Vulnerability assessment of existing bridges to scour, based on an indirect monitoring approach and machine learning tools

Abstract

Detecting scour is possible by locating sensors on a railway carriage and processing these signals. In the first stage of the simulation phase of this study, the bridge ‘apparent profile’(AP) is computed by a Finite Element model. The model is used to derive the Moving Reference Influence Lines. Reducing support spring stiffnesses will correspond to the scoured case. In the second phase, off-bridge conditions will be taken into account. Then, measured AP, i.e., signals from the instrumented in-service train carriage both for scoured and scour-repaired cases, will be processed. Lastly, the foundation stiffnesses of the bridge will be computed by minimizing the sum of squared differences between the calculated AP and the corresponding measured AP. Loss of foundation compared to a healthy condition will prove the existence of scour.