Adaptive Fault-Tolerant Control of Permanent Magnet Synchronous Motor Based on RBF Neural Network

Abstract

Permanent magnet synchronous motors (PMSM) have become one of the mainstream motors commonly used in the field of high-precision servo systems because of their advantages in power, size and fast response performance. In this paper, we design a RBF neural network adaptive fault tolerant control (RBF-AFTC) method to reduce this disturbance and makes the motor system more stable by addressing the problem that the collected feedback speed contains sinusoidal signal disturbance due to the failure of the speed sensor of the PMSM speed control system. The RBF neural network is applied to approximate the given speed, and the adaptive compensation control law is designed to replace the traditional PI algorithm. It is demonstrated that the designed Lyapunov function is bounded and tends to infinity with time, and the speed error gradually tends to zero. In the process of adaptive fault tolerance, the control law changes as the system speed signal occurs and can be adaptively reconfigured. The simulation results show that the RBF-AFTC has better steady-state performance and improved dynamic performance than the conventional PI control when a sinusoidal signal is superimposed on the collected speed under both no-load and load conditions, while the torque pulsation is smaller and the three-phase current is more uniform and smooth.