Neural Network Based Dynamic.pdf

8/12/2019 Neural Network Based Dynamic.pdf

1/13

Neural Network Based Dynamic

Performance of Induction Motor Drives

P. M. Menghal and A. Jaya Laxmi

Abstract In industries, more than 85 % of the motors are Induction Motors,

because of the low maintenance and robustness. Maximum torque and efficiency isobtained by the speed control of induction motor. Using Artificial Intelligence (AI)

techniques, particularly the neural networks, performance and operation of

induction motor drives is improved. This paper presents dynamic simulation of

induction motor drive using neuro controller. The integrated environment allows

users to compare simulation results between conventional, Fuzzy and Neural

Network controller (NNW). The performance of fuzzy logic and artificial neural

network based controllers are compared with that of the conventional proportional

integral controller. The dynamic modeling and simulation of Induction motor is

done using MATLAB/SIMULINK and the dynamic performance of inductionmotor drive has been analyzed for artificial intelligent controller.

Keywords Neuro network (NNW) PI controller Fuzzy logic controller(FLC) Sugeno fuzzy controller Hebbian learning algorithm

P. M. Menghal (&)

Faculty of Degree Engineering, Military College of Electronics and Mechanical

Engineering, Secunderabad 500015, India

e-mail: [email protected]

Department of EEE, Jawaharlal Nehru Technological University, Anantapur 515002,

Andhra Pradesh, India

A. Jaya Laxmi

Department of EEE, Jawaharlal Nehru Technological University, College of Engineering,

Kukatpally, Hyderabad 500085, Andhra Pradesh, India

e-mail: [email protected]

M. Pant et al. (eds.), Proceedings of the Third International Conference on Soft

Computing for Problem Solving, Advances in Intelligent Systems and Computing 259,

DOI: 10.1007/978-81-322-1768-8_48, Springer India 2014

539


2/13

1 Introduction

Three phase Induction Motor have wide applications as electrical machines. About

half of the electrical energy generated in a developed country is ultimately con-sumed by electric motors, of which over 90 % are induction motors. For a rela-

tively long period, induction motors have mainly been deployed in constant-speed

motor drives for general purpose applications. The rapid development of power

electronic devices and converter technologies in the past few decades, however,

has made possible efficient speed control by varying the supply frequency, giving

rise to various forms of adjustable-speed induction motor drives. In about the same

period, there were also advances in control methods and Artificial Intelligence (AI)

techniques. Artificial Intelligent techniques mean use of expert system, fuzzy

logic, neural networks and genetic algorithm. Researchers soon realized that the

performance of induction motor drives can be enhanced by adopting artificial-

intelligence-based methods. The Artificial Intelligence (AI) techniques, such as

Expert System (ES), Fuzzy Logic (FL), Artificial Neural Network (ANN), and

Genetic Algorithm (GA) have recently been applied widely in control of induction

motor drives. Among all the branches of AI, the NNW seems to have greater

impact on power electronics and motor drives area that is evident by the publi-

cations in the literature. Since the 1990s, AI-based induction motor drives have

received greater attention. Apart from the control techniques that exist, intelligent

control methods, such as fuzzy logic control, neural network control, genetic

algorithm, and expert system, proved to be superior. Artificial Intelligent Con-troller (AIC) could be the best controller for Induction Motor control [16]. Fuzzy

controller conventionally is totally dependent to memberships and rules, which are

based broadly on the intuition of the designer. This paper tends to show Neuro

Controller has edge over fuzzy controller. Sugeno fuzzy controller is used to train

the fuzzy system with two inputs and one output [711]. The performance of fuzzy

logic and artificial neural network based controllers is compared with that of the

conventional proportional integral controller.

2 Dynamic Modeling and Simulation of Induction Motor

Drive

The induction motors dynamic behavior can be expressed by voltage and torque

which are time varying. The differential equations that belong to dynamic analysis

of induction motor are so sophisticated. Then with the change of variables the

complexity of these equations decrease through movement from poly phase

winding to two phase winding (q-d). In other words, the stator and rotor variableslike voltage, current and flux linkages of an induction machine are transferred to

another reference model which remains stationary [16].

