Hybrid Artificial Intelligent Control For Speed Control Of Im

SICE-ICASE International Joint Conference 2006 Oct. 18-2 1, 2006 in Bexco, Busan, Korea

Hybrid Artificial Intelligent Control for Speed Control of Induction motor Jae-Sub Kol and Jung-Sik Choi2 and Dong-Hwa Chung3 Department of Electrical Control Engineering, Sunchon national University, Sunchon, Korea (Tel: +82-61-750-3543; E-mail: kokos22@naver . com) 2Department of Electrical Control Engineering, Sunchon national University, Sunchon, Korea (Tel : +82-61-750-3540; E-mail: 1108cj s@naver . com) 3Department of Electrical Control Engineering, Sunchon national University, Sunchon, Korea (Tel +82-61-750-3543; E-mail: hwa777@sunchon. ac. kr) 1

Abstract: This paper is proposed hybrid artificial intelligent controller for high performance of induction motor drive. The design of this algorithm based on fuzzy-neural network controller that is implemented using fuzzy control and neural network. This controller uses fuzzy rule as training patterns of a neural network. Also, this controller uses the back-propagation method to adjust the weights between the neurons of neural network in order to minimize the error between the command output and actual output. A model reference adaptive scheme is proposed in which the adaptation mechanism is executed by fuzzy logic based on the error and change of error measured between the motor speed and output of a reference model. The control performance of the hybrid artificial intelligent controller is evaluated by analysis for various operating conditions. The results of experiment prove that the proposed control system has strong high performance and robustness to parameter variation, and steady-state accuracy and transient response. Keywords: Induction Motor Drive, Fuzzy Control, Neural Network, Hybrid Artificial Intelligent Controller, High Performance

algorithms. To solve these problems, direct fuzzy control is developed, but it can not be obtained the response characteristic of robust and satisfaction performances about various load change and inertia change. [5] Neural network is evaluated very influential technique in parameter estimation and control of drive.[6]-[7] Neural network is excellent in more ability of adaptive control than generally vector control method. But it is have a problem in high performance and robust control that is indicated the characteristic of fuzzy control. In this paper is proposed hybrid artificial intelligent controller that is mixed fuzzy control and neural network for high performance of induction motor. Hybrid artificial intelligent controller is consisted the first part of fuzzy rule and the second part of clustering method, multilayer neural network. Hybrid artificial intelligent controller can be obtained high performance, robust control in advantage of fuzzy control, and high adaptive control ability in advantage of neural network. To improve more performance, hybrid artificial intelligent controller is applied adaptive mechanism method that based on reference model.[8] Final output is obtained add output of hybrid artificial intelligent controller and output of adaptive fuzzy control. In this paper, this controller applied drive system of induction motor, and then analyzed the parameter variation, characteristic of steady-state and response transient-state. And the validity of proposed controller is proof.

1. INTRODUCTION Recently, Artificial Intelligent control using Fuzzy control, neural network and genetic algorithm are recognized primary techniques that should be improved performance of power electronic system. These techniques are developed adaptive-artificial intelligent controller with mixed adaptive control technique. And these inter-mixture methods will high performance the induction motor drive that required adaptability and robustly. [1]-[3] Indirect vector control is applied extensively the drive system for high performance of induction motor. Because PI controller is indicated satisfactorily response of steady-state, industrial field is used a lot of PI controller. But PI controller can not expected satisfaction performance in transient-state because of nonlinear of induction motor. Especially, PI controller can not be obtained satisfaction performance about parameter variations such as disturbance, speed and load. If proportion gain kP of PI controller is big, rising time is short, overshoot is big and stability time is long and if kP of PI controller is small, rising time is long, overshoot is small and stability time is short. Thus although PI controller adjusted the gain coefficient, performance of drive can not more improved. To satisfaction performance of drive, adaptive is researched.[4] Adaptive control can be obtained more excellent performance than existing PI controller. Adaptive control technique is based on mathematical modeling and is very complicated because of numerous

89-950038-5-5 98560/06/$10 © 2006 ICASE

678

Fig. 1 shows that d q axis equivalent circuit at synchronously rotating reference frame. co,',

i,

List

L

r

0(w (0

i

vv-,

(a) d-axis

CO,e OdJc

Rs

LXi, (C"e ml OdX R,.

