基于小波多分辨滤波特性的结构损伤识别  

Structural Damage Identification Based on Wavelet Multi-resolution Filter

作　　者：唐和生[1] 薛松涛[1] 陈镕[1] 王远功[1] 

机构地区：[1]同济大学结构工程与防灾研究所,上海200092

出　　处：《力学季刊》2004年第1期152-155,共4页Chinese Quarterly of Mechanics

基　　金：国家杰出青年科学基金项目(59925820)

摘　　要：小波分析作为一个很好的损伤识别工具,可以看作是传统傅立叶变换的扩展,小波变换采取了可调整的时频窗口,因此小波变换的优势是它具有了“可变焦”性能对局部信号进行多尺度的刻画。小波基的伸缩和平移系列,使小波变换可看作是一组带通滤波器,本文全面分析了小波变换多尺度带通滤波器特性以及在结构在线损伤识别中的应用。结构损伤的出现体现在结构物理参数的改变,相对应的动力响应信号将会产生局部时变特性,利用小波分析的多尺度带通滤波器在不同尺度下对结构振动信号作滤波分析,通过观察不同带宽内振动信号的时频变化来判断结构损伤的存在。Wavelet analysis, as a promising tool for damage identification, may be viewed as an extension of traditional Fourier transform with adjustable time-frequency window location and size. The advantage of wavelet analysis lies in its ability to examine local data with a 'zoom lens having adjustable focus' to provide multiple levels of details and approximations of the original signal. With the translated and dilated forms of wavelet basis, the wavelet transform can be seen as a family of multi-band filters. The characteristics of multi-scale and band-pass filter for the wavelet transformation were analyzed; a new approach for structural online damage identification was proposed based on the characteristics. Damaged structures will show some slow or abrupt changes in their physical properties, which lead to the time variation of the dynamic signal. Using the band-pass filter of variant scales, the filtering analysis for dynamic signal of structure was carried out. In the case of abrupt damages, the occurrence of damage and the moment when it occurs can be clearly determined through the observation of signal changes within the variant band in time-frequency domain.

关 键 词：小波变换 多尺度带通滤波器 损伤识别 结构损伤 振动信号 

分 类 号：O346.5[理学—固体力学] TN713.5[理学—力学]