Q-MRAS-Based Speed Sensorless Permanent Magnet Synchronous Motor Drive with Adaptive Neural Network for Performance Enhancement at Low Speeds

Abstract

In this paper, a speed sensorless vector-controlled permanent magnet synchronous motor (PMSM) drive is discussed. The vector-controlled PMSM drive shows improved dynamic performance over the classical control technique of controlling the PMSM. From the viewpoint of cost, reliability, compatibility, and environmental issues, the PMSM is operated without speed sensor. Therefore, we require some speed estimation strategies to operate the motor in closed loop. There are number of speed estimation techniques available in the literature which compute the speed from the terminal variable (i.e., voltage and current). All the methods of speed estimation available in the literature have their own merits and demerits. The reactive power-based model reference adaptive system (Q-MRAS) speed estimator gives poor performance at low/near zero speeds, for low torques. In this paper, the performance of Q-MRAS is improved at these speeds by the use of artificial neural networks (ANNs) in the adjustable model of the Q-MRAS-based speed estimator. The proposed algorithm is simulated in MATLAB/Simulink, and the corresponding results are presented.