Studies And Comparisons Of The Methods For Structural Damage Identification

It is very important to study the methods of structural damage identification, because structural damage of architectural engineering may cause the serious accidents of the construction, influence the healthy operation of national economy, and endanger the security of people's life and property.The dynamic characteristic of structures has a direct relationship with structural parameters, and structural damage can cause dynamic characteristics shifts correspondently. Therefore, if the mapping relationship between structural damage and dynamic characteristic shifts can be established, the damage can be diagnosed by using dynamic measurements of the structures.In this thesis, surrounding the subject of structural damage identification, there would be the introduction, analysis, comparisons of the methods of structural damage identification, mainly including flexibility matrix method, wavelet transform method, neural network method etc.As to the flexibility matrix method for damage identification, the flexibility of the structure would be changed if the damage is occurred in any part of the structure, Thus, the damaged location and damaged degree of the structure can be identified due to variations of flexibility matrix. Numerical simulation results demonstrate that this method is capable of locating and quantifying structural damage.Wavelet transform method has the characteristic to pick out the singularity from the whole signal with singular property. A wavelet-based approach is proposed for structural damage identification. Occurrence of the damage and the locations of the damage when it occurs can be determined by the wavelet decomposition of dynamic signal. It shows that structural damage or the change in system stiffness may be detected by spikes in the details of the wavelet decomposition of the response data, and these spikes may accurately indicate the locations when the structural damage occurred. The method will have a broad application in engineering structural damage identification.With the merits of its strong nonlinear mapping ability, rapid computation and anti-interference capability, artificial neural network, which is used in the identification of structural damage based on vibration, becoming more popularly. However, for large scale complicated structures, the network structure is usually complicated and identification efficiency is low.The method based on BP neural network to identify the damaged location and damaged degree has been presented. The results show that this method can reduce the complexity of the network and improve learning efficiency effectively.Finally, there would be the detailed comparisons about the virtue and the disadvantage among flexibility matrix method, wavelet transform method, and BP neural network method, and would prospect the methods of structural damage identification.