An Adaptive Speed Sensorless Induction Motor Drive With Artificial Neural Network for Stability Enhancement

An artificial neural network (ANN) based adaptive estimator is presented in this paper for the estimation of rotor speed in a sensorless vector-controlled induction motor (IM) drive. The model reference adaptive system (MRAS) is formed with instantaneous and steady state reactive power. Selection of reactive power as the functional candidate in MRAS automatically makes the system immune to the variation of stator resistance. Such adaptive system performs satisfactorily at very low speed. However, it is observed that an unstable region exists in the speed-torque domain during regeneration. In this work, ANN is applied to overcome such stability related problem. The proposed method is validated through computer simulation using MATLAB/SIMULINK. Sample results from a laboratory prototype (using dSPACE-1104) have confirmed the usefulness of the proposed estimator.