A neural network model reference adaptive system for speed estimation of sensorless induction motor

Abstract

A neural network model reference adaptive system is designed for speed estimation of sensorless induction motor. The neural network identifier is used as reference model to replace the voltage model of rotor flux. This improves the accuracy of speed estimation under the interference of parameters such as winding resistance and mutual inductance. On the basis of model reference adaptive system, the single neuron adaptive algorithm is applied to increase the robustness of system compared with the PI regulation algorithm. The simulation results have shown that the neural network model reference adaptive system has better accuracy and stronger stability than ordinary model reference adaptive system for speed estimation of sensorless induction motor.