Damage Diagnosis Of Structure Based On Wavelet Analysis And Fractal Theory

Damage diagnosis technique is an active research area in current structure engineering and has widely engineering foreground. The correlative theory and technology are developing continuously. This thesis is concentrated on the research of the damage detection method by wavelet analysis and fractal theory. The main content is summarized as follows:According to multi-resolution characteristic of Wavelet transform, vibration signal of cantilever beam was decomposed by sym6 wavelet. Through the threshold value elimination and reconstruction, the high frequency noise was excluded effectively. Apply to theory of wavelet singularity detection, the vibration signal of damaged cantilever beam was decomposed and reconstructed. The position of the crack was located by the maximum modulus of the wavelet coefficients. The experiment results show that wavelet transform can locate the position of the crack exactly. This method is also in effect to light noisy signal. It demonstrates that wavelet transform has a certain characteristics of anti-noise.Base on detecting technology of wavelet packet frequency bnad energy, vibration signal of cantilever beam under different damage conditions was decomposed by wavelet packet. The energy components in different frequency were obtained, in which the sensitive one to damage was chosen as damage index. As experiments demonstrates, this method base on energy-damage is able to diagnose damage degree quickly even the mathematics model of structure is unknown.The fractal theory is introduced into area of damage diagnosis. This dissertation proposes correlation dimension as a indicator, which can be a proof of whether damage occurs or not. The main work of this part includes four aspects.1,The calculating method of box dimension for vibration signal was introduced. The program of box dimension was designed by MATLAB. The influence of noise on the computational precision of box dimension was discussed.2,The algorithm of correlation dimension (G-P algorithm) was interpreted and the method to choose parameters was given. Time-delay and Minimum Embedding Dimension...