Active Magnetic Bearing Controller Design based on Radial Basis Function Neural Network

This paper aims to design a high-performance and highly reliable controller, enabling stable levitation control of the active magnetic bearing (AMB) system. The distinctive features of AMB are open-loop instability and strong nonlinearity, which a PID controller can stabilize. PID controller is dependable and widely used, whereas it generates a large overshoot magnitude. Because of these problems, this paper utilizes the PID controller to stabilize the closed-loop system of AMB and uses a radial basis function neural network (RBFNN) to optimize PID manipulator parameters according to the operating conditions. And the optimization can reduce the impact of nonlinearity. Plus, Simulink simulation for PID controller based on RBFNN is carried out, which proves that in comparison to classical PID algorithm, the oscillation during the suspension is shortened substantially, and the performance of anti-interference can become stronger with the increase of learning time.