Magnetic Bearing Control Using Interval Type-2 Fuzzy Logic

Magnetic bearings are an exciting and innovative technology that has seen considerable advances in recent years. Being unstable by nature, these systems require active control. Most often linear techniques are used very successfully. On the other hand, there are applications where linear methods have limited effectiveness. Fuzzy logic control performs very well in nonlinear control situations where the plant parameters are either partially or mostly unidentified. Its effectiveness for nonlinear systems also offers advantages to magnetic bearing systems. Type-2 fuzzy logic systems represent significant advances over traditional fuzzy logic systems in general. These fuzzy logic systems are capable to deal with uncertainties which can be found in almost every practical system. Uncertainties stem from several sources; noise present in the position input signals, the location and shape of fuzzy sets and the fuzzy rule-base describing the operation of the fuzzy controller, among others. Since a mathe-matical model of the controlled plant is often only a conveniently close approximation of the real process at hand, a major challenge lies in the application of the control methods to real plants. Type-2 fuzzy logic and fuzzy logic systems in general tackle the control problem at hand using human reasoning based on rules and expert knowledge of the plant described by human expressions. The current work consist of model development, controller design, simulation and experimental validation. The basic simulation model consist of a horizontal shaft supported by a radial magnetic bearing. The magnetic bearing is modeled as a nonlinear element. The controller designs are implemented and tested using a bench-top rotor rig equipped with a radial magnetic bearing. Some representative results are presented and discussed.