Modeling and Simulation for the Electro-hydraulic

Servo System Based on Simulink

Junlan Cheng ,Wei Liu, Zengliang Zhang
Department of Mechanical Engineering
North China Institute of Aerospace Engineering
Langfang , China
lanlan0917@126.com

Abstract-This paper established mathematical model based on

control theory for electro-hydraulic position servo system, used II. ANALYSIS OF SERVO VALVE WITH ASYMMETRICAL
Simulink to simulate system model, and applied M files and CYLINDER
module mask function to simplify the system simulation process. Fig.1 is the diagram of servo valve controlled asymmetrical
This electro-hydraulic position servo system adopted PID
cylinder position control system [4]. The electric-hydraulic
controller. The PID control is the abbreviation of proportion,
integral and differential control, its value depends on its widely
servo valve is a zero opening and four-way slide valve. This
applicable for most of the control system, especially when the paper studies the load of the hydraulic system compose of the
mathematical model of the controlled object is known, PID inertial load, viscous load and linearization coulomb friction.
control could be particularly useful. This paper applied digital The three basic equations of valve control hydraulic cylinder
PID to get ideal system dynamic performance by adjusting the system are as follows [5].
PID parameters. This paper proposed a method that provides a
reference for other control systems analysis and design.

Keywords-mathematical model; system simulation; M files;

module mask

I. INTRODUCTION
Electro-hydraulic servo system refers to the control system
which combined two control modes of electrical and hydraulic.
Detecting, transmitting and processing the signal by use of
electric and electronic components, driving the load with
hydraulic transmission in the electro-hydraulic servo control
system. So it can make full use of electrical system for its
convenience and aptitude, make full use of hydraulic system
for its rapid response speed, big load stiffness and accurate Figure 1. The diagram of servo valve controlled asymmetrical
cylinder position control system
positioning characteristics to make the whole system more
adaptable. PID control is the abbreviation of proportion,
integral and differential control, its value depends on its widely Linearized flow equation of the zero opening and four-way
applicable for most of the control system, especially when the slide valve is
mathematical model of the controlled object is known, PID
control could be particularly useful. Simulink is a graphical
QL = K q X v − K c PL (1)
environment for dynamic system modeling, simulation and
analysis interactively [1]. Using Simulink environment, can The load flow continuity equation of the cylinder is
build complex system simulation model, using the mask Vt
function to simplify the model, using M files to initial the QL = Ctc PL + Ap sX p +sPL (2)
module variables, which provides great convenience for PID 4βe
parameters setting and change the variable value of the various The force balance equation effected on piston is
modules [2-3]. This paper establish the mathematical and
simulation model for the valve controlled asymmetrical Fg = PL A p = mt s 2 X p + ( B p +b) X p (3)
hydraulic cylinder position servo system, and optimize system Integrated (1), (2) and (3) deduce to the transfer function,
dynamic performance by setting PID parameters. which is output displacement of the piston rod to input
displacement of the valve

978-1-61284-459-6/11/$26.00 ©2011 IEEE

466
 X p Kq 1 (4) passes parameters. Because the variables that M files created
= ⋅ are variables of Matlab working space, when a program run to
X v Ap § s 2 2ξ h ·
s¨¨ 2 + s + 1¸¸ the end, the variables store in the working space, the other
© ωh ωh ¹ programs or simulation model can directly call these variables,
Where in order to achieve the purpose of transmitting parameters.
2 A complex Simulink model, in order to be versatile and
4β e A p β e mt K ce Vt Bp + b
ωh = ˈξ = ⋅ + ⋅ portable, each module parameter often set in parameter name
h
mt Vt Vt Ap β e mt 4 A p rather than specific value. Assign or initialize these variables
before simulation. There have many ways to initialize
The response speed of servo amplifier and displacement variables, commonly used the methods as follows.
sensor are far higher than servo valve controlled hydraulic
cylinder, so they all can be seen as a proportional link. So, A. Make a system parameter initialized data files
based on every part of the transfer function is known and
ignore leakage and linearize the dry friction [6-7], get Make a system parameter initialized data files (Matlab M
hydraulic position servo system simulation diagram as shown files), open and run this file manually before running Simulink,
in Fig.2. load the variable value to working space to initialize the
module parameters.

B. Put the variable statements in the form of text

Put the assignment of variable statements directly in the
form of text below Simulink figure, first copy these statements
into the command window and run them, and then run the
simulink model.

C. Load simulation model and automatic run initialized files

Load simulation model and automatic run initialized data
files. The method is click on the corresponding Simulink
Model ,then choice “File \Model properties \ Callback”, enter
Figure 2. Electro-hydraulic position servo system simulation diagram
the m file name to run in parameters box “PreLoadFcn “, so
Simulink model can save the variable to workspace when the
III. CONTROL SYSTEM SIMULATION WITH SIMULINK model is loaded.
In standard Simulink module library contains many
D. Run m files through double-click a module
modules, such as Sinks(output module group), Source(input
module group), Linear(Linear link module group), Nonlinear Run m files through double-click a module in Simulink
(non-linear links module group), Math(mathematical operation environment. The method is that first add a most common
module group), Continuous and Connections modules group SubSystem module to Simulink model, and delete all internal
and so on. Each module contains many sub-module [8]. Use content (general is an input and an output port and a line
these modules can easily convert the system simulation between them) in this Sub-system module, then return to the
diagram as Fig.2 shows to the simulation model shows as Fig.3. Simulink model, right-click on the Subsystem module and
choice the shortcut menu “Mask Subsystem”, and get into the
editor of mask Subsystem to complete Subsystem mask. In this
paper write “disp (‘Variable Initialization’)” in the Icon
drawing command bar. If execute the initial m files by double-
click on the module, need to select the module and choice
“Edit\ block properties \Callbacks” and write m file name in
OpenFcn dialogue.
In these methods, the method A needs to open and run m
files manually every time before simulation, so it is more
tedious. The method B is suit for less parameter. The methods
C load m files only when loading model, if modified data in m
Figure 3. Electro-hydraulic position servo system simulation model file, must shut down the Simulink model and open it again.
The methods D is the most effective method, double-click on
IV. APPLYING M FILE TO TRANSFER PARAMETERS
the module at any time according to need. So this paper selects
method D to establish a “svdata.m” file as the M file to
initialize variable in every module [11-12].
M file is the file with a letter M for its extension and is the
executable file in Matlab [9-10]. This paper introduces that
using M file to transfer parameters, which are not input or
output parameter but refers to the use of variable in M files to