Lis

(b) q-axis

Fig. 1 equivalent circuit at synchronously rotating reference frame

.rAdaptive

At equivalent circuit of Figure 1 evaluated the voltage equation is equal that below; RS + Lsp -ceLs ceLs RS + LSP

Vds Vqs 0

LmP

-(OiSLm

0

CosILm

LmP

tsl =WMe -

LmP Rr + Lrp

lLr

Rr

LmP

iqs

WicoslLr

'dr

LrP-

Fuzzy

ids

eLm i

(OeLm

iqr

-

This paper is consisted of hybrid artificial intelligent controller that is mixed adaptive control, fuzzy control and neural network. This controller control induction motor with high performance.

(1 )

Ids' lqs d, q axis current VdsI Vqs

L Rs Ls

d,

4. DESIGN OF FUZZY-NEURO CONTROLLER

q axis stator voltage

Resistance of stator, self inductance

Fuzzy-neuro controller is consisted of fuzzy control and neural network. It has the softness of strong expression and numerical value processing ability. This controller is consist that the first part of fuzzy rule and the second part of clustering method, multilayer neural network. And it has advantage of robust control that like fuzzy control, and high adaptive ability that like neural network. To control of induction motor drive, structure of direct fuzzy controller is equal that Figure 3.

Rr, Lr Resistance of rotor, self inductance Lm : Mutual inductance The mechanical

below that; Te =J

m+

dt

equation

of induction motor is

Bwm + T,

equal

(2)

p 2

(3)

ltrds

Origination torque, Tl load torque, J: Inertia coefficient, B the coefficient of friction e

Generally, dynamic action of fuzzy controller is characterized by group of language control rule that is based on professional knowledge. Language control rule is assumed that below.

Evaluated generation torque is equal below Te

jLm

(iqsdr -dsqr )

NN

Fig. 2 Research of hybrid artificial intelligent controller

slip angular speed

0r

;=Ts_,2.S|~ PWAIV

PID and adaptive control method are many used. But these methods are complicated that evaluated the d-q axis parameter and is very sensitive at parameter variation, load change. Fuzzy controller and neural network are developed for speed control of induction motor. These methods have the advantage about parameter variation, load change, system disturbance. But each controller contains problems. To develop more new method, the new way is tried. Mixture control that is inter-mixed adaptive, fuzzy control, neural network and genetic algorithm is predicted the influential method. Figure 2 schematically expresses about research of hybrid artificial intelligent controller.

2. MODELING OF INDUCTION MOTOR

(4)

If E is 4. and CE is 42 then U is Bi

3. HYBRID ARTIFICIAL INTELLIGENT CONTROLLER

(5)

E, CE and U is each indicated error, variation of error and control variable. An is indicated fuzzy variable that

To speed follow and control of induction drive, PI,

679

Indicated in figure 6, neural network between Al layer and A3layer is indicated realization of the first part about fuzzy rule. To effectively design membership function of first part, first part is applied clustering method that is improved the contraction speed and is simple structure. Table 1 shows set of fuzzy rule

is characterized by membership function 1A(XJ), and Bi is a constant that is composed of real. Fuzzy group of E, CE and U is {NL, NM, NS, ZE, PS, PM, PL}. Figure 4 shows the membership function of fuzzy group. To control robust and high performance of induction motor drive, dynamic characteristic is required satisfactory performance that can be adjusted various speed obeying ability, adapted load variation. Thus this paper is consist of fuzzy-neuro controller by inter-mixed fuzzy controller and neural network, because generally fuzzy controller is not satisfied these demands. Figure 5 shows fuzzy-neuro controller structure for induction motor drive control.

