Modeling and control of a magnetic bearing system

This paper presents the pedagogy of digital control design and experiment for a magnetic bearing system as part of a graduate level digital control class. The system identification verifies the system model structure, where the unstable MIMO system is kinematically de-coupled into SISO systems with model uncertainty bounds characterized for robustness analysis. Low order SISO plant models are used for control design while a high order MIMO model is used for verification before the implementation. Following a simple lead compensator, model based control design, analysis, verification, and implementation include state estimator feedback control and its augmentation with integrator and oscillator internal models; all stabilizing control and minimum variance control; approximate plant inversion for feedfoward tracking and repetitive control.