480 V
Temperature 450 °C
Irradiance 1000 W/m2
3 Electric Battery rating; 48 V, 24 Ah
vehicles Discharge current 10.4348 A
Capacity Nominal 21.7 Ah
Voltage
Range of initial [30%–80%]
SOC%
4 AC-DC Diodes 6
converters
5 Filters Resistor 1
Capacitor 10 μF
3 Modelling and Description of the Proposed Model
The schematic diagram for the proposed PV-Grid system is shown in Fig. 2. In the
proposed model a PV-Grid system has been used in which power is being supplied
from the PV module and AC grid to charge the electric vehicles connected to the
charging stations. The photovoltaic modules use light energy (photons) from the Sun
to generate electricity through the photovoltaic effect. Nominally rated maximum
power out (kWp) of a solar array of n modules, each with a maximum power of Wp
at STC is given by Eq. (2).
kW p = n × W p/1000 (2)
The available solar radiation (Ema) varies depending on the time of the year and
weather conditions. However, based on the average annual radiation for a location
and taking into account the efficiency (η) of the cell, the estimation of an average
PV system energy yield is obtained as per Eq. (3).
E p = Ema × kW p × η (3)
In the present model, four charging stations are being used to observe the results
where each charging station consists of eight charging points for connecting the
electric vehicles. Out of these eight charging points, six charging points can be
AC Grid
AC to DC converter
Charging Station Charging Station Charging Station Charging Station

3 4
1 2
Fig. 2 Schematic diagram for the proposed PV-Grid system for EV charging stations
connected to the grid or PV module. The reason for this is to give provision for
backup storage. The two charging points can be Energy Storage Units to prevent any
wastage of solar energy in case all the charging points are not being used to supply
energy. These Energy Storage Units can then be used to provide supply for charging
even when there is no solar energy generation for instance at night.
3.1 Flowchart
The flowchart depicting the entire process of working of the proposed PV-Grid hybrid
system is shown in Fig. 3. The process is described as follows:
Fig. 3 Flowchart depicting the scheduling process of the PV-grid system

• Initially, when the operation starts, the PV module is being used to charge the EVs.
The PV module makes use of irradiance and temperature as inputs to generate the
output power.
• As the irradiance is high and power generated is enough to charge the elec-
tric vehicle batteries but as time passes the irradiance starts decreasing and so
does the power and as a result PV alone is not sufficient to charge the electric
vehicle batteries. As a result, all the eight charging points in a charging station
are connected to the PV module supply.
• After some time when the power drops below the threshold limit of 10 kW, the
electric vehicle charging points are shifted from the PV module supply to the AC
grid. The AC supply from the AC grid is converted to DC current using AC-DC
converters to form a DC grid.
• Six out of eight charging points are connected to the DC grid whereas two
remaining charging points are connected to the PV module exclusively.
• The basis of this operating model is the switching between PV and AC grid. For
this purpose, a threshold has to be assigned to the system after which the shift
between PV and grid happens.
4 Simulations and Results
4.1 Real-Life Applications
The electric vehicle industry is on the rise and its increasing popularity has drawn
many major manufacturing companies to compete in the e-automobile market world-
wide. A number of major automobile companies have started implementing EV
technology and new electric vehicle products have started emerging. The vehicle
manufacturing companies like Hero Electric, Tunwal E-Vehicle, etc. are investing
in the EV domain and have come up with various EV products like EV cars, EV
bikes, EV scooters, etc. For example, Nyx E5 by Hero Electric is an electric vehicle
scooter and Mini Lithino 48 V by Tunwal E-Vehicle with 48 V/28 Ah and 12 V/28 Ah
batteries, respectively. The battery specifications used in the present model are same
as used in real-life vehicle batteries. Hence, such vehicles can easily be incorporated
with the proposed model to provide electric vehicle charging.
4.2 Effect of Different Internal Resistances on Battery

Performance
In the real world, all the electric vehicles do not have the same rating. The batteries
manufactured by the different manufacturers have varied specifications. To take this
into account, batteries with different internal resistances in the range from 0.002 to
Table 2 Internal resistance in terms of percentage of base value

S. no Battery no State of charge (%) Value of internal Percentage of base
resistance (ohms) value (%)
1 Battery 1 & 2 80 0.012 (base value) 100
2 Battery 3 50 0.002 16.67
3 Battery 4 50 0.006 50
4 Battery 5 40 0.010 83.34
5 Battery 6 40 0.014 116
6 Battery 7 30 0.018 150
7 Battery 8 30 0.022 183.34
0.022 have been selected and have been used for simulations. Also, different values
of state of charges for batteries have been considered. The base value considered for
the batteries is the value of the internal resistance as per default specifications in
the Simulink. The range of values taken for internal resistance of batteries and its
percentage with respect to the base value of the battery is shown in Table 2. The
batteries 1 and 2 with the internal resistance of 0.012 are supplied by the PV
alone.
Case 1: Batteries with the same State of Charge
Initially, the same state of charge (SOC) for all the electric vehicles i.e., 80% with
different internal resistance is considered. The range of values taken for internal
resistance of batteries and its effect on the battery state of charge, battery current and
battery voltage is shown in Table 3.
From Table 3 it can be seen that as the internal resistance of the battery increases:-
• The rate of charge of the battery decreases.
• The battery current decreases.
• The battery voltage remains almost constant.
Table 3 Battery performance with different internal resistances and same State of Charge
Internal resistance Battery SOC (%) Battery current (A) Battery voltage (V)
()
Battery 1 & 2 (0.012) 80.06 −22 50.21
Battery 3 (0.002) 80.14 −35 50.02
Battery 4 (0.006) 80.065 −24 50.02
Battery 5 (0.010) 80.025 −14.5 50.02
Battery 6 (0.014) 80.0175 −42.85 50.02
Battery 7 (0.018) 80.075 −8.5 50.02
Battery 8 (0.022) 80.0012 −6 50.02
Figure 4 shows the relation between internal resistances and the rate of charge of
batteries. It can be concluded that as the value of internal resistance increases, the
rate of charge of batteries decreases.
Case 2: Batteries with different State of Charge
In this case, electric vehicle batteries with different internal resistances and different
state of charge are considered. The range of values taken for internal resistances and
its effect on battery state of charge, battery current and battery voltage is shown in
Table 4.
Fig. 4 Rate of charge of batteries with different internal resistances at same SOC
Table 4 Battery performance with different internal resistances and different SOC
Internal resistance Battery SOC (%) Battery Current Battery Voltage
() (A) (V)
Battery 1 & 2 (0.012) 80.037 −13 50.065
Battery 3 (0.002) 50.155 −44 49.85
Battery 4 (0.006) 50.053 −20.55 49.85
Battery 5 (0.010) 40.041 −16 49.85
Battery 6 (0.014) 40.022 −11 49.85
Battery 7 (0.018) 30.028 −9.5 49.85
Battery 8 (0.022) 30.018 −7.4 49.85
4.3 Scheduling of Power for Charging Electric Vehicles

in the Proposed PV-Grid System
The proposed model has been run for 5 s (simulation time in Simulink) in Simulink
and the various outputs obtained are as follows.
Battery State of Charge (SOC)
Figure 5 shows the variation of State of Charge (SOC) with time for different values
of internal resistance of batteries of Electric Vehicles.
From the graph, it can be seen that as the internal resistance of the battery increases,
the rate of charge of the battery decreases. Also, it can be seen that the shift from the
PV module to the AC grid occurs at time 4.2 s (approximately) when power from
the PV module drops below the threshold value of 10 kW.
Battery Voltages
The rate of change of battery voltage w.r.t. time for different 8 values of internal
resistance of electrical vehicles is shown in Fig.6.
From Fig. 6 it can be seen that the Battery Voltage remains fairly constant. Batteries
1 & 2 show the graph for the charging point being supplied from the PV alone. It can
be seen also that the shift from the PV module to the AC grid occurs at time 4.2 s
(approximately) when power from the PV module drops below a threshold value of
10 kW.
Battery Currents
The following are the Time vs Current graphs, which show the rate of change of
current of EVs with time. The graphs here are plotted for 8 different values of
internal resistance.
Fig. 5 Battery SOC with Internal Resistance ranging from 0.002 to 0.012
Fig. 6 Rate of change of voltage of batteries 1–8 with different internal resistances
From Fig. 7 it can be seen that the Battery Current decreases as the internal
resistance of the battery increases. Batteries 1 & 2 show the graph for the charging
point being supplied from the PV alone. Here it can be seen that from Fig. 8, the shift
from the PV module to the AC grid occurs near 4.2 s (approximately) when power
from the PV module drops below 10 kW.
Fig. 7 Rate of change of current of batteries with different internal resistances

Fig. 8 THD (voltage distortions) injected in the grid caused due to the charging of electrical vehicles
4.4 Electric Vehicle Charging on AC Grid Alone
Electric vehicles are comprised of non-linear loads such as batteries which are
known to cause harmonic distortions in the electrical grid. By using an AC grid
designed in Simulink the Total Harmonic Distortions occurring in the system have
been calculated due to the charging of electric vehicles.
The acceptable THD limit as per standard IEEE 519 is 5%. The results show that
the Harmonic Distortion caused by charging electric vehicles through the AC grid is
14.65% for voltage distortions and 6.28% for current distortions as shown in Fig. 8
and Fig. 9, respectively. Hence it is important to reduce these distortions to a minimal
value as possible.
4.5 Reduction of THD in the Proposed PV-Grid System
Lower THD in power systems means higher power factor, lower peak currents and
higher efficiency. The low THD is an important feature in power systems. The Inter-
national Standards such as IEC 61,000-3-2 set limits on the harmonic currents of
various classes of power equipment. One of the objectives of this model is to control
the Total Harmonic Distortions occurring in the system.
To analyze the total harmonic distortions (THD) occurring due to the charging
of electric vehicles a PV-Grid system has been designed using MATLAB/Simulink
Fig. 9 THD (current distortions) injected in the grid caused due to the charging of electrical vehicles
software and the impact on the proposed system due to electric vehicle charging has
been studied.
It is observed that employing low pass filters and scheduling the charging period
of electric vehicles in the PV-Grid system helps in tackling THD. During the day
when there is an abundant amount of solar energy, the PV module is used to harness
this energy and further utilize it to charge the electric vehicles. As the day passes
the irradiance decreases and the power generated by the PV module also decreases.
When the power generated by the PV module drops below the threshold value of
10 kW the circuit breakers connecting the ac grid to the charging points in the electric
vehicle stations close. After the threshold value of 10 kW, both the PV and ac grid are
connected to the charging points on the electric vehicle charging stations to provide
the power required to charge the electric vehicles.
It is also observed that connecting the electric vehicles solely to the ac grid results
in large harmonics entering the system. On the other hand, by incorporating the PV
module into the system these harmonics are reduced to a large extent. In this system,
the electric vehicles are connected to PV module during the day when there is an
abundant amount of solar energy rather than connecting to the ac grid. As a result,
when PV supply is available the ac grid is not used for the electric vehicle charging
purpose and the load stress imposed on the system is decreased. Hence harmonics
entering into the system are reduced to a significant amount. The values of THD are
drastically dropped to 0.36% and 1.27% for voltage and current distortions as shown
in Fig. 10 and Fig. 11, respectively, when vehicles are charged on the proposed PV-
grid system as compared to the 14.65% and 6.28%, respectively, when charged on
ac grid alone.
Fig. 10 THD (voltage distortions) injected in the grid due to the charging of electric vehicles by
the proposed system
Fig. 11 THD (current distortions) injected in the grid due to the charging of electric vehicles by
the proposed system
5 Economic Benefits
The proposed PV-Grid system for charging of the electric vehicles can provide more
benefits as compared to charging the vehicles on the grid alone system. During the
peak times of the day when there is an abundant amount of solar energy available, the
vehicles are charged on the proposed PV-Grid system in which energy is supplied by
the solar modules. The solar energy has the least negative impact on the environment
as compared to any other type of energy source. PV modules do not create any noise
pollution, which is a major benefit. It does not produce greenhouse gasses and does
not pollute the environment and also has the least maintenance cost.
6 Conclusion
Ever-increasing pollution owing to emissions from ICE vehicles and limited fuel
sources has shown the importance of shifting to an alternate clean and renewable
energy resource. In the proposed work, a PV-Grid system has been developed for the
purpose of charging electric vehicles in MATLAB/Simulink. As the electric vehicles
manufactured by different industries may have different internal resistances so the
effect of different internal resistances on battery performance has been observed for
same SOC and different SOC conditions. The passive filters have been designed in
the proposed PV-GRID system to further enhance harmonic mitigation. The system
performance for various load conditions has been analyzed by employing different
charging stations. The system has been designed in such a way that the load on
the grid is scheduled for on-peak and off-peak hours i.e., during day and night.
This scheduling decreases the stress on the grid that could have been imposed if
the grid had been used for the purpose of electric vehicle charging for the entire
24 h period. It is observed that the proposed system has been successfully able to
suppress the total harmonic distortions (THD) to a great extent i.e. 14.65% to 0.36%
for voltage distortion and from 6.28% to 1.27% for current distortions, respectively.
The widespread use of hybrid renewable energy will not only solve the energy issues
but also ensure a green planet.
Acknowledgements The corresponding author wishes to thank Prof. Vijay Sood of Ontario Tech
University, Canada for his guidance and continuous support during the research work related to the
present paper.
References
1. Morgan JP (2018) Driving into 2025: the future of electric vehicles

2. Bloomberg (2018) Electric vehicle outlook
3. Kumar SA, Bharath A, Sukumar K (2010) The challenges and policy options for integrating
plug-in hybrid electric vehicle into the electric grid. Electr J 23(3):83–91
4. Sears J, Roberts D, Glitman K (2014) A comparison of electric vehicle level 1 and level 2
charging efficiency. In: IEEE conference on technologies for sustainability (SusTech). IEEE,
pp 255–258
5. Castaings W, Lhomme R, Trigui BA (2016) Comparison of energy management strategies of

a battery/supercapacitors system for electric vehicle under real-time constraints. Appl Energy
163:190–200
6. Chaouachi A, Bompard E, Fulli G, Masera M, De Gennaro M, Paffumi E (2016) Assessment
framework for EV and PV synergies in emerging distribution systems. Renew Sustain Energy
Rev 55:719–728
7. Fattori F, Anglani N, Muliere G (2014) Combining photovoltaic energy with electric vehicles,
smart charging and vehicle-to-grid. Sol Energy 110:438–451
8. Guo Y, Xiong J, Xu S, Su W (2016) Two-stage economic operation of microgrid-like electric
vehicle parking deck. IEEE Trans Smart Grid 7(3):1703–1712
9. Rao W, Zhang X, Xie J, Ju L (2015) Optimizing electric vehicle users’ charging behavior in
battery swapping mode. Appl Energy 155:547–559
10. Vander Kam M, Van Sark W (2015) Smart charging of electric vehicles with photovoltaic
power and vehicle-to-grid technology in a microgrid; a case study. Appl Energy 152:20–30
11. Van Roy J, Leemput N, Geth F, Buscher J, Salenbien R, Driesen J (2014) Electric vehicle
charging in an office building microgrid with distributed energy resources. IEEE Trans Sustain
Energy 5(4):1389–1396
12. Chandra Mouli GR, Bauer P, Zeman M (2016) System design for a solar powered electric
vehicle charging station for workplaces. Appl Energy 168:434–443
13. Schuller A, Flath CM, Gottwalt S (2015) Quantifying load flexibility of electric vehicles for
renewable energy integration. Appl Energy 151:335–344
14. Bhatti AR, Salam Z, Aziz MJBA, Yee KP (2016) A comprehensive overview of electric vehicle
charging using renewable energy. Int J Power Electron Drive Syst 7(1):114–123
15. Bhatti AR, Salam Z, Aziz MJBA, Yee KP, Ashique RH (2016) Electric vehicle charging using
photovoltaic: status and Technological review. Renew Sustain Energy Rev 54(1):34–47
16. Khan S, Shariff S, Ahmad A, Alam MS (2018) A comprephensive review of level 2 charging
system for electrical vechicals. J Smart Sci 6(3)
17. Sbordone D, Bertini I, Di Pietra B, Falvo MC, Genovese A, Martirano A (2015) EV fast
charging stations and energy storage technologies: a real implementation in the smart micro
grid paradigm. Electric Power Syst Res 120:96–108
18. Bansal P (2015) Charging of electrical vehicles: technology and policy implications 6(1)
19. Li X, Hui D, Lai X (2013) Battery energy storage station (BESS)-based smoothing control
of photovoltaic (PV) and wind power generation fluctuations. IEEE Trans Sustain Energy
4(2):464–473
20. Mortaz E, Valenzuela J (2017) Microgrid energy scheduling using storage from electric
vehicles. Electric Power Syst Res 143:554–562
21. Bhatti AR, Salam Z (2018) A rule-based energy management scheme for uninterrupted electric
vehicles charging at constant price using photovoltaic-grid system. Renew Energy 125:384–400
22. Bhatti AR, Salam Z, Ashique RH (2016) Electric vehicle charging using photovoltaic based
microgrid for remote islands. Energy Procedia 103:213–218
23. Bhatti AR, Salam Z, Sultana B, Rasheed N, Awan AB, Sultana U, Yuonas M (2019) Optimized
sizing of photovoltaic grid-connected electric vehicle charging system using particle swarm
optimization. Int J Energy Res 43(1):500–522
24. Gómez JC, Morcos MM (2003) Impact of EV battery chargers on the power quality of
distribution systems. IEEE Trans Power Delivery 18(3):975–981
25. Clement-Nyns K, Haesen E, Driesen J (2010) The impact of charging plug-in hybrid electric
vehicles on a residential distribution grid. IEEE Trans Power Syst 25(1):371–380
Development of the System of Initial
Excitation of the Autonomous Induction
Generator
Volodymyr Chenchevoi , Valeriy Kuznetsov , Iurii Zachepa ,

Oleksii Chornyi , Olga Chencheva , Vitaliy Kuznetsov ,
Rostyslav Yatsiuk , and Olha Luhova
1 Introduction
Despite numerous works on theoretical and practical studies of autonomous energy

source (AES) with induction generator (IG) [1–5], a number of technical issues
have not been fully resolved. For example, the factors influencing the conditions
of guaranteed excitation of IG are still unknown. The problem of correct choice
of parameters and correct design of AES determines the relevance of assessing the
impact of changes in generator parameters on the quality of the self-excitation process
in standard and emergency modes of operation of the IG [6, 7].
V. Chenchevoi (B) · I. Zachepa · O. Chornyi

Department of Systems of Automatic Control and Electric Drive Kremenchuk Mykhailo,
Ostrohradskyi National University Kremenchuk, Kremenchuk, Ukraine
V. Kuznetsov
Electric Power Department, Railway Research Institute, Chlopickiego str., 50, Warsaw, Poland
e-mail: vkuznetsov@ikolej.pl
O. Chencheva
Manufacturing Engineering Department, Kremenchuk Mykhailo Ostrohradskyi National
University Kremenchuk, Kremenchuk, Ukraine
V. Kuznetsov
Department of the Electrical Engineering and Electromechanic, National Metallurgical Academy
of Ukraine, Gagarina ave., 4, Dnipro, Ukraine
e-mail: wit1975@i.ua
R. Yatsiuk
Department of Automation and Computer-Integrated Technologies Kremenchuk Mykhailo,
Ostrohradskyi National University Kremenchuk, Kremenchuk, Ukraine
e-mail: RostyslavYatsiuk@protonmail.com
O. Luhova
Department of Electric Machines and Devices Kremenchuk, Mykhailo Ostrohradskyi National
University Kremenchuk, Kremenchuk, Ukraine
548 V. Chenchevoi et al.
The use of a generating power plant requires solving the problem of determining
the deployment time (commissioning)—to take into account the transient processes
of self-excitation when changing the parameters of IG to determine the recovery
time of critical infrastructure using inductive machine (IM) as IG. Also, the use of a
generating power plant requires solving the problem of determining the deployment
time (commissioning) after transportation to the connection point.
In power automated installations with IG, additional means are used to ensure
reliable self-excitation of the generator.
The most commonly used:
– forced short-term flow of electric current through the stator winding of a stationary
machine, by connecting an external power supply, or discharging the battery of
charged capacitors;
– connection of pre-charged capacitors to the stator winding during machine
rotation.
If the rotor is strongly demagnetized by short-circuit currents, for the occur-
rence of asynchronous oscillations, it is necessary to influence the IG. The impact
is possible in three ways:
– to amplify the shock created by the residual induction of the rotor by short-term
supply to the stator winding of a small three-phase or single-phase voltage;
– briefly increase the capacitance of the capacitor, and after setting the asyncronous
mode, reduce the capacitance to the initial value;
– increase the shock from the rotor, i.e. increase the residual induction of the rotor.
In the case where it is necessary to use an external power source (for example,
to power electronic control circuits), this source is also used to ensure reliable self-
excitation of the IG. In wind turbines, capacitors are pre-charged from special micro-
generators with permanent magnets, usually located on the axis of the aerometer.
Such microgenerators charge a bank of exciting capacitors, which is disconnected
from the generator at wind speeds below the lower limit of the wind turbine operating
speeds. It turns on when the turbine reaches the operating speed from the operating
speed limit.
To increase the residual induction of the rotor, some authors suggest placing a
permanent magnet on the stator or on the rotor IG [8–12].
In the absence of additional means to ensure reliable self-excitation of the gener-
ator, the main factor that causes the process and significantly affects its course is
the residual magnetism. In normal operating modes (absence of overloads and short
circuits), the value of the residual magnetism flux sufficient for the generator to re-
excite will always be preserved if the load is switched off first and only then the drive
is switched off. In the case of emergency modes (excitation caused by overload, short
circuit), the residual magnetism decreases many times, and the most unfavorable case
is the excitation of the generator caused by overload. In this case, the generator is
almost completely demagnetized and its re-excitation, with constant speed of rota-
tion of the rotor and the capacity of the excitation capacitors, is impossible. More
acceptable in this regard is a short circuit, after which the flux of residual magnetism
decreases by almost twenty times, but in most cases its value remains sufficient for
Development of the System of Initial Excitation … 549
re-excitation. In this regard, it is promising to create a system that provides reliable

self-excitation of the IG [13–17].
The purpose of the work is to develop a model of IG to assess the impact of
variations in generator parameters on the quality of the self-excitation process in
determining the basic and limit modes of operation and the initial excitation system
at constant generator parameters.
2 Research Method and Result
The speed of the generator self–excitation process depends on the capacitance of the
excitation capacitors, the value of the voltage to which they are charged, as well as
the number of pre–charged capacitors (in one, two or three phases).
To build these dependencies, the self-excitation process was simulated using the
developed mathematical model [18–22]. An IM with a short-circuited rotor was
modeled for the analysis of IG parameters. Main parameters:PAG = 2, 2 kW; p = 2;
Is = 4, 9 ; n = 1430 rpm; cos φ = 0.825; η = 81%. The simulation results are
shown in Figs. 1 and 2.
As a result of the simulation analysis, it is obvious that the rate of increase of the
voltage amplitude in the process of self-excitation of the IG significantly depends
on the number of phases of pre-charged capacitors. The difference in the duration of
the process of self-excitation of the generator for cases where the capacitors of all
phases and only one phase is pre-charged is about 200%.
This allows to limit the power of the additional power supply to the amount
required to pre-charge the capacitor to only one of the phases.
Studies of the IG self-excitation process have shown that the discharge of exci-
tation capacitors precharged from an external power supply (assuming that residual
magnetism is absent and that the rotor speed is constant) provides reliable IG
self-excitation.
There is a point of view that one of the reasons of excitation of IG is residual
magnetization of a rotor of the car. There are different views on the degree of influ-
ence of the residual magnetic flux on self-excitation, but experiments clearly show
that the residual magnetization of the rotor affects the excitation conditions: with a
magnetized rotor it is easier to excite an asynchronous machine than with its complete
demagnetization.
At least two causes leading to demagnetization of the rotor have been experi-
mentally established. First, demagnetization occurs after a three-phase short circuit;
secondly, demagnetization occurs after a complete stop of the asynchronous
generator without prior disconnection of the excitation capacities and the load.
It was also found that after demagnetization of the rotor it is not possible to
excite the IG without changing the previously applied excitation conditions. Further
studies have shown that additional magnetization of the generator is required for
re-excitation. As a result, it was decided to develop a system that would determine
463.064
370.426
277.788
185.15
92.512
U [V]
−0.125
−92.763
−185.401
−278.039
−370.677
−463.315
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
t [s]
a
583.064
466.4261
349.7882
233.1503
116.5124
U [V]
−0.1255
−116.7634
−233.4013
−350.0392
−466.6771
−583.315
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
t [s]
b
Fig. 1 Time diagrams of the generator output voltage at different excitation powers: a C = 100
µF; b C = 150 µF
the level of residual magnetization of hypertension and subsequent magnetization

with an additional power supply.
As one of the possible options for implementing the system, the scheme shown
in Fig. 3. Here, MD—magnetizing device; IG—induction generator; DM—drive
motor; C—capacitor block; Zw —load.
In Fig. 3, capacitor C is connected to the output of the AG generator driven by the
motor DM, in each phase, parallel to the load Zw . With an active load on consumers,
the reactive power coming from the capacitors must be equal to the reactive power
of the generator required to create a magnetic flux. When the generator is operating
250
225
200
175
150
U [V]
125
100
1
75
2
50
25
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t [s]
Fig. 2 Time diagrams of the effective value of the phase voltage for different voltage values of the
pre-charged capacitor: 1—one of the phases; 2—all phases
Fig. 3 Block diagram of the

IG magnetization system
MD
DM IG Zw
on an active-inductive load, the capacitors must compensate for the reactive power
of the machine and the load.
The magnetizing device (MD) must automatically determine the level of residual
magnetization (for example, monitors the residual EMF) and, if necessary, briefly
connect an additional power supply (battery or capacitor) to the stator windings of the
generator. In the case when the level of residual EMF is sufficient, the self-excitation
of the generator will occur in the normal mode.
The source of primary energy can be both a direct current source and a pre-
charged capacitor. In addition to the presented types of excitation, the source of
primary energy can be a solar battery.
Fig. 4 AXIOMA Energy

AX-30P solar module
The advantage of this method of initial excitation of an IG is.