540 P. M. Menghal and A. Jaya Laxmi


3/13

In Fig.1 stator inductance is the sum of the stator leakage inductance and

magnetizing inductance (Lls = Ls ? Lm), and the rotor inductance is the sum of

the rotor leakage inductance and magnetizing inductance (Llr = Lr ? Lm). From

the equivalent circuit of the induction motor in d-q frame, the model equations are

derived. The flux linkages can be achieved as:

1

xb

dwqs

dt vqs xe

xbwds Rsiqs 1

1

xb

dwdsdt

vds xexb

wqs Rsids 2

1

xb

dwqr

dt vqr xe xr

xbwdr Rsiqr 3

1xb

dwdrdt

vdr xe xrxb

wqr Rsidr 4

By substituting the values of flux linkages in the above equations, the following

current equations are obtained as:

iqswqs wmq

Xls5

ids wds

wmd

Xls 6

iqrwqr wmq

Xls7

Fig. 1 d q model of induction motor

Neural Network Based Dynamic Performance 541


4/13

idr wdr wmd Xls

8

where wmq and wmd are the flux linkages over Lm in the q and d axes. The flux

equations are written as follows:

wmq Xmlwqs

Xls wqr

Xlr

9

wmd XmlwdsXls

wdrXlr

10

Xml 11Xm

1Xls

1Xlr

11

In the above equations, the speed xr is related to the torque by the following

mechanical dynamic equation as:

Te Tload Jdxmdt

Tload J2p

dxr

dt12

then xr is achievable from above equation, where:

p: number of poles.J: moment of inertia (kg/m2).

In the previous section, dynamic model of an induction motor is expressed. The

model constructed according to the equations has been simulated by using

MATLAB/SIMULINK as shown in Fig. 2 in conventional mode of operation of

induction motor. A 3 phase source is applied to conventional model of an

induction motor and the equations are given by:

Va ffiffiffi2

p Vrmssin xt 13

Vbffiffiffi

2p

Vrmssin xt 2p3

14

Vcffiffiffi

2p

Vrmssin xt 2p3

15

By using Parks Transformation, voltages are transformed to two phase in the

d-q axes, and are applied to induction motor. In order to obtain the stator and rotor

currents of induction motor in two phase, Inverse park transformation is applied in

the last stage [4].



5/13

3 Fuzzy Logic Controller

The speed of induction motor is adjusted by the fuzzy controller. The membership

function of De, e and three scalar values of each triangle are applied into this

controller. In Table1, the fuzzy rules decision implemented into the controller are

given. The conventional simulated induction motor model as shown in Fig. 2 is

modified by adding Fuzzy controller and is shown in Fig. 3. Speed output terminal

of induction motor is applied as an input to fuzzy controller, and in the initial startof induction motor the error is maximum, so according to fuzzy rules fuzzy

controller produces a crisp value. Then this value will change the frequency of sine

wave in the speed controller. The sine wave is then compared with triangular

waveform to generate the firing signals of IGBTs in the PWM inverters. The

frequency of these firing signals also gradually changes, thus increasing the fre-

quency of applied voltage to Induction Motor [9].

As discussed earlier, the crisp value obtained from Fuzzy Logic Controller is

used to change the frequency of gating signals of PWM inverter. Thus the output

AC signals obtained will be variable frequency sine waves. The sine wave isgenerated with amplitude, phase and frequency which are supplied through a GUI.

Then the clock signal which is sampling time of simulation is divided by crisp

value which is obtained from Fuzzy Logic Controller (FLC). So by placing three

sine waves with different phases, one can compare them with triangular waveform

and generate necessary gating signals of PWM inverter. So at the first sampling

point the speed is zero and error is maximum. Then whatever the speed rises, the

error will decrease, and the crisp value obtained from FLC will increase. So, the

frequency of sine wave will decrease which will cause IGBTs switched ON and

OFF faster. It will increase the AC supply frequency, and the motor will speed up.

Figure3shows the Fuzzy logic induction motor model.

Torque

Speed

iqs

ids

iqr

idr

teta

Iabc

Ir-abc

Vqs

Vds

TL

iqs

ids

iqr

idr

Te

Wr

induction motor d-q model

Va

Vb

Vc

teta

Vqs

Vds

abc to d-qPark transformation

XY Graph

ICONR

To Workspace 2

ICONS

To Workspace 1

SCON

To Workspace

Scope1

Scope

1/s

Integrator

Final speed value

0

Constant 1

Va

Vb

Vc

AC Source

Fig. 2 Simulated induction motor model with conventional controller



6/13

4 Adaptive Neuro Fuzzy Controller

In the designing of a controller, the main criterion is the controllability of torque in

an induction motor with good transient and steady state responses. With certaindrawbacks, PI controller is able to achieve these characteristics. The main draw-

backs are (1) The gains cannot be increased beyond certain limit. (2) Non linearity

is introduced, making the system more complex for analysis. The shortcomings of

PI controller are overcome by artificial intelligent techniques. One such technique

is the use of Fuzzy Logic in the design of controller either independently or in

hybrid with PI controller. The draw-backs of Fuzzy Logic Control and Artificial

Neural Network are replaced by Adaptive Neuro-Fuzzy Inference System

(ANFIS). Adaptive neuro fuzzy combines the learning power of neural network

with knowledge representation of fuzzy logic. A neuro fuzzy system is based on a

fuzzy system which is trained by a learning algorithm derived from neural network

theory. Depending on the applications, one can use either ANN or FIS, or com-

bination of both. In this paper, the inputs will be e(k) and De(k) [9,12, 17]. A first-

order Sugeno fuzzy model has rules which are as follows:

