COMPARISON OF THREE TYPES OF NEURAL SPEED ESTIMATORS FOR THE INDUCTION MOTOR DRIVE

Abstract

Presented paper describes design and the comparison of different structures of artificial neural networks (ANN) and their applying to dynamical system modelling. Shaft sensorless control of induction motor drives requires knowledge of the instantaneous value of the rotor speed. This paper considers the problem of speed estimation, based on neural modelling approach, in the shaft sensorless field oriented control structure with the induction motor (IM). Various internal structures of neural speed estimators are simulated and their performances are compared. Structures of the ANN estimators are based on measurable motor variables: components of stator current and voltage.