Damage detection in railway bridges by means of train on-board sensors

To ensure a safe passing of a train over a railway bridge it is necessary that the structural health condition of the bridge is checked regularly. Up to now this is generally done by inspection teams performing a visual check. Since this way of working is very time consuming and can only been performed periodically, alternative monitoring methods are being developed. This work analyses the feasibility of a monitoring method using acceleration measurements taken on board train, during daily operation, to identify possible structural damages on railway bridges. In this way the health condition is checked continuously, every time a train equipped with the monitoring system crosses the bridge. The idea is to compare the acceleration signals measured from the passing train with a reference signal acquired on the same bridge in reference conditions. Damages on the bridge have an influence on the interaction between the train and the bridge, which will result in a different acceleration signals. The location of the damage will be identified as the location along the bridge where the maximal difference between the measured signal and the reference signal occurs. The model used in the research is a FEM representation of several type of bridges, both considered as simply-supported beams and as complete three-dimensional models, with ballasted track. Different types of trains are also modelled as multi body systems considering wheelsets, bogies and car bodies.