Novel PMSM Speed Control Scheme Based on Improved Wavelet Neural Network
Using PI control strategy in the PMSM speed control system cannot guarantee good dynamic re-sponse performance and anti-disturbance ability at the same time,and the parameters setting is complicated in practical application.To improve the above problems,a modified wavelet neural network PI controller was designed for the speed loop.It used adaptive learning rate gradient descent with an inertial term for up-dating network parameters online and expanded the PI parameters output range by applying weight factors.These improvements enhanced the controller's performance.In order to eliminate the influence of current loop proportional parameter on speed loop,a deadbeat predictive current controller was designed to further improve the dynamic response performance of the system.The simulation results show that the above con-trol scheme can realize online self-tuning of the PI parameters of speed loop,significantly improve the dy-namic response performance of the system,and guarantee good anti-disturbance ability.
PMSMwavelet neural networkadaptive learning ratePI controlmodel predictive control
万方数据
维普
