Structural condition assessment from recorded earthquake response data.

Monitoring of changes in vibration response has been widely used for assessment of structural integrity, performance and safety. In particular, on-line health monitoring and diagnostic assessment of critical infrastructure becomes vital after the occurrence of extreme events such as earthquakes and blast loads. However, it is found difficult, if not impossible, to characterize structural condition and detect time-varying, non-stationary spectral characteristics using the traditional frequency spectrum analysis method. In this paper, a wavelet signal analysis technique is implemented for post-earthquake damage assessment directly from recorded response data. The results obtained from numerical simulations and real response data from moderate to severe earthquakes indicate that wavelet analysis is a powerful tool that can be used to identify hidden transient characteristics and high energy bursts that may occur due to progressive structural stiffness degradation and/or pounding between structural components. The appearance of multiple closely spaced transient peaks in the details of the wavelet decomposition coefficients is used for the assessment of the occurrence of progressive damage in the structural system. Finally, the structural response behavior during earthquake excitation is characterized from a critical investigation of the details of the local high frequency signal energy variations in the time-frequency domain.