• availability and inexhaustibility of solar radiation as a source of energy
• cost-effectiveness of the process of using photocells
• minimum level of scheduled maintenance and high reliability.
Figure 4 shows the appearance of the solar module of the AXIOMA Energy
AX-30P series with a capacity of 30 kW and an output voltage of 18.3 V [23, 24].
Also Fig. 5 shows the I–V characteristics of the solar module and the dependence
of the scattering power in the load on the voltage of the module for the level of
illumination of the solar cell 100 W/m2 .
The system of initial excitation (Fig. 6) based on application of photoelectric
elements is offered. Special converter equipment is required to match the operation
of the solar power plant and the generator voltage. As such equipment the active
controlled rectifier (ACR) which initial characteristics correspond to the modern
standards regulating quality of the electric power is most often used.
The active controlled rectifier is a three-phase bridge on power IGBT transistors
with reverse diodes, the input of which is supplied by a solar battery to a DC-DC to
DC DC-DC converter. A capacitor of small capacity C1 is connected in parallel with
the output of the solar battery. The output clamps of the ACR are connected to the
output of the IG. The input circuit of the rectifier includes a capacitor C2, the voltage
at which must be higher than the amplitude of the linear voltage of the generator.
The energy generated by photocells is enough for the initial impulse of excitation of
IG.
Figure 7 shows a model in the software environment Matlab, which implements
the initial excitation of the IG. In this model, the generator is based on a three-phase
model of an inductive machine [25–28].
Fig. 5 I–V characteristics of

the solar module
IM A
B
C Load
Solar module
ω Сexc
С1 - С2 -
ACR DC/DC
+ +
Fig. 6 Autonomous power source based on an IG with an initial excitation system based on
photocells
Fig. 7 Simulink model for realization of initial excitation of induction generator by means of the
solar module
Fig. 8 Time diagram of the output voltage of the induction generator
Figure 8 presents the results of modeling the system of initial excitation of IG on

the basis of photocells. Figure 9 shows the dependence of IG excitation time on the
level of illumination.
As can be seen from Fig. 9 as the degree of illumination increases, the self-
excitation time decreases.
The obtained result of calculating the value of the illumination level for the emer-
gence of capacitor self-excitation can be used to estimate the parameters and require-
ments for the initial excitation systems capable of providing the required excitation
time and conditions.
3 Conclusion
Reliable self-excitation of the generator, regardless of its magnetic state, provides a

pre-charge from an external power supply of one of the phase capacitors of the discon-
nected battery of excitation capacitors and then connecting the whole battery to the
stator winding with a rotating rotor. The intensity of self-excitation also depends on
the number of pre-charged capacitors, as well as the value of the voltage to which they
Fig. 9 Dependence of induction generator excitation time on the level of illumination
were charged. The speed of the self-excitation process only largely depends on the
connection method and the polarity of the pre-charged excitation capacitors. Satis-
factory results are provided by a pre-charge of a single-phase capacitor. This limits
the power of the external power supply that is required to pre-charge the capacitors.
The system of initial excitation based on application of photoelectric elements that
allows to provide reliable self-excitation of IG in case of demagnetization of a rotor
is offered. The system allows to provide additional magnetization of the generator
after demagnetization of the rotor, without changing the previously applied excitation
conditions.
References
1. Vieira RJ, Sliaraf AM, Elgammal A (2011) Design of self-excited induction generators for
wind applications. CCECE, IEEE
2. Bansal RC (2005) Three phase self excited induction generators: an overview. IEEE Trans
Energy Convers 20(2)
3. Haque MH (2009) A novel method of evaluating performance characteristics of a self-excited
induction generator. IEEE Trans Energy Convers 24(2)
4. Malik NH, Al-Bahrani AH (1990) Influence of the terminal capacitor on the performance
characteristics of a self excited induction generator. IEE Proceedings 137
5. Murthy SS, Malik OP, Tandon AK (1982) Analysis of self-excited induction generators. IEE
Proc 129
6. Toroptsev ND (2004) Asinkhronnyye generatory dlya avtonomnykh energeticheskikh
ustanovok [Asynchronous generators for autonomous power installations]. NTF Energopro-
gres, Moskow, 87 p (in Russian)
7. Kitsis SI (2003) Self-excited asynchronous generator. Energoatomizdat, Moskow, 328 p
8. Diao T (2015) Study on dual-rotor permanent magnet induction motor and performance. Open
Electr Electron Eng J 9:584–590. https://doi.org/10.2174/1874129001509010584
9. Qu R, Lipo TA (2003) Dual-rotor, radial wound, permanent-magnet machines. IEEE Trans Ind
Appl 39(6):1665–1673
10. Xu L, Zhang Y (2006) Design and evaluation of a dual mechanical port machine and system.
In: IEEE 5th international power electronics and motion control conference, Shanghai, China
11. Qu R, Aydin M, Lipo TA (2003) Performance comparison of dual-rotor radial-flux and axial-
flux permanent magnet BLDC machines. In: Proceedings of IEEE electric machines and drives
conference, Madison, US, pp 1948–1954
12. Roy B (2017) A novel double rotor without stator electric motor: theoretical and functional
aspects. J Electr Eng Sci 3(2):1–10. https://doi.org/10.18831/djeee.org/2017021001
13. Nazir R, Syafii AP, Akbar F, Arfan Y (2020) Effect analysis of residual magnetism availability
level on the success of voltage generation processes in self-excited induction generators. Int J
Power Electron Drive Syst (IJPEDS) 11(3):1211–1219. https://doi.org/10.11591/ijpeds.v11.i3
14. Bansal RC (2005) Three-phase self-excited induction generators: an overview. IEEE Trans
Energy Convers 20(2)
15. Nazir R, Syafii AP, Akbar F, Dorinza A (2019) Differences in the impact of harmonic distortion
due to the installation of electronic load controller in self-excited induction generator and
synchronous generator. Int J Power Electron Drive Syst (IJPEDS) 10(1):104–116
16. Kadhum HH et al (2020) The influence of iron losses on selecting the minimum excitation
capacitance for self-excited induction generator (SEIG) with wind turbine. Indonesian J Electr
Eng Comput Sci 19(1):11–22
17. Zagirnyak M, Rod’kin D, Romashykhin I, Romashykhina Z, Nikolenko A (2017) Refined calcu-
lation of induction motor equivalent circuit nonlinear parameters by an energy method. Eastern-
Eur J Enterprise Technol 3(87):4–10. https://doi.org/10.15587/1729-4061.2017.104146
18. I. Romashykhin NR, Kuznetsov V (2017) The possibilities of the energy method for identifying
the parameters of induction motor. In: 2017 international conference on modern electrical and
energy systems (MEES), Kremenchuk, Ukraine, pp 128–131. https://doi.org/10.1109/MEES.
2017.8248869
19. Zagirnyak M, Zachepa Y, Chenchevoi V (2014) Estimation of induction generator over-
load capacity under connected direct current consumers. Acta Technica, Praga: Institute of
Thermomechanics AS CR, 2014, vol. 59, № 2, pp 149–169, ISSN 0001–7043
20. Zagirnyak M, Prus V, Rodkin D, Zachepa I (2019) A refined method for the calculation of steel
losses at alternating current. Arch Electr Eng 68(2):295–308
21. Zagirnyak M, Ogar V, Chenchevoi V (2014) Analysis of induction motors features taking into
account change of iron properties. Acta Tech 59(1):25–47
22. Chenchevoi V, Zachepa I, Chencheva O, Yatsiuk R (2020) Parameters of guaranteed self-
excitation of an induction generator for autonomous electric power sources. In: 2020 IEEE
problems of automated electrodrive. Theory and Practice (PAEP) 21–25 Sept. 2020
23. Available at: https://solar-tech.com.ua/solar-electricity/solar-panels/solnechnaya-batareya-axi
oma-energy-ax-30p.html
24. Available at: https://alfa.solar/ru/solnechnaya-batareya-axioma-ax-10p-id734.html
25. Perdulak J, Kovac D, Kovacova I, Ocilka M, Gladyr A, Mamchur D, Zachepa I, Vince T,
Molnar J (2013) Effective utilization of photovoltaic energy using multiphase boost converter
in comparison with single phase boost converter. Communications 15(3):32–38
26. Yevheniia K, Vitaliy K, Mykola T, Alisa K, Maksym T, Mykola B (2019) Development and
verification of dynamic electromagnetic model of asynchronous motor operating in terms of
poor-quality electric power. In: 2019 IEEE international conference on modern electrical and
energy systems (MEES), Kremenchuk, Ukraine, pp 350–353. https://doi.org/10.1109/MEES.
2019.8896598
27. Kuzenkov O, Kuznetsov V, Tryputen N (2019) Analysis of phase trajectories of the third - order
dynamic objects. In: 2019 IEEE 2nd Ukraine conference on electrical and computer engineering
(UKRCON), Lviv, Ukraine, pp 1235–1243. https://doi.org/10.1109/UKRCON.2019.8879819
28. Krishnan R, Krishnan R (2010) Electric motor drives—modeling analysis and control [text].
PHI Learning Private Limited, New Delhi, pp 626
Single Phase Five Level Step-Up
Switched Capacitor Based Grid
Connected Inverter with LCL Filter
1 Introduction
The concept of Reduced Device Count (RDC) Multilevel Inverters (MLIs) is very
popular in Renewable energy Source applications, and these produce stair case wave-
form with imitation of sinusoidal wave [1]. Conventionally, three types of MLIs
have been reported viz., Neutral Point Clamped (NPC), Flying Capacitor (FC), and
Cascaded H-Bridge (CHB). Generally for any MLI, the number of switch count,
input dc sources, gate drivers, discrete diodes, capacitors, and Total Blocking Voltage
(TBV/TSV) are important to decide the size of the inverter and requirement of filter
size in grid connection. For generating five level output, the NPC inverter takes
more number of diodes, more number of capacitors, and voltage gain becomes unity.
Additionally, to balance the capacitors, separate auxiliary circuits or complex control
schemes are required [2]. The number of requirement are more as compared with
NPC and in this for better operation, the capacitors must be pre-charged [3]. As
compared with NPC and FC, CHB MLI has higher modularity but requires more
number of isolated dc sources [4]. The TBV is more in [5] for generating five level
output with single dc source. Generally in most of MLIs, less TSV has been preferred
for reducing the cost. The topology [6] provides five level output with single dc source
and single capacitor but TBV becomes more. For generating five level output with
single dc source, it takes TSV of 11Vdc [7]. The topologies [8, 11] produces TBV
of 12Vdc with the same number of component count for generating five level output
and separate backend H-Bridge is required in [11] for generating negative voltage
levels. The topology [9] requires two dc sources with TSV of 10Vdc for generating
five level output but drawback of this topology is unity voltage gain that means
there is no boosting ability. For generating five level output, the topologies [10, 13]
requires two capacitors and single dc source. Here, both topologies gives TSV of
14Vdc. The topology [12] gives less TBV but requires two dc sources, and there is

Electrical Engineering Department, NIT Warangal, Telangana, India
no boosting ability. To alleviate aforementioned problems, a new five level output

inverter topology and its cascade connection have been introduced in this paper and
it provides self-balancing of capacitors and boosting ability with single dc source.
Conventionally, different types of modulation techniques [15, 16] are available and
in this paper a new URCS has been reported in RL load case by taking reference of
Nearest Level Control (NLC) technique. For higher ripple attenuation purpose, the
LCL filter has been integrated with grid [14].
The rest of the paper work is organized as follows: The operation and control
schemes of proposed topology have been presented in Sect. 2. In Sect. 3, the compar-
ative study has been described. The Sect. 4 gives the simulation results for both RL
load and grid connected cases. Finally, the conclusion has been presented in Sect. 5.
A. Operation of Proposed Inverter Topology
The schematic of single stage PV to grid connected system, RDC of five level
inverter topology with LCL filter, and cascade connection of two SC units with
1:5 source configuration have been depicted in Figs. 1, 2, and 3 respectively. In this
paper, the dotted portion in Fig. 1 has been considered for grid connection. The
Fig. 1 Single stage PV to VPV IPV

grid connected system Iinv Igrid
Switched
+ Filter
Capacitor Vgrid
PV
MLI
Panel
VPV
To Inverter
+ -
MPPT
IPV algorithm Control
Fig. 2 Proposed five level D1

RSC inverter topology with
LCL filter + C1
Vdc1 S3
S1 S5
S4 + C2
A B
S2 D2 S6
Iinv
LCL Filter
Igrid Vgrid
Single Phase Five Level Step-Up Switched Capacitor … 559
D1
+ C1
Vdc1 S3
S1 S5
S4 + C2
A
S2 D2 S6
LOAD
D3
B
+ C3
Vdc2 S9
S7 S11
S10 + C4
S8 D4 S12
Fig. 3 Cascade connection of two SC units with 1:5. Source configuration 25 level O/P
Table 1 Switching arrangement-five level SC inverter

S. no ON state switches C1 C2 Level
1 S1 − S4 − S6 C D +2VDC
2 S1 − S3 − S6 – C +1VDC
3 S2 − S3 − S6 – C 0
4 S2 − S4 − S5 C – −1VDC
5 S2 − S3 − S5 D C −2VDC
five level inverter topology comprises six unidirectional switches, eight diodes, two
switched capacitors (SCs), and one dc source. The operation of five level topology
and 25 level topology have been depicted in Tables 1 and 2 respectively. Gener-
ally the capacitors are charged in parallel and discharged in series then finally with
proper switching action, the capacitors have been self-balanced. From Table 1, it
is concluded that this topology provides voltage boosting of two with respect to
single stiff dc source. Generally, the voltage stress of each switch is indexed with
TBV/TSV. In this topology, the per unit value of TBV becomes four. Based on
maximum discharging period (MDP) method, it is possible to get proper switched
capacitance value within permissible limits of voltage ripple. However, in this paper
little more capacitance value has been considered for making successful operation
of grid. As compared with L, LC filters, the LCL filter provides higher ripple atten-
uation but it suffers from resonance problem and finally system becomes unstable.
To alleviate resonance problem, active damping method has been preferred and in
this method there is no additional power losses as compared with passive damping
method. The PLECS/ PSIM software are very helpful for finding switching, conduc-
tion losses of any MLI topology. Lastly, the advantages of proposed topology is that
it has less TBV, boosting ability, and self-balanced capacitor with single dc source
supply.
Table 2 Switching arrangement-25 level SC inverter

S1 to S6 C1 C2 S7 to S12 C1 C2 Level
S1 − S4 − S6 C D S7 − S10 − S12 C D 12VDC
S1 − S3 − S6 – C S7 − S10 − S12 C D 11VDC
S2 − S3 − S6 – C S7 − S10 − S12 C D 10VDC
S2 − S4 − S5 C – S7 − S10 − S12 C D 9VDC
S2 − S3 − S5 D C S7 − S10 − S12 C D 8VDC
S1 − S4 − S6 C D S7 − S9 − S12 – C 7VDC
S1 − S3 − S6 – C S7 − S9 − S12 – C 6VDC
S2 − S3 − S6 – C S7 − S9 − S12 – C 5VDC
S2 − S4 − S5 C – S7 − S9 − S12 – C 4VDC
S2 − S3 − S5 D C S7 − S9 − S12 – C 3VDC
S1 − S4 − S6 C D S8 − S9 − S12 – C 2VDC
S1 − S3 − S6 – C S8 − S9 − S12 – C VDC
S2 − S3 − S6 – C S8 − S9 − S12 – C 0
B. Proposed Control Schemes
(i) URCS-RL load case
The new URCS has been applied to the 25 level inverter topology with RL load,
and its strategy has been depicted in Fig. 4. The proposed control scheme is the
modified version of NLC technique, and it is also called as Unified Rounding Control
Scheme (URCS). The URCS has provides less computation time as compared with
conventional NLC and SHE PWM techniques.
In this scheme, modulating signal is given to rounding function and it generates
aggregated signal (AS). The Switching Table Generation (STG) is taken from Table
2. Because of truncation, there is no need of logic gate requirement for generating gate
pulses (GP). This control scheme is basically fundamental switching frequency tech-
nique there by switching losses greatly reduces, and it also provides self-balancing
of capacitors. It is universally applicable for any MLI topology with valid switching
table.
(ii) dq frame current control strategy-Grid Connected Case
Fig. 4 Unified rounding control scheme-URCS

Fig. 5 Proposed dq frame current control scheme
The control strategy for five level SC inverter in grid connected case has been
depicted in Fig. 5. Firstly, sense the grid voltage (let, Vgrid = Vβ ) and through proper
PLL, the Vd and Vq components have been generated. Here, PLL has been used to
generate reference variable for grid synchronization. Similarly, the Id and Iq compo-
nents have been generated by sensing inverter current (let, Iinv = Iβ ). To inject active
and reactive power into the grid, Id(ref) and Iq(ref) have been used. From decoupling
control strategy, the modulating β-component signal has been generated. By using
Proper Gain (PG ) value, it is possible to generate desired modulating signal. By
taking absolute function, it is possible to curtail the number of high frequency carriers
present in Repeating Table Generation (RTG) block. After truncation of ASG and
STG, the required switching pulses/gate pulses (GP) have been generated. By taking
Iq(ref) is equal to zero then it is possible to get UPF operation at grid side.
3 Comparative Study
Table 3 shows the comparative analysis with different conventional five level inverter
topologies. From this table, it is concluded that proposed topology gives less
TBV with single dc source supply there by the cost requirement of switches can
be decreased. Generally, switch requirement reduces means respective protection
circuit, gate driver circuit, and heat sink also reduces there by size of inverter can be
decreases.
Table 3 Comparative analysis with different conventional five level inverter topologies
S. no Description [5] [6] [7] [8] [9] [10] [11] [12] [13] P
1 N Levels 5 5 5 5 5 5 5 5 5 5
2 N Switches 12 6 6 8 6 8 8 8 10 6
3 N Drivers 12 6 6 8 5 8 8 8 9 6
4 N Diodes 12 8 7 8 6 10 8 10 10 8
5 N Sources 1 1 1 2 2 1 2 2 1 1
6 N Capacitors 2 1 1 0 0 2 0 0 2 2
7 TSV 20 12 11 12 10 14 12 10 14 8
A. RL Load Case
Table 4 shows simulation parameters-1:5 source configuration. The URCS has been
applied to the proposed cascaded topology with step change (at t = 0.05 s) in Modu-
lation Index (MI) values 1 and 0.7. Figure 6 shows simulation results for 1:5 source
configuration inverter with RL load. From this, the capacitors C1 & C2 are self-
balanced at 50 V and C3 & C4 are self-balanced at 250 V even if step changes in
MI values. With MI = 1 and 0.7, the inverter generates 25 level and 17 level outputs
respectively. Figure 7 shows harmonic spectrum of 25 level inverter output voltage
with MI = 1.0, and it achieves peak value of 596.6 V at 50Hz.
B. Grid Connected Case
Table 5 shows simulation parameters-grid connected case. The LCL filter values
have been calculated from [14]. Figure 8 shows simulation results for step changes
in Id(ref) from 20 to 10A: UPF operation mode. From Fig. 8, the capacitors C1 & C2
are self-balanced at 200 V and generates five level output then grid voltage and grid
current are maintains UPF operation even step changes in Id(ref). By the proposed
control strategy of Fig. 5, the grid current exactly reaches the peak value according
to Id(ref). Generally, Id(ref) is used to inject active power into the grid. Figures 9 and 10
show harmonic spectrums of inverter output voltage and grid current with respect to
Id(ref) = 20A.
Table 4 Simulation
parameters-1:5 source
configuration 1 Vdc1, Vdc2 50 V, 250 V
2 Voltage gain 2
4 R 100
5 L 30 mH
6 C1 = C2 3400 μF
Fig. 6 Simulation results 1:5 source configuration inverter with RL loadL URCS
Fig. 7 Harmonic spectrum

of 25-level inverter output:
MI = 1.0
Table 5 Simulation
parameters-grid connected
case 1 Grid voltage 230 V (RMS)
2 Switching frequency 10 kHz
5 Rated power(S) 2 kVA
6 Vdc 200 V
7 Inverter side inductance 3.98 mH
8 Grid side inductance 4 mH
9 Filter capacitance 5.95 μF
10 Switched capacitors 4700 μF
Fig. 8 Step changes in Id(ref). : UPF operation mode
Fig. 9 Harmonic spectrum-five level output voltage
5 Conclusion
The proposed five level SC inverter topology provides four per unit TSV and voltage
boosting ability of two with less number of switch count. With cascade connection
of two SC units provides 25 level output and it has been elucidated with URCS. In
this, the capacitors are self-balanced without need of auxiliary circuits and separate
complex control schemes. A comparative study is also done with other conventional
five level inverter topologies. With RL load and grid connected cases, the proposed
control schemes gives good results in the output. Even if step changes in direct
axis reference current values, the grid voltage and grid current still maintains UPF
Fig. 10 Harmonic spectrum-grid current with Id(ref) = 20A
operation. The drawback of this topology is that it has non-uniform blocking voltage
across the switches. This topology is well suitable for photovoltaic and fuel cell based
grid connected applications.
References
1. Rodriguez J, Lai J-S, Peng FZ (2002)Multilevel inverters: a survey of topologies, controls,

and applications. IEEE Trans Indus Electron 49(4):724–738. https://doi.org/10.1109/TIE.2002.
801052
2. Nabae A, Takahashi I, Akagi H (1981) A new neutral-point-clamped PWM inverter. In: IEEE
Trans Indus Appl IA-17(5):518–523. https://doi.org/10.1109/TIA.1981.4503992
3. Shukla A, Ghosh A, Joshi A (2007)Capacitor voltage balancing schemes in flying capacitor
multilevel inverters. In: 2007 IEEE power electronics specialists conference, Orlando, FL,
USA, pp 2367–2372. https://doi.org/10.1109/PESC.2007.4342381
4. Peng FZ, Lai J-S, McKeever J, VanCoevering J (1995) A multilevel voltage-source inverter
with separate DC sources for static VAr generation. In: IAS ‘95. Conference record of the 1995
IEEE industry applications conference thirtieth IAS annual meeting, vol 3, Orlando, FL, USA,
pp 2541–2548. https://doi.org/10.1109/IAS.1995.530626
5. Cheng C, He L (2016)Flying-capacitor-clamped five-level inverter based on switched-capacitor
topology. In: 2016 IEEE energy conversion congress and exposition (ECCE), Milwaukee, WI,
USA, pp 1–5. https://doi.org/10.1109/ECCE.2016.785512
6. Ye Y, Cheng KWE, Liu J, Ding K (2014) A step-up switched-capacitor multilevel inverter with
self-voltage balancing. IEEE Trans Indus Electron 61(12):6672–6680. https://doi.org/10.1109/
TIE.2014.2314052
7. Barzegarkhoo R, Kojabadi HM, Zamiry E, Vosoughi N, Chang L (2016) Generalized structure
for a single phase switched-capacitor multilevel inverter using a new multiple DC link producer
org/10.1109/TPEL.2015.2492555
8. Ruiz-Caballero DA, Ramos-Astudillo RM, Mussa SA, Heldwein ML (2010) Symmetrical
hybrid multilevel DC–AC converters with reduced number of insulated DC supplies. IEEE
Trans Indus Electron 57(7):2307–2314. https://doi.org/10.1109/TIE.2009.2036636
9. Hari Priya V, Sreenivasarao D, Siva Kumar G (2017) Improved pulse-width modulation scheme
for T-type multilevel inverter. IET Power Electron 10(8):968–976. https://doi.org/10.1049/iet-
pel.2016.0729
10. Niu D et al. (2019) A novel switched-capacitor five-level T-type inverter. In: 2019 2nd interna-
tional conference on smart grid and renewable energy (SGRE), Doha, Qatar, pp 1–6. https://
doi.org/10.1109/SGRE46976.2019.9021066
11. Su GJ (2005) Multilevel DC-link inverter. IEEE Trans Indus Appl 41(3):848–854. https://doi.
org/10.1109/TIA.2005.847306
12. Sivakumar K (2015) A fault-tolerant single-phase five-level inverter for grid-independent PV
systems. IEEE Trans Indus Electron 62(12):7569–7577. https://doi.org/10.1109/TIE.2015.245
5523
13. Ranjan A, Gupta KK, Kumar L, Jain S (2012)A switched-capacitors based multilevel boost
inverter with single input source. In: 2012 IEEE 5th India international conference on power
electronics (IICPE), Delhi, India, pp 1–6. https://doi.org/10.1109/IICPE.2012.6450373
14. Ali A, Shanmugham P, Somkun S (2017)Single-phase grid-connected voltage source converter
(iEECON), Pattaya, Thailand, pp 1–6. https://doi.org/10.1109/IEECON.2017.8075720
0861.2021.9404491
Single Stage Multiple Source
Bidirectional Converter for Electric
Vehicles
Siddhant Gudhe and Sanjeev Singh
1 Introduction
With the advancement in power electronics the dependence on the non-renewable

energy sources is decreasing exponentially and electric vehicles (EVs) are found
to be excellent choice for utilizing the renewable energy sources. Power electronic
switches are used to operate and control the EVs. Active research work is being
carried out for reconfiguration of the EVs [1]. Pure EV has only the battery as active
source while running therefore EVs with the two battery pack have been presented in
[1] so that auxiliary loads of the EVs such as AC, lights, horns, and power windows
may be operated separately. One battery is used to run traction motor, and other
battery is used to supply auxiliary loads. Combination of both the batteries can also
be used to supply the load during the high power requirement by the traction motor.
One battery replacement by the ultra-capacitor has been reported by [2]. Battery
supports high energy requirement and the Ultra-capacitor supports the instantaneous
power requirement of the motor. Bidirectional power flow is also possible in Multi-
Source Inverter (MSI) during regenerative operation stage of EV. The MSI for Hybrid
Electric Vehicles has been presented in [3, 4]. The battery is charged through ICE
using generator-rectifier and from main EV motor during regenerative mode.
The concept of using induction motor winding for charging of the EV batteries
has been presented in [5] which use single phase supply as shown in Fig. 1 through
the interleaved boost converter. Stator windings act as filter inductors. The single
phase is supplied through all the windings of the stator and the neutral point created
at the midpoint of the capacitors.
For boosting the voltage to charge Li-ion batteries, DC/DC converter has been
used. The circuit has the issues variable neutral voltage and the use of DC/DC
S. Gudhe (B) · S. Singh

S. Singh
e-mail: sschauhan@manit.ac.in
568 S. Gudhe and S. Singh
VSI
T1 T3 T5 C1
Bi-directional Li-ion
DC/DC Converter Battery Pack
House-hold C2
Outlet T4 T6 T2
Electric
Motor
Interleaved Boost Converter (IBC)
Fig. 1 Integrated charger using the stator winding of the traction motor [5]
Fig. 2 MSBC connected to Q1 Q2 Q3

stator windings and battery
P1
connections
L1
P2
R L2 Battery 1 +
Y _
L3
B
P3
+
_
Battery 2
T1 T2 T3
Q1 Q2 Q3
P1
L1
Single P2
Phase L2 Battery 1
+
AC L3 _
P3
+
_
Battery 2
T1 T2 T3
Fig. 3 Charging of batteries with single phase supply
converter. To tackle the issue of additional DC/DC converter, the inverter topology
named as Multi source inverter (MSI), patented in 2014 [6], may be used.
The Multi source Inverter (MSI) gives different voltage levels using different
combination of switches. The use of the battery for the reactive power compensation
is also obtained using one of the two batteries [7, 8]. Boost inverter topology is used
in single stage for both boosting and inverting which minimize the ripple components
Single Stage Multiple Source Bidirectional Converter … 569
which reduces the battery life [9]. But the requirement of the reactive power changes
with the weather conditions. The requirement of reactive power in summer is more
as compared to winter [10].
Integrated charging traction inverter supplied from single phase supply has been
patented in [11] but has not been researched extensively. This paper discusses
the charging of the multiples batteries using multi source bidirectional converter
(MSBC). The basic concept is that this topology connects different DC sources to
the single AC supply with not more than single stage of conversion. SO there is no
requirement for the DC/DC converter. During traction mode MSBC can act as three
level inverter and supply the appropriate voltage to the electrical motor. This power is
supplied with main battery which is called as primary battery. Other loads of EV such
as electronics loads and fans can be supplied with secondary battery. The voltage
rating and size of the secondary battery is almost half compared to main battery.
This reduces the converter complications and its control algorithm. Level 1 charger
has been used for charging through IM. The simulation to verify the single phase
charging of MSBC in MATLAB/Simulink environment has also been presented.
The paper is organized as follows. Operation principle of MSBC, Performance
Simulation of proposed MSBC topology, and conclusion are described in the Sects. 2–
4.
2 Operation Principle of Multi-Source Bidirectional