467
 V. MASK MODULE
Mask module is a powerful function in Simulink, it allows
users to define dialog and ICONS for subsystem. Using mask
module has the following functions.

A. Simplify the use of model.

Users can input all the parameters at the same time for sub-
system through masked dialog .

B. Provide users with a better interface

Provide users with a better interface by defining dialog  
module descriptions, identify of parameter fields and help text. Figure 6. The Simulink model after make subsystem

C. Avoid changing the content

Avoid changing the content of sub-system in mask VI. ANALYSIS THE SIMULATION RESULTS
interface. Because the tracking precision and the positioning
accuracy are two most important index of electro-hydraulic
In this paper, it is very convenient to adjust three position servo control system, so need to examine the variation
parameters of P, I, D after masking the PID controller. It can of these two indexes.
edit all parameters at the same time in one dialog box. The
model before masking is shown in Fig. 4. After masking, when The unit step signal is selected as Simulation test signal to
double-click the PID module, popup the dialog box as Fig.5 check localization accuracy of the system. The PID parameters
shows. after optimizing are as follows, proportional coefficient P is
3.5, Integral coefficient I is 2.5, differential coefficient D is
0.01,the simulation result show as Fig.7. The result shows that
the adjust time is 0.08s, the overshoots is 2% and the error at
steady-state is f0.02. It can better meet the system needs.
The ramp signal, which slope rate is 1, is selected as
simulation test signal to compare tracking precision of the
system. The PID parameters after optimizing are as follows,
proportional coefficient P is 14.5, Integral coefficient I is 76,
differential coefficient D is 0.01, the simulation result show as
Fig.8. The result shows that it can better meet the system
Figure 4. Internal structure of PID controller needs.

Figure 7. The system response curve for unit step signal

Figure 5. The dialog box after mask PID controller

In order to avoid changing the content of system model and

simplify model, make subsystem for the model that don’t
change constantly [13]. After make subsystem the Simulink
model is shown as Fig.6.
Figure 8. The system response curve for ramp signal

468
 [5] Mohammad Nuruzzaman, Modeling and Simulation in SIMULINK for
VII. CONCLUSIONS Engineers and Scientists, AuthorHouse,2005.
[6] Carlos Canudas de Wit and Pablo Lischinsky, “Adaptive friction
This paper establish the mathematic model of the electro- compensation with partially known dynamic friction model”, Journal of
hydraulic position servo system after analysis base on the Adaptive Control and Signal Processing, Special Issue on Adaptive
control theory, and create simulation model of this system Systems with Nonsmooth Nonlinearities, 1997, pp.65–80.
using Simulink base on actual model parameters to run system [7] Henrik Olsson and Karl Johan Astrom, “Observer-based friction
simulation, adopt M file to transfer value to the variables of compensation”, Proceedings of the 35th IEEE Conference on Decision
each module, and apply module package to simplify system and Control, 1996,pp. 4345–4350.
simulation process. Through adjusting the PID parameters, [8] Shihong Xie, Example Tutorial of MATLAB R2008 for Control System
Dynamic Simulation , Chemical Industry Press,2009 (In Chinese).
finally obtained the ideal dynamic performance of the system.
This paper proposed a method that provides a reference for [9] A. Pervozvanski and C. Canudas de Wit. Vibrational smoothing in
systems with dynamic friction. Technical report, Laboratoire
other control system analysis and design. d’Automatique deGrenoble,1997.
[10] Y.Tan ,I. Kanellakopoulos, “Adaptive nonlinear friction compensation
REFERENCES with parametric uncertainties,” Proceedings of American Control
Conference, vol. 4, 1999, pp. 2511-2515.
[1] Ying Li, Dynamic System Modeling and Simulation with Simulink(The
second edition), Xi’an University of Electronic Science and Technology [11] G. Liu, “On velocity estimation using position measurements,”
Press, 2009 (In Chinese). Proceedings of American Control Conference, Alaska, 2002.
[2] Zhonglin Huang, Automatic Control Principle Realize by MATLAB, [12] L. Xu and B.Yao, “Adaptive robust control of mechanical systems
Defense Industry Press ,2007 (In Chinese). with nonlinear dynamic friction compensation,” Proceedings of
American Conlrol Conference, vol. 4, 2000, pp.2595-2599.
[3] Bang, Brian J., Draxler, and James G., “Dynalic Hydraulic System
Simulation-An Integrated Approach”, SAE Technical Paper [13] G. Liu, “Decomposition-based friction compensation using a parameter
902003,1990. linear izalion approach,” Proceedings of IEEE Inlernational Conference
on Robotics and Automation, vol. 2, 2001, pp.1155-1160.
[4] Zhenshun Wu, The hydraulic control system, Higher Education
Press,2008(In Chinese).

469