A MIMO data driven control to suppress structural vibrations

Abstract

A novel data driven control method is presented to control vibrations of active flexible structures with multiple control inputs and multiple sensing outputs. The evaluation of vibration suppression performance and a MIMO stability criterion are formulated based on the sampled frequency response data, which results in a non-convex controller design problem solved by a global optimization technique. Such method is experimentally applied on an active cantilever beam, which demonstrates the low-order data driven controllers are very competitive with a model-based mixed sensitivity controller and can achieve promising vibration suppression for flexible structures. Without requiring dynamic models, significant advantages might be expected by the proposed method when the accurate reduced-order models of space structures are not available nor handy to be obtained.