Nonlinear Modeling and System Identification of a DC Gear Motor with Unknown Parameters

International Journal of Trend in Scientific Research and Development (IJTSRD)
Volume 3 Issue 5, August 2019 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470
@ IJTSRD | Unique Paper ID – IJTSRD26475 | Volume – 3 | Issue – 5 | July - August 2019 Page 737
Nonlinear Modeling and System Identification
of a DC Gear Motor with Unknown Parameters
Htet Htet Shin, Nay Min Tun
Department of Electronic Engineering, Mandalay Technological University, Mandalay, Myanmar
How to cite this paper: Htet Htet Shin |
Nay Min Tun "Nonlinear Modeling and
System Identification of a DC Gear Motor
with Unknown Parameters" Published in
International
Journal of Trend in
Scientific Research
and Development
(ijtsrd), ISSN: 2456-
6470, Volume-3 |
Issue-5, August
2019, pp.737-741,
https://doi.org/10.31142/ijtsrd26475
Copyright © 2019 by author(s) and
International Journalof Trendin Scientific
Research and Development Journal. This
is an Open Access article distributed
under the terms of
the Creative
CommonsAttribution
License (CC BY 4.0)
(http://creativecommons.org/licenses/by
/4.0)
ABSTRACT
Modeling and identification of industrial systems is an essential stage in
practical control design and applications. The paper presents linear, state
space, nonlinear modeling and identification of aDC gearmotor with real-time
experiments. The main aim of this research is to use the concept of modeling
and System Identification method for observing the greater accuracy and
better fitness system model, and validate it byapplyingvariousdatasetsofthe
hardware experiment. System Identification deals with the problem of
building mathematical models of dynamical systems based on observed data
from the systems. The methodology is based on results obtained from the
simulation of theoretical concepts, which are then validated by repeating
experiments on the motor. It is very important to do this validation because
sometimes these theoretical concepts are not able to fully capture the nature
of the physical elements, and both results may differ. Proceedingin thisway,it
can guarantee a greater extent that the results are correct.
KEYWORDS: Modeling, Linear, State Space, Nonlinear, System Identification
I. INTRODUCTION
Electromechanical devices (DC andalternatingcurrentmotors) are widelyused
as prime movers for mechanical systems and machines. In some cases where
control of the mechanical systems is required, a control unit is attached to the
electric motors and these type of motors are generally referred to as DC motor.
Due to the importance of the DC motors in the systems and
processes, research studies on the characterization,
mathematical modeling and parameter identification of
electromechanical devices have been published. Modeling
and simulation of physical systems are widely used in
engineering for better understanding of the characteristics
of systems in order to control the systems’ performance and
reduce costs by building and testing a prototype at the
preliminary stage instead of the exact machine [1].
When the physical structure and parametersof DCmotor are
unavailable conditions, a mathematical model representing
the system behavior may not be obtainable. For this case,the
system parameters should be obtained using a system
identification procedure. The concepts of system
identification are more useful during the modification of
existing systems when little or no information about the
existing system is available. Identification of linear systems
is a rather old field of study, and many methods areavailable
in the literature. However, the identification of nonlinear
systems is a relatively new topic of interest. In mechanical
systems with a DC motor, identification is an occasionally
employed procedure for examination and detection of the
system parameters. The nonlinear identification of DC
motors has also been of interest in recent years, together
with compensation for nonlinearities like Coulomb friction,
backlash and stick-slip effect [2].
System identification is proceeded through linear and
nonlinear models as to the linearity of the system. Linear
system identification that the input and the output of the
system stated with linear equations is mostly used because
of its advanced theoretical background. However, many
systems in real life have nonlinear behaviours. Linear
methods can be inadequate in the identification of such
systems and nonlinear methods are used. In nonlinear
system identification, theinput-outputrelationof thesystem
is provided through nonlinear mathematical assertions as
differential equations,exponentialand logarithmicfunctions
[3].
This paper basically focuses on linear, state space and
nonlinear modeling and proposes an innovative MATLAB
model to study the dynamic response of DC motors in open
loop. The results of the MATLAB model shall prove to be
very useful in designing the control strategy for applications
involving DC motors. The nonlinear system model is built
and a nonlinear Hammerstein Weiner structure is used for
the identification procedure.
II. Electromechanical Characteristics of DC Motor
The most important part of this section is the physical
reasoning behind the concept of transforming electrical
power in mechanical power. As a matter of fact, since the
magnetic field arises from the stator coils, not only the rotor
coils may rotate with respect to the stator,butalsothestator
supply may rotate by increasing the number of coilsand bya
more sophisticated way.
IJTSRD26475