Table 1 Modified fuzzy rule decision

De

NB NS ZZ PS PB

e PB ZZ NS NS NB NBPS PS ZZ NS NS NB

ZZ PS PS ZZ NS NS

NS PB PS PS ZZ NS

NB PB PB PS PS ZZ

Speed contrller

d-q to abc

Park transformation

stator current Scope

rotor current Scope

Continuous

powerguiiqs

ids

iqr

idr

teta

Iabc

Ir-abc

Vqs

Vds

TL

iqs

ids

iqr

idr

Te

Wr


Va

Vb

Vc

teta

Vqs

Vds

abc to d-qPark transformation Torque Scope

In

1

1 4 3 6 5 2

Speed Scope

IGBT1

IGBT4

IGBT3

IGBT6

IGBT5

IGBT2

Va

Vb

Vc

PWM ac source

0

Load Torque

1/s

Integrator

Feedback

Reference

U(k)

Fuzzy controller

Divide

1710

1

Fig. 3 Fuzzy control induction motor model



7/13

Rule1: If x is A1 and y is B1, then f1 = p1x ? q1y ? r1

Rule2: If x is A2 and y is B2, then f2 = p2x ? q2y ? r2.

In the Sugeno model ifthen rules are used, and output of each rule is linear

combination of inputs plus a constant value. The learning algorithm applied to this

model is Hebbian. This method is feed forward and unsupervised and the weights

will be adjusted by the following formula:

wi new wi old xiy 16The ANFIS layout is shown in Fig. 4. It states that if the cross product of output

and input is positive, then it results in increase of weight, otherwise decrease ofweight.

In layer 2 of ANFIS layout, the triangular membership function is same as that

of the fuzzy controller model. The output of layer 2 is given by:

O2 l1; l2; l3 17Layer 3 indicates the pro (product) layer and its output is product of inputs,

which is given by:

O3

li e

lj

De

18

Layer 4 represent Norm and it calculates the ratio ofith firing strength to sum of

all firing strengths. The obtained output is normalized firing strength, which is

given by:

Fig. 4 ANFIS layout



8/13

O4 wiPw

19

Layer 5 is an adaptive node with functionality as follows:

O5 wifi wi pi e qi De ri 20where pi, qi, ri are consequent parameters, which are initially are set to 0.48, 0.25

and 1 respectively. Then they are adaptively adjusted with Hebbian learning

algorithm. Layer 6 calculates the output which is given by:

O6P

wifiPwi

21

Figure5shows the overall structure of Adaptive Neuro-Fuzzy controller.

5 Neuro Controller

The most important feature of Artificial Neural Networks (ANN) is its ability to

learn and improve its operation using neural network training [13, 14]. The

objective of Neural Network Controller (NNC) is to develop a back propagation

algorithm such that the output of the neural network speed observer can track the

target one. The network structure of the NNC, indicates that the neural network has

three layered network structure. The first is formed with five neuron inputsD(xANN(K? 1)), D(xANN(K)), xANN, xS(K- 1), D(xS(K- 2)). The second

layer consists of five neurons. The last one contains one neuron to give the

command variation D(xS(K)). The aim of the proposed NNC is to compute the

Speed contrller

d-q to abc

Park transformation

stator current Scope

rotor current Scope

Continuous

powerguiiqs

ids

iqr

idr

teta

Iabc

Ir-abc

Vqs

Vds

TL

iqs

ids

iqr

idr

Te

Wr


Va

Vb

Vc

teta

Vqs

Vds

abc to d-q

Park transformationTorque Scope

In1

1 4 3 6 5 2

Speed Scope

IGBT1

IGBT4

IGBT3

IGBT6

IGBT5

IGBT2

Va

Vb

Vc

PWM ac source

0

Load Torque

1/s

Integrator

Feedback

Reference

U(k)

