Structural Damage Diagnosis Based on the Temporal Moment of Partially Measured Structural Responses

Abstract

Structural damage diagnosis is still a challenging task because most current methods are either insensitive to local structural damage or sensitive to measurement noise. A statistical moment–based structural damage detection (SMBDD) algorithm has been proposed to locate and detect damages, revealing superiority in noise immunity. However, it requests that the number of measured responses should be no fewer than that of unknown structural parameters. In this paper, to reduce the number of measurements required in the SMBDD algorithm, an improved method for damage diagnosis is proposed based on the temporal moment of partially measured structural responses. Firstly, structural partial acceleration responses are measured and split into several time segments. Then, the temporal moment in each segment of the measured acceleration response time history is estimated. Finally, an objective error function is established by the temporal moments of measured accelerations and calculated accelerations, and structural stiffness can be identified by minimizing the objective error function. The proposed method is simple and feasible with a robust antinoise property and can identify structural damage when the number of measured responses is lower than that of the structural stiffness. Numerical simulations and an experimental study are conducted, respectively, to verify the feasibility and effectiveness of the proposed method.