Converter
Figure 2 shows the diagram for MSBC with battery connection. RYB phase is
connected to the stator of the induction motor. Source inductance is represented
by L1 , L2 , and L3 . Two batteries are connected in particular configuration. Battery 1
is main battery which is used for supply to induction motor during motoring opera-
tion. Battery 2 is auxiliary battery to supply other applications of the EVs. Battery
1 is selected with the voltage rating approximately twice as compared to battery 1
voltage rating. Q1 , Q2 , and Q3 are the IGBT switches along with the anti-parallel
diode at the upper leg. T1 , T2 , and T3 form the lower leg switches. Similarly switch
P1 , P2 , andP3 forms middle leg.
During charging mode, single phase supply is supplied through any two stator
windings L1 , L2 , and L3 . Figure 3 shows the complete charging of both the batteries
along with all the switches. Single phase power is supplied through L1 and L2 wind-
ings. Battery 1 will charge through upper and lower leg switches. Battery 2 will
charge through middle leg and lower leg switches. Lower leg switches T1, T2, and
T3 forms the common negative path for both the batteries. During discharging mode,
winding connection is rearranged through to make stator windings in star connection.
The stator windings now will produce rotating magnetic field for the induction motor.
During positive half cycle, battery 1 will charge through Q1 -T2 . Battery 2 charges
through P1 -T2 while Q3 , T3 , P3 will be always off. Similarly during negative half
cycle, Q2 -T1 charges battery 1 and P2 -T1 charges battery 2. Voltage across battery 1
is V1, and Voltage across battery 2 is V2.
Figures 4 and 5 show the voltage waveforms during discharging of battery 1 and
battery 2. Battery 1 will discharge through Q1 , Q2 , and Q3 , while T1 , T2 , T3 . P1 , P2 ,
and P3 remains open. Battery 2 will discharge through P1 , P2 , and P3 , while T1 , T2 ,
T3 . Q1 , Q2 , and Q3 remains open. The discussion for the waveforms for the battery
1 discharging is completed while Fig. 1 holds true circuit while discharging with
single battery 1 connected in circuit.
Let Q1 -T2 -T3 are conducting initially, line voltage VRY will be positive, phase Y
and B will be short circuited, so the line voltage will be zero and line voltage VBR will
have negative voltage. After 60 degrees, T2 is switched off and the complementary
switch Q2 of the same leg is turned on. So the switches conducting are Q1 -T2 -T3 for
next 60 degrees. Now the Phase R and Y will be short circuited, line voltage VYB
will be positive and line voltage VBR will be negative.
Then after again 60 degrees, Q1 is switched off and complementary switch T1
will be turned on. Similar switching pattern will be followed, and output line voltage
as shown in Fig. 4 is obtained. It should be noted that each switch conducts for the
period of 180 degrees and turned on at 60 degree interval. Similar switching pattern
is obtained for battery 2 but the voltage is called V2, along with the middle and lower
leg switches.
When both the batteries are discharging simultaneously, the waveforms and
switching pattern obtained is shown in Fig. 5. For this operation, all the lower leg
switches T1 , T2 , and T3 shall remain open. The switching sequence is shown in
waveforms. The maximum voltage obtained when both the batteries are operating
simultaneously will be difference of both the voltage (V1-V2). Applying similar
switching pattern logic, the output line voltage waveforms can be obtained. Obvi-
ously, the maximum voltage obtained during this mode will be less than the required
Fig. 4 Voltage waveforms VRY
during discharging of +V1

Battery 1
0
-V1
VYB
+V1
-V1
VBR
+V1
0 60 120 180 240 300 360 60 120 180 240 300 360
-V1
Q1- Q1- T1- T1- T1- Q1- Q1- Q1- T1- T1- T1- Q1-
T2- Q2- Q2- Q2- T2- T2- T2- Q2- Q2- Q2- T2- T2-
T3 T3 T3 Q3 Q3 Q3 T3 T3 T3 Q3 Q3 Q3
Fig. 5 Voltage waveforms VRY
during discharging of +(V1-V2)
Battery 1 and Battery 2

0
simultaneously
-(V1-V2)
VYB
+(V1-V2)
-(V1-V2)
VBR
+(V1-V2)
0 60 120 180 240 300 360 60 120 180 240 300 360
-(V1-V2)
Q1- Q1- P1- P1- P1- Q1- Q1- Q1- P1- P1- P1- Q1-
P2- Q2- Q2- Q2- P2- P2- P2- Q2- Q2- Q2- P2- P2-
P3 P3 P3 Q3 Q3 Q3 P3 P3 P3 Q3 Q3 Q3
by induction motor. So, for EV application, only battery 1 is specially serving the
voltage required by induction motor and battery 2 is used for other application.
The similar charging and discharging principle as presented above is applied
this MSBC topology. The charging is provided through any two of the three phase
winding L1, L2, and L3. Battery 1 is discharged through the MSBC to create the
output voltage similar to the three-level inverter. Figure 6 presents the presented 3
level MSBC for running of the EV traction motor. The stator windings are connected
in star to create rotating magnetic field in the induction motor.
Q11 D11 Q12 D12 Q13 D13
Q21 D21 Q22 D22 Q23 D23

L1
D10 D30 D50
R L2 Battery 1
Y +
L3 _
B
Q31 D31 Q32 D32 Q33 D33
+
D20 D40 D60 _
Battery 2
Q41 D41 Q42 D42 Q43 D43
Fig. 6 Discharge of Battery 1 and Battery 2 through MSBC

State of Charge for Battery 1 State of Charge for Battery 2

50.002 50.5
State of Charge
50.001
50
50
49.999 49.5
49.998
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
50.0006 50.5
State of Charge
50.0004 50
50.0002
49.5
50
0.4 0.45 0.5 0.55 0.6 0.4 0.45 0.5 0.55 0.6
Time (seconds) Time (seconds)
Fig. 7 State of charge during charging and discharging
Charging and Discharging Currents

40
20
Amperes
-20
-40
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
40
R Phase
Y Phase
20
B Phase
Amperes
-20
-40
0.4 0.42 0.44 0.46 0.48 0.5 0.52 0.54 0.56 0.58 0.6
Time (seconds)
Fig. 8 Current during charging and discharging

3 Performance Simulation of Proposed MSBC Topology
The three-level MSBC is modeled in the MATLAB/Simulink environment. The three

phase stator winding are created using RL load with the value of R = 0.435 and L
= 4 mH which are the real life values of the 33 Hp, 400 V induction motor. During the
battery discharging mode, the load is connected in star to create the rotating magnetic
field for the induction motor. Main battery is called as battery 1, and auxiliary battery
is called as battery 2. Both the batteries are charged up to 0.5 s and then suddenly
discharged till 1 s. Figure 7 shows the simulated performance while charging and
discharging state of charge (SOC) of both the batteries. Battery 2 is charging at the
slower rate as compared to the battery 1. Battery 1 is discharged through RL load,
and battery 2 is discharged to small DC motor.
The current during charging and discharging is shown in Fig. 8. During charging
the third leg current is zero because only two legs are used for charging of both the
batteries while MSBC acts as a rectifier. During discharging, all the three legs of
MSBC, RL load is connected in star configuration. It is also important to observe
voltage across and current through the switches. During charging the current through
the switches Q13 , Q23 , Q33 , and Q43 is zero. Rest all the switches show the current
flowing through it. For the verification, the current through switch Q11 and voltage
across switch Q12 waveform for one cycle is shown in Fig. 9. It is important to note
that the voltage across and current through the switches does not exceed the rated
value. The voltage across the diode shows the similar results. These diodes are called
Current Through Switch Voltage Across Switch

5 300
<Switch current>
<Switch voltage>
200
0
100
-5
0
-10 -100
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
5 300
<Switch current>
<Switch voltage>
200
0
100
-5
0
-10 -100
0.2 0.205 0.21 0.215 0.22 0.2 0.205 0.21 0.215 0.22
Time (seconds) Time (seconds)
Fig. 9 Current through switch Q11 and voltage across switch Q12
as clamping diode, and the above topology is known as Neutral Point Clamped (NPC)
topology of the multi-level inverter. Space vector modulation control technique is
used during discharging of the batteries, i.e., operation of induction motor for EV.
4 Conclusion
This paper has presented a novel charging method for the two batteries using MSBC.
Battery 1 is used to run the traction (induction) motor, and battery 2 has been
connected to run a DC motor as an auxiliary load. The charged batteries are used to
run induction motor through this MSBC for EV applications.
MATLAB/Simulink results have been presented to validate the proposed concept.
The MSBC utilizes battery 1 directly to run induction motor without any additional
DC-DC converters. This decreases the magnetic in the electric vehicle. The losses
and controlling complexity is reduced. It is concluded that this MSBC shall be useful
for new and improved range of EVs.
References
1. Majid P, Jain P (2020) Soft-switching power electronics technology for electric vehicles: a
technology review. IEEE J Emerg Select Topics Ind Electron 1(1):80–90
2. Salari O, Nouri M, Zaad KH, Bakshai A, Jain P (2018) A multi-source inverter for electric
drive vehicles. In: 2018 IEEE Energy Conversion Congress and Exposition (ECCE)
3. Dorn-Gomba L, Magne P, Danen B, Emadi A (2018) On the concept of the multi-source inverter
for hybrid electric vehicle powertrains. IEEE Trans Power Electron 33(9):7376–7386
4. Dorn-Gomba L, Guo J, Emadi A (2019) Multi-source inverter for power-split hybrid electric
powertrains. IEEE Trans Veh Technol 68(7):6481–6494
5. Gupta J, Maurya R, Arya S (2020) Improved power quality on-board integrated charger with
reduced switching stress. IEEE Trans Power Electron 35(10):10810–10820
6. Emadi A, MAgne P (2014) Power-converter, Patent U.S. 011770 AI
7. Buddika D, Abeywardana W, Acuna P, Hredzak B, Aguilera R (2018) Single-phase boost
inverter-based electric vehicle charger with integrated vehicle to grid reactive power compen-
sation. IEEE Trans Power Electron 33(4):3462–3471
8. Sousa T, Hashemi S, Andersen P (2019) Raising the potential of a local market for the reac-
tive power provision by electric vehicles in distribution grids. IET Gener Transm Distrib
13(12):2446–2454
9. Janabi A, Wang B (2020) Switched-capacitor voltage boost converter for electric and hybrid
electric vehicle drive. IEEE Trans Power Electron 35(6):5615–5624
10. Yilmaz M, Krein PT (2013) Review of the impact of vehicle-to-grid technologies on distribution
systems and utility interfaces. IEEE Trans Power Electron 28(12):5673–5689
11. Rippel WE (1990) Integrated traction inverter and battery charger apparatus, U.S. Patent 4 920
475
Indirect Current Vector Controlled
Three-Level Inverter for Induction
Motor Driven Electric Vehicle
Paramjeet Singh Jamwal , Sanjeev Singh , and Shailendra Jain
1 Introduction
Electric vehicles (EVs) have number of advantages like good dynamic behavior,
higher efficiency, reduced emission, and sustainable alternative but faces number
of key issues like short driving range, long charging time, and reduced (passenger
and cargo) space [1–3]. Short driving range issue can be resolved by improving the
efficiency of battery, inverter, and electric motor of EVs.
Induction motor (IM), permanent magnet (PM) motor, and switched reluctance
motor (SRM) are three suitable option of electric motor for EV application as reported
in [2, 4]. SRM needs more research for its noise and vibration free operation in
EV application. PM motors are facing high-cost issues because of costly rare earth
magnet material. IM specially squirrel cage motor has poor efficiency, but still, it
is a strong competitor of PM motors and finds application in EVs due to brushless
construction and low cost.
Most EVs have two-level inverters (2LIs) for the conversion of battery power to
electric motor power. 2LI suffers with the problem like high voltage stress, current
stress, voltage total harmonic distortion (THDv ), and current total harmonic distortion
(THDi ). To mitigate the issues with 2LIs, an inductor-capacitor (LC) filter is used
[5]. Multilevel inverters (MLIs) are giving promising results to resolve the issues
faced with 2LIs [6]. Chang et al., has reported [7] the benefits of cascaded h-bridge
(CHB) MLI with Si MOSFET switch, over 2LI with IGBT and SiC MOSFET switch.
Kersten et al. has reported [8] the benefits of active neutral point clamped three-level
P. S. Jamwal (B) · S. Jain

Sant Longowal Institute of Engineering and Technology, Longowal, Sangrur, Punjab, India
e-mail: paramjeet_pei1803@sliet.ac.in
S. Singh
Maulana Azad National Institute of Technology, Bhopal, Madhya Pradesh, India
e-mail: sschauhan@manit.ac.in
576 P. S. Jamwal et al.
inverter (NPC-3LI) with MOSFET switch, over 2LI. Guan et al. has reported [9] the
loop construction method to provide high frequency isolation in NPC-3LI.
Scalar and vector controls are the two most common schemes to obtain variable
speed with IM [10]. Scalar control scheme is easy to implement and has good steady
state performance but shows poor dynamic performance. Jamwal et al. has reported
[11] the scalar control scheme to generate the switching sequence and operate the
switches of CHB-3LI and NPC-3LI for IM drive (IMD). Vector control (VC) is
difficult to implement but has good dynamic and steady state performance which
suits best for EV application. VC is categorized as direct vector control (DVC) also
known as direct field-oriented control (DFOC), indirect vector control (IVC) also
known as indirect field-oriented control (IFOC) and direct torque control (DTC) on
the basis of magnetic field sensing [12]. In DTC, there is requirement of flux and
torque measurement of IM to calculate the field angle [13]. Adigintla and Aware
have reported [14] the fractional order proportional-integral controller-based DTC
scheme for IMD. In DVC, there is requirement of flux measurement of IM to calcu-
late the field angle, while in IVC, field angle is estimated from voltage or current
measurements [15]. Kakodia and Dynamina have reported [16] the comparison of
DVC and IVC scheme to generate the switching sequence and operate the switch
of NPC-3LI for IMD. IVC is further categorized as indirect voltage vector control
(IVVC) and indirect current vector control (ICVC). Nair and Barai have reported
[17] the IVVC scheme with space vector modulation (SVM). Krishnatheeram et al.
[18] presented the IVVC scheme with continuous SVM and discontinuous SVM in
which reference voltage is used to generate the switching sequence and operate the
switch of NPC-3LI for IMD.
In this paper, the ICVC scheme in which reference current is used to generate the
switching sequence and operate the switch of CHB-3LI and NPC-3LI are used to
drive IM for EV application and performance is compared with 2LI. The paper is
divided in five major parts. The basic research reported for EVs has been described
in the first section. Inverter topology with description is presented in second section.
MATLAB Simulink logic of ICVC scheme is described in third section. Performance
simulation with discussion is presented in fourth section. Conclusion drawn from the
research work is given in fifth section.
2 The Indirect Current Vector Controlled Scheme Fed IMD
The basic block diagram of ICVC scheme for the operation of inverter fed IMD is
given in Fig. 1. The control scheme begins with the rotor speed (ωr ) measurement
of IM. This measured rotor speed is fed to reference rotor flux generator to estimate
rotor flux reference (λr ). This measured speed is also compared with reference speed
to obtain speed error (ωe ).
This speed error is fed to reference torque generator for the estimation of elec-
tromagnetic torque reference (Te ). On the basis of estimated electromagnetic torque
reference, rotor flux reference, and measured rotor speed, vector controller generates
Indirect Current Vector Controlled Three-Level Inverter … 577
Vdc
Reference
ωe Te iT* ia* iae
ωref Torque
Switching Signal
ia
Generator Vector if*
Stator
ib* ibe PWM ib
Current Inverter ic
IM
Reference Control Synthesizer Controller
λr θf ic* ice ωr
ωr Rotor Flux
Generator
Indirect Estimation
ωr
Fig. 1 Block diagram of ICVC scheme operated inverter fed IM
reference torque current (iT * ), reference field current (if * ), and field angle (θf ) which
is fed to stator current synthesizer to generate three-phase reference current (ia * ,
ib * , ic * ). This three-phase reference current is compared with measured three-phase
motor current (ia , ib , ic ) to obtain error in three-phase current (iae , ibe , ice ). On the
basis of three-phase current error, PWM controller generate the switching signal to
operate the 2LI as given in Fig. 2a, CHB-3LI as given in Fig. 2b, and NPC-3LI as
given in Fig. 2c.
3 The Indirect Current Vector Controller
The logic used for indirect current vector controller is shown below in Fig. 3, as
created in MATLAB Simulink.
The indirect current vector control scheme begins with the generation of rotor
flux reference from rotor speed. When the rotor speed is between -1430 to 1430 rpm,
rotor reference flux generator produces constant rotor flux of 0.85 value. When
the speed is greater than 1430 rpm, reference rotor flux generator produces rotor
flux of 0.85*(1430/ωr ) value. When the speed is less than −1430 rpm, reference
rotor flux generator gives rotor flux of 0.85*(−1430/ωr ) value. Reference torque
generator consists of proportional-integral (PI) controller and gain of 9000 value
which generates reference electromagnetic torque from speed error. The value of
proportional gain (Kp ) and integral gain (Ki ) is calculated with Ziegler–Nichols
method as reported in [19]. These reference rotor flux, reference electromagnetic
torque and sensed rotor speed values are fed to vector controller (Eqs. 1–4) which
generates reference torque current (iT * or iT_ref), reference field current (if * or if_ref),
and field angle (θf or theta_f).
Te∗ L r 2 2
i T∗ = (1)
λr∗ L m 3 P

∗ L r λr∗
if = 1+ p (2)
Rr L m
L m · Rr i T∗
ωsl∗ = · ∗ (3)
Lr λr
Two-Level Inverter
S1 S3 S5
a
Vdc b IM
c ia ib ic
ωr
S2 S4 S6 ωref
ωe
S1 S2 S3 S4 S5 S6
iae
PWM Controller for ibe ICVC
2LI ice
(a) ICVC operated 2LI fed IM.
CHB-A CHB-B CHB-C
S1a S4a S1b S4b S1c S4c

a b IM
Vdc/2 Vdc/2 Vdc/2 c
n ia ib ic
ωr
S2a S3a S2b S3b S2c S3c ωref

ωe
S1a S2a S3a S4a S1b S2b S3b S4b S1c S2c S3c S4c iae
PWM Controller for 3LI ibe ICVC
ice
(b) ICVC operated CHB-3LI fed IM.
Fig. 2 Schematic diagram of various topologies for IMD

Fig. 3 Logic of ICV controller in MATLAB Simulink

θ ∗f = ∫ ωr + ωsl∗ dt (4)
These reference torque current, reference field current, and field angle values are
fed to stator current synthesizer which generate three-phase reference currents (ia * ,
ib * , ic * ), which are compared with sensed three-phase currents (ia , ib , ic ) to generate
current error (iae , ibe , ice ). This current error is divided by a constant (120) to generate
unit three-phase current template (I_temp). This unit current template is fed to pulse
width modulation (PWM) controller to generate switching sequence and operate the
switches of 2LI and 3LIs of IMD [20]. Parameters used in the logic of ICV controller
is given in Table 1.
Table 1 ICV controller

Variable Parameter Value
parameters
Kp Proportional gain 0.003
Ki Integral gain 0.04
Lr Rotor inductance (H) 0.005839
Lm Mutual inductance (H) 0.1722
P Number of Poles 4
4 Performance Simulation
The presented ICVC scheme fed IMD are modeled in MATLAB-SIMULINK envi-
ronment for performance simulation of the complete drive for EV application. Speed-
THDv characteristics at 85, 100, and 105% torque with 2LI and 3LIs are obtained
and shown in Fig. 4. These torque values are selected to realize the real time loading
of the EV as curb weight (80%) of the vehicle with one passenger (5%), with four
passengers (20%) and with five passengers (25%). The speed variation has been
considered from standstill to rated speed.
From speed-THDv characteristics it is observed that THDv is lesser with both
3LIs as compared with 2LI at all three torques. Speed-Efficiency characteristics at
85, 100, and 105% torque with 2LI and 3LIs is given in Fig. 5.
From speed-efficiency characteristics of IM it is clear that efficiency of IM is
better with both 3LIs as compared with 2LI at 85% torque. While efficiency of IM is
better with CHB-3LI and same with NPC-3LI as compared with 2LI at 100% torque.
Efficiency is same with 3LIs and 2LI at 105% torque. Speed-PF characteristics at
85, 100, and 105% torque with 2LI and 3LIs is given in Fig. 6.
From speed-PF characteristics of inverters it is clear that PF is better with both
3LIs except at 75% speed as compared with 2LI at 85% torque. PF is better with both
800 2LI 800 800

2LI 2LI
THDv (%)
THDv (%)
THDv (%)
600 CHB-3LI 600 CHB-3LI 600 CHB-3LI

400 NPC-3LI 400 NPC-3LI 400 NPC-3LI
200 200 200

0 0 0
25 50 75 100 25 50 75 100 25 50 75 100
Speed (%) Speed (%) Speed (%)
(a) 85% Torque (b) 100% Torque (c) 105% Torque
Fig. 4 Speed-THDv characteristics of inverters
100 100 100

IM Efficiency (%)
IM Efficiency (%)
IM Efficiency (%)
80 80 80
60 60 60
2LI 2LI 2LI
40 40 40
CHB-3LI CHB-3LI CHB-3LI
20 NPC-3LI 20 NPC-3LI
20 NPC-3LI
0 0 0
25 50 75 100 25 50 75 100 25 50 75 100
Fig. 5 Speed-efficiency characteristics of IM

1.0 1.0 1.0
Power Factor
Power Factor
Power Factor
0.8 0.8 0.8
0.6 0.6 0.6
0.4 2LI 0.4 2LI 0.4 2LI
CHB-3LI CHB-3LI
0.2 NPC-3LI 0.2 CHB-3LI 0.2 NPC-3LI
NPC-3LI
0.0 0.0 0.0
25 50 75 100 25 50 75 100 25 50 75 100

Fig. 6 Speed-PF characteristics of inverters
200 200 200

Current Stress
Current Stress
Current Stress
150 150 150
100 100 100

2LI 2LI
50 2LI 50 CHB-3LI 50 CHB-3LI
CHB-3LI
NPC-3LI NPC-3LI NPC-3LI
0 0 0
25 50 75 100 25 50 75 100 25 50 75 100
Fig. 7 Speed-current stress characteristics of inverters
3LIs as compared with 2LI at 100% and 105% torque while PF is better with NPC-
3LI as compared with CHB-3LI at 100% and 105% torque. Speed-Current Stress
characteristics at 85, 100, and 105% torque with 2LI and 3LIs is given in Fig. 7.
From speed-current stress characteristics of inverters it is clear that current stress
(kA/sec) through switch is lesser with both 3LIs as compared with 2LI at 85, 100, and
105% torque. Current stress is lesser with CHB-3LI as compared with NPC-3LI at
85% torque. While current stress is lesser with NPC-3LI as compared with CHB-3LI
at 100% and 105% torque. Voltage stress bar graph with 2LI and 3LIs is given in
Fig. 8.
From voltage stress bar graph of inverters, it is clear that voltage stress (kV/sec)
across the switch is lesser with both 3LIs as compared with 2LI. Line voltage, line
current, rotor speed, and electromagnetic torque performance of 2LI fed IMD system,
CHB-3LI fed IMD system, and NPC-3LI fed IMD system is given in Fig. 9a, b, and
c respectively.
From the speed performance it is clear that ICVC operated 2LI and 3LI fed IMD
system is able to follow reference speed at 100% speed, 75% speed, 50% speed, and
25% speed. From the torque performance it is clear that there is reduction in torque
ripples with both 3LIs as compared to 2LI fed IMD system.
Fig. 8 Voltage stress bar
Voltage Stess (kV/sec)

8000 6500
graph of inverters
6000
4000
1610 1615
2000
0
2LI CHB-3LI NPC-3LI
Inverters
From the performance simulation results, it is observed that the three level
inverters improve the overall performance of the system along with efficiency,
thereby, increasing the range of the electric vehicle which is one of the key
requirements at present.
5 Conclusions
This paper has presented an indirect current vector controlled three-level inverter
for induction motor drive fed electric vehicle with a focus on efficiency and range
improvement. The performance of two commonly used topologies i.e., CHB-3LI
and NPC-3LI have been compared with 2LI fed IMD system for EV application.
The presented performance has demonstrated reduction in total harmonic distortion
of inverter output voltage and voltage stress through switch of inverter with both
CHB-3LI and NPC-3LI. There is considerable improvement in efficiency of IM with
CHB-3LI as compared to NPC-3LI. The improvement in power factor of inverter
output with NPC-3LI is more as compared to CHB-3LI. The cost is also low for
the squirrel cage Induction motor and the indirect vector control is implementable
on low-cost digital signal processors. This shall result in a low cost, efficient and
improved range complete drive for induction motor based EVs. It is concluded that
NPC-3LI have better performance as compared to CHB-3LI fed IMD system for EV
application.
(a) Performance with 2LI fed IMD system.
(b) Performance with CHB-3LI fed IMD system.
Fig. 9 Line voltage, line current, rotor speed and electromagnetic torque
Appendix
Battery Voltage (Vdc): 645 V, Frequency: 50 Hz, Friction Factor: 0.002985 N-m-
s, Inertia: 0.0131 kg-m2 , Line Voltage: 400 V, Motor Power: 4 kW, Motor Speed:
1430 rpm or 149.75 rad/sec, Motor Torque: 26.71 Nm, Rotor Resistance: 1.395 ,
and Stator Resistance: 1.405 .
(c) Performance with NPC-3LI fed IMD system.
Fig. 9 (continued)
References
1. Nicoletti L, Brönner M, Danquah B, Koch A, König A, Krapf S, Pathak A, Schockenhoff F,