Fuzzy controller

Divide1710

1

Fig. 5 Adaptive neuro-fuzzy controller simulation model



9/13

command variation based on the future output variation D(xANN(K? 1)). Hence,

with this structure, a predictive control with integrator has been realised. At time k,

the neural network computes the command variation based on the output at time

(k? 1), while the later isnt defined at this time. In this case, it is assumed that

xANN(K? 1) : xANN(K). The control law is deduced using the recurrent

equation given by,

xS K xS K 1 GD xS K :It can be seen that the d axis and q axis voltage equations are coupled by the

terms dEand qE. These terms are considered as disturbances and are cancelled by

using the proposed decoupling method. If the decoupling method is implemented,

the flux component equations become

Udr G s vdsUqr G s vqs

Large values of g may accelerate the ANN learning and consequently fastconvergence but may cause oscillations in the network output, whereas low values

will cause slow convergence. Therefore, the value ofghas to be chosen carefully

to avoid instability. The proposed neural network controller is shown in Fig. 6.

6 Simulation Results and Discussion

Modeling and simulation of Induction motor in conventional, fuzzy and adaptiveneuro fuzzy are done on MATLAB/SIMULINK. A complete simulation model

and dynamic performance for inverter fed induction motor drive incorporating the

proposed FLC, adaptive neuro fuzzy controller and Neuro controller has been

2

is

1

wo

Wo

idq

Vdq

X1

X2

X3

X4

X5

X6

X7

Y1

Y2

Y3

Y4

Y5

Y6

Y7

Sub system 1

Pure linear

Neuron2

Pure linear

Neuron1

Ia

Ib

Va1

Vb1

TL

W

Mech_ANN

1

sIntegrator6

1

s

Integrator1

u

2

TL

1

V

Fig. 6 Neural network controller



10/13


11/13

0 500 1000 1500-5

0

5

10

15

20

Time

Torque

0 0.5 1 1.5 2 2.5 3

x 105

-60-40-20

02040

6080

100120

Time

Tor

que

0 0.5 1 1.5 2 2.5 3x 105

-60-40-20

02040

6080

100120140

Time

Torque

0 2 4 6 8 10 12 14x 104

-60-40-20

020406080

100120

Time

Torq

ue

(a)

(b)

(c)

(d)

Fig. 8 Torque characteristics. a Conventional controller. b Fuzzy controller. c Adaptive neuro

fuzzy controller. d Neuro controller

0 0.5 1 1.5 2 2.5x 106

0200

400

600

800

1000

1200

1400

1600

1800

Time

SPEED

0 0.5 1 1.5 2 2.5 3

x 105

-2000

200400600800

10001200140016001800

Time

Speed

0 0.5 1 1.5 2 2.5 3x 10

5

-2000

200400

600800

1000

12001400

1600

1800

Time

p

0 0.5 1 1.5 2 2.5 3x 10

5-200

0200400600800

1000

1200

1400

16001800

Time

Speed

(a)

(b)

(c)

(d)

Fig. 9 Speed characteristics. a Conventional controller. b Fuzzy controller. c Adaptive neuro

fuzzy controller. d Neuro controller



12/13

controller, FLC and adaptive neuro controller under dynamic conditions which are

shown in Fig. 8. With the neuro controller, speed reaches its steady state value

faster as compared to Conventional, FLC and adaptive neuro fuzzy controller.

7 Conclusion

In this paper, comparison of simulation results of the induction motor are pre-

sented with different types of controller such as conventional, fuzzy control,

Adaptive neuro fuzzy and neuro controller. From the speed waveforms, it is

observed that with adaptive fuzzy and neuro controller the rise time decreases

drastically, in the manner which the frequency of sine waves are changing

according to the percentage of error from favourite speed. According to the directrelation of induction motor speed and frequency of supplied voltage, the speed will

also increase. Fuzzy controller has better performance than the conventional

controller. By comparing neuro controller, Adaptive neuro fuzzy model with FLC

model, it is apparent that by adding learning algorithm to the control system will

decrease the rising time more than expectation and it proves neuro controller has

better dynamic performance as compared to FLC and Adaptive neuro fuzzy

controller.

Appendix A

The following parameters of the induction motor are chosen for the simulation

studies:

V = 220 f = 60 HP = 3 Rs = 0.435 Rr = 0.816 Xls = 0.754

Xlr = 0.754 Xm = 26.13 p = 4 J = 0.089 rpm = 1,710

References

1. Chan, T.F., Shi, K.: Applied Intelligent Control of Induction Motor Drives. IEEE Willey

Press (2011)

2. Krause, P.C.: Analysis of Electrical Machinery and Drives System. IEEE Willey Press (2000)

3. Mohan, N.: Advanced Electric Drives: Analysis, Control Modeling using Simulink.

MNPERE Publication (2001)

