一种基于共轭梯度法的动态载荷识别方法  

A multisource dynamic loads reconstruction method based on conjugate gradient method

作　　者：王林军[1] 曹慧萍[1] 

机构地区：[1]三峡大学机械与动力学院,湖北宜昌443002

出　　处：《舰船科学技术》2016年第2期69-73,共5页Ship Science and Technology

基　　金：国家自然科学基金资助项目(11202116);三峡大学人才科研启动基金资助项目(KJ2011B033)

摘　　要：在载荷识别过程中,由于噪声和结构矩阵病态的问题,对测量的响应数据有影响。本文运用共轭梯度法最优化方法,来解决多源动态载荷识别中的不适定问题。将该方法应用在加筋板模型和平面桁架结构上进行数值仿真,仿真算例结果表明所提出的载荷识别方法在测量位移响应数据含有5%的噪声水平下,有效地得到稳定近似解,实现了载荷识别。该方法抗噪能力强,有效解决反问题的不适定性。The influence of measurements of the response data in the process of load identification because of noise and inverse problem of structure matrix. In this paper,conjugate gradient method is applied to solve an ill-posed problem of muti-source dynamic loads reconstruction. The loads were identified on a stiffened plate model and plane truss structure. The numerical examples indicate that the proposed load reconstruction method can effectively get stable approximate solution and realize the load identification using displacement response data under 5% noise. This method has a strong anti-noise ability and effective to solve the inverse problem of the ill-posedness.

关 键 词：载荷识别 共轭梯度法 反问题 不适定性 

分 类 号：TB123[理学—工程力学]