Table I Fuzzy rule table

NL

j

NM NS ZEPS PM

PL

+1

NM NS ZEPS PM

~Izv v \/ C1 C2 0-cl -C2 (b)

PL

Lcewrlpu]

\

1

,uA(Aiqs) NL

NM NS ZEPS PM

PL

Ai*qAA]

+1

"Il i2

0

(c)

i-"I-2

-1

Fig. 4 Membership function of the fuzzy variables

Ti

Fig. 5 Structure of fuzzy-neuro controller Figure 6 shows organization of fuzzy-neuro controller. Input variable of two is error e, variation of error ce and output variable is U. A3

PS

NM NS

NL

NL

NL

NL

NL

NM NS

NS

NL

NL

NM NS

ZE

NL

NM NS

PM

PL ZE

ZE

PS

ZE

PS

PM

ZE

PS

PM

PL

PS

NM NS

ZE

PS

PM

PL

PL

PM

NS

ZE

PS

PM

PL

PL

PL

PL

ZE

PS

PM

PL

PL

PL

PL

i

=

A2

ZE

NL

2 2 (1T

A1

NS

Input space is divided space of 49 at fuzzy rule base. If rule is equal same action, doing clustering with same input space. And then rule that united new clustering is redesigned by expert and realized neural network. The number of fuzzy rule can be very decreased by clustering method. This realization can be designed the nonlinear function and is divided input space of fuzzy by sigmoid characteristic of neuron. Each output of neuron at A3layer is truth value of fuzzy rule at each part space. To realize fuzzy clustering and learning, fuzzy-neuro controller that obtained the same effect at 49 rules is needed function of error. Error function that is determined through achievement of requirement partition in fuzzy-neuro controller is frame as

Sl(ceo) ) NL

NM

NM NL

NL

',l(eCor)

NL

1, =

(xI,x2)eR'

0, otherwise

6)) (6) (7)

Where, r is number of cluster, Ti is function that determine whether any input date is included required cluster. And °i is output of neuron at A3 layer. After defined error function, To minimize error, next step is adjust weight Wik and Wi between Al layer and A3 layer by back-propagation algorithm. Through weight control, neural network can be completely realized the first part of fuzzy rule.

A4

Fig. 6 The construction of fuzzy-neuro controller

A Wjk=

5. REALIZATION OF FUZZY-NEURO CONTROLLER

AWij = -77

5.1. The first part realization

Where

680

(8)

a Wjk

8i

(8)

-qOiOj

(9)

(5.

=

(T 0,)f'(U,)

(10)

6j f'(Uj)Y 6,Wj =

(11)

°i is output of neuron atA2layer, 17is learning rate, f ( ) is differentiation of sigmoid function, Ui and U is indicated total input about each neuron at between A2 layer and A3 layer. Finally, to improve contraction speed and avoid vibration in learning course, new adjustment can be assumed momentum term.

Wjk (t + 1) Wjk (t) + AWJk + [Wjk (t) -Wjk (t- 1)] W,j (t + 1) = WXj (t) + A\Wj + a[W,, (t) Wxy (t 1)] -

-

0Or(k)

Output by fuzzy-neuro algorism is

(12) (13)

2

(U* U)2

iqs(k) output

by AFC is iqs2(k) . After sum of two outputs, order q axle current obtained trough integrator. Output of motor 6r(k) is compared with output of reference model Mm (k) and error enm(k) is accomplished by AFC. Reference model is used primary system for requirement performance that is satisfied design reference like overshoot and stability time. Figure 8 is indicated the AFC by reference model. AFC and fuzzy-neuro loop is connected in parallel. A

In Fig. 6, neural network between A3 layer andA4 layer is indicated second part of fuzzy rule. While learning, weight wci is adjusted minimize next error function -

{Ss

Fig. 7 Algorithm of proposed hybrid artificial intelligent controller

5.2 Realization of the second part

E

FNN IntgrIo NN)LL-

~~T Or ( kw (k)

ceW

r

(14)

O

AFC

(k)

ee m(k)

+Xp

U and U is indicated demand value and real value

(k)

COr(k)

of fuzzy-neuro controller. Variation ciof Weight ci using general delta rule can be minimized equation (14) and The second part of fuzzy rule redefine and determine the below the equation.

