Damage detection of bridge structures under moving loads based on CEEMD and PSD sensitivity analysis

This study proposes a new damage identification method based on a combination of complete ensemble empirical mode decomposition (CEEMD) and power spectrum density (PSD) sensitivity analysis to analyze the acceleration signals of bridge structures under moving loads and achieve damage detection of bridge structures. This paper has achieved the ability to accurately identify the location of cracks and the extent of the damage along a girder with only one acceleration sensor arrangement. The measured data is processed by the CEEMD method. The damage location is revealed by directly examining the first-order intrinsic mode function corresponding to the highest-order pseudo-frequency component, which presents an abrupt change at the damage location. Secondly, after determining the damage location of the bridge, only the power spectrum sensitivity analysis of the crack parameters at the damage location is required to obtain the damage level, avoiding the need to blindly solve the power spectrum for all elements. Finally, the identification method is validated by considering environmental noise, damage locations, and crack depths. The numerical simulation results and experiments for various working conditions show that the method adopted in this paper has good identification capability in identifying cracks in bridge structures.