4. Menghal, P.M., Laxmi, A.J.: Adaptive neuro fuzzy based dynamic simulation of inductionmotor drives. IEEE International Conference on Fuzzy Systems, pp. 18 (2013)

5. Menghal, P.M., Laxmi, A.J.: Neural network based dynamic simulation of induction motor

drives. IEEE International Conference on Power, Energy and Control (2013). doi:10.1109/

ICPEC.2013.6527722

http://dx.doi.org/10.1109/ICPEC.2013.6527722http://dx.doi.org/10.1109/ICPEC.2013.6527722http://dx.doi.org/10.1109/ICPEC.2013.6527722http://dx.doi.org/10.1109/ICPEC.2013.6527722


13/13

6. Menghal, P.M., Laxmi, A.J.: Adaptive neuro fuzzy interference (ANFIS) based simulation of

Induction motor drive. Int. Rev. Model. Simul. (IRMOS) 5(5), 20072016 (2012)

7. Shi, K.L., Chan, T.F., Wong, Y.K., Ho, S.L.: Modeling and simulation of the three phase

induction motor using SIMULINK. Int. J. Elect. Eng. Educ. 36, 163172 (1999)

8. Dandil, B., Gokbulut, M., Ata, F.: A PI type fuzzyneural controller for induction motor

drives. J. Appl. Sci. (2005). doi:10.3923/jas.2005.1286.1291

9. Kumar, R., Gupta, R.A., Surjuse, R.S.: Adaptive neuro-fuzzy speed controller for vector

controlled induction motor drive. APEJ (2009). doi:14.79e41505757a2a8cab

10. Denai, M.A., Attia, S.A.: Fuzzy and neural control of an induction motor. Int. J. Appl. Math.

Comput. Sci. (2002). doi:10.1.1.135.303

11. Subudhi, B., Anish Kumar, A.K., Jena, D.: dSPACE implementation of fuzzy logic based

vector control of induction motor. IEEE TENCON (2008). doi:10.1109/TENCON.2008.

4766502

12. Bose, B.K.: Neural network applications in power electronics and motor drivesan

introduction and perspective. IEEE Trans. Ind. Electron. (2007). doi:10.1109/TIE.2006.

888683

13. Uddin, M.N., Hafeez, M.: FLC-based DTC scheme to improve the dynamic performance ofan IM drive. IEEE Trans. Ind. Appl. (2012). doi:10.1109/TIA.2011.2181287

14. Uddin, M.N., Wen, H.: Development of a self-tuned neuro-fuzzy controller for induction

motor drives. IEEE Trans. Ind. Appl. (2007). doi:10.1109/TIA.2007.900472

15. Uddin, M.N., Radwan, T.S., Rahman, A.: Performance of fuzzy logic based indirect vector

control for induction motor drive. IEEE Trans. Ind. Appl. (2002). doi:10.1109/TIA.2002.

802990

16. Uddin, M.N., Huang, Z.R., Chy, M.M.I.: A simplified self-tuned neuro-fuzzy controller based

speed control of an induction motor drive. IEEE Power Engineering Society General Meeting

(2007). doi:10.1109/PES.2007.385720

http://dx.doi.org/10.3923/jas.2005.1286.1291http://dx.doi.org/14.79e41505757a2a8cabhttp://dx.doi.org/10.1.1.135.303http://dx.doi.org/10.1109/TENCON.2008.4766502http://dx.doi.org/10.1109/TENCON.2008.4766502http://dx.doi.org/10.1109/TIE.2006.888683http://dx.doi.org/10.1109/TIE.2006.888683http://dx.doi.org/10.1109/TIA.2011.2181287http://dx.doi.org/10.1109/TIA.2007.900472http://dx.doi.org/10.1109/TIA.2002.802990http://dx.doi.org/10.1109/TIA.2002.802990http://dx.doi.org/10.1109/PES.2007.385720http://dx.doi.org/10.1109/PES.2007.385720http://dx.doi.org/10.1109/TIA.2002.802990http://dx.doi.org/10.1109/TIA.2002.802990http://dx.doi.org/10.1109/TIA.2007.900472http://dx.doi.org/10.1109/TIA.2011.2181287http://dx.doi.org/10.1109/TIE.2006.888683http://dx.doi.org/10.1109/TIE.2006.888683http://dx.doi.org/10.1109/TENCON.2008.4766502http://dx.doi.org/10.1109/TENCON.2008.4766502http://dx.doi.org/10.1.1.135.303http://dx.doi.org/14.79e41505757a2a8cabhttp://dx.doi.org/10.3923/jas.2005.1286.1291

Neural Network Based Dynamic.pdf

Documents

Transcript of Neural Network Based Dynamic.pdf