Finite Element Model Based Predictive Control OfSmart Structures

Abstract

Model based control for smart structures requires an accurate enough representation of the coupled dynamics of the structure between the locations of the sensors and actuators. The model based predictive control proposed in this paper has the advantage of generating actuators commands taking into account the current measurements as well as a prediction of the immediate and future states and control commands. In the first part of the paper, various models of the structure are analysed from the viewpoint of efficiency and real time computability for active control of position and damping of structural vibrations. Single structural stiffness matrix, mass-stiffness matrices and finite elements model for non-minimal phase and multiple natural frequencies representations are compared. In the second part, model based predictive control is proposed for achieving a stable non-collocated position control for a flexible structure. The effects of the sampling time and the lengths of the receding horizon are analysed in terms of the efficiency of the feedforward and feedback parts of the controller. Simulation results for simple flexible structures are presented for illustrating the performance of the proposed model and controller.