A wavelet transform and substructure algorithm for tracking the abrupt stiffness degradation of shear structure

Time-varying parameter identification is an important research topic for structural health monitoring, performance evaluation, damage diagnosis, and maintenance. Practical civil engineering structures usually contain multiple degrees of freedom; however, damage often locally occurs. In this study, a discrete wavelet transform and substructure algorithm is presented for tracking the abrupt stiffness degradation of shear structures. A substructure model is built by the extraction of the local structure which may contain damaged region. Time-varying stiffness and damping are expanded into multi-scales using discrete wavelet analysis. An optimization method based on Akaike information criterion is introduced to select the decomposition scale. The expanded scale coefficients are evaluated using least square method, then the original time-varying stiffness or damping parameter is identified by reconstructing from the scale coefficients. To validate the proposed method, a numerical example of seven-story shear structure with time-varying stiffness and damping is proposed. Experiment for a three-story shear-type structure with abrupt stiffness degradation is also tested in the laboratory. Both numerical and experimental results indicate that the proposed method can effectively identify the abrupt degradation of stiffness parameter with a satisfactory accuracy