Output is occurred a Aqs2(k) by AFC. After the sum Ai*I (k) and A*i2(k) , Ai qs (k) is obtained. And of As

AWCi = -aE -qdcO0 WJ(t+1)=WWi(t)+AWWi +a[WZi(t)-WZi(t-1)]

then iqS(k) is occurred by integrator, and apply the plant. AFC input use input of reference model, error of real speed (eOJm(k)), variation of error (cean (k))

Fig. 8 AFC with reference model

(15) (16)

6c is error signal in fuzzy-neuro controller

ec6m(k) = Mtm (k) -etim(k)

(17) (18)

cea,m (k) = ea, (k) - ea (k -1)

6. DESIGN OF HYBRID ARTIFICIAL INTELLIGENT CONTROLLER

Command q axis current of fuzzy controller can be obtained that sum of

Induction motor drive is required high performance and tenacity although various speed assumption ability, load and inertia, parameter is changed in excess characteristic. Hybrid artificial intelligent controller is designed reference model in fuzzy-neuro controller and is consist of adaptive mechanism as fuzzy inference. Figure 7 is consisted of hybrid artificial intelligent controller that is designed fuzzy-neuro controller and connected (Adaptive Fuzzy Controller) AFC in parallel. In order to fuzzy-neuro error compensation, used AFC. AFC is indicated adaptive mechanism that considered reference model.

A\iq$I(k)

and

A\iq>2(k)

iqs(k) = i* (k -1) + [Ai'* (k) + Ai'*2(k)]

(19)

Controlled induction motor drive system by hybrid artificial intelligent controller is same figure 9. And Inverter used SV PWM method. H

Controller

A

Currelnt f

Control

V.IM bs

SV PWM In vrter

Int rator

Fig. 9 Configuration diagram of induction motor drive

681

7. PERFORMANCE RESULT OF SYSTEM 1800

Figure 10 is indicated that manufactured induction motor operation system that made for experiment in this paper.

(

CH1 CHZ

12, 00 t1 12~00Vt

CH1 CHZ CH2

-4:

<< Mn,in >>

[rpm] 20

(b)iqsI[A]

00Vt -4:. 00VW 6, 0 0Vt

... ..

0 ..

20 (c)Te[N m] CH4

..

00 -5000 - 0"

V1

5.00

Fig. 12 Response characteristics of HAI controller with step command speed

IBM PC

Fig. 10 System configuration CHI

A variety drive condition of Induction motor was compared that the results of experiment HAI controller indicated with FNN controller in this paper Fig. 11 and 12 are conclusion that the response features of FNN controller and HAI controller when the command speed drive a rated 1720[rpm] in no load condition. Fig. (a) is command speed and real speed, fig. (b) is indicated q axis current, fig. (c) is indicated generation torque. Experiment results about step command speed is highly reduced HAI controller speed more than FNN controller and the faster rising time, the faster reached steady-state. Fig. 13 and 14 is indicated that response characteristics of FNN controller and HAI controller when drives 1000[rpm] at no load and bring load-torque 1O[N.m]. While at a driving of constant speed, speed change is reduced and reached command speed faster the response characteristics of HAI controller than FNN controller. Fig. 15 and 16 is indicated that response characteristic of FNN controller and HAI controller when step command speed make the forward-reverse operation about -1200[rpm] and 1200[rpm]. CHI

1800

0

[A] 0

M.,i,,

0

)>

0, 00 V O.. 00 V 6.. 00 V V

CHI CH3 CH2 CH4

30. 00

(c) Te [N m]0o CH4

. 1 0 00

-5,. 00

-

1O', 00

v

5. 00

Fig. 13 Response characteristics of FNN controller with change of load torque Ik

S.. 00 V ol 00 v

CH1 CH3

n.

