Damage Detection in Base-Isolated Steel Structure Using Singular Spectral Analysis

The role of structural health monitoring (SHM) in detection of post-earthquake damage to buildings is well established, and the same is extended to seismically isolated buildings in this research. In this study, a steel structure was considered with the isolation system which consists of elastomeric bearings. Numerical models were developed using SAP2000. The damage in the building was introduced by reducing the stiffness of all isolators in both horizontal directions unequally. After nonlinear response history analysis, the responses at two different floor levels of the healthy and damaged isolated superstructure were obtained, which are filtered using singular spectral analysis (SSA) and reconstructed using the dominant time series. The reconstructed responses were used to form a matrix that is decomposed, and singular vectors are obtained. The difference in the angle made by the singular vector in two-dimensional space obtained for healthy and damaged cases is the parameter for damage detection. SSA has successfully filtered the output responses, and the proposed algorithm has efficiently used the reconstructed responses to identify the damage in the structure.