Research On Fuzzy Modeling And Fuzzy Control Of Active Magnetic Bearing Systems

Active magnetic bearings (AMB) are novel supporting instruments that have been attracting scholar's great attention. Their bearing structure which is made up of controlled electric magnet with current coils is different from traditional sliding bearing and rolling bearing. There is no mechanical contact between the bearings and the rotor, and the electromagnetic force can suspend the rotor at the expecting position in the air. The structural speciality make AMB system have advantages of no contact, no friction, no lubrication, longevity of service, and so on, but also make AMB system have the characteristic of nonlinearity, instability and parameter uncertainty which brings difficulty to the design of controller.Because of its instability under open-loop state, AMB system can work well only with the help of the closed-loop controller that is very important to the application of AMB system in practice. At present, the usual method is to get the linearization model at the balance point of AMB system, then to design the linear controller. But under control of the linear controller, the system has long settling time and big over-shoot;at the same time, the rotor's stable operating range is small. Thereof, this paper tries to design a controller that not only can improve the system's dynamic performance but also can extend the rotor's stable operating range.In order to design a good controller, a suitable modeling method , that can well discribe AMB system's nonlinear characteristic , must be found. Thus the first main part work is to model AMB system with T-S fuzzy model. Two different models are founded separately when the rotor is near to and far away from the balance point, then the two models' effective combination is taken as the system's final fuzzy model. The simulation results demonstrate that the fuzzy model not only can approach the high nonlinear relation of AMB system, but also can fit the system's various input-output data precisely.The next main part work is to design the controller based on the founded T-S fuzzy model. According to the principle of PDC controller, linear controller is designed separately with regard to AMB system's various subsystem described by the consequence of the T-S fuzzy model, then the weigted combination of every subsystem controller's output is made as the total control current of AMB system. The simulation results in Matlab demonstrate that the PDC controller's effect is much superior to that of the traditional PID controller.