Damage identification in shear-type structures using a proper orthogonal decomposition approach

Proper orthogonal modes (POMs) obtained through proper orthogonal decomposition (POD), as a statistical pattern analysis technique, have been physically demonstrated to represent the dominant structure of the dynamic response data in previous study. In this paper, a novel POMs-based damage identification approach for shear-type buildings is developed. First, POMs of acceleration dynamic response of a shear-type building under Gauss White Noise (GWN) ground motion are obtained using singular value decomposition; then, the dominant POMs of acceleration response are used to identify the damage locations and severities through particle swarm optimization (PSO) algorithm. This proposed approach is applied to two three-story shear-type buildings in numerical simulation and a three-story shear-type frame in experimental study. The results demonstrate that the locations and severities of structural damage in shear-type buildings can be effectively identified by using the proposed method.