ABSTRACT
Drive-by methodologies have emerged as a feasible and cost-effective monitoring solution for accurately and early detect damage on railway bridges while minimizing train operation interruptions. This study presents a novel approach for drive-by damage detection in high-speed railway (HSR) bridges. The methodology relies on acceleration records collected from multiple bridge crossings by an operational train equipped with onboard sensors. Log-Mel spectrogram features derived from the acceleration records are used together with sparse autoencoders for computing statistical distribution-based damage indexes. Numerical simulations were performed on a 3D vehicle-track-bridge interaction system (VTBI) model implemented in Matlab® to evaluate the robustness and effectiveness of the proposed approach. The model considers several damage scenarios, vehicle speeds, and environmental and operational variations (EOVs), such as multiple track irregularities and varying measurement noise. The results showed that the proposed approach can successfully detect and localize damages, as well as characterize their severity, especially for very early-stage damages.