International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
Figure1. Electromechanical Characteristics of DC
Motor
The next important thing is to deal with the mechanical
representation of the motor. Fig. 1 shows the model along
with the electrical characteristics and it directly provides
rotary motion and coupled with wheelsor drums,and cables
can provide translational motion [4, 5]. The equivalent
circuit for a DC motor is represented in Fig.1 and Fig. 2
represents a 12V DC gear motor used in this research.
Figure2. 12V DC Gear Motor
And, the motor transfer function is calculated by equation
(1).
Where, T(s) = transfer function of DC motor
Ke = motor back emf constant
Kt = motor torque constant
R = resistance of motor rotor (armature)
L = inductance of armature
B = motor friction coefficient
J = motor of inertia
III. Modeling of a System
System modeling is the process of developing abstract
models of a system, with each model presenting a different
view or perspective of that system. There are three
categories of mathematical modeling: white-box (physical)
modeling, gray-box modeling and black-box (experimental)
modeling [6].
A. White-box Modeling
If the physical laws governing the behavior of thesystem are
known, it is called a white-box model in which all
parameters and variables can be interpreted in terms of
physical entities and all parameters are known.
B. Black-box Modeling
Figure3. Black Box Model
In Fig. 3, a black-box model is simply the functional
relationship between system input and system output
without any knowledge of its internal workings.
Experimental or black-box modeling also called system
identification, is based on measurements. Advantages of
black-box modeling are to develop easier than theoretical
models and applicable over wide ranges of operating
conditions.
C. Gray-box Modeling
In many practical cases, it often occursthatoneknowsonlya
little bit about the system, that is, the system modeling is
based on the recorded input and output data with some
prior knowledge about the system, e.g., the structure and
order of the system. By analyzingandextractinginformation
from the system and using the identification methodsforthe
black-box model, a gray-box model will be constructed.
IV. SYSTEM IDENTIFICATION
System identification uses the input and output signals that
measured from a system to estimate the values of adjustable
parameters in a given model structure. The main idea of
system identification is studying the behavior of existing
structures by recording the output or input-output
relationship of the system. The input-output description of a
discrete-time system consists of a mathematical expression
which explicitly defines the relationship between the input
and output signals. The system is assumed to be a "black-
box" to the user. So this philosophy for identifying the
specifications of the system (structure) is system
identification. On the other hand, each system (structure) is
the same as a filter to convert the input signals with the
specific frequency and characteristics to the output signals
with filtered frequencies due to system parameters. So the
process of constructing models from experimental data is
called system identification shown in Fig. 4. And an example
of input and output data arraysarepresented inequation (2)
and (3) respectively.
Figure4.System Identification
)](),...,3(),2(),([ ssssmeas NTFTFTFTFu = Equation (2)
)](),...,3(),2(),([ ssssmeas NTyTyTyTyy = Equation (3)

