Vibration Control of Composite Beams Using Adaptive Positive Position Feedback

This paper presents a new method of active vibration control (AVC) of flexible composite cantilever beams. A recursive least mean squares (LMS) based filter is used to perform on-line system identification, updating a positive position feedback (PPF) controller in real-time. With conventional time-invariant PPF control the damping ratio and natural frequency are selected based on the structural response, with the feedback path taking the form of a second-order filter; therefore having performed an off-line system identification of the plant dynamics, this controller is relatively straightforward to implement. Some structures undergo parameter changes over time and actuation in the control sequence can contribute to this change in the system dynamics. PPF control is optimized when an accurate estimation of the natural frequency is available. It is demonstrated that the recursive LMS algorithm can be used to provide a continual on-line estimation of the plant dynamics, for robust control of a slender glass fiber beam. The result is an increase in vibration attenuation, which is robust to variations in the beam properties. It is shown that this controller delivers a high performance settling time when compared with the standard PPF controller for free vibration in the first mode. Robustness of the technique is further demonstrated by altering the fundamental frequency of the beam, via the addition of a lumped mass. The adaptive controller re-tunes itself to this parameter variation, ensuring a high level of vibration suppression is upheld.