Abstract:

Permanent magnet synchronous motors have nonlinear and strong coupling characteristics, and it is difficult to accurately control them with conventional vector control methods. Besides, the motor system is susceptible to load disturbances, resulting in speed and electromagnetic torque fluctuations. In view of the problem of slow system response and large overshoot caused by fixed speed loop parameters, this study proposes a fuzzy radial basis function neural network PID control strategy to replace the PID control of speed loop in the vector control system. Based on the incremental PID control method, this strategy combines neural network and fuzzy control, and uses the gradient descent optimization algorithm to dynamically adjust the PID parameters in the speed loop. The simulation results show that the overshoot of the motor system controlled by the fuzzy neural network PID is small. Compared with conventional PID control, the proposed method has reduced the start-up time of low-speed and high-speed operation by 66.7% and 75.9%, respectively, and has faster dynamic response, better robustness and anti-interference ability. The experimental platform is built using DSP, and the experimental results prove the effectiveness of the control method.

Key words: permanent magnet synchronous motor, server system, fuzzy logical controller, vector control, radial basis function neural network, increment PID, gradient descent, torque pulsation