Control Of Magnetic Bearing Levitated Flexible Rotor System: A Convex Optimization Approach

This paper reports on an investigation of the control of active magnetic bearings (AMB) in suspended rotor systems. The AMB model is nonlinear inherently, and parameter uncertainty has to be taken into consideration for effective control design. For a lightly damped flexible structure rotor, the closed-loop system stability and performance are very sensitive to errors in the natural frequencies, while the high speed spinning rotor displays strong elastic characteristics and gyroscopic effects, which entail a linear parameter varying (LPV) system model. The effective control system design is a challenge. To meet robust control objectives and to design the controller in an LPV framework, we propose a robust LPV control design based on convex optimization. The overall LPV system model with uncertainty characterizations is formulated as a convex hull. A new H1 analysis condition which utilizes an extra variable on LMI conditions is formulated. A parameter dependent Lyapunov function for the closed loop convex hull can be constructed to greatly reduce controller design conservatism and take advantage of the acceleration/deceleration information of the rotor.