On-Board Monitoring of In-Service Trains for the Structural Health Monitoring of Railway Bridges

Railway systems are expanding rapidly worldwide as a 'green' means of transportation. Such rapid expansion poses a challenge toguaranteeing safety and ensuring the quality of railway systems. Continuous monitoring of tracks and rail bridges can help diagnoseand prognose damage timely, increasing railway transportation reliability. In this project, we focus on rail bridges, whose conditionassessment traditionally relies on a static network of sensors. Although well-established, this technique lacks portability, entailsconsiderable cost and effort, and, crucially, inhibits continuous monitoring, as the lifespan of sensors is generally smaller than that ofthe bridge. To tackle these limitations, indirect monitoring practices have recently gained attraction. Such methods extract thedynamic properties of bridges via measurements from traversing vehicles by exploiting the Vehicle-Bridge Interaction (VBI) principle.This way they offer mobility and lower installation and maintenance costs. However, the majority of studies on indirect monitoring ofbridges focus on highways. Despite the similarities of VBI modelling between highways and railways, distinct differences (e.g., vehicleand contact models, moving speed, loading patterns) do not allow the direct application of current indirect monitoring methods torailway bridges. Another challenge is the blurring effect of surface roughness, which, simplistically, is typically considered as known oris eliminated. To this end, this project proposes a consistent methodology to continuously monitor railway bridges via On-BoardMonitoring of in-service trains, with a simultaneous characterization of the involved roughness profiles. Accordingly, it proposes ahybrid modelling procedure that considers a VBI model (which represents actual train-bridge systems and considers the roughness ofthe involved track system) fused with vibration measurements collected from traversing trains.