Sethuraman G, Wolff S, Lienkamp M (2020) Review of trends and potentials in the vehicle
concept development process. In: 2020 IEEE fifteenth international conference on ecological
vehicles and renewable energies (EVER), Monte-Carlo, Monaco
2. Chan CC, Bouscayrol A, Chen K (2010) Electric, hybrid, and fuel-cell vehicles architectures
and modeling. In: IEEE Trans Vehicular Technol 59(2):589–598
3. Jain S, Kumar L (2018) Fundamentals of power electronics controlled electric propulsion. In:
Rashid MH (eds) Power electronics handbook (4th ed) Butterworth-Heinemann an imprint of
Elsevier, pp 1023–1065
4. Rind SJ, Ren Y, Hu Y, Wang J, Jiang L (2017) Configurations and control of traction motors
for electric vehicles—a review. Chinese J Elect Eng 3(3):1–17
5. Mishra P, Maheshwari R (2019) Design analysis and impacts of sinusoidal LC filter on pulse
width modulated inverter fed induction motor drive. IEEE Trans Ind Electron 67(4):2678–2688
6. Poorfakhraei A, Narimani M, Emadi A (2021) A review of modulation and control techniques
for multilevel inverters in traction applications. IEEE Access 9:24187–24204
7. Chang F, Ilina O, Lienkamp M, Voss L (2019) Improving the overall efficiency of automotive
inverters using a multilevel converter composed of low voltage Si MOSFETs. IEEE Trans
Power Electron 34(4):3586–3602
8. Kersten A, Grunditz E, Thiringer T (2018) Efficiency of active three-level and five-level NPC
inverters compared to a two-level inverter in a vehicle. In: 2018 20th European conference on
power electronics and applications (EPE’18 ECCE Europe), Riga, Latvia
9. Guan Y, Li L, Zhang J (2020) A novel high frequency isolated three-level inverter and its
topological derivation method. In: 2020 15th IEEE conference on Industrial Electronics and
Applications (ICIEA), Kristiansand, Norway
10. Ellabban O, Mierlo JV, Lataire P (2011) A comparative study of different control techniques for
an induction motor fed by a Z-source inverter for electric vehicles. In: International conference
on power engineering, energy and electrical drives, Malaga, Spain
11. Jamwal PS, Singh S, Jain S (2021) Three-level inverters with volt per Hz control for induc-
tion motor driven electric vehicles. In: Tiwari M, Shrivastava A, Dwivedi S, Singh S (eds)
Proceedings of conference on flexible electronics for electric vehicles, lecture notes in electrical
engineering, in press, Springer
12. Mohan H, Dwivedi SK, Pathak MK (2018) Control of induction motor drives—technological
advancements. In: 2018 8th IEEE India International Conference on Power Electronics (IICPE),
Jaipur, India
13. Naganathan P, Srinivas S (2020) Direct torque control techniques of three-level H-bridge
inverter fed induction motor for torque ripple reduction at low-speed operations. IEEE Trans
Ind Electron 67(10), 8262–8270
14. Adigintla S, Aware MV (2020) Position control of the induction motor using fractional
order controllers. In: 2020 international conference on Power, Instrumentation, Control and
Computing (PICC), Thrissur, India
15. Krishnan R (2001) Electric motor drives modeling, analysis, and control. Prentice Hall
16. Kakodia SK, Dynamina G (2020) A comparative study of DFOC and IFOC for IM drive.
In: 2020 First IEEE International Conference on Measurement, Instrumentation, Control and
Automation (ICMICA), Kurukshetra, India
17. Nair NS, Barai M (2020) A three-level NPC fed induction motor drive with closed loop
dc-link voltage balancing algorithm. In: 2020 IEEE international conference on Power,
Instrumentation, Control and Computing (PICC), Thrissur, India (2020)
18. Krishnatheeram S, Peter J, Mohammed Shafi KP, Ramchand R (2018) Current error boundary
analysis of constant switching frequency voltage controlled three level VSI fed IM drive.
In: 2018 IEEE international conference on Power, Instrumentation, Control and Computing
(PICC), Thrissur, India
19. Singh S (2019) DC motor drives. In: Rahman MF, Dwivedi SK (eds) Modeling, simulation and
control of electrical drives. IET Control, robotics and sensors series vol 118, pp 101–136
20. Jamwal PS, Singh S, Jain S (2021) Three-level inverters for induction motor driven electric
vehicles. In: 2020 IEEE 3rd International Conference on Energy, Power and Environment
towards clean energy technologies (ICEPE), Shillong, Meghalaya, India
Comparative Analysis of Conventional
and Sliding Mode Control Techniques
for DC-DC Boost Converter for PV
System Under Transient Conditions
Pankaj Swarnkar, Suresh Kumar Gawre, and Gagnesh Akodiya
1 Introduction
The global demand of energy is increasing enormously due to industrialization and

urbanization. The existing fossil fuel is limited and causes the global warming effect,
this leads to searching for alternate renewable energy sources. PV sources are one
of the important power suppliers than other renewable energy sources due to clean,
pollution free, and endless. Many methods have been presented to model and simu-
late a photovoltaic array and to find parameters of the current–voltage equation
by adjusting the characteristics at three points: open circuit, maximum power, and
short circuit. Effects of irradiation and temperature change on characteristics of PV
are also considered to model the PV array [1–3]. To obtain maximum power from
solar panel, it needs to operate at maximum operating point. Maximum power point
tracking (MPPT) for photovoltaic systems using perturb and observe (P&O) algo-
rithm is the most common technique for this purpose [4]. When MPPT technique is
not used then the maximum available power is not being extracted and as a result
it leads to the low efficiency of the solar panel and hence the cost per kW of the
solar panel is increased. Hence in order to minimize the cost per kW of the solar
panel and to improve efficiency various MPPT algorithms are used. Ronn Raedani
et al. [5] explore the ‘Incremental Conductance’ and ‘P&O’ techniques of MPPT.
A comparative analysis on the basis of cost of two techniques is also conducted to
perform the comparative analysis between the MPPT techniques. In order to get the
constant desired output from a PV array suitable DC-DC boost converter must be
designed. Selection of appropriate components of boost converter is also an essential
requirement [6, 7]. In order to regulate the output of boost converter specific control
technique must be designed. Many conventional and advanced control techniques
have been designed and implemented so far to enhance the dynamic characteristic
P. Swarnkar · S. K. Gawre (B) · G. Akodiya

588 P. Swarnkar et al.
of DC-DC boost converter which is capable of achieving better transient response is

more steady and robust [8–11].
Under classical PI control of DC-DC boost converter, increasing the proportional
gain Kp of the controlled variable, advances the system performance and improves the
accuracy. However, a high Kp also increases the overshoot and reduces the stability
of the system. Increment in integral gain Ki improves the accuracy by reducing
the steady-state error but has negligible effect on transient performance. Finally,
increasing the derivative control gain Kd reduces the magnitude of oscillation of the
response. Kd although does not improve the steady-state accuracy but may increase
the noise [11, 12]. Such disadvantages of linear control method made designers to
think about nonlinear control methods such as sliding mode control [13–17] which
offer better performance under system uncertainties and large perturbations. DC-DC
power converters are inherently variable structure systems since at the two switching
states, i.e., ON and OFF, they exhibit two different structures. Many adaptive and
advance controllers based on AI techniques have also been suggested to overcome
the nonlinearity and time varying effect of system [18–20].
In this work the main aim is on regulating the output voltage of boost converter
to the desired value by using conventional control technique that is PI control and
advanced control technique that is Sliding Model Control. The two control techniques
applied for boost converter are simulated for variable input voltage, where input
voltage is varied by approx. 10 V difference or so while keeping the load resistance
constant. System is also tested for variable load conditions to validate the feasibility
of robust control technique over the conventional control.
2 Design of PV Array
The system is designed for normal household application that requires a power of
2000 watts for normal usage. So the PV array designed should be capable of gener-
ating 2000 watts of power. The PV module used in simulation is ‘1Soltech 1sth-235-
WH’. Table 1 shows the specifications of a PV array used to feed the boost converter
which is considered as plant for designing the control schemes.
Table 1 Specification of PV
Parameters Value
array
Peak power of PV Array 2000 W
Voltage at MPP 104 V
Current at MPP 22.5A
S.C. current 24A
O.C. voltage 135 V
Temperature 60 °C
Comparative Analysis of Conventional and Sliding Mode Control … 589
2.1 Design of Boost Converter
A boost converter steps up the input voltage. A boost converter with a power MOSFET
is shown in Fig. 1; Power for the boost converter is taken from solar panel which is
denoted by a battery in figure [6–11].
Characterized by switching, DC-DC power converters are inherently variable
structure systems since at the two switching states, i.e., ON and OFF, they exhibit
two different structures. The working of boost converter can be classified as Operation
1 and Operation 2. Operation 1 is defined for transistor switched on at time t = 0.
The input current increases and passes through inductor L and transistor (Fig. 2).
Operation 2 is described for transistor switched off at time t = t1 . The input current
now flows through L, C, load, and diode D as shown in Fig. 3.
L
D
iL iC iR
Vi G S C R Vo
Fig. 1 DC-DC boost converter topology
iL i C i R
V i C R V o
Fig. 2 DC-DC boost converter topology in ON state
iL iC iR
V i C R Vo
Fig. 3 DC-DC boost converter topology in OFF state

2.2 Calculation of Boost Converter Parameters
The system is designed to give output power of 2000 watts for normal household
use. As the aim of this work is to get the constant output voltage of 220 V under any
external changes in input voltage or output load resistance, so the value of output
voltage can be considered constant at 220 V.
From PV curve of solar module shown in Fig. 4, it is observed that the maximum
power is achieved by solar panel at around 100 V so taking minimum output voltage
of solar panel to be around 90 V, hence the minimum input voltage of boost converter
is also taken as 90 V. Also the switching frequency is taken as 10 kHz for system
operation.
So by concluding all some initial values are taken as
Po (out put power ) = 2000watts, f s (switching f requency) = 10khz
Vi (input voltage) = 90volts, Vo (out put voltage) = 220volts
The other parameters of boost converter are given in Table 2.

Hence the value of load resistance is obtained as 24.2 ohms which can be
approximated as 25 ohms. This is the rated value of load resistance for 2000 watts
system. Using the values of resistance, inductance and capacitance the DC-DC boost
converter is realized and further used to implement the control techniques PIC and
SMC.
Fig. 4 P–V characteristic of PV array

Table 2 Parameters of boost converter

Parameters Description Value
Vi
Duty cycle d =1− Vo =1− 90
220 0.59
(1−d)2 ×d×R
Inductance (minimum value to keep the boost converter L min = 2× f s 0.15 mH
in continuous conduction mode)
d×Vo
Capacitance (for producing the required output voltage C= R× f s ×Vo 0.244 m F
ripple)
Vo2 (220)2
Load Resistance R= Po = 2000 24.2
2.3 Mathematical Modeling of DC-DC Boost Converter
In this section mathematical model of DC-DC boost converter is presented using in

state-space method. This is a generalized time domain representation of system [10].
Dynamics of the converter operating in continuous conduction mode is obtained by
applying KVL in the loop containing inductor, DC source, and switch.
d IL Vo Vi
= −(1 − u) + (1)
dt L L
Now the KCL is applied on node connecting the capacitor branch, following
equation is obtained
d Vo IL Vo
= (1 − u) − (2)
dt C RC
By defining state variables as, x1 = I L and x2 = Vo , the state equations can be
obtained as
x2 Vi
ẋ1 = −(1 − u) + (3)
L L
x1 x2
ẋ2 = (1 − u) − (4)
C RC
Finally in matrix form,
1
ẋ1 0 −(1 − u) L1 x1
= + L Vi (5)
ẋ2 (1 − u) C1 − RC
1
x2 0
where Vo is output capacitor voltage, I L is inductor current, Vi is input voltage. The

control input ‘u’ represents the ON or OFF state of switch in boost converter, it
is a discrete valued function and its values are in set (0,1). It is assumed that the
continuous conduction mode is maintained by keeping inductor value greater than

minimum value of inductor as mentioned in Eq. (1).
3 Design of PI Controller
Figure 5 shows the schematic diagram of control scheme for boost converter using
PI controller.
In conventional control scheme error signal is generated as the difference between
actual and required set point and feed to PI controller. The controller tries to mitigate
the error by adjusting the process by introducing a manipulated variable which in
boost converters case is its duty cycle. The detection of error is in terms of current,
voltage, angle, movement, temperature, etc. for observing the effect of proportional
controller only, the integral gain K i is set to zero.
P-controller tends to generate an offset value and also increase the maximum
overshoot of the system. Also too high a value of K p will lead to the oscillation.
The PI controller confirms that the boost converter will deliver a required amount of
voltage to the load. The PI control action is defined by the following relation:
t
u(t) = K p e(t) + K i e(t)dτ (6)
0
Fig. 5 Schematic diagram of boost converter with PI control

Fig. 6 Duty signal generation of boost converter by PI controller
Here as shown in Fig. 5 variable input voltage is taken from the solar panels for the
boost converter. To implement the PI control on boost converter the output voltage
of boost converter is compared with the reference value of voltage which in this case
is set to 220 V. Error in voltage generated by the comparison is amplified and then
it is fed to the PI controller which generates an equivalent control signal as shown
in Fig. 6, which is then compared with the saw tooth wave to produce an equivalent
duty cycle which is used to drive the switching device of the converter. As long as
the control signal is greater than saw tooth signal, switch is in ON state and inductor
current rises, and when control signal is less than the saw tooth signal the switch is
in OFF state.
4 Sliding Mode Control Design For Boost Converter
The state-space model developed in previous section is used to implement sliding

mode control (SMC) on a boost converter. Inductor current and output voltage are
chosen as state variables. So state variables x1 = I L and x2 = Vo . Now the aim is to
achieve a desired constant output voltage Vr e f . Thus,
x2 = Vr e f (7)
ẋ2 = V̇r e f = 0 (8)
From the conventional SMC theory, the sliding surface is defined as
S = x1 − x1∗ = 0 (9)
x1∗ is the reference value.

In order to force the initial variables on the sliding surface S = 0, the ideal control
law for the switch is
1
u= (1 − sign(S)) (10)
2
Here the objective is to confirm that the state trajectory of the initial state of
system is moved toward the sliding surface S = 0 and slides over it, until it reaches
to a stable equilibrium. This is obtained by appropriate design of control law using
the reaching condition,
S Ṡ < 0 (11)
Since
S = x1 − x1∗
At the equilibrium (generally origin of phase plane), the final values of state
variables coincide with the corresponding reference values then
x1∗ = 0
By Eqs. (9) and (11) we get
x2 > Vi
This means existence condition in sliding mode satisfies if the output voltage is
greater than the input voltage of boost converter.
Block Diagram of boost converter with SMC is shown in Fig. 7 where Vr e f
represents the reference voltage which is desired at the output of boost converter.
The system basically consists of two feedback loops; one is the voltage loop and
other one is another current loop. The voltage loop is the slower loop since the
change in voltage is slower as compared to the change in current. The system works
in two stages, in first the output voltage is compared with the reference voltage and
hence the error is generated which is fed to the PI controller which generates the
reference inductor current. Hence error (difference between reference and actual
inductor current) generated is nothing but the sliding surface for boost converter,
now the control signal ‘u’ of boost converter is achieved by Eq. (10). And hence the
sliding mode control of boost converter is achieved.
Figure 8 shows the output voltage of boost converter without any control (open
loop). It can be seen that overshoot is 34.09% which is quite large and can cause the
instability. It can also be observed that boost converter without any control is able
to give the output voltage of 220 V when duty cycle is correctly adjusted. But as the
output voltage of solar panel won’t remain constant as irradiance and temperature
Fig. 7 Schematic Diagram of boost converter with SMC
Fig. 8 Simulation results of open-loop boost converter for Vi = 90 V
on solar panels are constantly changing. So in that case it is really necessary to have
some controller to regulate the duty cycle to get the desired Vo .
Figure 9 shows the outcomes of boost converter with SMC and PIC when input
voltage is set at 90 V and the desired output voltage is set at 220 V. Curves present the
comparative analysis of PI control and SM control, Hence the performance can be
analyzed in terms of Rise Time (RT), Peak Overshoot (PO), Settling Time (ST), and
Steady-state error (SSE) for both control techniques. From figure it can be clearly
observed that boost converter with PI control is able to track the Vo of 220 V, and
the steady-state value is also close to the desired output voltage. The peak overshoot
Fig. 9 Simulation results of boost converter with SMC and PI Controller for Vi = 90 V
Table 3 Comparison of different control scheme (input voltage is constant)

Input voltage Specification Open loop PI controller SMC
90 V RT 1.5 ms 1.3 ms 0.9 ms
PO 34.09% 24.9% 5.4%
ST 0.1 s 0.15 s 0.03 s
SSE 0.7 V 0.9 V 0.08 V
is reduced as compared to open-loop characteristic hence the system is stable. But

curve is having oscillations with large settling time. Now with the application of
SMC the overshoot is reduced to permissible limit of 5.4% with no undershoot. Also
the settling time is reduced to large extent. So it can be concluded that sliding mode
controller performs much better than the PI controller. Table 3 gives the summary of
all time response specifications obtained by simulation study with different control
schemes.
4.1 Effect of Transients
In this section, the robustness of proposed controller designs is tested under transient
conditions. First the system is tested for variation in input voltage (with constant
load) and then it is tested for variation in load resistance as well as input voltage. The
performance of boost converter is analyzed without any control, with PI control and
with SMC for transient condition. In these results the input voltage of boost converter
is varied between 90 V and 140 V and the load resistance is varied by ± 5 ohms of
its rated value, i.e., 25 ohms (rated). Also the reference value of output voltage is set
at 220 V.
4.2 Variation in Input Voltage
The input voltage is varied from 100 to 140 V while keeping the load resistance
constant at its rated value of 25 ohms. In previous section the input voltage is kept
constant at 90 V, now the further study is done by varying the input voltage to observe
the performance of boost converter with different control schemes.
Table 4 gives results of transient analysis for variation in input voltage from Vi =
90 V to Vi = 140 V. It can be observed that when the input voltage is increased the
performance of boost converter is deteriorated very fast with open loop and PIC. The
overshoot of the system is increased beyond the limit and making system unstable.
On the other hand, SMC is still able to control the system keeping the specifications
at the satisfactory limits.
Table 4 Comparison of different control schemes (input voltage is variable and output resistance
Ro = 25 ohms)
Input voltage Specification Open loop PI controller SMC
90 V RT 1.5 ms 1.3 ms 0.9 ms
PO 34.09% 24.9% 5.4%
ST 0.1 s 0.15 s 0.03 s
SSE 0.7 V 0.9 V 0.08 V
100 V RT 1.07 ms 1.16 ms 0.82 ms
PO 35.45% 34.09% 8.63%
ST 0.08 s 0.22 s 0.03 s
SSE 0.65 V 0.8 V 0.2 V
110 V RT 0.96 ms 1.08 ms 0.76 ms
PO 40% 43.63% 16.36%
ST 0.09 s 0.12 s 0.025 s
SSE 0.65 V 0.8 V 0.15 V
120 V RT 0.86 ms 1 ms 0.71 ms
PO 45% 53.63% 23.18%
ST 0.1 s 0.14 s 0.05 s
SSE 1.6 V 0.8 V 0.08 V
130 V RT 1 ms 0.95 ms 0.67 ms
PO 49.54% 62.72% 30.9%
ST 0.1 s 0.19 s 0.045 s
SSE 0.62 V 0.8 V 0.07 V
140 V RT 0.71 ms 0.89 ms 0.64 ms
PO 55.45% 72.27% 38.18%
ST 0.1 s 0.17 s 0.045 s
SSE 0.55 V 0.7 V 0.07 V
Input voltage Specification PI controller SMC
different control schemes
(input voltage is variable and 90 V RT 0.129 ms 0.9 ms
output resistance Ro = 20 PO 24.54% 4.09%
ohms)
ST 0.25 s 0.029 s
SSE 2.5 V 0.1 V
100 V RT 1.17 ms 0.82 ms
PO 33.86% 8.63%
ST 0.22 s 0.07 s
SSE 0.9 V 0.1 V
110 V RT 1.1 ms 0.76 ms
PO 44.31% 15.9%
ST 0.12 s 0.035 s
SSE 0.9 V 0.13 V
125 V RT 0.98 ms 0.7 ms
PO 58.63% 26.81%
ST 0.14 s 0.035 s
SSE 0.9 V 0.08 V
140 V RT 0.89 ms 0.65 ms
PO 71.81% 38.18%
ST 0.16 s 0.05 s
SSE 0.9 V 0.09 V
4.3 Variation in Load Resistance
It is important to check the behavior of the boost converter when the load varies from
its rated value. Hence load resistance of the boost converter is varied by 20% above
and below the rated value of load resistance of 25 ohms. The system is tested with
PIC and SMC, when input voltage is also varied for under and over load conditions.
The input voltage is varied from 90 to 140 V. The results advocate the superiority of
SMC over PIC.
4.4 Load Resistance is Set at 20 Ohms
To study the performance of output voltage when it is loaded under the rated value,
the value of load resistance is set at 20 ohms which is 20% below the rated value.
Table 5 shows the results for variable input voltage Vi = 90 V to Vi = 140 V with
load resistance Ro = 25 ohms. Table 5 compares control schemes on the basis of rise
time (RT), peak overshoot (PO), settling time (ST) and steady-state error (SSE).
Input voltage Specification PI controller SMC
different control schemes
(input voltage is variable and 90 V RT 1.28 ms 0.89 ms
output resistance Ro = 30 PO 25% 6.136%
ohms)
ST 0.12 s 0.03 s
SSE 0.65 V 0.06 V
100 V RT 1.2 ms 0.82 ms
PO 34.54% 9.09%
ST 0.14 s 0.02 s
SSE 0.65 V 0.2 V
110 V RT 1.085 ms 0.75 ms
PO 45% 16.36%
ST 0.15 s 0.045 s
SSE 0.65 V 0.12 V
125 V RT 1 ms 0.69 ms
PO 59.09% 27.27%
ST 0.15 s 0.05 s
SSE 0.6 V 0.06 V
140 V RT 0.88 ms 0.64 ms
PO 72.27% 38.63%
ST 0.13 s 0.05 s
SSE 0.7 V 0.065 V
5 Load Resistance is Set at 30 Ohms
To study the performance of output voltage when it is loaded above the rated value,
the value of load resistance is set at 30 ohms which is 20% above the rated value. Table
6 gives the comparative analysis of both the control schemes under such variations
where input voltage is varied from Vi = 90 V to Vi = 140 V and converter is
overloaded with Ro = 30 ohms.
6 Conclusion
This paper presents the PI and Sliding Mode control techniques for solar penal fed
boost converter. The system is tested both under steady state and transient conditions.
The study shows that both control schemes work satisfactorily under steady-state
condition when the system parameters are set at their rated values. As the transients
are introduced in the system, the performance of PI controller deteriorates and crosses
the acceptable limits while SMC still shows the robustness and keeps the system in
stability limit. So it can be concluded that sliding mode control is a better option
because it generated less disturbance in the system and also at the input side.
References
1. Villalva MG, Gazoli JR, Filho ER (2009) Comprehensive approach to modelling and simulation
of photovoltaic arrays. IEEE Trans Power Electron 24(5)
2. Krismadinataa, Rahima NA, Pinga HW, Selvaraja J (2013) Photovoltaic module modelling
using simulink/matlab. Proc Environ Sci 17:537–546
3. Bellia H, Youcef R, Fatima M (2014) A detailed modeling of photovoltaic module using Matlab.
Nriag J Astron Geophys 3(1):53–61
4. Al Sadi S, Al Sayid B (2012) Maximum power point tracking simulation for photovoltaic
systems using perturb and observe algorithm. Int J Eng Innov Technol 2(6):80–85
5. Raedani R, Hanif M (2014) Design, testing and comparison of P&O, ICand VSSIR MPPT
techniques. Renew Energy Res Appl (ICRERA), pp 322–330
6. Masri S, Chan PW (1990) Design and development of a Dc-Dc converter with constant output
voltage. IEEE Trans Power Electron 5(3):323–330
7. Hasaneen BM, Elbaset Mohammed AA (2008) Design and simulation of Dc/Dc boost converter.
In: 12th international middle-east power system conference, pp 335–340
8. Mohan N, Undeland TM, Robbins WP (2003) Power electronics: converters, applications, and
design, 3rd edn. Wiley, USA
9. Jung D, Ji Y, Park S, Jung Y, Won C, Member S (2011) Interleaved soft-switching boost
converter for photovoltaic power-generation system. IEEE Trans Power Electron 26(4):1137–
1145, 9
10. Hosseini SH, Haghighian SK, Danyali S, Aghazadeh H (2012) Multi-input Dc boost converter
supplied by a hybrid PV/wind turbine power systems for street lighting application connected
to the grid. In: Universities Power Engineering Conference, pp 1–6
11. Bakar AA, Utomo WM, Taufik T, Aizam S, Jumadril (2015) Dc/Dc boost converter with PI
controller using real-time interface. ARPN J Eng Appl Sci 10(19)
12. Dave MR, Dave KC (2012) Analysis of boost converter using PI control algorithms. Int J Eng
Trends Technol 3(2)
13. Abutbul O, Gherlitz A, Berkovich Y, Ioinovici A (2003) Step-up switching-mode converter
with high voltage gain using a switched capacitor circuit. IEEE Trans Circuits Syst I, Fundam
Theory Appl 50(8):1098–1102
14. Gueldemir H (2005) Sliding mode control of boost converter. J Appl Sci 5(3):588–592
15. Hijazi A, Di Loreto M, Bideaux E, Venet P, Clerc G, Rojat G (2009) Sliding mode control of
boost converter: application to energy storage system via super capacitors. In: Electronique,
Barcelone, Spain
16. Das S, Qureshi MS, Swarnkar P (2018) Design of integral sliding mode control for DC-DC
converter. Material Proc Charact Assoc Mater Today Proc Elsevier J 5(2):4290–4298
17. Qureshi MS, Das S, Swarnkar P, Gupta S (2018) Design and implementation of sliding mode
control for uncertain systems. Material Proc Charact Assoc Mater Today Proc Elsevier J
5(2):4299–4308
18. Guo L, Hung JY, Member S, Nelms RM (2009) Evaluation of DSP-based PID and fuzzy
controllers for Dc—Dc converters. IEEE Trans Ind Electron 56(6):2237–2248
19. Harsh G, Pankaj S (2019) Investigations on metaheuristic algorithm for designing adaptive
PID controller for continuous stirred tank reactor. MAPAN-J Metrol Soc India 34(1):113–119
20. Swarnkar P, Jain S, Nema RK (2014) Adaptive control schemes for improving the control
system dynamics: a review. IETE Techn Rev 31(1):17–33
Development of Fault Indicators
for Stator Inter-turn Fault Diagnosis
of a Synchronous Generator Using
Kalman Filter
P. V. Sunil Nag and C. Santhosh Kumar
1 Introduction
With the advancement in civilization, the need for reliable electric power has
increased drastically. The workhorse for a typical power system is a synchronous
generator and hence the reliability of the synchronous generator is paramount. Reli-
ability can be enhanced using a fault diagnostic system that can identify incipient
faults so that early remedial action can be taken, avoiding costly downtime. One
of the major faults in the synchronous generator is the stator inter-turn fault which
accounts for 30 to 40% of the faults in a machine [1]. This is a seed fault for most of
the electrical faults in the synchronous generator. Hence, detecting and diagnosing
this fault improves the reliability of the synchronous generator and thus of the power
system.
Many strategies have been developed for the fault diagnosis of the stator inter-turn
fault. The methods for fault diagnosis, in general, can be classified into 3 categories:
signal-based, data-based, and model-based [2]. The model-based approach is superior
as it readily makes use of the knowledge of the physics of the machine unlike the
other two classes of methods which ignore the physics of the machine involved. One
important class of the model-based methods is based on the popular observer/Kalman
filter and its variants. One of the important requirements of a fault diagnostic system
is its ability to distinguish between a fault and similar disturbing factors [3]. For the
stator inter-turn fault of a synchronous generator, one important disturbing factor is
the load imbalance [4]. Hence, a fault diagnostic system should be able to produce
a fault indicator that can distinguish between the two.
In the recent literature, Kalman filter/observer-based methods have been used for
fault diagnosis of the inter-turn fault of electrical machines [3–8] uses a fault indicator
P. V. Sunil Nag (B) · C. Santhosh Kumar

