Experiments in fuzzy logic based control of a magnetic bearing system

The paper deals with the control of a levitated object with position dependent nonlinearity. The authors applied fuzzy modeling and control to a nonlinear magnetic bearing system to obtain uniform desired performance over the entire clearance. They represented the nonlinear magnetic bearing by the Takagi-Sugeno-Kang (TSK) fuzzy model, in which the nonlinear global model is approximated by a set of linear local models. Then a model-based fuzzy controller, the so-called parallel distributed compensation (PDC), is employed. Experimental results demonstrate that the controller provides excellent compensation for unstable/nonlinear characteristics of the magnetic bearing system and maximizes the controller's valid region of operation, while guaranteeing pre-specified transient performance.