A Wavelet Based Damage Identification For Shipbuilding Gantry Cranes

With the rapid development of China’s economy, shipbuilding industry has experienced significant growth and advancement in the last decade. Shipbuilding gantry cranes are large, massive and complex structures that are used for lifting extremely heavy loads and are subject to harsh and corrosive working environment, which makes them susceptible to damage and deterioration. These damages accumulate gradually during the lifetime of these structures, and the damage accumulation may lead to catastrophic failures, collapse of these structures and fatal accidents. Hence, the development of a reliable and effective damage identification method can help provide viable damage information that can be used to prevent further damage accumulation and to repair the present damage.In this research a comprehensive overview of several existing damage identification methods is presented. Subsequently, considering the many useful features of wavelet analysis, as a time-frequency method, over other techniques, it is selected as a basis for the development signal analysis tool for damage identification. Given the type and the nature of the damage in gantry cranes, and considering several wavelets, proper and most suitable wavelet functions are considered and a specific wavelet based methodology is proposed.The proposed wavelet based time-frequency technique is capable of detecting singularities in the mode shapes of a structure. Finite element models of beam structures, single layer frames and scaled gantry cranes are analyzed to obtain the response signals of the structure. Based on these mode shapes obtained via corresponding acceleration and displacement response data, damages are successfully identified.By utilizing long-gage fiber optic strain sensors, a modified wavelet packet energy rate (MPWER) is introduced. This approach is based on the fact that long-gage strain signal is sensitive to damage. Based on the analytical results and the comparison of wavelet packet energy rate and envelope area of strain-time curvature, this MPWER is proposed as a viable damage identification technique. Computational results are utilized to identify the feasibility of MPWER. The results show that MPWER is capable of identifying the location as well as the severity of damage. Moreover, the results demonstrate that MPWER is robust to noise.Finally, experimental study is conducted on a simply supported steel beam. The strain response signals are used to calculate the MWPER. Based on the MWPER, the location of damage is detected with a high degree of accuracy and the severity of the damage from the computational results closely match the experimental ones. This experiment confirms the feasibility of MWPER as a viable technique for detecting the location as well as the severity of the damage.The comprehensive study carried out in this research, consisting of analytical, computation and an experimental study verifies the viability of the proposed wavelet based method for damage identification and provide the basis for the construction of a comprehensive structural health monitoring tool for gantry cranes.