International Journal of Power Electronics and Drive System (IJPEDS)
Vol.2, No.4, December 2012, pp. 402~408
ISSN: 2088-8694 402
Journal homepage: http://iaesjournal.com/online/index.php/IJPEDS
Artificial Neural Network Based Controller for Speed Control
of an Induction Motor using Indirect Vector Control Method
Ashutosh Mishra *, Prashant Choudhary **
* Deptt. Of Electrical Engineering, RCET, Bhilai
Chhattisgarh, Bhilai - India
Article Info ABSTRACT
Article history:
Received Jun 17, 2012
Revised Nov 10, 2012
Accepted Nov 23, 2012
In this paper, an implementation of intelligent controller for speed control of
an induction motor (IM) using indirect vector control method has been
developed and analyzed in detail. The project is complete mathematical
model of field orientation control (FOC) induction motor is described and
simulated in MATLAB for studies a 50 HP(37KW), cage type induction
motor has been considered .The comparative performance of PI, Fuzzy and
Neural network control techniques have been presented and analyzed in this
work. The present approach avoids the use of flux and speed sensor which
increase the installation cost and mechanical robustness .The neural network
based controller is found to be a very useful technique to obtain a high
performance speed control. The scheme consist of neural network controller,
reference modal, an algorithm for changing the neural network weight in
order that speed of the derive can track performance speed. The indirect
vector controlled induction motor drives involve decoupling of the stator
current in to torque and flux producing components.
Keyword:
FOC
IM
IVCIM
NN
PI
Copyright 2012 Institute of Advanced Engineering and Science.
All rights reserved.
Corresponding Author:
Ashutosh Mishra
Deptt. Of Electrical Engineering, RCET, Bhilai Chhattisgarh, Bhilai - India
1. INTRODUCTION
An induction motor is an asynchronous AC (alternating current) motor. The least expensive and most
widely used induction motor is the squirrel cage motor. The interest in sensor less drives of induction motor
(IM) has grown significantly over the past few years due to some of their advantages, such as mechanical
robustness, simple construction, and less maintenance. These applications include pumps and fans, paper and
textile mills, subway and locomotive propulsions, electric and hybrid vehicles, machine tools and robotics,
home appliances, heat pumps and air conditioners, rolling mills, wind generation systems, etc. So, Induction
motors have been used more in the industrial variable speed drive system with the development of the vector
control technology. This method requires a speed sensor such as shaft encoder for speed control.
However, a speed sensor cannot be mounted in some cases such as motor drives in a hostile
environment and high-speed drives [1]. In addition, it requires careful cabling arrangements with attention to
electrical noise. Moreover, it causes to become expensive in the system price and bulky in the motor size. In
other words, it has some demerits in both mechanical and economical aspects. Thus current research efforts
are focused on the so called sensor less vector control problem, in which rotor speed measurements are not
available, to reduce cost and to increase reliability. The control and estimation of ac drives in general are
considerably more complex than those of dc drives, and this complexity increases substantially if high
performances are demanded. The main reasons for this complexity are the need of variable-frequency,
harmonically optimum converter power supplies, the complex dynamics of ac machines, machine parameter
variations, and difficulties of processing feedback signals in the presence of harmonics. The selection of
drive for motor control is based on several factors such as [2]:
One-, two- or four-quadrant drive,
403 ISSN: 2088-8694
IJPEDS Vol. 2, No. 4, December 2012 : 402 408
Torque, speed, or position control in the primary or outer loop,
Single- or multi- motor drive,
Range of speed control Does it include zero speed and field-weakening regions, Accuracy and response
time,
Robustness with load torque and parameter variations,
Control with speed sensor or sensor less control,
Type of front-end converter,
Efficiency, cost, reliability, and maintainability consideration,
And Line power supply, harmonics, and power factor consideration.
2. Over view of DIFFERENT CONTROLLING SCHEMES for Speed Control of Three Phase
Induction Motor
2.1 Scalar control
Scalar control as the name indicates, is due to magnitude variation of the control variable only, and
disregards the coupling effect in machine. For example, the voltage of machine canbe controlled to control
the flux, and frequency or slip can be controlled to control the torque.However flux and torque are also
function of voltage and frequency respectively.