Department of Electronics and Communication Engineering, Amrita School of Engineering,
Amrita Vishwa Vidyapeetham, Coimbatore 641112, Tamil Nadu, India
e-mail: pv_sunil@cb.amrita.edu
602 P. V. Sunil Nag and C. Santhosh Kumar
namely residual current vector obtained using an observer and sequence component
decomposition. Aubert et al. [5] used a joint extended Kalman filter (EKF) to estimate
the fault parameters (ratio of the number of shorted turns to the healthy turns) that are
used as indicators of the fault. The above works address the issue of load imbalance
on the performance of the fault indicator and use a permanent magnet synchronous
generator. The model used for the fault diagnosis, in the above works, is capable of
detecting a fault in any of the 3 phases and uses the dq0 transformation. El Sayed et
al. [6] is similar to the above work but compares the performance of the EKF with
unscented Kalman filter but this work does not address the issue of load imbalance.
Idrissi et al. [7] uses a linear Kalman filter to diagnose the stator inter-turn fault in a
doubly-fed induction generator. The model used is based on a stationary reference (α,
β)frame. The proposed diagnostic system can find fault in any of the three phases, but
it does not address the issue of load imbalance. Here it is mentioned that the linear
Kalman filter is more suitable for hardware implementation. Nadarajan et al. [4]
uses a joint extended Kalman filter for fault diagnosis of a Brushless synchronous
generator. This work addresses the issues of the effect of load imbalance, further
a dq0 model is used. Here fault in the R phase is only addressed. Here it should
be mentioned specifically that a brushless synchronous generator is more complex
than a permanent magnet synchronous generator, hence the modeling exercise is
more difficult. This work uses a combination of time-domain and frequency-domain
processing to diagnose the fault. Jaiswal et al. [8] extended the work of [4] for failure
prognosis and it uses FEM analysis of the BLSG. This work uses a dq0 model with
a fault in a single phase along with an extended Kalman filter.
From the above survey of the recent literature on fault diagnosis of inter-turn fault
using Kalman filter and its variants, it could be observed that there is a necessity of a
model for fault diagnosis that can accommodate fault in all phases for a synchronous
generator. Considering hardware implementation requirements it is desirable if a
linear Kalman filter is used for diagnosis. The Kalman filter should also generate a
fault indicator that can distinguish between inter-turn fault and load imbalance.
For testing the efficacy of a fault diagnostic algorithm, a fault injection capa-
ble machine is required [9]. Developing such a system for a generator with a high
current/voltage rating is difficult. The alternative is to develop a simulated system
[10, 11]. The best simulation that almost exactly matches the real system is a finite
element simulation.
Considering the above conclusions, this work discusses (1) the development of
a finite element model-based fault injection capable synchronous generator using
ANSYS Maxwell hereafter referred to as a Maxwell. (2) The dq0 model of the
synchronous generator along with a fault parameter in all the 3 phases. The model is
in the state-space form and this model can be used by a linear Kalman filter for state
and parameter estimation. The model developed here is an extension of [4], where
the model has been derived for fault in one phase. (3) The process of generating
a fault indicator that can identify an inter-turn fault and at the same time exhibit a
distinct behavior in case of a load imbalance. Here an important assumption is that
a fault can occur in only one phase at a time.
Development of Fault Indicators for Stator Inter-turn Fault Diagnosis … 603
The proposed contributions of this work are the dq0 model with a fault in all
three phases an extension of the work given in [4]. Further, a new fault indicator is
proposed to identify inter-turn fault using the Kalman filter. This indicator also helps
in identifying a load imbalance condition.
Section 2 describes the proposed architecture, Sect. 3 deals with the procedure for
obtaining a fault injection capable Maxwell model. Section 4 gives an overview of
the derivation of the dq0-based state-space model for the synchronous generator with
a fault in all the 3 phases. Section 5 provides the relevant details of how the stan-
dard linear Kalman filter algorithm is used to generate the proposed fault indicator.
Section 6 gives the results and discussion and the last section provides the conclusion
and future work.
2 Block Diagram of the Proposed System
The block diagram of the proposed system is shown in Fig. 1. The fault injection
capable synchronous generator is a system in which fault can be injected by closing
a switch. It has been designed with split coils.
Terminals from the split coils are brought out and a switch can be used to short the
terminals of the particular split. This generator is implemented in Maxwell software.
The data from the generator, namely the stator voltages and currents are collected
and given to a Joint Kalman filter (a Kalman filter that estimates both states and
parameters simultaneously). The Kalman filter uses a dq0-based state- space model.
Fig. 1 Block diagram of the

proposed system
The Kalman filter in turn estimates the states and indicators for fault diagnosis. The
next section describes the fault injection capable generator.
3 Fault Injection Capable Synchronous Generator
As mentioned earlier, building a fault injection capable generator in hardware requires

a lot of customization and hence will be costly. A viable alternative for this would
be to build a Finite element model (FEM) of the machine. This FEM model is the
closest we can get to the actual machine in the simulation domain. Maxwell software
was chosen to build the model in this work.
To obtain the model, a 15 kVA synchronous generator was designed using standard
procedures [12]. These designed values were entered into the ANSYS RMxprt™
software which will enable easy generation of the FE model in the Maxwell software.
After obtaining the Maxwell model of the generator, the FE model was made fault
injection capable following the same lines as given in [13]. Each phase of the 15 kVA
generator was made up of 12 coils and each coil had 16 turns. Each coil in the phase
had a taping at 10 turns. This taping along with the end terminals of each coil was
brought out as terminals. The terminals could be shorted externally to simulate an
inter-turn short. The next section describes the development of the analytical model
called as dq0 model hereafter.
4 Derivation of the Dq0 Based State-Space Model
The schematic shown in Fig. 2 contains the various coils (3 phase coils with inter-turn
fault belonging to the stator and 3 coils belonging to the rotor) in the synchronous
generator and their relative orientation. Here each phase in the stator winding is split
into two parts for example A phase is split into a1 healthy and a2 faulty. To arrive
at the state-space model, two types of equations, namely the flux equations and the
voltage equations have to be obtained from the coil configurations. A typical flux
equation for the coil say a1 is given in Eq. (1), similar equations can be written for
other coils, namely a2, b1, b2, c1, c2, fd, kd, and kq. Here the Kirchhoff’s current
law equations and flux equations for each phase can be written as given in Eqs. (2)
to (7). The Eqs. (2) to (7) can be used to manipulate the flux equations that were
written for each coil to obtain modified equations. These equations are shown in the
matrix form in Eq. (8).
Where ψa , ψb , and ψc are the flux linkages associated with the phases a, b, and
c, respectively. Flux linkages of other coils like a1, a2, b1, b2, fd, kd, and kq can
be similarly represented. i a , i b , and i c are the currents in the phases a, b, and c,
respectively. The currents through other coils like a1, a2, b1, b2, fd, kd, and kq
can be similarly represented. Here the inductance L x1x2 represents the inductance
between coils x1 and x2, for example, L a1a2 represents the inductance between the
Fig. 2 Schematic of a synchronous generator with a fault in all phases
coils a1 and a2. Parameters μ1 , μ2 , and μ3 are the ratios of the faulty turns to the
total number of turns for the phases A, B, and C, respectively (here A,B,C are used
to represent R,Y,B phases interchangeably).
Here it can be observed in Eq. (8) that all the terms of the inductance matrix
are time varying. To overcome this park’s transformation or dq0 transformation is
used for shifting the frame of reference to the rotor reference frame. The simplified
inductance matrix obtained after applying the park’s transformation in block matrix
form is as shown below in Eq. (9)
ψa1 = −L a1a1 i a1 − L a1a2 i a2 − L a1b1 i b1 − L a1b2 i b2 − L a1c1 i c1 − L a1c2 i c2

(1)
+ L a1 f d i f d + L a1kd i kd + L a1kq i kq
i a2 = i a1 − i a f (2)
i b2 = i b1 − i b f (3)
i c2 = i c1 − i c f (4)
ψa = ψa1 + ψa2 (5)
ψb = ψb1 + ψb2 (6)
ψc = ψc1 + ψc2 (7)

⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ψa L aa L ab L acL a f d L akd L akq L aa f L ab f L ac f ia
⎢ ψb ⎥ ⎢ L ba L bb L bcL b f d L bkd L bkq L ba f L bb f L bc f ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ib ⎥
⎢ ψc ⎥ ⎢ L ca L cb L cc
L c f d L ckd L ckq L ca f L cb f L cc f ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ic ⎥
⎢ψ f d ⎥ ⎢ L f da L f db L f dc
L f d f d L f dkd L f dkq L f da f L f db f L f dc f ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥ ⎢i f d ⎥
⎢ ψkd ⎥ = ⎢ L kda L kdc f ⎥ ⎢ ⎥
⎢ ⎥ ⎢ L kdb L kdc
L kd f d L kdkd L kdkq L kda f L kdb f ⎥ ∗ ⎢ i kd ⎥
⎢ ψkq ⎥ ⎢ L kqa L kqb L kqc
L kq f d L kqkd L kqkq L kqa f L kqb f L kqc f ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥ ⎢ i kq ⎥
⎢ ψa2 ⎥ ⎢ L a2a L a2b L a2c
L a2 f d L a2kd L a2kq L a2a f L a2b f L a2c f ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥ ⎢ia f ⎥
⎣ ψb2 ⎦ ⎣ L b2a L b2b L b2c
L b2 f d L b2kd L b2kq L b2a f L b2b f L b2c f ⎦ ⎣ i b f ⎦
ψc2 L c2a L c2b L c2c
L c2 f d L c2kd L c2kq L c2a f L c2b f L c2c f ic f
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ (8)
ψsdq0 Z1 Z2 Z3 i sdq0
⎣ ψrabc ⎦ = ⎣ Z 4 Z 5 Z 6 ⎦ ∗ ⎣ irabc ⎦ (9)
ψ f dq0 Z7 Z8 Z9 i f dq0
where
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ψsd ψfd ψd f
ψsdq0 = ⎣ ψsq ⎦ ; ψrabc = ⎣ ψkd ⎦ ; ψ f dq0 = ⎣ ψq f ⎦ ;
ψs0 ψkq ψ0 f
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
i sd id f i fd
i sdq0 = ⎣ i sq ⎦ ; i f dq0 = ⎣ i q f ⎦ irabc = ir dq0 = ⎣ i kd ⎦
i s0 i0 f i kq
Z 1 toZ 9 in equation (9) can be obtained as follows
⎡ ⎤ ⎡ ⎤
−L d 0 0 μ1 L d 0 0 (10)
Z 1 = ⎣ 0 −L q 0 ⎦ ; Z 3 = −Z 7 = ⎣ 0 μ2 L q 0 ⎦
0 0 −L 0 0 0 μ3 L 0
⎡ ⎤
L a f d L akd 0
Z2 = ⎣ 0 0 L akq ⎦
0 0 0
⎡ 3 ⎤ ⎡3 ⎤
− 2 La f d 0 0 L μ
2 af d 1
0 0
Z 4 = ⎣ − 23 L akd 0 0 ⎦ Z 6 = ⎣ 23 L akd μ1 0 0⎦
0 − 2 L akq 0
3
0 3
2
L akq μ2 0
⎡ ⎤ ⎡ ⎤
L f d f d L f dkd 0 L a f d μ1 L akd μ1 0
Z 5 = ⎣ L f dkd L kdkd 0 ⎦ Z 8 = ⎣ 0 0 L akq μ2 ⎦
0 0 L kqkq 0 0 0
⎡ 2 ⎤ (11)
μ1 (L d − L 0 ) + μ1 L 0 0 0
Z9 = ⎣ 0 μ2 2 (L q − L 0 ) + μ2 L 0 0 ⎦
0 0 μ3 L 0
⎡ ⎤
−(L ad + L l ) 0 0 L ad L ad 0 μ1 (L ad + L l ) 0 0
⎢ 0 −(L aq + L l ) 0 0 0 L aq 0 μ2 (L aq + L l ) 0 ⎥
⎢ ⎥
⎢ 0 0 −L 0 0 0 0 0 0 μ3 L 0 ⎥
⎢ ⎥
⎢ −L ad L ad μ1 0 ⎥
⎢ 0 0 L f d f d L f dkd 0 0 ⎥
⎢ ⎥
Z =⎢ −L ad 0 0 L f dkd L kdkd 0 L ad μ1 0 0 ⎥
⎢ ⎥
⎢ 0 −L aq 0 0 0 L kqkq 0 L aq μ2 0 ⎥
⎢ ⎥
⎢−μ1 (L ad + L l ) 0 0 L ad μ1 L ad μ1 0 P 0 0 ⎥
⎢ ⎥
⎣ 0 −μ2 (L aq + L l ) 0 0 0 L aq μ2 0 Q 0 ⎦
0 0 −μ3 L 0 0 0 0 0 0 R
wher e P = μ1 2 (L ad + L l − L 0 ) + μ1 L 0 ; Q = −μ2 2 (L aq + L l − L 0 ) + μ2 L 0 ; R = μ3 L 0
(12)
Converting the above set of Eq. (9) into the per-unit form and writing Z (where Z
is the matrix containing Z 1 to Z 9 ) the expression in Eq. (12) is obtained. So the Eq.
(9) with the value of Z given by Eq. (12) forms the final set of flux equations. The
various terms in the Eq. (12) are defined in Table 1 in Appendix A
Next, the equations for the voltage induced in the coils are derived. Figure 2 can
be redrawn as shown in Fig. 3. A typical voltage equation for the coil say a1 is given
in Eq. (16). Similar equations can be written for other coils, namely a2, b1, b2,
c1, c2, fd, kd, and kq. Here the Kirchhoff’s current law equations for each phase
can be written as given in Eqs. (2) to (4). Other Eqs. (13) to (15) can be used to
manipulate the voltage equations that were written for each coil to obtain modified
equations. These equations are shown in the matrix form in Eq. (17). Applying the
park’s transformation to Eq. (17), we obtain the Eq. (19).
ra = rb = rc = rs ; ra1 = (1 − μ1 )rs ; ra2 = μ1rs ;

(13)
rb1 = (1 − μ2 )rs ; rb2 = μ2 rs ; rc1 = (1 − μ3 )rs ; rc2 = μ3rs ;
va2 = −ra f i a f ; vb2 = −rb f i b f ; vc2 = −rc f i c f ; (14)
va = va1 + va2 ; vb = vb1 + vb2 ; vc = vc1 + vc2 ; (15)
Fig. 3 Modified Schematic

of a synchronous generator
with fault in all phases
da1
va1 = −ra1 i a1 + (16)
dt
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
vsabc −Rs O μabc Rs i sabc ψsabc
⎣vrabc ⎦ = ⎣ O Rr O ⎦ ∗ ⎣ irabc ⎦ + d ⎣ ψrabc ⎦ (17)
0 −μabc Rs O μabc Rs + R f i f abc dt ψ
f abc
where
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ψa ψfd ψa2 ia
ψsabc = ⎣ψb ⎦ ; ψrabc = ⎣ ψkd ⎦ ; ψ f abc = ⎣ψb2 ⎦ ; i sabc = ⎣i b ⎦ ;
ψc ψkq ψc2 ic
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
i fd ia f va
irabc = ⎣ i kd ⎦ ; i f abc = ⎣i b f ⎦ ; vsabc = ⎣vb ⎦ ;
i kq ic f vc (18)
⎡ ⎤ ⎡ ⎤
vfd 0
vrabc = ⎣ vkd ⎦ ; 0 = ⎣0⎦ ; Rs = diag[rs rs rs ]; Rr = diag[r f d rkd rkq ];
vkq 0
R f = diag[r f r f r f ]; μabc = diag[μa μb μc ]; O is 3X3 nullmatrix
⎡ ⎤ ⎡ ⎤⎡ ⎤
vsdq0 P O O −Rs O μdq0 Rs
⎣vr dq0 ⎦ = ⎣ O I3 O ⎦ ⎣ O −Rr O ⎦∗
0 O O P −μdq0 Rs O μdq0 Rs + R f dq0
⎡ −1 ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤ (19)
P O O i sdq0 P O O ψsabc
⎣ O I3 O ⎦ ⎣ ir dq0 ⎦ + ⎣ O I3 O ⎦ d ⎣ ψrabc ⎦ ;
O O P −1 i f dq0 O O P dt ψ f abc
Where μdq0 = P −1 diag[μa μb μc ]P; R f dq0 = diag[r f r f r f ] (20)
Where
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ψa ψfd ψa2 i sd
ψsabc = ⎣ψb ⎦ ; ψrabc = ⎣ ψkd ⎦ ; ψ f abc = ⎣ψb2 ⎦ ; i sdq0 = ⎣i sq ⎦ ;
ψc ψkq ψc2 i s0
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ifd id f vsd vfd 0
ir dq0 = ⎣ i kd ⎦ ; i f dq0 = ⎣i q f ⎦ ; vsdq0 = ⎣vsq ⎦ ; vr dq0 = ⎣ vkd ⎦ ; 0 = ⎣0⎦ ;
i kq i0 f vs0 vkq 0
(21)
⎡ ⎤ ⎡ ⎤⎡ ⎤
vsdq0 −P Rs P −1 O Pμdq0 Rs P −1 i sdq0
⎣vr dq0 ⎦ = ⎣ O −Rr O ⎦ ⎣ ir dq0 ⎦
−1 −1
0 −Pμdq0 Rs P O P(μdq0 Rs + R f dq0 )P i f dq0
⎡ ⎤ ⎡ ⎤ (22)
P O O ψ
d ⎣ sabc ⎦
+ ⎣ O I3 O ⎦ ψrabc ;
O O P dt ψ f abc
simplifying further the Eq. (22) is obtained. The derivative part of the Eq. (22) can
be simplified to an expression as given in Eq. (23)
⎡ ⎤ ⎡ ⎤⎡ ⎤
vsdq0 −P Rs P −1 O Pμdq0 Rs P −1 i sdq0
⎣ vr dq0 ⎦ = ⎣ O −Rr O ⎦ ⎣ ir dq0 ⎦
0 −Pμdq0 Rs P −1 O P(μdq0 Rs + R f dq0 )P −1 i f dq0
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ (23)
ψ Ṗ P −1 O O ψsdq0
d ⎣ sdq0 ⎦ ⎣
+ ψr dq0 + O O O ⎦ ⎣ ψr dq0 ⎦ ;
dt ψ
f dq0 O O Ṗ P −1 ψ f dq0
.
From the parks transformation matrices, the value of P P −1 can be obtained as below
in Eq. (24)
⎡ ⎤
.
− sin θ − sin θ − 2π − sin θ + 2π
2 ⎣ 3 3
⎦∗
P P −1 = ω − cos θ − cos θ − 2π 3
− cos θ + 2π3
3
0 0 0
⎡ ⎤ (24)
cos θ 2π −sin θ 1
⎣cos θ − − sin θ − 2π 1⎦
3 3
cos θ + 3 − sin θ + 2π
2π
3
1
Using standard identities [14], we can simplify the Eq. (24) as in Eq. (25). Using Eq.
(25) in Eq. (23), the Eq. (26) is obtained
⎡ ⎤
.
0 ω0
P P −1 = ⎣−ω 0 0⎦ (25)
0 00
.
V = R I + ψ + Qψ (26)
where
⎡ ⎤
0 ω 0 0 0 0 0 0 0
⎢−ω 0 0 0 0 0 0 0 0⎥
⎢ ⎥
⎢ 0 0 0 0 0 0 0 0 0⎥
⎢ ⎥
⎢ 0 0 0 0 0 0 0 0 0⎥
⎢ ⎥
Q=⎢
⎢ 0 0 0 0 0 0 0 0 0⎥⎥
⎢ 0 0 0 0 0 0 0 0 0⎥
⎢ ⎥
⎢ 0 0 ω 0⎥
⎢ 0 0 0 0 0 ⎥
⎣ 0 0 0 0 0 0 −ω 0 0⎦
0 0 0 0 0 0 0 0 0
⎡ ⎤ (27)
−rs 0 0 0 0
0 μ1 rs 0 0
⎢ 0 −rs 0 0 0
0 0 μ2 rs 0 ⎥
⎢ ⎥
⎢ 0 0 −rs 0 0
0 0 0 μ3 rs ⎥
⎢ ⎥
⎢ 0 0 0 rfd 0
0 0 0 0 ⎥
⎢ ⎥
R=⎢
⎢ 0 0 0 0 rkd 0 0 0 0 ⎥
⎥
⎢ 0 0 0 0 0 rkq 0 0 0 ⎥
⎢ ⎥
⎢−μ r 0 0 μ1 rs + r f ⎥
⎢ 1 s 0 0 0 0 0 ⎥
⎣ 0 −μ2 rs 0 0 0 0 0 μ2 rs + r f 0 ⎦
0 0 μ3 rs 0 0 0 0 0 μ3 rs + r f
Substituting the value of ψ from Eq. (9) in Eq. (26) and collecting terms, the
Eq. (28) is obtained. When PU conversion is done, the form of the Eq. (28) remains
the same. The final set of differential equations or the dq0 model describing the
synchronous generator is given in Eq. (28). This set of equations can be further
rearranged to obtain the state-space model shown in Eq. (30). Here it should be
mentioned specifically that the model used here is similar to the model in [4], but the
model in the quoted work was done for a fault in a single phase the same has been
extended here for fault in all three phases using the fault parameters μ1 , μ2 and μ3 .
The next section describes the Kalman filter used in this work.
V = Z I˙ + (R + Q Z )I (28)
Fig. 4 Flux map during healthy and faulty conditions

Fig. 5 Generator winding connection with the load
Fig. 6 Plot of the stator currents actual and estimated with fault introduced at 0.8 s
Where
R + QZ=
⎡ ⎤
−rs −ω(L aq + L l ) 0 0 0 L aq ω μ1 r s μ2 ω(L aq + L l ) 0
⎢ ω(L ⎥
⎢ ad + L l ) −rs 0 −L ad ω −L ad ω 0 −μ1 ω(L ad + L l ) μ2 r s 0 ⎥
⎢ −rs μ3 r s ⎥
⎢ 0 0 0 0 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 0 rf 0 0 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 0 0 r 0 0 0 0 ⎥
⎢ d ⎥
⎢ 0 0 0 0 0 rq 0 0 0 ⎥
⎢ ⎥
⎢ −μ1 rs −μ2 ω(L aq + L l ) L aq μ2 ω ra f +μ1 rs ⎥
⎢ 0 0 0 N 0 ⎥
⎢ ⎥
⎣μ1 ω(L ad + L l ) −μ2 rs 0 −L ad μ1 ω −L ad μ1 ω 0 M ra f +μ2 rs 0 ⎦
0 0 −μ3 rs 0 0 0 0 0 ra f +μ3 rs
where M = −μ1 ω(L 0 +(L ad + L l )μ1 − L 0 μ1 ) and N = μ2 ω(L 0 +L q μ2 − L 0 μ2 )

(29)
ẋ(t) = Ax(t) + Bu(t)

Where
⎡ ⎤−1
−(L ad + L l ) 0 0 L ad L ad 0
⎢ 0 −(L aq + L l ) 0 0 0 L aq ⎥
⎢ ⎥
⎢ 0 0 −L 0 0 0 ⎥
A = −⎢ ⎢ 0 ⎥
⎢ −L ad 0 0 L f d f d L f dkd 0 ⎥ ⎥
⎣ −L ad 0 0 L f dkd L kdkd 0 ⎦
0 −L aq 0 0 0 L kqkq
⎡ ⎤
−rs −ω(L aq + L l ) 0 0 0 L aq ω
⎢ω(L ad + L l ) −rs 0 −L ad ω −L ad ω 0 ⎥
⎢ ⎥
⎢ 0 0 −r 0 0 0 ⎥
⎢ s ⎥
⎢ 0 0 0 rf 0 0 ⎥
⎢ ⎥
⎣ 0 0 0 0 rd 0 ⎦
(30)
0 0 0 0 0 rq
⎡ ⎤−1
−(L ad + L l ) 0 0 L ad L ad 0
⎢ 0 −(L aq + L l ) 0 0 0 L aq ⎥
⎢ ⎥
⎢ 0 0 −L 0 0 0 0 ⎥
B=⎢ ⎢ −L ad
⎥
⎢ 0 0 L f d f d L f dkd 0 ⎥ ⎥
⎣ −L ad 0 0 L f dkd L kdkd 0 ⎦
0 −L aq 0 0 0 L kqkq
⎡ . ⎤ ⎡ ⎤
i sd − μ1 i d f vsd
⎢i sq − μ2 i q f ⎥ ⎢ vsq ⎥
⎢ ⎥ ⎢ ⎥
⎢i s0 − μ3 i 0 f ⎥ ⎢ ⎥
x(t) = ⎢ ⎥; u(t) = ⎢ vs0 ⎥
⎢ ir d ⎥ ⎢v f d ⎥
⎢ ⎥ ⎢ ⎥
⎣ irq ⎦ ⎣ vkd ⎦
ir 0 vkq
5 Linear Kalman Filter and the Proposed Indicator
The standard linear Kalman filter has two steps, namely the prediction step and
correction step [15, 16]. In this work, the linear Kalman filter uses the state-space
model given in (30), but here the measurements are the stator currents. So to obtain
the stator currents from the state vector in the model, a joint Kalman filter with the
extended state vector as given in (31) is required. The resultant Am, Bm, and Cm,
matrices of the state-space model used by the joint kalman filter are given in (32). The
above model is a continuous-time state-space model, to obtain a discretized model,
the standard procedure for discretization is used. Here it should be noted that the new
system with the extended state vector is linear as the matrices Am, Bm, and Cm are
independent of the state vector, hence a linear Kalman filter can be used making the
system computationally efficient.
T
x(t) = i sd − μ1 i d f i sq − μ2 i q f i s0 − μ3 i 0 f ir d irq ir 0 μ1 i d f μ2 i q f μ3 i 0 f (31)
A0 B
Am = ; Bm = ; Cm = I O I (32)
0 I 0
Fault Indicator 1=μ1 i d f cos(θr ) − μ2 i q f sin(θr ) + μ3 i 0 f

