Estimation and rejection of unknown sinusoidal disturbances using a generalized adaptive forced balancing method

This paper presents a method for estimation and rejection of sinusoidal disturbances with unknown amplitudes and frequencies arising in feedback control systems. A method known as adaptive forced balancing has been used successfully to reject sinusoidal disturbances with known frequency and unknown amplitude in magnetic bearing control systems. The occurrence of such a disturbance is a result of a common problem to all mechanical systems with rotating shaft known as mass unbalance. Because mass unbalance consists of a sinusoidal sensor disturbance within the bandwidth, conventional methods suffer from instability or fail to preserve desired bandwidth. AFB overcomes these problems and offers unique features which preserve stability and performance robustness. In this paper, we extend AFB with a frequency tracking capability. This is achieved by using two different adaptive observer structures. We show that the modified version of AFB is capable of estimating both amplitude and frequency of the disturbance signal and demonstrate its rejection in the feedback loop.