V. Implementation of the Research
In this section, the relationship of the input and output of
motor taken from the experimental hardware setup is
described in section A. In section B, it is described about
Matlab System Identification Toolbox, insertion of input
dataset and estimation of models for the system. And, about
the model estimation of linear, state-space and nonlinear of
the system are expressed in details in section C, D and E.
A. Taking Input and Output Data from Experiment
Firstly, the input (pwm) and output (cycles per seconds)
data of DC gear motor are taken fromtheexperimentalsetup
as shown in Fig. 5. And, different input types such as square
wave, sine wave and sawtooth wave are applied in the real
experiment and all of the output are noted. By this way, the
relationship between the input and output of the system are
taken. By applying these input-output data, dynamic model
of the system can be observed by using the System
Identification of MATLAB.
Figure5. Hardware Experimental Setup
B. MATLAB System Identification Toolbox
The toolbox provides MATLAB functions, SimulinkMATLAB
System Identification blocks, and an application for
constructing mathematical models of dynamicsystemsfrom
measured input-output data. It lets create and use modelsof
dynamic systems not easily modeled from first principles or
specifications. Time-domain and frequency-domain input-
output data can be used to identify continuous-time and
discrete-time transfer functions, processmodels,state-space
models, nonlinear models and so on. The toolbox also
provides algorithms for embedded online parameter
estimation.
Figure6. MATLAB System Identification Toolbox
Window
The toolbox provides identification techniques such as
maximum likelihood, prediction-error minimization (PEM),
and subspace system identification. To represent nonlinear
system dynamics, it can estimate Hammerstein-Weiner
models and nonlinear ARX models with wavelet network,
tree-partition, and sigmoid network nonlinearities. The
toolbox performs gray-box system identification for
estimating parameters of a user-defined model. It can use
the identified model for system response prediction and
plant modeling in Simulink. The toolbox also supports time-
series data modeling and time-series forecasting [7].
As first, the System Identification Toolbox window is found
by using ident tools command as shown in Fig. 6. In Fig. 7
and Fig. 8, time-domain input and output data from
workspace is imported with 0.01 samplingtime.In there, the
square wave data set is used for finding the system model of
linear, state-space and nonlinear. When being observed all
types of model, other different datasets (sawtooth and sine
wave) is applied to these model for validation of these
models. And finally, the best model will be chosen by
comparing all these models.
Figure7. Importing Input-output Data
0 5 10 15 20 25
0
100
200
300
y1
Input and output signals
0 5 10 15 20 25
0
100
200
300
Time
u1
Figure8. Input and Output Signals
C. Identifying Linear Transfer Function Model
The general transfer function model structure is:
)()(
)(
)(
)( sEsU
sden
snum
sY += Equation (4)
Y(s), U(s) and E(s) represents the Laplace transforms of the
output, input and error, respectively. num(s) and den(s)

represent the numerator and denominator polynomialsthat
define the relationship between the input and the output.
The roots of the denominator polynomial are referred to as
the model poles. The roots of the numerator polynomial are
referred to as the model zeros. The System Identification
Toolbox estimates the numerator and denominator
polynomials, and input/output delays from the data [7].
It must specify the number of poles and zeros to estimate a
transfer function model. Here, one zero and two poles
transfer function is estimated and transfer function with
71.31% fitness of simulated modelwith themeasured model
is taken and expressed in equation 7 and Fig. 9.
05838.0461.2
0536.0773.1
2
++
+
=
ss
s
TF Equation (5)
0 5 10 15 20 25
0
50
100
150
200
250
Time
PWM
Measured and simulated model output
simulated output
measured output
Figure9. Measured and Simulated Output for Linear
Transfer Function Model
D. Identifying State Space Model
The general state-space model structure (innovation form)
is:
)()()( tKetButAx
dt
dx
++= Equation (6)
)()()()( tetDutCxty ++= Equation (7)
Where, y(t) = the output at time t, u(t) = the input at time t,
x(t) = the state vector at time t, and e(t) = the white-noise
disturbance. The System Identification Toolbox estimated
the state space matrices A, B, C, D, and K from the data. And,
Fig. 10 illustrated the fitness of the simulatedmodel withthe
measured model is 69.2%. The estimated A, B, C, D and E are
in following.
0 5 10 15 20 25
0
50
100
150
200
250
Time
PWM
Measured and simulated model output
simulated output
measured output
Figure10. Measured and Simulated Output for State-
Space Model
E. Identifying Nonlinear Model
Of actually, all physical systems are nonlinear to an extent.A
system in which the input-output steady-state relation is
nonlinear is called nonlinear system. Because nonlinear
models are able to describe the system behaviour in a much
larger operating region than corresponding linear models,it
is reasonable and necessary to characterize or predict the
behavior of real nonlinear processesdirectlyusingnonlinear
models to improve identification performance over their
whole operating range. Therefore, it leads to the
development of approaches for nonlinear modeling and
analyzing nonlinear systems [7].
In System Identification Toolbox, there are two model types
as nonlinear ARX andHammerstein-Wiener.Inthisresearch,
the Hammerstein-Wienermodelisestimated. Hammerstein-
wiener structure consists of two nonlinear blocks in series
with a linear block as presented in Fig. 11 and the fitness of
Hammerstein-Wiener model is 94.2%bycomparingwiththe
measured model as shown in Fig. 12.
Figure11. Hammerstein-Wiener Model
0 5 10 15 20 25
0
50
100
150
200
250
Time
PWM Measured and simulated model output
simulated output
measured output
Figure12. Measured and Simulated Output for
Hammerstein-Wiener Model
VI. Validation of Simulated Models and Results
COMPARISON
A. Validation of Simulated Models
Fig. 13 shows a Simulink model for model validation that
consists of three different transfer function models. Thefirst
is for linear, the second is for state space and the last is for
nonlinear Hammerstein-Wiener. These all models are taken
from square wave input-output dataset using System
Identification Toolbox as presented in section V.
In this section, sine wave and sawtooth waves are imported
as inputs to these three models. And then, the simulated
model's outputs are compared with real time experiment
outputs for the validation of these model whether themodel
taken from MATLAB System Identification Toolbox is
identical or different with real time plant or system. In other
words, how much the simulated model representstheactual
model.

