小波變換是近年來發展起來的一種新的時頻分析方法。它具有多分辨率分析(Multiresolution Analysis)的特點，而且在時頻兩域都具有表徵信號局部特徵的能力，所以成為眾多領域的研究重點。傳統梁構件損傷檢測是利用損傷前與損傷後之模態振型向量來進行各種指標計算，如模態尺度因子(MSF；Modal Scale Factor)、坐標模態確認準則(COMAC；Coordinate Modal Assurance Criterion)、模態振型差法(Modal Shape Deflection Method)與模態振型曲率變化法(Method of Modal Shape Curvature Change)等…。但在實際的工程應用中較難取得損傷前的模態振型向量，因此本論文結合了小波變換具有表徵信號局部特徵的優點和古典結構模態參數辨識方法-分量分析法，對梁結構損傷後之模態振型向量進行小波係數奇異性檢測，並進行室內實驗，對此新的損傷定位方法進行討論。 Wavelet analysis is a new time-frequency analytic technique which is developed in recent years. It becomes a key area of many researching due to its “Multiresolution Analysis” and its capability to reveal characters of partial signals in both time and frequency domain. In traditional beam member damage location identification, mode shapes of before and after damaged are used to calculate indices such as MSF(Modal Scale Factor), COMAC(Coordinate Modal Assurance Criterion), Modal Shape Deflection Method, MMC(Method of Modal Shape Curvature Change) etc.. But in real circumstance, it is difficult to obtain the mode shapes before damaged. Therefore this discourse presents a new damage location identification method, which combines both classical system identification method–Component Analysis and wavelet analysis. In this new method, the mode shapes of damaged structures are obtained by Component Analysis. Then the mode shapes are analyzed by wavelet transformation to get the positions of the damages by showing a peak in the distributions of the wavelet coefficients. We conduct lab experiment to recognize and discuss this new damage location identification method in this discourse.
