Damage identification for high-speed railway bridges based on instantaneous energy

The structural damage affects the operation and safety of high-speed railway bridges and induces serious accidents. Hence it is necessary to identify the existing damage in time. With the increase of the train speed, the vehicle-bridge interaction (VBI) effect, which significantly influences the bridge dynamic responses, should be considered in damage identification. Therefore, this paper proposed a damage identification methodology for the high-speed railway bridge based on the instantaneous energy of train-induced bridge dynamic responses considering the VBI effect. The instantaneous energy is obtained by the Variational Mode Decomposition (VMD) method and the Hilbert transform (HT), which can represent the time-varying characteristics of the VBI system. Furthermore, the AutoEncoder is utilized for feature extraction of instantaneous energy. An AutoEncoder prediction error based-damage index is proposed. Then, the damage index and the defined early warning threshold are used to identify the operational state of high-speed railway bridges. A FEM model of a 32m length high-speed railway bridge is adopted as a numerical simulation to validate the effectiveness of the proposed approach. Simulated results suggest that the damage identification approach has relatively high computational efficiency and precision and less influence by the train speed.