Figure13. Validation of Simulated Models
B. Comparison of Results
The results comparison of above Simulink model for
sawtooth and sine wave inputs are expressed in Fig. 14 and
Fig. 15 respectively. For both figures, the blue is input
waveform and the black line is output result from the
hardware experiment. And the other lines are estimated
model simulation results. By comparing hardware actual
outputs with simulated results, system identification or
black-box modeling can take out well the models of the
unknown system. All of them, it has been found that the
nonlinear model is the best accuracy model as shown in
Table1.
0 5 10 15 20 25 30
0
50
100
150
200
Cycles per seconds
PWM
input
linear
state space
nonlinear
actual
Figure14. Model Validation with Sawtooth Wave Input
0 5 10 15 20
0
50
100
150
200
Cycles per seconds
PWM
input
linear
state space
nonlinear
actual
Figure15. Model Validation with Sine wave Input
TABLE I. Model Fitness Comparison
Fitness of
Simulated Model
Linear Transfer
Function Model
71.31%
State Space Model 69.2%
Nonlinear Hammerstein-
Wiener Model
94.2%
VII. CONCLUSION
The model of DC motor is estimated by using System
Identification or black-box modeling without knowing any
parameters of the system or its internal working. It can also
estimate the nonlinear model and ithas betteraccuracythan
any other model. In this research, three models of linear,
state space and nonlinear are taken out. When these model
outputs are validated with actual hardware results, the
simulated result of the nonlinear model is nearly identical
with actual output.
ACKNOWLEDGMENT
The author is highly grateful to her supervisor, Dr. Nay Min
Tun, Lecturer, Department of Electronic Engineering,
Mandalay Technological University, for his supervision,
patient guidance, support, encouragement and tolerance
helped in all the time of this research work.
REFERENCES
[1] Modeling and Parameter Identification of a DC Motor
Using Constraint Optimization Technique, Surajudeen
Adewusi1, Mechanical Engineering Department,Jubail
University College
[2] Nonlinear Modeling and Identification of a DC Motor
for Bidirectional Operation with Real Time
Experiments, Tolgay Kara, IIlyas Eker, Department of
Electrical and Electronics Engineering, Division of
Control Systems, University of Gaziantep, 27310
Gaziantep, Turkey
[3] System Identification Application Using Hammerstein
Model, SABAN OZER1, HASAN ZORLU1, and SELCUK
METE2D, department of Electrical and Electronic
Engineering1, Erciyes University, 38039 Kayseri,
Turkey, Kayseri Regional Office2, Turk Telekom A.S.,
38070 Kayseri, Turkey
[4] Identification and Control of DC Motors, Darshan
Ramasubramanian, September 2016, Automatic
Control and Robotics, Escola Tècnica
Superiord’Enginyeria Industrial de Barcelona
[5] Real Time Model Validation and Control of DC Motor
Using MATLAB and USB, MOHAMMED FAROOQ
MOHAMMED SALIH, Faculty of Electrical Engineering,
Universiti Teknologi Malaysia, JUNE
[6] Black-box Models from Input-Output Measurements,
Lennart Ljung, Div. of Automatic control, Linkoping
University, Sweden
[7] http://www.mathworks.com

Nonlinear Modeling and System Identification of a DC Gear Motor with Unknown Parameters

More Related Content

What's hot (20)

Similar to Nonlinear Modeling and System Identification of a DC Gear Motor with Unknown Parameters (20)

More from ijtsrd (20)

Recently uploaded (20)

Nonlinear Modeling and System Identification of a DC Gear Motor with Unknown Parameters