Abstract
In a rotary machine, the mass imbalance of the rotor often causes vibration problems. Therefore, balancing of the rotor mass is generally required. Magnetic bearing support can provide forces which counteract disturbance arising from the mass imbalance and can keep the rotor in the vicinity of a desired position.
The main contribution of this paper is to develop a method of compensation for mass imbalance using the periodic learning control method. Using the periodic learning control method makes the compensation accurate even under the condition that identification of parameters of the system is incomplete. In this study, inverse transfer function compensation is combined with the periodic learning control method. The error converges to zero with less repetitions. Experimental study with an instrument incorporating gyro mechanism assures the effectiveness of this control scheme.