Fault Indicator 2=μ1 i d f cos(θr − 2π/3) − μ2 i q f sin(θr − 2π/3) + μ3 i 0 f (33)
Fault Indicator 3=μ1 i d f cos(θr + 2π/3) − μ2 i q f sin(θr + 2π/3) + μ3 i 0 f
In [4], a fault indicator μi f s has been proposed. This indicator can detect the fault
in only one phase since, in this work, the model has been extended to all the 3 phases,
3 sets of indicators are proposed as shown in Eq. (33). To get a proper indicator a 200
pt moving RMS was taken. These indicators can thus indicate a fault in any one of
the three phases and in addition, it has been observed that these three indicators react
differently to load imbalance in the 3 phases and hence can be also used as indicators
for load imbalance. The experiments performed and the results with conclusions are
provided in the next section.
Simulation experiments were done with the fault injection capable generator that
was built in Maxwell. As shown in Fig. 5, the generator had a star connected stator,
and each phase was split to create fault injection capability as described earlier. The
stator windings were connected to a star connected resistive load. For the simulation
experiments, the value of the load current was fixed at around 1.5 A. A set of nine
experiments were conducted. One set of 3 experiments involved shorting 10 turns
of each phase (namely R, Y, and B) in turn and collecting the stator currents, stator
voltages, and field current. The next set of 3 experiments involved shorting of 6
turns for each phase and collecting the data. The final set of 3 experiments involved
introducing a load imbalance in each phase and collecting the data. Each experiment
involved running the generator for 1.6 s and fault or load imbalance was introduced
at 0.8 s. The 2D transient FE simulation was run on a High performance computing
cluster. A time step of 0.0001 s was used. The data obtained from the simulations
were given to the linear Kalman filter implemented in MATLAB. The Kalman filter
generated fault indicators that were described in the previous section
The first thing that has to be proved is the correctness of the FEM. Figure 4
shows the flux map captured during the run of the simulation in healthy and faulty
conditions. It shows clearly that asymmetry has been introduced in the flux pattern
due to the inter-turn fault. This asymmetry leads to variation in the currents observed
due to the inter-turn fault.
Next, to check the validity of the proposed dq0 model, the estimated stator currents
from the Kalman filter and the actual stator currents are plotted. It can be seen from
Figs. 6 and 7 that the Kalman filter using the dq0 model was able to track the actual
stator currents in both the healthy and faulty conditions. As a further confirmation, the
frequency spectra of the actual stator currents and the estimated stator currents were
obtained and the same is shown in Figs. 8 and 9. The Fig. 8 contains the frequency
spectra of the stator currents for the no-fault case both estimated and actual infact
the estimated and actual current spectra overlap perfectly. The Fig. 9 contains the
frequency spectra of the stator currents for 10 turn fault in R-phase, both estimated
and actual, here too both the spectra overlap confirming that the dq0 model was able
to exactly capture the features of the system. Further, by comparing Figs. 8 and 9,
Fig. 7 Zoomed version of Fig. 6

Fig. 8 Frequency spectrum of the stator currents estimated and actual No-fault case (a) R-phase
(b) Y-phase (c) B-phase
Fig. 9 Frequency spectrum of the stator currents estimated and actual With 10 turn fault in R-phase
(a) R-phase (b) Y-phase (c) B-phase
we see that there is a significant change in the third harmonic component due to fault
confirming that the FE simulation data confirms the theoretical behavior of inter-turn
faults [17].
Next, the data from the 9 sets of simulation experiments mentioned earlier were
fed to the Kalman filter and the indicators developed in Eq. (33) were plotted. The
results are shown in Figs. 10, 11, and 12. Figure 10 is a subplot showing the three
Fig. 10 Fault indicators with 6 turn fault at 0.8 s introduced in (a) R-phase (b) Y-phase (c) B-phase
Fig. 11 Fault indicators with 10 turn fault at 0.8 s introduced in (a) R-phase (b) Y-phase (c) B-phase
indicators when a 6 turn fault is introduced in the various phases. The subplot (a)
shows the three fault indicators when 6 turn fault is introduced in R phase. It can
be seen that all the indicators increase in value from the no-fault conditions but the
Fault indicator-3 has the minimum deviation. Similarly, from the subplots(b) and
(c), it can be concluded that for a fault in the Y phase the increase will be minimum
in the Fault indicator-1, and for a fault in the B phase increase will be minimum in
the Fault indicator-2. Figure 11 leads to the same conclusions but for 10 turn fault.
Fig. 12 Fault indicators with load imbalance at 0.8 s introduced in (a) R-phase (b) Y-phase (c)
B-phase
Figure 12 (a) shows the indicators when there is a load imbalance in the R phase
where it can be observed that almost all the indicators reduced in magnitude. The
maximum reduction occurs in the Fault indicator-1. Similarly, from subplots (b) and
(c), it can be concluded that imbalance in the Y phase will produce a maximum
decrease in Fault indicator-2 and imbalance in the B phase will produce a maximum
decrease in Fault indicator-3. Hence, we can conclude that this novel indicator can
identify an inter-turn fault and at the same time distinguish the condition from a load
imbalance condition.
7 Conclusions and Future Work
Considering the importance of stator inter-turn fault diagnosis in the reliable oper-
ation of a synchronous generator, this work proposed an improved way to model a
generator so that fault can be diagnosed in all three phases. Further, fault indicators
are proposed to identify the phase in which the inter-turn fault has occurred. This
indicator also indicates the presence of a load imbalance which can often lead to a
faulty diagnostic conclusion regarding the presence of an inter-turn fault. This work
assumes that only one of the conditions, either the fault or load imbalance, can occur
at a given instant, it is also assumed that fault can occur in one of the phases. These
can be taken up for future research.
8 Appendix
Table 1 Terms defined

Variable Meaning
L ad mutual inductance between stator and d-axis rotor ckts
L aq mutual inductance between stator and q-axis rotor ckts
Ll leakage inductance
L fdfd self -inductance of the field winding
L f dkd mutual inductance between field and d-axis damper
L kdkd self-inductance of the d-axis damper
L kqkq self-inductance of the q-axis damper
rs armature winding resistance
rf field winding resistance
rd d-axis damper winding resistance
rq q-axis damper winding resistance
i sd d-axis stator current
id f faulty d-axis stator current
i sq q-axis stator current
iq f faulty q-axis stator current
i s0 0-axis stator current
i0 f faulty 0-axis stator current
ir d d-axis damper current
irq q-axis damper current
ir 0 0-axis damper current
vsd d-axis stator voltage
vsq q-axis stator voltage
vs0 0-axis stator voltage
vfd field voltage
vkd d-axis damper voltage
vkq q-axis damper voltage
References
1. Kliman G, Premerlani W, Koegl R, Hoeweler D (1996) A new approach to on-line turn fault
detection in ac motors. In: IAS ’96 Conference Record of the 1996 IEEE industry applications
conference thirty-first IAS annual meeting, vol 1, pp 687–693. https://doi.org/10.1109/IAS.
1996.557113
2. Gao Z, Cecati C, Ding SX (2015) A survey of fault diagnosis and fault-tolerant techniques-part
i: fault diagnosis with model-based and signal-based approaches. IEEE Trans Ind Electron
62(6):3757–3767
3. Mazzoletti MA, Bossio GR, De Angelo CH, Espinoza-Trejo DR (2017) A model-based strategy
for interturn short-circuit fault diagnosis in pmsm. IEEE Trans Ind Electron 64(9):7218–7228.
https://doi.org/10.1109/TIE.2017.2688973
4. Nadarajan S, Panda SK, Bhangu B, Gupta AK (2016) Online model-based condition monitoring
for brushless wound-field synchronous generator to detect and diagnose stator windings turn-
to-turn shorts using extended kalman filter. IEEE Trans Ind Electron 63(5):3228–3241
5. Aubert B, Regnier J, Caux S, Alejo D (2015) Kalman-filter-based indicator for online interturn
short circuits detection in permanent-magnet synchronous generators. IEEE Trans Ind Electron
62(3):1921–1930. https://doi.org/10.1109/TIE.2014.2348934
6. El Sayed W, Abd El Geliel M, Lotfy A (2020) Fault diagnosis of pmsg stator inter-turn fault
using extended kalman filter and unscented kalman filter. Energies 13(11). https://doi.org/10.
3390/en13112972
7. Idrissi I, El Bachtiri R, Chafouk H, Khanfara M (2019) Fault diagnosis of stator inter-turn short
circuit in doubly fed induction generator of wind turbine. In: 2019 6th international conference
on control, decision and information technologies (CoDIT), pp 1313–1318. https://doi.org/10.
1109/CoDIT.2019.8820697
8. Jaiswal V, Wang D, Kumar PS (2020) Winding fault diagnosis and failure prognosis technique
for brushless synchronous generator. Electr Power Compon Syst 48(11):1159–1170. https://
doi.org/10.1080/15325008.2020.1831652
9. Gopinath R, Nambiar T, Abhishek S, Pramodh SM, Pushparajan M, Ramachandran K, Kumar
CS, Thirugnanam R (2013) Fault injection capable synchronous generator for condition based
maintenance. In: 2013 7th international conference on intelligent systems and control (ISCO),
pp 60–64. https://doi.org/10.1109/ISCO.2013.6481123
10. Naveen P, Praveen Kumar N, Sriganesh K, Rajesh T, Sushmitha K (2019) Stator fault analysis of
permanent magnet synchronous motor using finite element method. In: 2019 4th international
conference on recent trends on electronics, information, communication technology (RTEICT),
pp 655–660. https://doi.org/10.1109/RTEICT46194.2019.9016771
11. Prasob K, Kumar NP, Isha T (2017) Inter-turn short circuit fault analysis of pwm inverter fed
three-phase induction motor using finite element method. In: 2017 international conference on
circuit, power and computing technologies (ICCPCT), pp 1–6. IEEE.
12. Sawhney A (2014) Electrical machine design, 6th edn. Dhanpat Rai & Sons, New Delhi (2014)
13. Chen Y, Wang L, Wang Z, Rehman AU, Cheng Y, Zhao Y, Tanaka T (2015) Fem simulation and
analysis on stator winding inter-turn fault in dfig. In: 2015 IEEE 11th international conference
on the properties and applications of dielectric materials (ICPADM), pp 244–247. IEEE
14. Krause PC, Wasynczuk O, Sudhoff SD (2013) Analysis of electric machinery and drive systems,
3rd edn. IEEE Press, New York (2013)
15. Simon D (2006) Optimal State Estimation: Kalman, H Infinity, and Nonlinear Approaches.
Wiley-Interscience, USA
16. Nag PS, Kumar CS, Thampatty KS, Isha T (2018) A modified approach for application of
augmented extended kalman filter for stator interturn fault diagnosis of a synchronous generator.
In: 2018 4th international conference on electrical energy systems (ICEES), pp 465–469. https://
doi.org/10.1109/ICEES.2018.8442354
17. Sottile J, Trutt F, Leedy A (2006) Condition monitoring of brushless three-phase synchronous
generators with stator winding or rotor circuit deterioration. IEEE Trans Ind Appl 42(5):1209–
1215. https://doi.org/10.1109/TIA.2006.880831
Grid Integration of Wind Energy
Conversion System Through Parallel
Inverter Topology for Power Quality
Enhancement
M. E. Aswathi, C. M. Nirmal Mukundan, M. V. Manoj Kumar,

and P. Jayaprakash
1 Introduction
Development of renewable energy sources (RES) has been enlarged in recent years
[1]. Among these RES wind energy is one of the widest rising technologies. Because
of their high energy efficiency, wind energy conversion systems (WECS) with
variable-speed and its maximum power point tracking (MPPT) control have attracted
a lot of attention [2]. The efficient operation of wind energy conversion system is
dependent on the proper design of power converters and controllers (WECS) [1].
The intermittent nature, as well as equipment damage due to abrupt source and load
variations, are challenging the reality of wind energy converters. For regulating volt-
age, frequency, real and reactive powers, and MPPT, a diode rectifier with a DC-DC
converter is used [3].
Voltage source inverter (VSI) is used to regulate voltage and frequency, load
balancing and leveling, and harmonics mitigation at PCC. To attach to the grid, it
must first filter the harmonic content for the feeding of a sinusoidal current to the grid.
The grid-connected WECS offers additional functions such as load compensation,
power quality enhancement, active, and reactive power flow control with suitable
control strategies. Since the entire operation is performed by a single inverter at a
time, the inverter would be burdened [3].
Parallel inverter technology is a fine and reliable scheme and some favor points of
parallel scheme are briefly defined, such as how each inverter’s design, development,
installation, maintenance become much more simple and adaptable [3]. The current
study looks at how parallel inverters can be used to compensate for reactive and
harmonic currents. The following are some of the benefits of the suggested method:
(i) using a lower switching frequency or reducing current ripple, (ii) because of the
shared connection, there will be less stress on the DC link capacitor, (iii) paralleling
M. E. Aswathi (B) · C. M. Nirmal Mukundan · M. V. Manoj Kumar · P. Jayaprakash

Department of Electrical and Electronics Engineering, Government College of Engineering
Kannur, Kannur, Kerala, India
622 M. E. Aswathi et al.
allows for more effective performance for high-power applications, (iv) use of smaller
passive components and various modulation techniques to minimize electromagnetic
interference (EMI).
2 System Configuration
The schematic diagram of the grid-tied wind energy conversion system is shown
in Fig. 1a. A wind turbine unit is connected with a permanent magnet synchronous
generator. The output voltage of PMSG is transformed from AC to DC by operating
an uncontrolled diode bridge rectifier. The output side of the rectifier is connected
to a DC link capacitor. The main action of this capacitor is to reduce the ripples
in the signal and also helps to keep a constant DC link voltage. Employment of a
DC-DC converter that is a boost converter helps to increase the voltage magnitude
for grid integration and also tracks maximum power point. A perturb and observe
(P&O) algorithm controls the boost converter to extract the maximum power point.
Rather than sensing the wind speed, a sensorless MPPT technique is implemented.
Electrical parameters such as DC link voltage and DC link current are sensed here,
and the duty ratio is varied by using the P&O algorithm. This technique helps to bring
out the maximum wind power and also promotes keeping the power co-efficient C p
[4]. This paper deals with a parallel inverter structure that is used to increase the
performance of the system and empowers the grid to exchange the power generated
by the wind energy conversion system. The main purposes of parallelly coupled
inverters (inverter 1 and inverter 2) are active power injection and reactive power
compensation, respectively. This topology helps to reduce the burden of a single
inverter and increase the life span of inverters. Synchronous reference frame theory
(SRFT)-based control algorithms are used to function the grid-tied parallel inverter
scheme. The design of DC-DC converter and parameters selection are employed [5]
and [6].
2.1 Wind Turbine Unit (WTU)
Wind turbine rotor blades are used for the transformation of wind force to mechanical
power. PMSG is used to transform mechanical power into electrical power. Through
a gearbox, WTU is connected to the PMSG shaft.
The following is the mathematical relationship that shows the extraction of
mechanical power from the wind:
ρA 3
Pt = C p (λ, β) V (1)
2
Grid Integration of Wind Energy Conversion System Through … 623
Fig. 1 a Schematic diagram of grid-connected WECS b SRF control of parallel inverter for WECS
where Pt is the extracted power from the wind, ρ is the air density (kg/m 3 ), and C p
represents the power co-efficient, which is a function of both tip speed ratio λ and
blade pitch angle β (deg) [1].
The power co-efficient C p is given by
C2 C5
C p (λ, β) = C1 ( − C3 β − C4 ).ex p( ) (2)
λi λi
1 0.035 −1
λi = [( − 3 )] (3)
λ + 0.089 β + 1
The coefficients of the wind turbine characteristic C1 to C5 are C1 = .5, C2 = 98,

C3 = .4, C4 = 5, C5 = 16.5.
Tmech is the rotor mechanical torque can be expressed as Tmech = ( Pmech
ωr
) ωr is the
rotor angular velocity.
3 Control Configuration of WECS
Converter and inverter control are critical for the operation of the grid integrated
WECS. For the control of boost converter, P&O algorithm-based MPPT control
is used, and it extracts the maximum available wind power through the WTU. To
realize the grid synchronization through the inverter, an SRF-based control technique
is used. These control algorithms are explained in this section. Figure 1b shows block
diagram representation of the inverter control.
3.1 MPPT Control
The perturb and observe algorithm is the widely used MPPT technique [7]. This
algorithm produces the duty ratio that is used to operate the converter. A wind speed-
sensorless MPPT technique is implemented instead of conventional MPPT logic.
Here, electrical parameters like DC link voltage and current are measured for boost
converter control.
3.2 Parallel Converter Control
The topology is made up of two identical three-phase PWM inverters coupled in

parallel. A DC-capacitor that is shared for the inverters [8–11]. Figure 1b shows the
SRF-based control diagram of the parallel inverter scheme. Inverter 1 and inverter 2
are two inverters that are connected in parallel, and SRF control is used to generate
gating pulses. Inverter 1 is used to inject active power, and inverter 2 is used to com-
pensate for the reactive power and harmonics. The corresponding control techniques
are applied to those inverters.
For inverter 1, first find out the difference of reference DC link voltage (Vdcr e f )
and DC link voltage (Vdc ). Vdcr e f is considered as 750 V [4]. Only the direct axis
component is used to control inverter 1, while the quadrature axis component is
ignored. As a result, active power injection takes place. As a result of the d-axis
component being converted from dqo to abc frame, reference inverter current is
generated. The reference inverter current is compared with current through inverter
1, the gating pulse for controlling inverter 1 is created.
For inverter 2, the three-phase load current (i L ) is extracted and converted from
abc to dq0 frames. A low-pass filter, which helps to extract the low-frequency compo-
nent from the load current and gives out the fundamental component. The reference
fundamental component is generated by converting the fundamental DC component
from dq0 to abc frame. The subtraction of a fundamental component from the load
current aids in the development of harmonics and reactive power. These harmonics
and reactive power are compared with current through inverter 2, the gating pulses
for controlling inverter 2 are produced [12–15].
The wind energy conversion system has been simulated in the MATLAB/Simulink
for various conditions such as active power injection, reactive power and harmonic
compensation, sag, swell, load unbalance whose results are listed below. Wind speed
is continuously changed from 12 to 10 m/s.
Grid Integration of Wind Energy Conversion System Through … 625
Fig. 2 Waveforms of grid-connected WECS: a Power characteristics of WECS under steady state
operation b Performance under the varying load c Performance at grid voltage sag and swell
As the wind speed changes, wind power is also changed. Figure 2a shows the
power characteristics of WECS under steady-state condition.The power co-efficient
(C p ) is a measure of a system performance. C p ranged from 0.41 to 0.49 depending
on the design. The plot value of C p is around 0.4. From this figure, it is clear that input
power is considered as 40kW and it is given by the wind turbine unit. By observing
the grid power, it indicates that the power is injected to the grid.
The performance characteristics of WECS under steady-state operation and
together the unbalanced load condition is shown in Fig. 2b. Even at the load is unbal-
anced, the grid currents must be balanced. Any one of the inverters can generate an
unbalanced current in order to achieve balanced grid currents. Inverter 2 assists in
the processing of unbalanced currents and ensures that grid currents are balanced.
During this time, inverter 1 will inject active power without causing any disturbance.
The inverter 2 of WECS compensates for the unbalance in the load which in turn
gives the balanced sinusoidal grid currents. So during the unbalanced load condition,
a satisfied performance is achieved without altering the maximum extracted power
point of the wind energy conversion system.
Grid side voltage variations were applied in sequence of sag, normal voltage, and
then swell. Sag is applied for 0.3 to 0.35 s, then normal grid voltage occur. Then
swell was applied for the next seconds as shown in Fig. 2c. Grid side currents rise as
the voltage difference increases in a voltage sag, and vice versa in a swell state.
During the entire operation, parallel inverter system maintains the grid voltage
and current quality within the IEEE 519 standard even at adverse load condition.
5 Conclusion
A parallel inverter topology-based wind energy conversion system has been presented
in this work. The active power injection from the wind turbine unit has been achieved
along with power quality improvement. The system performances are tested at vari-
ous operating conditions such as grid voltage sag, swell and reactive nonlinear and
unbalanced loads. Wind turbine unit with speed-sensorless MPPT technique is used
to enhance the power capturing capacity of the whole system. The theoretical claims
are validated with simulation proofs with MATLAB/Simulink software. The wind
energy conversion system is well suited for the long-term use of renewable energy
grid integration. One of the major failure reasons of WECS is the malfunction of the
mechanical anemometer, which is avoided here. Moreover, the active power injection
and nonlinear power support are achieved through different inverters, increases the
service life of the system.
References
1. George T, Pychadathil J (2020) Maximum power tracking and power sharing in grid connected
WECS using modified PFC rectifier and PR controlled inverter. Electr Power Compon Syst
48(14–15):1584–1597
2. Ibrahim AM, Gawish SA, El-Amary NH, Sharaf SM (2019) STATCOM controller design and
experimental investigation for wind generation system. IEEE Access 7:150453–150461
3. Asiminoaei L, Aeloiza E, Kim JH, Enjeti P, Blaabjerg F, Moran L, Sul SK (2006) Parallel
interleaved inverters for reactive power and harmonic compensation. 1–7
4. Busawon K, Jovanovic M, Dodson L (2006) Power coefficient estimation in a wind energy
conversion system. 1873–1877
5. Hasaneen BM, Elbaset Mohammed AA (2008) Design and simulation of DC, DC boost con-
verter. In: 12th international middle-east power system conference, vol 2008. Aswan, Egypt,
pp 335–340
6. Sridhar JP, Prakash R (2017) Power quality issues and its mitigation by unified power quality
conditioner. 887–894. https://doi.org/10.21647/ICCTEST/2017/49081
7. Thongam J, Ouhrouche M (2011) MPPT control methods in wind energy conversion systems.
https://doi.org/10.5772/21657
8. Broeck H, Boeke U (1998) A simple method for parallel operation of inverters. 143–150
9. Abuhamdia A (2018) Grid connected three phase inverter. 1–18
10. Zhong Q-C, Hornik T (2012) Parallel PI voltage-current control of a neutral leg. In: Control of
power inverters in renewable energy and smart grid integration, IEEE, pp 219–231
11. Zhong Q-C, Hornik T (2012) Parallel operation of inverters. In: Control of power inverters in
renewable energy and smart grid integration, IEEE, pp 297–333
12. Wu B, Narimani M (2017) PWM current source inverters. In: High-power converters and AC
drives, IEEE, pp 225–256
13. Tuladhar A, Jin H, Unger T, Mauch K (2000) Control of parallel inverters in distributed AC
power systems with consideration of line impedance effect. IEEE Tran Ind Appl 36(1):131–138
14. Zhang C, Guerrero JM, Vasquez JC, Coelho EAA (2016) Control architecture for parallel-
connected inverters in uninterruptible power systems. IEEE Tran Power Electr 31(7):5176–
5188
15. Jin L, He Z, Zhang Y, Nie J, Zhang M (2016) A new virtual impedance method for parallel
inverters with droop control. IEEE Vehicle Power Propuls Conf (VPPC) 2016:1–4
Improvement in Quality of Power
of Grid-Connected Wind Energy System
Through STATCOM—A Review
Manisha Priyadarsani Nayak, Suresh Kumar Gawre, and Shailendra Kumar
1 Introduction
Wind energy is known to be a clean and environmentally friendly renewable energy

source as it reduces our dependency on fossil fuels. Nowadays, it has become very
famous and is being utilized in every sector as it does not emit any harmful gases such
as CO2 [1]. The global wind report 2019 highlights a positive growth of wind energy
having a total installed capacity of 650GW and the wind installation surpassed 60GW
in 2019. The individual units present in the distribution network have a capacity of
about 2 MW. Nowadays, more than 3,41,320 wind turbines are successfully working
worldwide. From the total number; 14,384 offshore wind turbines are there across
the world.
When a wind energy system is coupled with an existing power system, it creates
many difficulties. Wind turbine gives rise to a changeable and irregular output power
due to the fluctuation in wind speed, and it results in fluctuation in the electrical power
and in voltage [2]. The power quality has a vital role in the power system and it is
greatly affected by the addition of renewable energy like wind into the grid. Power
quality is defined as “Any issue and quality degradation of voltage, current, and
frequency which results in improper functioning or operational error in equipment”
[3].
These issues of power quality can be classified as voltage swell, sag, voltage dips,
flickers, harmonics, variation in real and reactive power, etc. Induction generators are
the most preferable machine to run a wind generating system due to these advantages.
• Simple and rugged construction
• Low maintenance
• Cheaper as compared to other motors
• Self-starting nature
M. P. Nayak (B) · S. K. Gawre · S. Kumar

Department of Electrical Engineering, Maulana Azad National Institute of Technology, Bhopal,
Madhya Pradesh, India
628 M. P. Nayak et al.
To establish magnetic flux, reactive power is required by an induction gener-

ator. The active power, reactive power and terminal voltage vary a lot, and system
gets disturbed due to the varying wind nature. So to balance the reactive power,
the application of FACTS device is introduced. Here, we will use a VSC-based
STATCOM to improve the power system stability and performance as it has much
better performance and transient response than the static var compensator [4]. A
battery energy storage (BESS) is needed to manage the fluctuations caused by wind
turbines. Importance of STATCOM in a grid-connected system is discussed below
[5]:
• Power factor improvement(unity) at source side
• Control of dynamic voltage, active power and reactive power with help of BESS
• Power oscillation damping can be improved in system
• Voltage flicker control issue can be resolved
• Fast dynamic response due to Bang Bang controller
• No requirement of additional shunt reactor and capacitor bank
In paper [1], the power quality issue, consequences, grid coordination rule,
power quality improvement and its topology, control scheme, and its performance
are discussed. In paper [2, 3, 5], the voltage unbalance, voltage fluctuation, inter
harmonics and controller of STATCOM are discussed. Paper [4] describes the
topology for power quality improvement, types of controller of STATCOM, i.e.
Bang Bang controller and fuzzy logic controller. In paper [6, 7, 8], the power quality
issue, role of STATCOM and equivalent representation are discussed. Paper [9] repre-
sents the role of different FACT devices in improving power quality in wind energy
systems. Paper [10, 11] describe the use of STATCOM and PID controller for stability
enhancement in grid-connected wind farms.
Using D-STATCOM, how the power quality is improved in a hybrid power system
is discussed in [12, 13, 14]. In paper [15, 16], power quality improvement using a
doubly-fed induction generator is discussed. This generator faces severe voltage
instability problems during grid faults, and STATCOM is mainly helpful to resolve
this problem [17]. We can also compensate for the real and reactive power without
adding any type of storage system [18, 19].
Power quality improvement by new techniques of STATCOM is developing day
by day is discussed in [20, 21]. The reactive power flow management in renewable
energy with low and high voltage ride through is discussed [24]. Using sinusoidal
PWM technique, the voltage sag and swell of D-STATCOM can be controlled very
easily and getting popular nowadays [25]. To maintain the voltage within the limit
mainly corrective, coordinated and preventive voltage control methods are used [27].
By using cascaded STATCOM, the efficiency to improve the voltage stability will
increase very much [29].
Huge energy storage devices are needed in case of more fluctuation so to solve
this problem STATCOM with a low pass filter can be used for the generation of line
power reference [33]. Hybrid energy storage system along with supercapacitor can
also be used to smooth the variation of wind as it has a longer life cycle [34].
Improvement in Quality of Power of Grid-Connected Wind Energy … 629
Wind farm connected with permanent magnet synchronous generator needs a