2.2 Vector Control or Field Orientated Control (FOC)
In DC machine the field flux is perpendicular to the armature flux. Being orthogonal, these two
fluxes produce no net interaction on one another. Adjusting the field current can therefore control the DC
machine flux, and the torque can be controlled independently of flux by adjusting the armature current [9].
An AC machine is not so simple because of the interactions between the stator and the rotor fields, whose
orientations are not held at 90 degrees but vary with the operating conditions. We can obtain DC machine-
like performance in holding a fixed and orthogonal orientation between the field and armature fields in an
AC machine by orienting the stator current with respect to the rotor flux so as to attain independently
controlled flux and torque. Such a control scheme is called flux-oriented control or vector control. Vector
control is applicable to both induction and synchronous motors.
The cage induction motor drive with vector or field oriented control offers a high level of dynamics
performance and the closed-loop control associated with this derive provides the long term stability of the
system .Induction Motor drives are used in a multitude of industrial and process control applications
requiring high performances. In high-performance drive systems, the motor speed should closely follow a
specified reference trajectory regardless of any load disturbances, parameter variations, and model
uncertainties. In order to achieve high performance, field-oriented control of induction motor (IM) drive is
employed. However, the controller design of such a system plays a crucial role in system performance. The
decoupling characteristics of vector-controlled IM are adversely affected by the parameter changes in the
motor. So the vector control is also known as an independent or decoupled control [10].
2.3 Proportional Integral (PI) control
In this project complete mathematical model of FOC induction motor is described and simulated in
MATALAB for studies a 50 HP(37KW) induction motor has been considered .The performance of FOC
drive with proportional plus integral (PI) controller are presented and analysed. One common linear control
strategy is proportional-integral (PI) control. The control law used for this strategy is given by
T = K
p
e + K
i
e dt
Its output is the updating in PI controller gains (K
p
and K
i
) based on a set of rules to maintain
excellent control performance even in the presence of parameter variation and drive nonlinearity
nonlinearityThe use of PI controllers for speed control of induction machine drives is characterized by an
overshoot during tracking mode and a poor load disturbance rejection. This is mainly caused by the fact that
the complexity of the system does not allow the gains of the PI controller to exceed a certain low value. At
starting mode the high value of the error is amplified across the PI controller provoking high variations in the
command torque. If the gains of the controller exceed a certain value, the variations in the command torque
become too high and will destabilize the system. To overcome this problem we propose the use of a limiter
ahead of the PI controller [11]. This limiter causes the speed error to be maintained within the saturation
limits provoking, when appropriately chosen, smooth variations in the command torque even when the PI
IJPEDS ISSN: 2088-8694 404
Artificial Neural Network Based Controller for Speed Control of An Induction Motor (Ashutosh Mishra)
controller gains are very high. The motor reaches the reference speed rapidly and without overshoot, step
commands are tracked with almost zero steady state error and no overshoot, load disturbances are rapidly
rejected and variations of some of the motor parameters are fairly well dealt with [20].
2.4 Fuzzy Logic Control
Due to continuously developing automation systems and more demanding small Control
performance requirements, conventional control methods are not always adequate. On the other hand,
practical control problems are usually imprecise. The input output relations of the system may be uncertain
and they can be changed by unknown external disturbances. New schemes are needed to solve such
problems. One such an approach is to utilize fuzzy control. Since the introduction of the theory of fuzzy sets
by L. A. Zadeh in 1965, and the industrial application of the first fuzzy controller by E. H. Mamadani in
1974, fuzzy systems have obtained a major role in engineering systems and consumers products in 1980s
and 1990s. New applications are presented continuously. A reason for this significant role is that fuzzy
computing provides a flexible and powerful alternative to contract controllers, supervisory blocks, computing
units and compensation systems in different application areas [12]. With fuzzy sets nonlinear control actions
can be performed easily. The transparency of fuzzy rules and the locality of parameters are helpful in the
design and maintenances of the systems. Therefore, preliminary results can be obtained within a short
development period.Fuzzy control is based on fuzzy logic, which provides an efficient method to handle in
exact information as basis reasoning. With fuzzy logic it is possible to convert knowledge, which is
expressed in an uncertain form, to an exact algorithm. In fuzzy control, the controller can be represented with
linguistic if-then rules [13].
2.5 Neural Network Control:
We introduce the multilayer perceptron neural network and describe how it can be used for function
approximation. The back propagation algorithm (including its variations) is the principal procedure for
training multilayer perceptrons, it is briefly described here. Care must be taken, when training perceptron
networks to ensure that they do not over fit the training data and then fail to generalize well in new situations.
