Repetitive Control of Magnetic Bearings

Abstract

A magnetic bearing has merits of no friction and no energy loss because it can support a rotor without physical contact. However, there are some problems such as poor system robustness and vibration caused by rotor imbalance. In this paper, we introduce two control methods to resolve these problems; disturbance cancellation control and repetitive control. First, the disturbing torque is estimated by the observer and fed back to the plant along with the position and velocity signals. This technique also cancels the difference between the actual control plant and its nominal model. As a result, the control system has accurate positioning and robustness from parameter variations. Repetitive control is a kind of servo control that follows a constant periodic command. Hence it is applied to a real magnetic bearing in order to remove the imbalance vibration which has the same period as the rotor speed. However, the rotor does not always rotate at a constant speed. Therefore, angular domain repetitive control is derived and applied to the magnetic bearing. The stability problem of the system is solved by an optimal regulator theorem. An experimental setup is made using rotational interruption, and the proposed method is tested.