STATCOM as it reduces the chance of losing a great portion of wind generation
[36]. To analyze the performance of static compensator; a proportional resonant
controller can be connected to it [37, 38]. Sub-synchronous resonance (SSR) is
the major problem of wind power systems and STATCOM is very much efficient to
mitigate this problem. It increases the grid stiffness by decreasing the grid impedance
and resolving the SSR issue [39].
But nowadays, UPFC is getting more popular than STATCOM because UPFC
can improve the low voltage ride through (LVRT) better than STATCOM [40].
Contents and the summary of all the papers are discussed in Table 1.
2 Issues in Power Quality
2.1 Voltage Fluctuation
Voltage variation takes place due to error in pitch, error in yaw and fluctuation of
wind speed. It is related to variations in real and reactive power. Fluctuation in voltage
should be within + /5%. The classifications in voltage variation are [1]:
• Voltage swell
• Voltage sag
• Voltage dips
• Long duration voltage variation
• Voltage variation in short duration
Frequency range for voltage fluctuation is 10–35 Hz.
2.2 Harmonics
Harmonics results due to the addition of converter and nonlinear load into the system
[2]. Effect of harmonics is a major cause to degrade the power. As per the IEC-61400-
36 guideline, the harmonic voltage should be within limit and each harmonic current
source can allow only a limited value. The continuous operation of switch reduces
lower order harmonics, but high frequency current is present in output current, which
can be filtered out easily.
Table 1 Summary of papers

Reference paper Contents discussed Summary
[1] STATCOM based on Bang Bang • Cancel out the load current with
controller to improve power quality harmonics and improve the power
in grid factor
• No limit to the switching frequency
is a great disadvantage
[2, 3] Hysteresis current-controlled • Provides fast switching signal to the
technique is used STATCOM
• The hysteresis error is more than the
maximum error
[4] Comparison of Bang Bang controller • THD in the Fuzzy logic controller is
and Fuzzy logic controller is less than the Bang Bang controller
discussed • Fuzzy logic controller is more
simple and faster than Bang Bang
controller
[5, 8] Power quality issues and • Voltage sag, swell, harmonics are
consequences are discussed the major issues in wind energy
systems connected to grid
[6, 7] Power quality improvement using • STATCOM, UPFC, IPFC and SSSC
different FACTS devices play an important role in better
power quality
[9] Power quality improvement • The System connected to UPQC
comparison between STATCOM and provides better performance than
UPQC STATCOM, because THD in UPQC
is less than STATCOM
[10, 11] STATCOM and PID controller-based • PID controller and PWM technique
stability enhancement are used for control of STATCOM
[12, 13, 14] Active power compensation using D • D-STATCOM reduces the power
STATCOM is discussed loss and improves the stability of
system
[16, 17] Power quality improvement in • STATCOM and BESS help in
doubly-fed induction generator. - compensation of generator rotor
angle oscillation
• It improves the overall stability and
dynamic behavior
[19] Power quality improvement without • A system without any storage
adding a storage system system is less effective than system
having a storage system because
BESS stores the energy and
stabilizes the system
(continued)
Table 1 (continued)
Reference paper Contents discussed Summary
[20–23] Different control techniques of • LVRT method, nonlinear adaptive
STATCOM are discussed coordinating neuro control, optimal
power flow, resonant regulators and
DFT synchronization algorithm are
very popular control methods of
STATCOM
[25] Modeling and simulation of • VSC is developed using the SPWM
D-STATCOM using SPWM technique
technique • D-STATCOM provides good
voltage regulation and robust
[26, 2728] Application of STATCOM in wind • To avoid number of STATCOM;
energy generated system and optimal optimal voltage control can be used.
voltage control of it It is cost- effective. This is
applicable in the offshore cable
system
[29, 30] Wind Farm voltage stability using • Cascaded STATCOM is quicker and
cascade STATCOM is discussed more reliable and has better
efficiency
[31, 32] Stabilization of multi-machine power • In a multimachine power system,
system using STATCOM STATCOM reduces the frequency
oscillation
[33] Regulation of terminal voltage using • A low pass filter when integrated
STATCOM with STATCOM increases the
power factor
[34] Role of HESS in wind power • Super capacitor with hybrid energy
generation storage system reduces variation of
wind
[35, 36] Use of STATCOM with high • In permanent magnet synchronous
penetration of wind generation generator, STATCOM helps to
improve power
[38] Mitigating power quality issues by • Resonant controller can analyze the
STATCOM with resonant controller performance of static compensators
[39] Control strategy of STATCOM to • SSR can be reduced by increasing
improve sub-synchronous resonance the grid stiffness. By adding
(SSR) STATCOM, the grid impedance
decreases and stiffness improves. So
SSR can be reduced
[40] Role of Cascaded H Bridge (CHB) • CHB STATCOM can maintain the
STATCOM grid voltage by regulating the
Negative sequence Output
Admittance (NSOA)
2.3 Reactive Power
When induction generator and capacitor used for compensation of reactive power are
added to the wind energy; the wind turbine may get self-excited, after disconnecting
the wind turbine from the nonlinear load. The capacitor connected helps in reactive
power compensation.
2.4 Switching Operation
Voltage fluctuation such as voltage sag swell caused due to switching operation.
According to IEC 61,400-3-7, the switching operation should be within 10 min to
2 h.
3 Consequences of Issues
The voltage deviation and dips, transient, flicker, harmonics make the following
dysfunction of equipment [5].
• Tripping of protection devices and contractors
• Computer, PLC system may get stop and the process may stop working and even
can damage many pieces of equipment.
• Due to these issues power quality degrades in the grid.
4 Components of Grid-Connected System to Improve

Power Quality
The wind energy system connected to grid is shown in Fig. 1, consists of a nonlinear
load, induction generator, wind generating system and STATCOM with a shunt
connected BESS.
To settle issues involved in power quality and for refinement of power quality,
STATCOM along with its control scheme and with a BESS is connected. The
STATCOM provides a current that will nullify the reactive and harmonic compo-
nent of nonlinear load and induction generator current so that the source current is
free from harmonics, and it has a desired value of phase angle with the STATCOM
and the power factor improves, and hence quality of power also improves in the
system.
Fig. 1 Grid-connected wind energy System with its component [3]
Vac
AC System
Coupling Transformer
Iac
Voltage Source AC Terminal

Converter
Idc
Vdc DC terminal
EnergySource
Fig. 2 Block diagram of STATCOM [1]

4.1 STATCOM and BESS
The STATCOM shown in Fig. 2, is a three-phase voltage source converter connected

at the common coupling point. The voltage source converter is connected with a DC
capacitor source and converts the DC energy to AC energy, which is again fed to the
coupling transformer. This coupling transformer is a constant frequency device and
it transfers the AC energy from STATCOM to the AC system line and neutralizes
harmonics. When the control scheme gives a switching signal to the STATCOM,
it will supply a harmonic free current (I) of changing magnitude and frequency
component. Also a harmonic filter is present in a STATCOM that attenuates the
harmonics and high-frequency component [18].
The battery energy storage system (BESS) and DC capacitor are connected in
parallel and it helps in storage of energy. The main functions of BESS are:
• To stabilize the grid system, BESS injects the reactive power to the system.
• In case of need when the reactive power of system is less than the STATCOM, it
supplies power and it absorbs power when reactive power of system is more than
static compensator; by this charging and discharging operation.
• It balances the voltage regulation.
• It maintains the DC capacitor voltage constant.
• It controls the transmission and distribution system very rapidly.
4.2 Wind Energy Generating System
Wind turbines in wind generating system have a great role as it turns the kinetic
energy into clean electrical energy. Wind generating system has a rectifier which
converts the AC to DC power and that DC energy stored in battery and supplied to
controller in case of need. Then inverter converts the DC energy to AC energy. This
energy is supplied to the ac system box or grid. When wind turbine blades rotate due
to flow of wind the rotor captures the kinetic energy and converts this mechanical
energy to electrical energy. The power equation of wind energy system is formulated
as [9] (Fig. 3).
pwind = 0.5ρAV 3 (1)
Wind turbine Recifier(AC-DC) Controller Inverter(DC-AC) AC BOX

generator(AC)
Baery
Fig. 3 Wind energy generating system

where
ρ(kg/m3 ) = Density of air

A(m2 ) = area of turbine blade
V (mtr/s) = Speed of wind
The mechanical power can be derived from the kinetic power is given by:
pmech = Cp ∗ pwind (2)
where (Cp) is the power coefficient. Now the mechanical power generated by wind
turbine is [9]
pmech = 0.5ρπ r 2 V 3 Cp (3)
where r = Radius of blade.
4.3 System Operation
STATCOM with parallel connected battery system is connected at the point of

common coupling with induction generator and non-linear load in grid system. The
output of STATCOM is varied to maintain a good quality of power in grid system.
The hysteresis current control strategy has a crucial role in the control scheme, which
operates the switching ON/OFF mechanism of the STATCOM compensator in the
grid. The block diagram of control scheme of the system is given in Fig. 4 [5].
The current limiter used in this scheme does not allow signal above a specified
value and similarly the voltage controller limits the voltage for a predetermined value.
Source
Voltage
Unit vector
Usa,Usb,Usc
Vdc I*sabc
Sa
Sb
Sc
Isabc
Voltage controller Limiter Hysteresis current
controller
Fig. 4 Control Scheme [5]

5 Control Scheme
The control scheme used here to supply suitable harmonics free currents into the
grid is known as “Bang Bang controller”, which is also called as on–off controller. In
this controller, hysteresis current controlled technique is used, which is a nonlinear
method. This technique has a specified hysteresis area and it maintains the variable
between its boundary and gives desired signal to turn on the STATCOM to inject the
current into grid.
With the help of sensor, we need to measure the source current is and DC voltage
Vdc. The current control block that is shown in Fig. 3 receives the reference current i
sabc , which we have to calculate from the source phase voltage, and actual current isabc
and the difference between the two currents are provided to the hysteresis current
controller to start up the operation of STATCOM.
5.1 Controller of STATCOM
In a three-phase balanced system, by using the source phase voltages (Vsa, Vsb,
Vsc), the amplitude of RMS source voltage is calculated and known as sampled peak
voltage (Vsm):

V sm = {2/3(V sa2 + V sb2 + V sc2 )} (4)
By using the voltage of each phase and RMS value of unit vector, the in-phase
unit vectors are derived as shown below [3].
Usa = V sa/V sm (5)
Usb = V sb/V sm (6)
Usc = V sc/V sm (7)
Due to in-phase unit voltage template, the in-phase reference current is generated
and shown below [2].
i ∗ sa = I ∗ Usa, i ∗ sb = I ∗ Usb, i ∗ sc = I ∗ Usc (8)
where ‘I’ has a proportionate relation with the source voltage magnitude for respec-
tive phases, which indicates that the source current is sinusoidal to make the ref
current a sinusoid one.
5.2 Bang Bang Controller
The Bang Bang current controller working activates the STATCOM by providing an
error into the system, which is calculated by adjusting a variable x to some reference
value or setpoint x set [5] and keeping it there and error is calculated as
Error = X − XS (9)
If Error > HB output u = 1

If Error < HB output u = 0
To provide the error of the ref current and source current into the hysteresis
controller, the reference current is calculated using phase RMS voltage, and actual
current is sensed by a current sensor and it is compared with the hysteresis band to
turn ON/OFF the STATCOM.
To turn off the IGBT of STATCOM the condition is:
If isa < (i∗sa − HB) then SA = 0
To turn on the IGBT of STATCOM to inject a current; the condition is:
If isa > (i∗sa − HB) then SA = 1
Similarly, S B andS C can be obtained for phase b and phase c and HB can be taken
as small as possible for less ripple in the input current. Here, in this configuration,
HB can be taken as 0.08.
6 Simulation and Result
In this paper, to analyze the working of STATCOM in wind farm, wind turbines are
taken that uses squirrel cage induction generator. The design comprises six wind
turbines of 1.5 MW. A distribution system of 25 kV is connected to the wind turbine
and supplies power to a 120 kV grid through a feeder. The system is simulated using
Simulink tool of Matlab 2017B (Figs. 5, 6, and 7; Table 2).
Capacitor bank connected to the wind turbine compensates the reactive power,
but not completely. So STATCOM of 3-Mvar is there to maintain the reactive power
at bus.
To realize the importance of STATCOM, the model is simulated first without
connecting the STATCOM and then it is added to the system and run for 20 s.
From Fig. 8, we can see how the STATCOM supplies the reactive power in a wind
farm.
Fig. 5 Simulation model
Fig. 6 Active power, reactive power, wind speed and pitch angle variation at wind turbine without
STATCOM
• Figures 6 and 9 represent the waveform of active power, reactive power, wind
speed and pitch angle of wind turbine without STATCOM and with STATCOM
respectively.
• By comparing these two figures, we can conclude that the reactive power and
active power improve by using STATCOM (refer Fig. 9). Also, STATCOM is
very efficient to control the system during wind speed variation.
• Figures 7 and 10 show the waveform of voltage, current and power at the bus
without STATCOM and with STATCOM, respectively.
Fig. 7 Voltage, active power, reactive power at bus without STATCOM
Table. 2 Wind turbine

Parameters Symbol Value
specification
Base power of wind turbine S 1.5 MW
Nominal wind speed V 9 m/s
Nominal mechanical power P 1pu (3 MW)
Grid frequency F 60 Hz
Pitch angle controller gain Kp. Ki 5.25
Pitch angle (maximum) θ 45°
Simulation time T 20 s
Fig. 8 Voltage and reactive power injected by static compensator

Fig. 9 Active power, reactive power, wind speed and pitch angle variation at wind turbine with
STATCOM
Fig. 10 Voltage, active power, reactive power at bus with STATCOM
• The voltage, current and power have minimum deviation and have a smooth
conduction by using STATCOM at bus, which is shown in Fig. 10.
7 Conclusion and Future Scope
STATCOM injects current to nullify reactive and harmonic component of the load
and improves the power factor and quality of the overall system.
Hysteresis control helps in generating a switching signal for STATCOM and the
choice of hysteresis band should be narrow to improve the current quality in the
system and to reduce the ripple in the system. Hysteresis current control provides a
good reliability and stability. It has good speed and correct accuracy so it is widely
accepted. Although it is popular for its easy implementation, it has no limit to the
switching frequency; so the upper and lower limits of frequency make the hysteresis
band error very high as hysteresis band error is twice the maximum error.
By using capacitors in wind farm, the sub-synchronous resonance (SSR) increases.
So in future, we can use PV-STATCOM to mitigate this problem. In this technique,
the solar farm will act as a PV-STATCOM, which will maintain the voltage stability
and compensate the power very efficiently.
References
1. Sunil TP, Loganathan N (2012) Power quality improvement of a grid-connected wind energy
conversion system with harmonics reduction using FACTS device. In: IEEE-International
Conference on Advances in Engineering, Science and Management (ICAESM-2012), Nagap-
attinam, Tamil Nadu, pp 415–420
2. Devabalaji KR, Ravi K (2014) Power quality improvement in wind farm connected to grid
using STATCOM. In: 2014 International Conference on Advances in Electrical Engineering
(ICAEE), Vellore, 2014, pp 1–5. https://doi.org/10.1109/ICAEE.2014.6838508
3. Nirmala N, Suresh Kumar V (2013) A STATCOM-control scheme for wind energy system to
improve power quality. In: 2013 International Conference on Information Communication and
Embedded Systems (ICICES), Chennai, pp 1108–1113. https://doi.org/10.1109/ICICES.2013.
6508175
4. George SK, Chacko FM (2013) Comparison of different control strategies of STATCOM for
power quality improvement of grid connected wind energy system. In: 2013 International Mutli-
Conference on Automation, Computing, Communication, Control and Compressed Sensing
(iMac4s), Kottayam, pp 650–655. https://doi.org/10.1109/iMac4s.2013.6526394
5. Mohod SW, Aware MV (2010) A STATCOM-control scheme for grid connected wind energy
system for power quality improvement. IEEE Syst J 4(3):346–352. https://doi.org/10.1109/
JSYST.2010.2052943
6. Yuvaraj V, Raj EP, Mowlidharan A, Thirugnanamoorthy L (2011) Power quality improvement
for grid connected wind energy system using FACTS device. In: Proceedings of the Joint
INDS’11 & ISTET’11, Klagenfurt, pp 1–7. https://doi.org/10.1109/INDS.2011.6024828
7. Patel DM, Nagera AR, Roy KC (2012) Application of static compensator to improve the power
quality of grid connected induction generator based wind farm. In: IEEE-International Confer-
ence On Advances In Engineering, Science And Management (ICAESM-2012), Nagapattinam,
Tamil Nadu, pp 1–4
8. Shanthini K, Verappan N (2012) Power quality enhancement of wind generators connected
to grid. In: 2012 International Conference on Emerging Trends in Electrical Engineering
and Energy Management (ICETEEEM), Chennai, pp 398–403. https://doi.org/10.1109/ICE
TEEEM.2012.6494474
9. Amita A, Sinha AK (2018) Power quality comparison of grid connected wind energy system
with STATCOM and UPQC. In: 2018 International Conference on Intelligent Circuits and
Systems (ICICS), Phagwara, pp 355–360. https://doi.org/10.1109/ICICS.2018.00079
10. Fayek A et al (2019) STATCOM and PID controller based stability enhancement of a grid
connected wind farm. In: 2019 International Conference on Energy and Power Engineering
(ICEPE), Dhaka, Bangladesh, pp 1–4. https://doi.org/10.1109/CEPE.2019.8726728
11. Mienski R, Pawelek R, Wasiak I (2002) Application of STATCOM controllers for power quality
improvement - modeling and simulation. In: 10th International Conference on Harmonics and
Quality of Power. Proceedings (Cat. No.02EX630), Rio de Janeiro, vol 2, pp 620–625. https://
doi.org/10.1109/ICHQP.2002.1221507
12. Parija B, Behera S, Pattanayak R, Behera S (2019) Power quality improvement in hybrid power
system using D-STATCOM. In: 2019 3rd International Conference on Computing Method-
ologies and Communication (ICCMC), Erode, India, pp 564–567. https://doi.org/10.1109/
ICCMC.2019.8819656
13. Saradva PM, Kadivar KT, Pandya MH, Rana AJ (2016) Reactive and real power compensation
in distribution line using D-STATCOM with energy storage. In: 2016 International Conference
on Computation of Power, Energy Information and Communication (ICCPEIC), Chennai, pp
726–732. https://doi.org/10.1109/ICCPEIC.2016.7557316
14. Kulkarni HR, Virulkar VB (2017) Mitigation of flicker in a distribution-connected wind farm
with STATCOM. In: 2017 International Conference on Energy, Communication, Data Analytics
and Soft Computing (ICECDS), Chennai, pp 2208–2212. https://doi.org/10.1109/ICECDS.
2017.8389844
15. Lei L, Shengtie W, Guizhen T (2016) Grid power quality improvement with STATCOM/HESS
for wind turbine with squirrel-cage induction generator. In: 2016 IEEE 11th Conference on
Industrial Electronics and Applications (ICIEA), Hefei, pp 2552–2557. https://doi.org/10.1109/
ICIEA.2016.7604023
16. Ghorbanian MJ, Goodarzvand F, Poudaryaei A, Mahadi WNL (2014) Power quality improve-
ment of grid connected doubly fed induction generator using STATCOM and BESS. In: 2014 4th
International Conference on Engineering Technology and Technopreneuship (ICE2T), Kuala
Lumpur, pp 110–115. https://doi.org/10.1109/ICE2T.2014.7006229.
17. Fazli M, Shafighi AR, Fazli A, Shayanfar HA (2010) Effects of STATCOM on wind turbines
equipped with DFIGs during grid faults. In: 2010 World Non-Grid-Connected Wind Power
and Energy Conference, Nanjing, 2010. https://doi.org/10.1109/WNWEC.2010.5673205
18. Muthusamy M, Kumar CS (2014) New STATCOM control scheme for power quality improve-
ment in wind farm. In: 2014 International Conference on Green Computing Communication and
Electrical Engineering (ICGCCEE), Coimbatore, pp 1–5. https://doi.org/10.1109/ICGCCEE.
2014.6922323
19. Rekik M, Abdelkafi A, Krichen L (2013) Power quality improvement of a wind farm without
adding any type of storage system. In: 14th International Conference on Sciences and Tech-
niques of Automatic Control & Computer Engineering—STA’2013, Sousse, 2013, pp 132–135.
https://doi.org/10.1109/STA.2013.6783118
20. Aboul-Seoud T, Jatskevich J (2008) Improving power quality in remote wind energy systems
using battery storage. In: 2008 Canadian conference on electrical and computer engineering,
Niagara Falls, ON, 2008, pp 001743–001746. https://doi.org/10.1109/CCECE.2008.4564842
21. Rezaeipour R, Kiani B (2009) Review of novel control techniques for STATCOM and its
effects on a wind farm. In: 2009 International Conference on Sustainable Power Generation
and Supply, Nanjing, , pp 1–5. https://doi.org/10.1109/SUPERGEN.2009.5348152
22. Ma Y, Huang A, Zhou X (2015) A review of STATCOM on the electric power system. In:
2015 IEEE International Conference on Mechatronics and Automation (ICMA), Beijing, pp
162–167. https://doi.org/10.1109/ICMA.2015.7237475.
23. Subramanian SB et al (2018) Control of STATCOMs—a review. In: 2018 IEEE Power &
Energy Society General Meeting (PESGM), Portland, OR, pp 1–5. https://doi.org/10.1109/
PESGM.2018.8586479
24. Sarkar MNI, Meegahapola LG, Datta M (2018) Reactive power management in renewable rich
power grids: a review of grid-codes, renewable generators, support devices, control strategies
and optimization algorithms. IEEE Access 6:41458–41489. https://doi.org/10.1109/ACCESS.
2018.2838563
25. Madhusudan R, Ramamohan Rao G (2012) Modeling and simulation of a distribution
STATCOM (D-STATCOM) for power quality problems-voltage sag and swell based on Sinu-
soidal Pulse Width Modulation (SPWM). In: IEEE-International Conference On Advances In
Engineering, Science And Management (ICAESM -2012), Nagapattinam, Tamil Nadu, 2012,
pp 436–441
26. Castaneda J, Enslin J, Elizondo D, Abed N, Teleke S (2010) Application of STATCOM with
energy storage for wind farm integration. In: IEEE PES T&D 2010, New Orleans, LA, 2010,
pp 1–6. https://doi.org/10.1109/TDC.2010.5484308
27. Wu Q, Solanas JIB, Zhao H, Kocewiak ŁH (2016) Wind power plant voltage control optimiza-
tion with embedded application of wind turbines and STATCOM. In: 2016 Asian Conference on
Energy, Power and Transportation Electrification (ACEPT), Singapore, 2016, pp 1–5. https://
doi.org/10.1109/ACEPT.2016.7811534
28. Roy C, Sengupta A † (2015) Implementation of STATCOM to overcome disturbances in grid
connected Windfarm. In: 2015 international conference on energy, power and environment:
towards sustainable growth (ICEPE), Shillong, 2015, pp 1–4. https://doi.org/10.1109/EPETSG.
2015.7510150
29. Shengqing L, Wenxiang X, Weizhou, Huanyue Z (2013) Wind farm grid voltage stability
researching based on cascade STATCOM. In: 2013 3rd international conference on intelligent
system design and engineering applications, Hong Kong, 2013, pp 1610–1613. https://doi.org/
10.1109/ISDEA.2012.387
30. Bubshait AS, Mortezaei A, Simões MG, Busarello TDC (2017) Power quality enhancement for
a grid connected wind turbine energy system. IEEE Trans Ind Appl 53(3):2495–2505. https://
doi.org/10.1109/TIA.2017.2657482
31. Muyeen SM, Ali MH, Takahashi R, Murata T, Tamura J (2007) Stabilization of wind farms
connected with multi machine power system by using STATCOM. In: 2007 IEEE Lausanne
power tech, Lausanne, 2007, pp 299–304. https://doi.org/10.1109/PCT.2007.4538333
32. Chavhan ST, Bhattar CL, Koli PV, Rathod VS (2015) Application of STATCOM for power
quality improvement of grid integrated wind mill. In: 2015 IEEE 9th International Conference
on Intelligent Systems and Control (ISCO), Coimbatore, 2015, pp 1–7. https://doi.org/10.1109/
ISCO.2015.7282295
33. Sheikh MRI, Eva F, Motin MA, Hossain MA (2012) Wind generator output power smoothing
and terminal voltage regulation by using STATCOM/SMES. In: 2nd International Conference
on the Developments in Renewable Energy Technology (ICDRET 2012), Dhaka, Bangladesh,
2012, pp 1–5
34. Han Y, Xu Y, Li Y (2015) Research on application of STATCOM/HESS in wind power inte-
grated to the grid. In: 2015 IEEE 2nd International Future Energy Electronics Conference
(IFEEC), Taipei, Taiwan, 2015, pp 1–6. https://doi.org/10.1109/IFEEC.2015.7361566
35. Jamil E, Qurratulain, Hameed S (2017) STATCOM-Based Voltage Regulation in Grid Inte-
grated Wind Farm under Variable Loading Conditions. In: 2017 14th IEEE India Council Inter-
national Conference (INDICON), Roorkee, India, pp 1-6https://doi.org/10.1109/INDICON.
2017.8488062
36. Pereira RMM, Pereira AJC, Ferreira CMM, Barbosa FPM (2016) STATCOM to improve the
voltage stability of an electric power system with high penetration of wind generation. In: 2016
51st international Universities Power Engineering Conference (UPEC), Coimbra, Portugal, pp
1–5. https://doi.org/10.1109/UPEC.2016.8114142
37. Tanaka T, Ma K, Wang H, Blaabjerg F (2019) Asymmetrical reactive power capability of
modular multilevel cascade converter based STATCOMs for offshore wind farm. IEEE Trans
Power Electron 34(6):5147–5164. https://doi.org/10.1109/TPEL.2018.2866398
38. Avudai Lakshmi B, Karpagam R (2015) Resonant controller based STATCOM used in wind
farms to mitigate power quality issues. In: 2015 IEEE international conference on electrical,
computer and communication technologies (ICECCT), Coimbatore, India, 2015, pp 1–5.
https://doi.org/10.1109/ICECCT.2015.7226003
39. Li G, Chen Y, Luo A, Wang H (2020) An enhancing grid stiffness control strategy of
STATCOM/BESS for damping sub-synchronous resonance in wind farm connected to weak
grid. IEEE Trans Ind Inf 16(9):5835–5845. https://doi.org/10.1109/TII.2019.2960863
40. Peng Y et al (2021) Coordinated control strategy of PMSG and cascaded H-Bridge STATCOM
in dispersed wind farm for suppressing unbalanced grid voltage. IEEE Trans Sustain Energy
12(1):349–359. https://doi.org/10.1109/TSTE.2020.2995457
A Computational Intelligence-Based
Novel Bearing Defect Detection Method
Arpana Singh, K. R. Satyajit, and Papia Ray
1 Introduction
A machine element has a bearing that constrains the relative motion only to the
desired motion, hence decreases friction between the moving parts. Prevent motion
by controlling forces of normal vector in moving parts is one of the features of the
design of bearing. Bering design also helps in a linear movement or free rotation
around an axis. Hence, the elements of rolling are the vital components. Working
under different operational and environmental conditions is the job of rotary elements,
which leads to numerous defects over the operating surfaces of the components of
bearing. In this paper, two techniques are implemented to reduce the fault occurring
at the bearing. The fault is diagnosed by the method bispectrum analysis, EMD and
EEMD.
In this paper, it has been seen that noise is of two types to be eliminated, i.e.
Gaussian and non-Gaussian. The Gaussian noise is removed easily because of its
zero mean value but the non-Gaussian noise cannot be removed easily. Hence, two
computational intelligence methods (EMD and EEMD) have been adopted to eradi-
cate the non-Gaussian noise from the bearings. Further, it has been noticed that EMD
possesses issues like mode mixing, which makes it difficult to swipe noise from the
signal whereas EEMD and masking EMD overcome this issue. So, in this paper,
EEMD and masking EMD have been suggested to determine bearing defect.
A. Singh · K. R. Satyajit · P. Ray (B)

