Study On Control Schemes Of Stability For Active Magnetic Bearings

The active magnetic bearing can be called AMB briefly. There is no any mechanical contact between the rotor and stator in AMB system. AMB system does not need the lubrication and relevant sealing device, and it can be kept on the equilibrium position which is set beforehand by using AMB controller. Compared to the conventional bearings, the AMB has the advantages of low noise, less wear, long life, working in vacuum environment and wide working temperature.However, AMB in essence is kind of typical nonlinear system which is reflected on the nonlinear relationship between control current and rotor displacement. This characteristic will lead to the failure of keeping the balance of rotor if the AMB system does not use the closed-loop controller. Besides, the AMB system may be subjected to the disturbance of noise which probably affects the practical control accuracy and further increases the working noise. In addition, as to major rotating parts, the effect of mass unbalance is inevitable. Nevertheless, the mass unbalance will cause the periodical vibration of bearing base and increase working noise.The study object of this paper is eight-pole radial active magnetic bearings. As to the problems illustrated above, this paper firstly develops the improved fuzzy PID controller which is based on the thought of variable universe of proportion scaling to realize the stability of AMB system. Then, in order restrain the effect of noise and mass unbalance, the AMB controller has been further optimized by introducing Kalman filter and LMS (Least Mean Square) filter. The simulation results show that:1) Compared to the conventional bearings, the fuzzy PID controller can adjust the PID parameters online and realize the nonlinear control of system. As to the complex working conditions which are subjected to alternate load and accidental impact load simultaneously, the fuzzy PID controller can adjust the PID parameters based on the current system state to make sure that the AMB system can adapt the complex working conditions.2) Compared to the conventional fuzzy controller, the improved fuzzy PID controller has better identification accuracy of rotor displacement. In this method, the interval ranges will be scaled proportionally along with the practical variable values. Specifically, the interval ranges will be enlarged if variable values are much larger than initial values. On the contrary, the interval ranges will be shrunk. As this process does not increase the inference rules, it indirectly improves the control accuracy of identifying tiny displacement error.3) In the relevantly short time, LMS filter can effectively eliminate the periodical mass unbalance signal from rotor displacement signal in order to avoid the periodical vibration of bearing base. As LMS filter does not affect the system stability, there is no need to modify the controller structure and control scheme, which actually makes the controller easy to use.4) By introducing the Kalman filter, the system noise of AMB system can be effectively restrain. The simulation result shows that the impacts of noise disturbance can be significantly limited by connecting with Kalman filter, and the maximum steady-state error can be decreased to about 85%. This capacity is highly beneficial to practical applications because it provides an alternative to the use of costly sensors with good signal-to-noise ratio.