Adaptive rejection of sinusoidal disturbances of known frequency acting on unknown systems

Adaptive feedback algorithms for the rejection of periodic disturbances are useful in many practical applications involving rotating equipment. The simplest algorithms require knowledge of the disturbance frequency as well as of the system's transfer function or frequency response. This paper considers the case when the plant dynamics are unknown. An existing approach consists in obtaining an estimate of the plant frequency response in a first step and in using the estimate to update the control signal in a second step. The paper presents a new algorithm that enables simultaneous estimation of the frequency response and control signal. The parameters of the controller are updated continuously, as opposed to existing techniques where the parameters are held constant during blocks of time.