A neural network based stator current MRAS observer for speed sensorless induction motor drives

Abstract

This paper presents a novel model reference adaptive system (MRAS) speed observer for induction motor drives based on stator currents. The measured currents are used as reference model for the MRAS observer to avoid the use of a pure integrator. A two layer Neural Network (NN) stator current observer is used as the adaptive model which requires the rotor flux information. This can be obtained from the voltage or current model but instability and dc drift can downgrade the overall observer performance. To overcome these problems another off-line trained multilayer feedforward NN is proposed here as a rotor flux observer. Speed estimation performance of the MRAS scheme using the three different rotor flux observers is studied and compared when applied to an indirect vector control induction motor drive. Promising results have been obtained when using the NN flux observer with less sensitivity to parameter variation and stability in the regenerating mode of operation.