DESIGN DEVELOPMENT AND PERFORMANCE INVESTIGATIONS OF SENSORLESS CONTROL OF INDUCTION MOTOR DRIVES USING NEURAL NETWORK STATE OBSERVERS

A novel method is described which considerably improves the performance of rotor flux MRAS based sensorless drives in the critical low an d zero speed region of operation. It is applied to a vector controlled induction motor drive and is experimentally verified. This new technique uses a neural network (NN) as a rotor flux observer to rep lace the conventional voltage model. This makes the reference model free of pure integration and less sensitive to stator resistance variations. Experimental and simulation results reveal some int eresting features of NN and shows that the network has good potential for use as an alternative to voltage model used in classical MRAS speed observer