A NUMERICAL INVESTIGATION OF NEW ALGORITHMS FOR THE DRIVE-BY METHOD IN RAILWAY BRIDGE MONITORING

Railway Infrastructure Manager increasingly demand for new and effective method for the monitoring of infrastructure health, so that a relevant effort is being placed in the development of condition based monitoring and maintenance techniques for railway infrastructure. This includes track geometry, overhead contact line and the structural health of railway bridges. This paper proposes an algorithm for the bridge monitoring of railway bridges, based on accelerations measured on board train (drive-by method). The diagnostic algorithm is developed and validated based on the results of a 3D numerical model simulating the dynamic interaction of the bridge/track/train structure. The analysis is focused on steel truss bridges, on which the degradation of the connection of one diagonal element is simulated (e.g. lose bolts, deteriorated welding). In comparison with previous research works, this paper tries to overcome the need of comparison to a reference signal corresponding to the healthy bridge, which is the technique on which most of literature papers rely on. This is achieved by exploiting the roll component of the train (difference between right and left acceleration), excited as a consequence of the presence of the defect which generates an asymmetry in the bridge behavior. The numerical results show good accuracy in damage identification and location, good robustness to vehicle speed, to the presence of electrical noise and, in the case of a relevant damage, to track irregularity.