几种模态参数盲辨识方法的比较研究  

COMPARATIVE STUDY OF SEVERAL BLIND MODAL PARAMETER IDENTIFICATION TECHNIQUES

作　　者：樊江玲[1] 张志谊[1] 华宏星[1] 

机构地区：[1]上海交通大学振动,冲击,噪声国家重点实验室,上海200030

出　　处：《振动与冲击》2006年第5期153-157,165,共6页Journal of Vibration and Shock

基　　金：国家自然科学基金资助项目(10302019)

摘　　要：对一多自由度振动系统的模态参数辨识进行了仿真研究。由于该系统具有弱小模态特征,各阶特征能量相差悬殊,不同辨识方法的可辨识性、辨识精度及数值运算量都不相同。分别利用频域分解法、时频域Gabor展开方法、传统子空间辨识法、改进子空间辨识法以及基于高阶统计量的盲方法对该多自由度系统进行了模态参数辨识,分析和比较了各种盲辨识方法的辨识性能。A vibration system of multiple-degrees-of-freedom is used for the comparative study of several blind techniques such as Peak Picking, Frequency Domain Decomposition(FDD), Gabor Expansion, Stochastic Subspace Identification (SSI, traditional and improved) and Higher Order Statistics(HOS). Due to the great disparity in mode energy, weak modes of the vibration system are in fact difficult to be identified with common techniques. The performances of these blind techniques can be assessed in terms of their capability in extracting weak modes as well as the accuracy of estimated parameters. Each blind technique is tested with this system. According to the theoretical results, the improved SSI technique seems to be the most favourable one that can obtain all modes with higher accuracy and reasonable time consume.

关 键 词：模态参数辨识 盲方法 FDD GABOR展开 子空间方法 高阶统计量 

分 类 号：O324[理学—一般力学与力学基础] TH113[理学—力学]