The direct method for adaptive feed-forward vibration control of magnetic bearing systems

As the actuators of active magnetic bearings (AMB) are open loop unstable a feedback control system is essential for their successful operation and for ensuring good performance. In addition, when rotor systems utilizing AMBs are operated at high speeds there will be synchronous vibrations due to unavoidable rotor mass unbalances, which generate synchronous disturbance forces. These forces and vibrations can often lead to severe problems in rotating machines including collision with the touchdown bearings at high speed and the possible risk of causing damage to the internal parts of the machines. In this paper, a new ''direct" adaptive feedforward algorithm is proposed for minimizing selected vibration performance measures which is based on the filtered-x LMS adaptive filter algorithm. This new adaptive algorithm has been used for adjusting the amplitude and phase of a synchronous signal injected at the summing junction of the bearing feedback control loop. The experimental results reported in this work show the effectiveness of the proposed method.