Abstract
Adaptive feedforward broadband vibration (or noise) compensation requires a reliable correlated measurement with the disturbance (an image of the disturbance). The reliability of this measurement is compromised in most of the systems by a “positive” internal feedback coupling between the compensator system and the correlated measurement of the disturbance. The system may become unstable, if the adaptation algorithms do not take into account this positive feedback. Instead of using classical infinite impulse response (IIR) or finite impulse response (FIR) feedforward compensators, this paper proposes and analyses an IIR Youla-Kucera parametrization of the feedforward compensator. A model-based central IIR stabilizing compensator is used, and its performance is enhanced by the adaptation of the parameters (Q-parameters) of an IIR Youla-Kucera filter. Adaptation algorithms assuring the stability of the system in the presence of the positive internal feedback are provided. Their performances are evaluated experimentally on an active vibration control system. Theoretical and experimental comparisons with FIR Youla-Kucera parameterized feedforward compensators and IIR feedforward compensators are provided.

This publication has 22 references indexed in Scilit: