Although it has long been recognized that when a train passes over a bridge there is frequently an increase in noise levels, accurate prediction of this phenomenon is often neglected. Using a modified version of AEA Technology Rail's ‘re-radiated noise’ model and data on existing bridges, this paper shows how the noise from a bridge can be predicted. In addition, the paper considers how the noise from bridges can be predicted using statistical energy analysis, and shows that reducing the vibrational input or providing baffles around the structure is likely to be the most effective way of reducing the noise from a bridge.