Several techniques for improving generalization are discussed [18]. Three neural Network control techniques
are model reference adaptive control, model predictive control, and feedback linearization control. These
controllers demonstrate the variety of ways in which multilayer perceptron neural networks can be used as
basic building blocks. But in this project we are used model predictive control for speed regulation of
induction motor [14]. There are a number of variations of the neural network predictive controller that are
based on linear model predictive controllers [19]. The neural network predictive controller that is discussed
in next chapter [34] uses a neural network model of a nonlinear plant to predict future plant performance. The
controller then calculates the control input that will optimize plant performance over a specified future time
horizon. The first step in model predictive control is to determine the neural network plant model (system
identification). Next, the plant model is used by the controller to predict future performance [15].
3. MATLAB Model of Indirect Vector Control IM Drive
3.1 Hysteresis Current Regulator:
The current regulator, which consists of three hysteresis controllers, is built with Simulink blocks.
The motor actual currents are provided by the measurement output of the Asynchronous Machine block. The
actual motor currents and reference current are compared in hysteresis type relay.
Figure 3.1. Hysteresis Current Regulator
405
IJPEDS Vol. 2, No. 4, December
3.2 MATLAB Simulation IVCIM based on Neural Network Predictive controller:
In this Matlab simulation NN controller is different from the PI and Fuzzy controller. NN controller
take two input one is reference input and another input plant or IM speed output
Figure 3.2. Matlab Simulink block diagram of indirect vector control using Neural Network controller
4. RESULTS AND ANALYSIS
4.1 Performance of Indirect Vector Control IM Using Neural Network predictive Control
I have used neural network predictive control for speed
used which shown in figure 3.15(a). First identification of plant has been performed by use N N toolbox.
After identification training data was generated which was accepted depending on comparison of plant output
and plant input? Network was trained using this data to obtained optimum value of weight and biases using
trainlm function (Levenberg Markquardt back propagation).weight and biases values were applied to NN
Predictive controller. I have used 20 hidden layer
converged after 12 epochs when the sum squared error is 3.23681e
Then simulation of IVCIM was performed using NN Predictive controller and results were recor
motor current, speed and torque.
reference speed at t =2.1 sec and it draw a low starting current 307.5 amp then the PI and fuzzy control. The
Motor torque is also good then the PI and Fuzzy control which is 80.3 N
PI control results:
Neural Network predictive control results
Case-I Results at Initial starting value at time t= 15.0 m sec for no load.
Motor current (I
abc
) = 307.371 amps
Torque (Te) = 20.7 N/m
Case-II Results at no load for speed reach 120 red/sec at time t=2.0sec
Motor current (I
abc
) = 21.24 amps
Torque (Te) = 20 N/m
.Case-III Results after applying 25N
Motor current (I
abc
) = 22.12 amps
Torque (Te) = 21 N/m
December 2012 : 402 408
IVCIM based on Neural Network Predictive controller:
In this Matlab simulation NN controller is different from the PI and Fuzzy controller. NN controller
take two input one is reference input and another input plant or IM speed output
ab Simulink block diagram of indirect vector control using Neural Network controller
ND ANALYSIS
Performance of Indirect Vector Control IM Using Neural Network predictive Control
I have used neural network predictive control for speed control of IVCIM. Simulink plant model is
3.15(a). First identification of plant has been performed by use N N toolbox.
After identification training data was generated which was accepted depending on comparison of plant output
nd plant input? Network was trained using this data to obtained optimum value of weight and biases using
trainlm function (Levenberg Markquardt back propagation).weight and biases values were applied to NN
Predictive controller. I have used 20 hidden layers, 8000 training sample and 200 epochs. The network has
converged after 12 epochs when the sum squared error is 3.23681e-005 was obtained at learning rate of 0.05.
Then simulation of IVCIM was performed using NN Predictive controller and results were recor
motor current, speed and torque. The reference speed is 120 rad/sec, t is observed that motor pick up the
reference speed at t =2.1 sec and it draw a low starting current 307.5 amp then the PI and fuzzy control. The
he PI and Fuzzy control which is 80.3 N-m and 14.3 N
Neural Network predictive control results:
I Results at Initial starting value at time t= 15.0 m sec for no load.
