Drive-by structural health monitoring of railway bridges using train mounted accelerometers

Bridge damage can be detected by observing changes in its spectral properties. In its infancy, bridge health monitoring involved monitoring physical properties via direct instrumentation, i.e. sensors attached to the bridge. In recent years many authors have investigated the ability of indirect methods to assess the structural health of bridges, i.e. the vehicles traversing the bridges are fitted with sensors. This has the potential of reducing monitoring costs as the vehicle may be used to monitor many bridges on the network. Most of the investigation in this relatively new field of study has been on road bridges and road vehicles. A method is proposed in this paper for the detection of the bridge damage through an analysis of vehicle accelerations resulting from the train/track/bridge dynamic interaction. In a train/track/bridge interaction there are additional complications which do not exist on road bridges. The signal generated by the train as it traverses the bridge is normally short in duration. Studies on railway bridges are complicated by the addition of rails, sleepers and sometimes ballast between the tracks and the bridge deck. However, the weight of the train relative to the bridge is considerably larger than previous studies using road vehicles and this will excite the bridge to a higher degree. Numerical validation of the drive-by concept is achieved by using a 2-dimensional dynamic vehicle model with 10 degrees of freedom. The finite element interaction model is implemented in MATLAB. The track is modelled as a continuous beam, supported at 0.545m centres on three layers of springs and masses representing sleepers, and ballast lying on a simply supported bridge beam. This paper reports the results of the numerical simulations and the plans that are underway to test the concept in field trials.