结构控制中输出反馈鲁棒极点配置优化方法

OPTIMAL METHOD OF OUTPUT FEEDBACK ROBUST POLE ASSIGNMENT PROBLEM IN STRVCTURAL CONTROL

给出了输出反馈鲁棒性极点配置的数学描述和数值解法，基于任何结构控制都要求极点 对系统矩阵元素的改变不敏感，通过选择鲁棒度量，并以其尽可能小为判据，用优化方法寻求 近似解．最后将该方法应用于一个大跨度桥在确定的运动载荷作用下的控制，计算结果表明方 法合理、有效．

in this paper, the solvability of Output Feedback Pole Assignment Problem is discussed. The numerical method of Output Feedback Robust Pole Assignment is also described. By computing the QR decomposition of the control location matrix and observation matrix, the orthogonormal bases of state vectors can be constructed. A desirable property of any system design is that the poles should be insensitive to perturbations in the coefficient matrices of the system equations. We use robust pole assignment technique to design congtrollers. We choose the measure of robustness baized on the consistent relations between two selected vectors. Minimizing the measure of robustness of the closed-loop system thus ensure other desirable properties of the system. This criterion may be used to restrict the degrees of freedom in the pole assignment problem. Since the measure of robustness is expected to be as small as possible,we consider this problem as an unconstrained optimization problem. To:find the approximate solution, we use BFGS optimal algorithm here. This algorithm was presented by Broyden, Fletcher, Goldfarb and Shanno in 1970. Its numerical stability is better than DFP algorithm. BFGS optimal algorithm can be proved to be superlinear convergence. Finally, the robust.pole assignment method and its application to the control of a long span bridge under a deterministic moviing load are analysed. The midspan displacement and acceleration descend rapidly after control. The performance of the method is demonstrated adequately. Aside from specific advantages such as maintained specified performance, this method warrants the robustness of the controlled system. The numerical method is reasonable and valid.