) = 307.371 amps
Torque (Te) = 20.7 N/m
II Results at no load for speed reach 120 red/sec at time t=2.0sec
) = 21.24 amps
III Results after applying 25N-m load torque at time t=2.2 sec.
) = 22.12 amps
ISSN: 2088-8694
IVCIM based on Neural Network Predictive controller:
In this Matlab simulation NN controller is different from the PI and Fuzzy controller. NN controller
ab Simulink block diagram of indirect vector control using Neural Network controller
Performance of Indirect Vector Control IM Using Neural Network predictive Control
control of IVCIM. Simulink plant model is
3.15(a). First identification of plant has been performed by use N N toolbox.
After identification training data was generated which was accepted depending on comparison of plant output
nd plant input? Network was trained using this data to obtained optimum value of weight and biases using
trainlm function (Levenberg Markquardt back propagation).weight and biases values were applied to NN
s, 8000 training sample and 200 epochs. The network has
005 was obtained at learning rate of 0.05.
Then simulation of IVCIM was performed using NN Predictive controller and results were recorded for
The reference speed is 120 rad/sec, t is observed that motor pick up the
reference speed at t =2.1 sec and it draw a low starting current 307.5 amp then the PI and fuzzy control. The
m and 14.3 N-m.
IJPEDS ISSN: 2088-8694 406
Artificial Neural Network Based Controller for Speed Control of An Induction Motor (Ashutosh Mishra)
Figure 4.1. Performance of IVCIM with Neural Network control at no loads with reference speed 120
rad/sec.
Figure 4.2. Neural Network control Response of IVCIM at no load with speed 120 red/sec
Figure 4.3. Neural Network control Response of IVCIM with Appling load torque =25N-m,at time t=2.2 sec
Figure 4.4. Neural Network control Response of IVCIM with Appling load torque =25N-m, at time t=2.1 sec
to 3 sec
407 ISSN: 2088-8694
IJPEDS Vol. 2, No. 4, December 2012 : 402 408
The performance of NN control has been compared with the dynamic loading conditions at different
values of torques. Figure 4.12 shows the performance for aload torque variation of 0N-m to 25 N-m as
against the previous values of no-load. It was observed that at 25N-m load torque average estimated value of
speed decrease from 119.5 rad/sec (at no-load condition) to 114 rad/sec, after 0.2 sec. it gains the reference
speed as figure 4.2.
5. CONCLUSION
This paper has successfully demonstrated and a properly designed PI, Fuzzy logic and Neural
Network predictive controller.
The NN predictive controller is more robust than the PI and fuzzy logic controller when load
disturbances occurred.
The NN predictive controller performance when certain motor parameters (i.e. current and motor torque)
were increased by a factor was still quite good and far better than the PI and fuzzy logic controllers
performance when the same parameters.
NN predictive controller base makes the superior to PI and fuzzy logic control techniques.
Required numerous trials and constant retuning to get reasonable performance
ACKNOWLEDGEMENTS
This project is made under the guidance of Mr. Prashant choudhary (Asscociate professor) in RCET,
BHILAI. Mr. Prashant choudhary sir is a fantastic personality & he is my guide of my project. I would like to
thanks my parents, teachers and my guide & my dear friends for helping me in this project and making it
wonderful.
REFERENCES
[1] Li Zhen and Longya Xu, On-Line Fuzzy Tuning of Indirect Field-Oriented an Induction Machine Drives, IEEE
Transactions on Power Electronics, Vol. 13, No. 1, pp. 134-138, January 1998.
[2] L. A. Zadeh, fuzzy theory, university of California, Berkely,1965
[3] Hassan Baghgar Bostan Abad, Ali Yazdian Varjani, Taheri Asghar, Using Fuzzy Controller in Induction Motor
Speed Control with Constant Flux, Proceedings of World Academy of Science, Engineering And Technology, Vol.
5, pp. 307-310, 2005.
[4] N. Mariun, S. B Mohd Noor, J. Jasni, and O. S. Bennanes, A Fuzzy Logic Based Controller For An Indirect Vector
Controlled Three-Phase Induction Motor, IEEE IECON Conf. Rec., Vol. 4, pp. 1-4, Nov. 2004.
[5] Vinod Kumar and R. R. Joshi, Hybrid Controller based Intelligent Speed Control of Induction Motor, Journal of
Theoretical and Applied Information Technology, pp. 71-75, 2005.
