ABSTRACT
During the last two decades a large amount of research work has been conducted to develop and deploy bridge health monitoring (BHM) frameworks based on a sensory system directly mounted on a bridge, or so-called direct monitoring. Although, direct monitoring has shown promise, it suffers from the huge cost associated with the hardware installation, data transmission and regular maintenance of the system, to name a few. Alternatively, developing BHM frameworks based on vehicle-mounted sensing or indirect monitoring, is currently a rapidly growing research area as they overcome many of the problems with the direct approach. The premise of these methods is to monitor the vibration responses measured from sensors installed on a vehicle passing over a bridge to identify any change in the structural integrity of that bridge. In this study, successful implementation of such an idea is numerically presented. A novel data-driven approach using a nonlinear dimensionality-reduction technique using Uniform Manifold Approximation and Projection (UMAP) is proposed to analyse the vibration response collected from a moving vehicle. For the purpose of validation, a simply supported bridge model is considered. To create multiple states of the bridge, in addition to the intact state, the structural state of the bridge is altered. For each bridge state, the dynamic response of a moving vehicle while slowly passing over is collected. The proposed method is then applied to identify changes that occur in the bridge. The results for damage characterization demonstrate that various bridge states can successfully be separated from one another. The research presented in this work can open up new opportunities for the sustainable condition monitoring of bridge network.
