Predictive control based on fuzzy model: a case study

The implementation of the fuzzy predictive functional control (FPFC) on the magnetic suspension system is presented. The magnetic suspension system is in our case the pilot plant for magnetic bearing and is an open-loop unstable process, therefore a lead compensator is used to stabilize it. The high quality control requirements include a periodical step response and zero steady-state error. Adding the integrator to a feedback causes overshoot. The solution to the problem is a cascade control with fuzzy predictive functional controller in the outer loop. To cope with the unknown model parameters and the nonlinear nature of the magnetic system, a fuzzy identification based on the FNARX model is used. After a successful validation the obtained fuzzy model is used for the controller design. The FPFC is compared with a cascade PID control. The results obtained with the FPFC are very promising and comparable with the conventional control techniques.