基于VMD-SSI的结构模态参数识别  

Structural modal parameter identification based on VMD-SSI

作　　者：殷红[1] 董康立 彭珍瑞[1] YIN Hong;DONG Kangli;PENG Zhenrui(School of Mechanical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China;College of Biomedical Engineering&Instrument Science,Zhejiang University,Hangzhou 310027,China)

机构地区：[1]兰州交通大学机电工程学院,兰州730070 [2]浙江大学生物医学工程与仪器科学学院,杭州310027

出　　处：《振动与冲击》2020年第10期81-91,共11页Journal of Vibration and Shock

基　　金：国家自然基金项目(51768035);甘肃省高校协同创新团队项目(2018C-12);兰州市人才创新创业项目(2017-RC-66)。

摘　　要：将变分模态分解(VMD)和随机子空间法(SSI)结合,提出了基于VMD-SSI的结构模态参数识别新方法。针对VMD中的模态分层数K值确定困难的问题,提出模态重复比率准则,保证了模态信息的有效分解。依据模态重复比准则确定测量信号的最优分层数K;利用VMD方法进行信号并行分解,用奇异值分解(SVD)去噪,以提高模态参数的识别精度。用该研究提出的VMD-SSI方法识别模态固有频率和阻尼,用VMD方法辨识模态振型,将VMD-SSI法应用于外伸梁模型的模态参数识别,并利用统计理论分别检验识别的模态频率、模态阻尼和模态振型的精度。结果表明, VMD-SSI法识别模态参数的精度高于传统SSI法。In order to improve the modal parameter identification precision,a VMD-SSI modal identification method was proposed, which is based on variational mode decomposition(VMD) and stochastic subspace identification(SSI). A modal repetition ratio criterion was proposed for the optimization of mode number K, which ensures the effective decomposition of modal information. Firstly, the optimal K of measurement signal was determined according to the proposed modal repetition ratio criterion. Secondly, singular value decomposition(SVD) was used to denoise, which further improves the accuracy of modal identificati on. Thirdly, VMD-SSI method was proposed to realize the identification of structural modal parameters. Finally, the VMD-SSI method was applied to the modal identification of the overhanging beam model.The validity of the modal frequency, modal damping and mode shape was tested by statistical theory. The results show that the modal identification precision by VMD-SSI method is statistically higher than that by traditional SSI method.

关 键 词：变分模态分解(VMD) 随机子空间法(SSI) 模态参数识别 统计检验 

分 类 号：TU311.3[建筑科学—结构工程]