Estimation of bridge noise during the passage of a high-speed train with local wears on its wheel treads based on rail vibration in low-speed running

Bridge noise may increase when a train with local wears on wheel treads runs at higher speed. In this study we developed the estimation method for bridge noise at a high-speed section based on permanent accelerometers at a low-speed section. We proposed combining outputs of three accelerometers to measure precisely the influence of the wear of each wheel on rail vibration at the low-speed section, since the influence of the wear is attenuated depending on distance between a position of an accelerometer and that of where the local wear contacts rail. Composite measured data at the low-speed section showed that rail vibration affected by the local wear can be measured independent of distance between them. We also proposed an estimation method that converts the rail vibration at the low-speed section into that at the high-speed section to predict the bridge noises during train running at the high-speed section. In the estimation, the vibration properties of a wheel, a track and a contact spring at each section are used, and it is assumed that a wheel roughness is greater than a rail roughness at the spatial frequencies where the effect due to the local wear appears. The impulse excitation tests were carried out to obtain these vibration properties and transfer functions in vibration from the rail to the bridge at the high-speed section. The transfer functions are used to estimate bridge vibration using the rail vibration induced by one wheel. The bridge noise generated by one wheel is also estimated from the bridge vibration by taking account of the radiation area of the viaduct. Summing the vibrations and noise generated by all wheels in one train results in the overall bridge vibration and noise. The estimated results at the high-speed section show a good agreement with the measured one at the frequencies where the local wear has a stronger effect.