ABSTRACT
Indirect field orientation (IFO) induction machine drives are increasingly employed in industrial drivesystems, but the drive performance is often degrades. Motor works on best performance at certain voltageand frequency for certain loads. In this paper artificial neural network is used to predict the operatingvoltage and frequency when the load torque and speed going changed so motor efficiency is increased.Simulation and experimental results are shown to validate the scheme.
I. INTRODUCTION
Because of their easy implementation and lowcost, indirect field oriented (IFO) inductionmachine drives are finding numerous industrialapplications. Most of the industrials motors areused today are in fact induction motors. Inductionmotors have been used in the past mainly inapplications requiring a constant speed becauseconventional methods of their speed control haveeither been expensive or highly inefficient.Various methods have been designed to achievethe best performance of the motor and theapproaches used can be classified into threedifferent types. The first, so called “loss modelcontroller” (LMC). The second type, named“search controller” (SC). The third one uses “lookup tables”.This type of control scheme uses moremathematical calculations and algorithms, whichinvolves heavy computing and needed efficientand costly controllers. Here we introduce a noveltheory to improve the performance of the motorrunning it at optimum voltage and frequency foroptimum motor efficiency at different points. Thisis an offline method and not for online and realtimecontrol.The present paper proposed the new approach tooptimize the efficiency by a control technique withthe help of Artificial Neural Network (ANN). TheANN based controller gives the ratio of optimumvalue of voltage and frequency, which is used todetermine the optimum flux operates the motor atits maximum efficiency for given value of torqueand speed.Like the human brain an NN may beconstructed with many artificial neurons. A neuroncan be modeled to perform a mathematicalfunction such as a pure linear function, stepfunction, tan-sigmoid function etc. The attractivefeature of NN is that it can be trained to solvecomplex nonlinear function with variableparameter, which may not be attainable, byconventional mathematical tools [5].A computer mat lab program is developed forgetting the ratio of optimum voltage andfrequency. Neural Network is automaticallytrained itself by compare the actual voltage andfrequency with the sample corresponding data. Themotor flux is in the ratio of the ratio of optimumvoltage and frequency. So this flux is comparedwith the reference flux and generates the errorsignal and control the action of the inductionmotor
.II. PERFORMANCE OF INDUCTIONMOTOR
Energy supplied to the induction motor isdistributed in the two parts, the first is in the formof mechanical output and second one is in the formof losses. For the high performance of the motorthe motor losses should be small, so the output ofmotor goes high. An efficient motor not only savesthe energy, hence money, but will also generateless internal heat, and run cooler and more quietly[6]. It is also likely to last longer and more reliablethan a less efficient motor.


download full report
http://googleurl?sa=t&source=web&cd=1&ve...ol3no3.pdf&ei=UNmSTYLQGcewcZqUuIkH&usg=AFQjCNGVAOGxuzSumQQ-2TNVtFMhSMxD7A&sig2=TaW6bGcrQYrTZ_JqYhg48Q