Structural health monitoring using adaptive wavelet functions / Seyed Hossein Mahdavi

• Main Author: Seyed Hossein, Mahdavi
• Format: Thesis
• Published: 2016
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: http://studentsrepo.um.edu.my/6178/4/hossein.pdf; http://studentsrepo.um.edu.my/6178/

In this study, two aspects of structural dynamic problems have been considered, involving direct structural dynamics as well as inverse problems. The first part of this research is directed towards improving an explicit and indirect time integration method for structural dynamic problems capable of using adaptive wavelets. The developed scheme is comprehensive enough for use with any wavelet basis function. To investigate the applicability of different wavelet functions for different problems, in particular, the simple family of Haar wavelets, the complex and free-scaled Chebyshev wavelets of the first (FCW) and second kind (SCW) and Legendre wavelets (LW) have been evaluated. A detailed assessment is carried out on the stability, accuracy and computational efficiency of responses calculated by Haar wavelet, FCW, SCW and LW. The proposed method lies on an unconditionally stable scheme, hence, there is no requirement on the selection of the time interval. This allowed the numerical procedure to be performed on long time increments. Practically, an efficient structural health monitoring strategy is the resultant of the implementation of an enhanced structural simulation through inverse problem approach. As a consequence, the computational performance of structural health monitoring strategies will be directly influenced by the higher computational competency and convergence rate of the proposed wavelet-based method for structural simulation. Accordingly, in the second part of this research, the procedure of structural identification and damage detection has been developed by employing the wavelet-based method through the modified genetic algorithms (GAs) to optimally solve inverse problems. For this purpose, a wavelet-based GAs strategy is improved by using free-scaled adaptive wavelets to optimally identify unknown structural parameters. The appropriateness and effectiveness of the proposed strategy have been evaluated both numerically and experimentally. The iv numerical assessment demonstrated the robustness of the proposed technique for identification and damage detection of large-scaled structures with the best performance. For the experimental validation, three test setups were conducted for identification and damage detection, including two different MDOF systems and a 2-dimentional truss structure. Consequently, it was shown that the computational efficiency of structural identification and relatively, damage detection strategies were significantly enhanced. This led the optimum results with the highest accuracies and provided the sufficiently reliable strategy in assessing the structural integrity, safety and reliability.