Identification And Nonlinear Control On Magnetic Floating Bearing

The Magnetic Bearing, which floats depend on magnetic force, is a new type of high performance bearing without friction between rotor and stator. It is widely used in fields of mechanism, turbine, vacuum and so on. But it is strongly nonlinear for vortex, hysteresis and magnetic saturation, so can hardly be identified by traditional identification model. Fortunately, the Neuro Network is proved superior in approximating the arbitrary nonlinear system. Then, this article identify the Magnetic Bearing with Neuro Network(BP and ELMAN network) on the hardware base which designed and made before. Finally, after the model completed, a kind of nonlinear PID algorithm is discussed to control the bearing, and simulated on MATLAB.