Department of Electrical Engineering, Veer Surendra Sai University of Technology, Burla,
Sambalpur, Odisha, India
e-mail: papiaray_ee@vssut.ac.in
646 A. Singh et al.
1.1 Literature Review
Only Gaussian can be completely removed from bispectrum analysis [1]. The higher
order spectrum of the signal gets impeded due to the presence of non-Gaussian noise.
It also adversely influences the examination of default features. Hence, the elimina-
tion of noise that is not Gaussian in the present signal is done before the bispectrum
analysis. The Ensemble Empirical Mode Decomposition (EEMD) is a technique
that is a self-adaptive signal decomposition method [2]. Like EMD, EEMD can also
break the signals into a series of higher-to-lower frequency bands [3]. The various
components of frequency of the given signal are firstly segregated then removal of the
non-Gaussian noise present in the signal is done. EEMD is one of the noise-assisted
data analysis methods that easily prevents the mode mixing process caused by EMD.
To prevent this mode, mixing phenomenon addition of white noise is done to the
signal. In contrast to the EMD, the intrinsic property of the signal can be disclosed
better in IMFs generated by EEMD. Over time, the diagnosis of the mechanical fault
is done extensively by the EEMD [4–6]. Wang et al. [7] also use EEMD with the
method of correlation dimension for distinguishing the failures of diesel engines.
This paper shows the advantages of EMD and EEMD done for detecting the noise in
the rolling bearing. Firstly, the decomposition of signal is carried out by EMD and
numerous IMFs are generated, but EMD has a mode mixing phenomena problem
so we use the FFT method to eliminate the mode mixing phenomena. Further again
EEMD is performed, and we get the true IMFs as well as false IMFs. Now a day,
audits and diagnosis of faults in the equipment are of huge discussion in each indus-
trial sector. Maintenance costs for emergencies used for the fault diagnosis saved a
huge amount of money. Rolling element bearings is an important element due to their
excessive wear and tear, which causes the prevention of failures in the whole system.
The accurate working of these elements of a machine is crucial in the industry to
completely eradicate long-term brutal accidents. To prevent machinery performance
degradation, faults, or devastating failures, online machinery condition monitoring
system is used [8]. Therefore, we need upgraded technologies to detect the perfor-
mance of bearings. Bearings are of two types, i.e. sliding and rolling bearing. Both
the bearings are called rolling bearings and are extensively seen in the elements of
machines. The bearings of rolling elements permit the rotary motion to the shafts in
the machinery for huge applications such as bicycles, electric motors, power trans-
mission aircraft gas turbines, rolling mills, handpieces, polishing filling, etc. The
methods used for defect detection are bispectrum analysis, EMD and EEMD. These
methods will help in eliminating the Gaussian and non-Gaussian signals in fixed
and nonlinear signals. EMD is an adjustable time, space analysis method used for
analyzing the series that are fixed and non-linear. Noise from the signal is hence
eliminated [9]. Whereas EEMD is a noise-assisted data analysis method in which a
white noise signal is added. White noise is important to force the ensemble to exhaust
all possible solutions in the shifting process, thus making the different scale signals
collate in proper intrinsic mode functions.
A Computational Intelligence-Based Novel Bearing … 647
The contribution of the paper is to obtain a self-changeable technique to study the

characteristics of the simulated signal based on the EMD method. The analysis done
in the paper is to show the removal of non-Gaussian noise in the rolling bearing of the
machine with the method EMD. As we get the mode mixing problem so FFT masking
is performed for the removal of it. Another method, i.e. EEMD is also obtained to
reduce the number of frequency scales in the EMD. EEMD also deals with the mode
mixing phenomenon problem.
The first part of the paper (Sect. 2) talks over the method of the decomposition of
signals, i.e. Empirical Mode Decomposition and Ensemble Empirical Mode Decom-
position. The second part (Sect. 3) deals with the analysis and performance evaluation
of the decomposition methods of the simulated signal. The third part (Sect. 4) shows
the masking technique to remove the mode mixing phenomenon and lastly (Sect. 5)
deals with the conclusion.
2 Methods of Defect Decomposition
2.1 Empirical Mode Decomposition
IMFs can be reduced in components called Intrinsic Mode Functions because of its
self-changeable data decomposition method EMD. This method helps in decaying
a nonlinear and fixed signal. Basically, in the EMD algorithm, IMFs are pulling
out from a given time series through a shifting process. Hence, it allows decaying
the nonlinear and fixed signal into various IMFs. It is extensively used in denoising
in certain areas like biomedical signals, acoustic signals. However, the potency of
the EMD method is influenced by the mode mixing phenomenon (modal aliasing)
[10–12]. While the breakdown of the signals into various components, EMD can
be realized with other methods such as Fourier transform and Wavelet transform. In
the decomposition methods, the first IMF has the most oscillating (high-frequency)
components, it can be rejected to eradicate the high-frequency components (e.g.,
random noise). EMD is the most adjustable and systematic method for decomposi-
tion. The EMD method decomposed a simulated signal into several IMFs that must
satisfy to conditions, which are also shown in Fig. 1.
Step 1—Find out the local minima and maxima of the signal f(t).
Step 2—Interpolate between maxima to get an upper envelope and maxima to get
a lower envelope.
Step 3—Compute the mean of the upper and lower envelope.
Step 4—Extract IMF.
648 A. Singh et al.
Fig. 1 EMD Algorithm
2.2 Ensemble Empirical Mode Decomposition
The mode mixing problem can be eradicated by the EEMD by inducing white noise
at various times in it. In a background space that is filled with white noise, a decom-
posed signal is applied, then calculating the average of white noise followed by
calculating the zero-mean noise making the signal stable by canceling the noise with
each other [13–15]. Ensemble empirical mode decomposition (EEMD) is one of
the most important tools for processing signals with intermittency. But, a problem
exists in the EMD method, which is the mode mixing phenomena made better by the
addition of white noise should be enough for the simulated signal. EEMD method
was formulated to deal mainly with the signals containing noise. The next step is
to measure the performance of the EEMD method using a noisy simulated signal.
The results proved that the EEMD method could lessen the problem to a large extent
concerning the EMD method. In addition to the above, the reports very well pointed
that the EEMD method could act as an improvement for the EMD method. The steps
for the calculation of EEMD (as shown in Fig. 2) are as mentioned below:
1. The addition of white noise with normal distribution is carried into the
decomposed signal.
2. On the signal, EMD is done, and several white noises are added to acquire IMFs.
3. Repeat the above process by adding white noise in every step.
4. The average value of IMFs calculated is considered as the final output.
The basic principle of the EEMD can be described as: the whole time–frequency
space is uniformly populated by the white noise that is added with the comprising
components of variable scales set apart by the filter bank. In addition to the signal to
this white background that is uniformly distributed, the bits of the signal of variable
scales are automatically extrapolated onto the definite scales of reference, which is
Fig. 2 EEMD algorithm
established by the background white noise. Undoubtedly, each trial produces very
noisy results, and each noise-added decomposition consists of the signal and white
noises that are added. As separate trials consist of variable noise in each trial, it gets
canceled out in the ensemble mean of most of the trials.
3 Analysis of Decomposition Methods
The non-Gaussian noise is to be removed before the bispectrum analysis. HUANG

proposed the ensemble empirical mode decomposition method, which is a changeable
signal decomposition method. Bispectral analysis, EMD, EEMD based on the signal,
can decay signals into a band of different frequencies. The mode mixing phenomenon
is removed by EEMD, which occurs by the EMD method if white noise is added to
the signal. The intrinsic characteristic of the signal is revealed by the IMFs pulled
out by EEMD as compared to EMD. The areas of mechanical fault correction and
diagnosis are using this method EEMD [16]. The methods will be analyzed step by
step, that is firstly the bispectrum analysis will be done to analyze the Gaussian Noise
(as in Fig. 3) and then the EMD will be performed to extract the IMFs and reduce the
non-Gaussian noise. This method deals will problems of different frequency modes,
i.e. mode mixing problems that are again then dealt with the ensemble empirical
mode detection method.
Rayleigh distribution as given in Fig. 4 is a continuous probability distribution
for nonnegative valued random variables.
Superposition of the three signals is taken in the signal X(t). The non-Gaussian
Noise is eradicated by the analysis of the simulated signal.
The proposed signal is given as.
X (t) = f (t) + g(t) + r (t)

650 A. Singh et al.
Fig. 3 Gaussian distribution
Fig. 4 Rayleigh distribution
= sin(20 ∗ pi ∗ t) + sin(60 ∗ pi ∗ t) + sin(100 ∗ pi ∗ t) + n(t) + r (t) (1)
Equation (1) consists of:

f(t) : It has three sine wave signals frequencies of 10Hz, 30Hz and 50Hz,
g(t) : consists of random signal with Gaussian Distribution
r(t) : consists of random signal with Rayleigh distribution
The sampling rates of f(t), g(t) and r(t) are 1000 Hz.
Figure 5 shows the superposition of the signal of three sine-wave signals, i.e.
10, 30 and 50 Hz. It also shows the signal with Gaussian distribution and signal
with Rayleigh distribution and the summation of all the signals is the output of the
simulated signal.
3.1 Empirical Mode Decomposition
The signal X(t) is decomposed into Intrinsic Mode Functions (IMFs) by intrinsic
mode signal decomposition (EMD) algorithm as shown in Fig. 6. The signal X(t) has
three sinusoidal frequencies 10 Hz, 30 Hz and 50 Hz. The ratio of the above IMFs
in the figure is much larger than the below IMFs. The real IMFs can be effectively
identified, while the false IMFs are removed.
Fig. 5 Simulated signal x(t)
Fig. 6 Signal Decomposition by EMD
In Fig. 6, we have taken the output of the simulated signal X(t) and decomposed
the signal using EMD. From the fig, we can extract information from IMF 1, 2, 3, 4
and 5 and hence these are true IMFs and from IMF 6, 7, no information is extracted.
652 A. Singh et al.
3.2 Ensemble Empirical Mode Decomposition
The IMFs of the EMD method consist of coupled oscillations in each IMF. Hence, it
also suffers from the mode mixing problem. To solve this problem and identify the
time scale, EEMD is used. The proposed method is explained in Fig. 7. The results
shown in Fig. 8 make it clear that the synchronizations between corresponding IMF
pairs have improved a lot, specifically for the IMF components.
Fig. 7 Proposed method
Fig. 8 Signal decomposition by EEMD

In Fig. 8, we can see that IMF has fewer frequency components than in the case of
IMF in EMD, and hence noise is eliminated. Also, we can extract information from
the IMF1, 2, 3, 4 & 5 but unable to extract it from the IMF 6, 7 and 8 because they
are false IMFs.
3.3 Performance Evaluation of Decomposition Methods
Three different sinusoidal signals with frequencies of 10 Hz, 30 Hz and 50 Hz are

produced.g(t) is Random signal with Gaussian distribution, r(t) is random signal
with Rayleigh distribution. f(t), g(t) and r(t) have sampling rates of 1000 Hz. We
denoise the signal for eliminating the non-Gaussian noise, and non-Gaussian noise
is eliminated by using EMD. Due to the issue caused by mode mixing phenomena,
the noise was not eliminated effectively in this method. A sudden change occurs in
the time-scaled decomposed signal containing huge noise. Single IMF has a large
number of frequency components. So we choose a different decomposition method
EEMD affecting the decomposition by the addition of the white noise, which cancels
each other. IMFs produce both false and true one, the false one gets eliminated and
true significantly reduces the chance of mode mixing. EEMD helps for reaching the
ultimate goal of data analysis for extracting the true signals. The EEMD utilizes the
characteristic of noise and effectively is a better method than EMD.
4 Modified Masking in Empirical Mode Decomposition
Separation of mode frequencies by EMD (Empirical Mode Decomposition) present

in an octave is not possible due to the problem of mode mixing. Thus, to effectively
improve the mode mixing issue, the technique of noise-assisted EMD is imple-
mented. The following section discusses the estimation methods of closely spaced
low-frequency modes through masking signals and the mode mixing problem of
standard EMD [16].
4.1 FFT-Based Masking Technique
Extension of standard EMD by using masking signal refines the differentiation of the
original EMD yet appears to be insignificant to separate the closely spaced modes.
Different modes of oscillations can be distinguished by the method of masking
signals. It filters out the frequency signal of weak frequencies and later it is removed
out during the process of EMD. Selecting the masking signal properly improves
the mode extraction potential of the masking method. Thus, the masking technique
through FFT improves the mode separation capability of EMD.
654 A. Singh et al.
4.2 FFT of Standard EMD
Assume a test signal “S” representing two components of a very low frequency 10
and 30, where the magnitude of the higher frequency component is notably lower
than the low-frequency component as given in Fig. 9.
S(t) = sin(2 × pi × 30 × t) + 5 × sin(2 × pi × 10 × t) (2)
Examination of mode mixing of EMD is carried out by a synthetic test signal. The
figure given below represents the extracted IMFs form of the signals as mentioned
in (10) with the help of the original EMD. It is observed that the IMF 1 and 2 shown
in Fig. 8 do not show any duplication with the component of the original signal.
For better physical interpretation, Hilbert transform can be used because IMF 1
suffers mode mixing problem and is not monotonic. The mode separation process of
EMD is rectified by the masking signal. The method is implemented to remove any
uncertainty and mode mixing of unseparated and closely spaced frequencies. Hence,
in-depth analysis and technical usage of the masking method with the EMD are
proposed. The drawback of the method yet observed is that it is not able to examine
multiple low-frequency components of oscillating signals.
Step 1: To locate the components of different frequencies f1, f2, f3,… fn from the
given FFT S(t), the signal.
Step 2: Construct the masking signals: mask2, mask3, mask4, mask5, …, maskn,
such that: maskn(t) = Mn × [sin (2π (f n + f n − 1) t].
Step 3: Construct the signal by addition and subtraction of the masking signals
from S(t). Where, S(t) + maskn & S(t) − maskn.
Step 4: Evaluate EMD as shown in the steps (1to7) in the expression: S(t) +
maskn and S(t) − maskn to get the value of first IMF Hence after calculating C(t)
= [(IMF + ) + (IMF − )]/2.
Step 5: Calculate the residue, r(t). Where r(t) = [S(t) − C(t)].
Step 6: The steps from 3 to 4 are repeated by replacing S(t) with residue r(t)
until the second last IMF containing components of frequency f2, f3, …, fn are
obtained and the final residue comes in the form of f1 component.
Fig. 9 Sum of 10 and 30 Hz freq sine signal

4.3 FFT of EMD by Masking Technique
In Fig. 10, EMD is performed on the signal given in Fig. 9, and IMF 1 IMF2 and
IMF 3 are true IMFs, i.e. we can extract the information about the signal and the rest
IMFs ignored. For the enhancement of the mode separation phenomenon ability of
the EMD, the masking signal-based approach was used to mask the signal, and hence
the FFT frequency spectrum is used. Imprecise and inexact information of frequency
is seen for the FFT of the signal that is distorted. Analyzing the signal that exhibits
non-linearity and non-stationary is a major concern. Hence, the IMF we got with
masking signal done with FFT has two frequency components, called mode mixing
problem. The mode estimation by FFT-based masking signal is performed below for
seeing the mode mixing problem.
We take the signal S(t) and obtain the frequency spectrum. The frequency spectrum
in Fig. 11 gives the value 9.766 and 29.79 as we have the signal of a frequency of 10
and 30 Hz.
Figure 11 shows the amplitude spectrum of the sine wave signal having frequencies
29.79 Hz and 9.799 Hz.
In Fig. 12, we can observe that the different frequency modes of the signal S(t)
and again by decomposing IMF1 obtained from the EMD we can see that in Fig. 12a
and b, two different frequencies can be obtained causing the mode mixing problem
and then by masking technique, we apply the FFT to remove the different frequency
mode and mode mixing.
Again now in Fig. 13a and b, we note that the mode mixing is eradicated and we
get one frequency in both the figures. In Fig. 13a, we get a frequency of 4 Hz and
in Fig. 13b, we get the frequency in 8 Hz. These figures mention that the different
frequency that is mode mixing issue is eradicated from the signal to get the picture
of true IMFs. We perform FFT on IMF 1 to obtain the Fourier spectrum of the IMF 1
evaluated by the EMD. Hence, the extracted signal has two component frequencies
Fig. 10 EMD of the signal S(t)

656 A. Singh et al.
Fig.11 Amplitude spectrum of S(t)
Fig. 12 Fourier spectrum of IMF (a) IMF1 (b) IMF 2
in which one frequency higher than (8 Hz) and one less than that (5 Hz). As the
frequencies are packed closely so the process is more evident. In the second figure,
we can see that by the masking technique, we can remove this mode mixing problem
that arose in the EMD method and got just one frequency, i.e. (4 Hz), and again the
same is repeated with the IMF 2. This masking EMD separates modes of closely
spaced frequencies efficiently. The method reconstructs the analysis signal so that
the masking method works effectively. Preprocessing of original data using EMD is
done before masking EMD on the given data. Original signal EMD produces IMFs
and trend signal as residue. Henceforth, all of the significant amplitude IMFs were
summed. To obtain different modes of frequencies in the signal, FFT was applied
to the IMFs. It suggests extracting a single mode at a time rather than extracting all
modes at a single period.
Fig. 13 FFT-based masking technique of IMF (a) IMF1 (b) IMF2
4.4 Performance Evaluation of Methods
The Fast Fourier Transform method employing masking signal upgrades the EMD
ability for mode separation. We take the signal S(t) and obtain the frequency spectrum.
The frequency spectrum in Fig. 11 gives the value 9.766 and 29.79 as we have a signal
658 A. Singh et al.
of a frequency of 10 and 30 Hz. We perform FFT on IMF 1 to extract from the Fourier
spectrum of the IMF 1 came from the EMD. As a result, the signal has two component
frequencies that are one higher frequency (8 Hz) and one less than that (5 Hz) as
given in Fig. 12. This phenomenon is seen when the frequencies are closely packed.
In Fig. 12, we can see that by the masking technique we can remove this mode mixing
problem that arose in the EMD method and got just one frequency, i.e. (7 Hz) and
again the same is repeated with the IMF 2. Hence, the masking of EMD separates
closely spaced modes of frequencies systematically. Similarly in Fig. 13, we can
observe that a single value of frequency of 8 and 4 Hz of IMFs is obtained.
5 Conclusion
This paper proposes the decomposition methods to eradicate the non-Gaussian noise.
The output of the simulated signal is therefore evaluated with the decomposition
methods EMD and EEMD. With the EMD method, true IMF information is extracted
and the rest of the IMFs are referred to as false IMFs. With the addition of the
white noise to EMD, EEMD was performed and here it is seen that signal with
fewer frequency components is present. Hence, noise is eliminated successfully by
EEMD and better IMFs are produced. So by the noise added in the EEMD eases,
the mode mixing problem is caused by EMD and hence produces a signal with
effectively reduced noise. Further, we can also go with other better methods for noise
assistance for true IMF signals, i.e. Complete Ensemble Empirical Decomposition
with Adaptive Noise (CEEMDAN). Further, in the paper FFT method, the signal is
taken to obtain the frequency spectrum. A masking signal was used to eradicate the
mode mixing. FFT masking technique was done to give the single component for the
IMFs. As a result, the extracted signal has two components of frequency in which
one is a higher frequency and the other one is less. By masking one single frequency
is found and the mode mixing problem is eradicated. It is concluded that by EEMD
and masking technique, we can remove this mode mixing problem that arose in the
EMD method and got just one frequency for the IMFs.
References
1. Liu H, Liang J, Hu YH (2005) Research on parametric bispectrum of heart sound signal analysis
method based on wavelet transform domain. Signal Process 20(1):5–9
2. Huang NE, Shen Z, Long SR, Wu MC, Shih HH, Zheng Q (1998) The empirical mode decom-
position and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proc
Math Phys Eng Sci 454(1971):903–995
3. Wu ZH, Huang NE (2012) Ensemble empirical mode decomposition: a noise-assisted data
analysis method. Adv Adapt Data Anal 1(1):1–41
4. Wei ZL, Robbersmyr KG, Karimi HR (2017) An EEMD aided comparison of time histories
and its application in vehicle safety. IEEE Access 5:519–528
5. Qi YM, Shen CQ, Wang D, Shi JJ, Jiang XX, Zhu ZK (2017) Stacked Sparse autoencoder-based
deep network for fault diagnosis of rotating machinery. IEEE Access 5:15066–15079
6. Xiang JW, Zhong YT (2017) A fault detection strategy using the enhancement ensemble
empirical mode decomposition and random decrement technique. Microelectron Reliab
75:317–326
7. Wang X, Liu C, Bi F, Bi X, Shao K (2013) Fault diagnosis of diesel engine based on adaptive
wavelet packets and EEMD-fractal dimension. Mech Syst Signal Process 41(1–2):581–597
8. Gu F, Shao Y, Hu N, Naid A, Ball AD (2011) Electrical motor current signal analysis using
a modified bispectrum for fault diagnosis of downstream mechanical equipment. Mech Syst
Signal Proc 25(1):360–372
9. Li W, Zhang G, Shi T, Yang S (2004) Gear crack early diagnosis using bispectrum diagonal
slice. Chinese J Mech Eng 16(2):193–196
10. Jiang Y, Tang C, Zhang X, Jiao W, Li G, Huang T (2020) A novel rolling bearing defect detec-
tion method based on bispectrum analysis and cloud model-improved EEMD. IEEE Access
8:24323–24333
11. Ray P, Lenka RK, Mohanty BK (2017) A modified noise assisted EMD to extract low frequency
modes present in a WAMS data of dynamic power system. IEEE Calcutta Conf CALCON
2017:2–3
12. Tang B, Dong S, Ma J (2013) Study on the method for eliminating mode mixing of empirical
mode decomposition based on independent component analysis. Chinese J Sci Inst 33(7):1477–
1482
13. Li L, Dang R, Fan Y (2014) Modified EEMD de-noising method and its application in
multiphase flow measurement. Chinese J Sci Inst 35(10):2365–2371
14. Cheng Y, Wang ZH, Chen BY, Zhang WH, Huang GH (2019) An improved complementary
ensemble empirical mode decomposition with adaptive noise and its application to rolling
element bearing fault diagnosis. ISA Trans 91(21):218–234
15. Liu D, Xiao ZH, Hu X, Zhang CX, Malik OP (2019) Feature extraction of rotor fault based on
EEMD and curve code. Measurement 135(73):712–724
16. Han T,Jiang D, Wang N (2016) The fault feature extraction of rolling bearing based on EMD
and difference spectrum of singular value. Hindawi 5957179:1–14

Recent Advances in Power Electronics and Drives: Shailendra Kumar Bhim Singh Arun Kumar Singh Editors

Uploaded by

Copyright:

Available Formats

Recent Advances in Power Electronics and Drives: Shailendra Kumar Bhim Singh Arun Kumar Singh Editors

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

Recent Advances in Power Electronics and Drives: Shailendra Kumar Bhim Singh Arun Kumar Singh Editors

Uploaded by

Copyright:

Available Formats

Lecture Notes in Electrical Engineering 852

More information about this series at https://link.springer.com/bookseries/7818

Recent Advances in Power

Arun Kumar Singh

ISSN 1876-1100 ISSN 1876-1119 (electronic)

Fuzzy Controlled Multi-converter-UPQC for Power Quality

A Novel Current Control Scheme for Three-Phase Three-Level

Self-balanced Symmetric Source Configuration of Nine Level

PV-Array-Integrated UPQC for Power Quality Enhancement

Comparative Analysis of Conventional and Sliding Mode Control

Dr. Shailendra Kumar is an Assistant Professor in the Department of Electrical

Prof. Bhim Singh is a Professor in the Department of Electrical Engineering at

Dr. Arun Kumar Singh is a Professor in the Department of Electrical Engineering,

S. Joyal Isac and S. Srinath

S. Joyal Isac (B)

quality identification is composed of performing load balancing, eradicating the

2.1 Related Works

simulation outcomes revealed the efficiency of the MC-UPQC-oriented 3P4W distri-

Table 1 Features and challenges of state-of-the-art MC-UPQC methods

Fig. 1 Diagrammatic illustration of MC-UPQC

3 Power Quality Improvement Using Improved MC-UPQC

3.2 Control Strategies

The primary direct axis component current is given to the DC as portrayed in

4 Fuzzy Controller-Based MC-UPQC for Power Quality

4.1 Proposed Fuzzy Controller

Fig. 2 Block diagram for fuzzy controller

In the above equation, the term Q O = 1, 2, 3 · · · O I , where O I denotes the

verr f uzz = vdc − vdc

5 Results and Discussions

5.1 Simulation Setup

5.2 MC-UPQC Performance in Connection with Feeder 1

5.3 MC-UPQC Performance in Connection with Feeder 2

5.4 Harmonic Analysis

1. Mohammadi HR, Varjani AY, Mokhtari H (2009) Multiconverter unified power-quality

Debanjan Mukherjee and Sourav Mallick

D. Mukherjee (B) · S. Mallick

2 Description of the Photovoltaic System

Advantages of photovoltaic system as a viable source of energy motivate this work

Table 1 Description of the

3 Proposed Topology of Single Phase 5-Level MLI

4 Selective Harmonic Elimination as an Optimization

Expression of output voltage by expanding through Fourier series is necessary for

(cos(kα1 ) + cos(kα2 )) = 0 (4)

where MI can be computed as Eq. (5).

where Vdesir ed is the desired voltage magnitude at fundamental frequency.

5 Brief Discussion on Biogeography-Based Optimization

The biogeography-based optimization (BBO) technique was developed by D. Simon

Fig. 2 Flow chart of BBO technique

6 The Proposed Technique

where l is the decay parameter.

Hi (S I V ) = η ∗ S*(Hi (S I V )−Hbest (S I V )) (8)

Fig. 3 Pseudo-code of the modified mutation phase in the MBBO technique

6.1 Flowchart for Firing Angle Optimization

7 Hardware in Loop (HIL)

Execute proposed, BBO, BOA, GSA

Fig. 5 Experimental setup of hardware in loop

9 Results and Discussions

9.1 Comparison of Reduced Switch MLIs with the Proposed

9.2 Comparison of Results of Different Optimization