Detection of Damage in Simply-Supported Plates by Discrete Wavelet Transform of Reconstructed Modal Data

Localized singularities caused by changes in the stiffness or mass of the damaged region cannot be simply visible through smooth mode shape curves. However, the wavelet transform of input signal can identify the location of defects by sudden changes in the spatial variation of transformed response. The aim of this study is to present a new method for damage detection in a squarely damaged plate with element of reduced thickness with damage ratio 3% and to decrease the possible random noises that are come from different sources such as numerical solutions or physical situation. In this approach, the value of each point of mode shape data is subtracted from or is added with its symmetric point with respect to the plate’s point of symmetry (point O). The obtained value is located at its first point, again. The result of this process is a symmetrical mode shape matrix of the damaged plate, which is utilized as the input signal function in wavelet analysis.The results demonstrated that using reconstructed modal data is superior to the original modal data and they provide a better crack indicator than the result of the DWT of the original mode shape data. The effects of crack location and sampling interval are examined. The simulated results show that this approach is more effective in damage detection of plate-like structures, as the method requires the knowledge of only the response of the damaged plate as a baseline for crack detection.