<<

>>

un

1000 (a) oJ[rpm]

(b)'qs[A]

0 20 0

0, 00 V

CH1

CH3

0. 00 V

CH2 CH4

6. 00 V V

30. 00

20

(c)Te[N-m] 0 CH.I .H

4

CH2 CH4

0 V

6.

4

0-

-

-5:. 00

0 0 0 'V 00 V

5. 0 0

Fig. 14 Response characteristics of HAI controller with change of load torque

j

20

(c) Te [N m]

20 0 20

(b)'qs [A]

20

(b)'qs

<<

(a) ,4rpm]

. ..

(a) C [rpm] 0

0

81 00 v

1000

<< M1ain >>

12, 00 V 12~00 V

CH3

8-. 00 v

CH3

t .H -5 0

..... ......... ~ ~ ~ ~ . . . . .

1

.0

Fig. 11 Response characteristics of FNN controller with step command speed

682

CH1

1200

8-. 00 v

CH3

81 00 v

<<

M.,m

REFERENCES

>)

[1] E. Cerruto, A. Consoil, A.Raciti and A. Testa, "adaptive fuzzy control of high performance motion systems," in Proc. IEEE IECON Conf. Rec., San Diego, CA, Nov. 9-13, pp. 88-94, 1992. [2] H. Hong, et al., "A design of auto-tuning PID controller using fuzzy logic," in Proc, IEEE IECON Conf, Rec., San Diego, CA, Nov. 9-13, pp 971-, 1992 [3] E. Cerruto, A. Consoil, P. Kucer and A. Testa, "A fuzzy logic quasi sliding mode controlled motor drive," in Proc. IEEE ISIE Conf. Rec, Budapest, Hungrary, June 1-3, pp. 652-657, 1993. [4] K. J. Astrom and B. Wittenmark, "Adaptive control," Addison-Wesley, 1989 [5] D. H. Chung, "Fuzzy control for high performance vector control of PMSM drive system," KIEE, vol. 47, no. 12, pp. 2171-2180, 1998. [6] M. G. Simoes and B. K. Bose, "Meural network based estimation of feedback sigmals for a vector controlled induction motor drive," IEEE Trans. IA, vol. 31, no. 3, pp. 620-629, 1995. [7] M. T. Wishart and R. G. Harley, "Identification and control of induction machines using neural networks," IEEE Trans. IA, vol. 31, no.3, pp. 612-619, 1995 [8] D.H. Chung, et al., "MRAC fuzzy control for high performance control of induction motor," The Trans. of KIPE, vol. 7, no. 3, pp. 215-223, 2002

(a) af[rpm] -1200 20

(b)'qs [A]

-8: 00 v -8-. 00 V 6-. 00 V v

CH1 CH3 CH2 CH4

20 20

(c) Te[N m]0 CH2 CH4

-5. 00

.10 00 V - 1O', 00 v

5. 0 0

Fig. 15 Response characteristics of FNN controller with change of step command speed 8,:00 v

CHI

<<Mni

>>

1200

()[rpm]

20

20

(c) Te [N m]

CHI

-.

....--------..1..

0

Fig. 16 Response characteristics of HAI controller with change of step command speed

8. CONCLUSION In this paper is proposed hybrid artificial intelligent controller for high performance control of induction motor drive. Fuzzy-neuro controller is consisted that term and conclusion part of fuzzy rule using clustering method and multilayer neural network. In this controller is obtained high performance, robust control that is indicated in fuzzy control's advantage and ability of high adaptive control that is indicated advantage of neural network. To improve more performance, we applied mechanism method that is based on reference model. Hybrid artificial intelligent controller is applied to induction motor drive system which the response characteristics were analyzed such as parameter change, steady-state and transient-state. Hybrid artificial intelligent controller rising time is faster than fuzzy-neuro controller and overshoot is indicated more small and can be assumed high performance with sampling time, speed, load torque and change of inertia. In this paper, hybrid artificial intelligent controller can be obtained satisfactory performance that response characteristics of parameter variation, transient-state and estimation of speed.

683