PERFORMANCE EVALUATION OF VECTOR CONTROLLED INDUCTION MOTOR DRIVE BASED ON SPEED ESTIMATION

Abstract

With the advancements of power electronics and digital signal processor technology, electrical drives for induction machines, based on indirect vector control strategy are now widely used. This control strategy allows high performance to be achieved from an induction motor. Vector control or field-oriented control is the first method that enables the application of a linear strategy for the torque control of an induction motor. However the prerequisite for the correct operation of the vector controller is an accurate knowledge of the rotor velocity, which requires speed transducers such as shaft mounted tacho-generators, resolvers, or digital shaft position encoders. These degrade the system's reliability, and reduce the advantage of an induction motor drive system. In some cases it is difficult (e.g. a compact drive system) or extremely expensive (e.g. submarine applications), to use sensors for speed measurement. This has led to speed sensor less vector control in which the speed is estimated using some of the estimation methods rather than using speed sensors. The present work presents simulation of indirect vector control of induction machine based on speed estimation using Matlab/Simulink. ANFIS (Adaptive Network based Fuzzy Inference System) is employed to estimate the speed of induction motor. The technique is based on speed expression obtained from the induction motor dynamic equations. The input-output data patterns are obtained from computer simulation of a vector control system, which are used to train ANFIS (offline) using hybrid-learning algorithm. ANFIS is trained to obtain the speed information so that it can replace the speed transducer in the indirect vector control scheme.