SVPWM control of permanent magnet synchronous motor based onneural network dynamic tuning

Abstract

To sovle high torque ripple at low speed, high frequence noise and slow response caused by torque ripple for conventional control systems of permanent magnet synchronous motor(PMSM), the scheme of Digonal Recurrent Neural Network(DRNN) Dynamic speed Controller for space vector PWM(SVPWM) control of a permanent magnet synchronous motor(PMSM) is presented. The structure of dynamic PID speed Controller based on digonal recurrent neural network(DRNN-PID), and an dynamic learning algorithm are analyzed. The results of simulation are provided to demonstrate system superior characteristics with low torque ripple, few harmonics and the effectiveness of the proposed DRNN dynamic speed Controller under the occurrence of parameter variations and external disturbance.