SENSOR-LESS FIELD ORIENTED SPEED CONTROL OF INTERIOR PERMANENT MAGNET SYNCHRONOUS MOTOR DRIVES USING MRAS OBSERVER AND NEURAL ADAPTIVE CONTROLLER

Abstract

In this thesis, develops a Model Reference Adaptive System (MRAS) is used to estimate speed for interior permanent magnet synchronous motor (PMSM) using Neural Network (ANN) .The algorithm is based on field oriented control theory. The proposed control scheme was a model reference adaptive system (MRAS) speed observer based on stator currents with space vector pulse width modulation technique. The objective of the proposed sensor-less control is to improve the speed control performance and robustness of PMSM drive under load variations. For speed estimation performed at low speed, pi-radial basis function neural network observer is proposed instead of the traditionally used pi controller in adaptation mechanism of the MRAS.In this work, the traditionally used pi-controller in adaptation mechanism of the MRAS observer is replaced by pi-radial basis function neural network observer. The improvement in Speed response during load variations is shown by comparative study between conventional PI and neural PI. Simulation result showed the performance of speed estimator under various operating conditions using MATLAB/SIMULINK environment