Research On Fuzzy PID Control Of AMB

Active magnetic bearing (AMB) is a novel supporting component, and attracting great attentions because of the advantage of no mechanical touch, no frays and free lubrication compared with the traditional bearing. AMB can only work well with the help of control system. The controller affects the dynamic performance of AMB and the control precision of the rotor directly.The most important research field of AMB focuses on the research on the controller. At present, the usual idea is to get the linearization model at the balance point of the AMB, then to design the traditional PID controller based on linear theory. However, AMB has the characteristic of instability, non-linearity and parameter incertitude, the classical linear controller can not satisfy the control requirements. What's more, the unknown load disturbances will affect the control effect, and result in the bad even invalid control results.In this article, the theory of operation for AMB is introduced. Mechanical model of differential model is deduced and motion equation of rotor is built up. MATLAB simulation rerult of PID control shows that it is only suitable for small range near by the equivalent point. Owing to the nonlinear characteristic of AMB, fuzzy control can expand the levitating range. Fuzzy-PID control method, including parameter self-tuning Fuzzy-PID control and Fuzzy-PID double mode control, is researched and simulation results indicate that the dynamic performance is improved.Making use of the high mapped ability of neural networks, a strategy which combines the fuzzy control with single neuron control (F-N) is proposed for the AMB control system: At the beginning, the response of the rotor produces a very high system error, the main controller at this point is fuzzy controller in order to improve the dynamic performance. When system is approaching to the steady state, fuzzy controller and single neuron controller are combined to overcome the defect of the fuzzy control. Simulation results demonstrate that the combined controller has improved not only the dynamic performance of AMB but also the disturbance rejection ability. Subsequently, PID control based on the BP neural network is discussed, and then combined with fundamental fuzzy control to constitute the NNF-PID controller. Simulation results show that this controller has better performance compared to the traditional PID controller.According to the digital control system of AMB, this article represents the design of hardware components, such as controller, displace sensor, analog filter, A/D converter, D/A converter and power amplifier. Furthermore, the realization of control arithmetic is introduced. Finally, experimental results are also illustrated.