On-line rotor resistance identification for induction motor drive with artificial neural networks supported by a simple PI stator resistance estimator

Abstract

This paper presents a new method of on-line estimation for the rotor and stator resistances of the induction motor in the indirect vector controlled drive, using artificial neural networks and a simple PI compensator. The back propagation algorithm is used for the training of the neural networks for rotor resistance identification. The error between the desired state variable of an induction motor and the actual state variable of a neural model is back propagated to adjust the weights of the neural model, so that the actual state variable tracks the desired value. A PI compensator is used to estimate and update the stator resistance used in the rotor resistance estimation in order to eliminate the error in rotor resistance estimation. The performance of the induction motor drive with the above rotor and stator resistance estimators is investigated in detail for torque and flux responses, to analyze the effects of stator resistance observer on rotor resistance identification, for variations in the stator and rotor resistances from their nominal values, conducting both simulations and experiments.