[6] M. N. Uddin, T. S. Radwan, and M. A. Rahman, Performances of Fuzzy-Logic-Based Indirect Vector Control for
Induction Motor Drive, IEEE Transactions On Industry Applications, Vol. 38, No. 5, pp. 1219-1225, Sept./Oct.
2002.
[7] P. Vas, Vector control of AC machines, New York: Clarendon, 1990
[8] B. K. Bose, power electronics and AC Derive. Englewood Cliffs, NJ. Prentic hall, 1986
[9] J. W Finch, DJ Atkinson and PP Acarnleg University of Newcastle up on Tyne, UK, General Principles of Modern
Induction motor control
[10] F. Biaschke, The principle of field oriention as applied to new transvector closed loop control system for rotating
field machine, Siemens Rev, vol. 34, pp. 217-220, May 1972
[11] R. Krishna and A.S. Bharadwaj, A Review of Parameter Sensitivity and Adaptation in Indirect Vector Controlled
Induction Motor Drive Systems, IEFE Transactions On Power Electronics, Vol. 6, No 1, pp. 695-703, Oct 1991.
[12] Bhim Singh, B. N. Singh, and B. P. Singh, Performance Analysis of a Low Cost Vector Controlled Induction
Motor Drive: A Philosophy for Sensor Reduction, IEEE IAS Annu. Meet. Conf. Rec., pp. 789-794, 1997.
[13] J. S. R. Jang, Fuzzy Controller Design Without Domain Experts, IEEE International Conference on Fuzzy
Systems, 8-12 March 1992 Page(s): 289 296.
[14] Brian Heber, Longya Xu, and Yifan Tang, Fuzzy Logic Enhanced Speed Control of an Indirect Field-Oriented
Induction Machine Drive, IEEE Transactions on Power Electronics, Vol. 12, No. 5, pp. 772-778, September 1997.
[15] A. Miloudi and A. Draou, Variable Gain PI Controller Design For Speed Control and Rotor Resistance Estimation
of an Indirect Vector Controlled Induction Machine Drive, Conference Record of the IECON 02 Sevilla, Spain,
Vol. 1, pp. 323-328, Nov 2002.
[16] Adel Merabet, Mohand Ouhrouche and Rung-Tien Bui, Neural Generalized Predictive Controller for Induction
Motor. Volume 1 Number 1 pp. 83100, University of Quebec at Chicoutimi.2006
[17] K. S. Narendra and K. Parthasarathy, Identification and control of dynamic systems using neural networks, IEEE
Trans. Neural Networks, vol. 1, pp. 4-27, Jan. 1990.
IJPEDS ISSN: 2088-8694 408
Artificial Neural Network Based Controller for Speed Control of An Induction Motor (Ashutosh Mishra)
[18] Seyed Hossein HOSSEINIand Mohamad Reza BANAEI Neural network speed controller for induction motor
based on feedback linearization.
[19] Jianrong Bu, Longya Xu, A High Performance Full Fuzzy Controller for Induction Machine Drives, IEEE APEC
Conf., Vol. 2, pp. 592-596, 1998.
[20] P. Vas, and J. Li, Simulation Package for Vector Controlled Induction Motor Drives, Oxford University press,
1993.
BIOGRAPHIES OF AUTHORS
Ashutosh Mishra was born in 14 of july in the year 1978. He recieved his batchers degree (BE)
from LNCT from barkatullah university bhopal in 2000 in ELECTRICAL ENGG. He is
pursuing his Masters (ME) in POWER ELECTRONICS ENGG from rungta college of engg &
technology, bhilai. Presently he is working as a SR. LECTURER in RSR rungta college of
engg & technology, bhilai in E&E deptt. Ashutosh has total teaching experience is
approximately about of 7 years in different organizations in bhilai since 2004 in csvtu & he is
teaching various subjects of electrical & electronics engg .
Prashant Kumar Choudhary was born in 22 may 1982 in jamshedpur (tata nagar). He recieved
his batchelor's degree (BE) from BIT mesra in 1999. He has completed his Master's (M TECH)
from BIT durg from chattisgarh swami vivekanand technical university from chaattisgarh in
control system engg. Presently he is working as asscociate professor in rungta college of engg
& technology, bhilai since 2004 in the deptt of electrical engg & teaching various subjects of
electrical engg .
