In recent years, considerable effort has been spent at a European level to establish comprehensive methods for the experimental assessment of
rolling noise emission of rail-bound vehicles and tracks. This work was concentrated in the European METARAIL and STAIRRS projects. The objective of these was to improve the accuracy and the reproducibility of pass-by
noise measurements compared to the standards that were current at that time. A further aim was to develop experimental methods separately to identify the contributions to
rolling noise of the vehicles and the tracks.
In these projects, measurement methods were developed that could determine the combined wheel/rail roughness and the ‘transfer functions’ for the vehicle and the track, that is, the separate noise contributions per unit roughness. The roughness and transfer function spectra provide a powerful basis by which vehicles and tracks can be characterized by measurement, to a high extent, independent of the running speed and site conditions. Such a description of the track and rolling stock allows the prediction of rolling noise spectra for different combinations of vehicles and track from those at which the characteristics have been measured. The measurement effort is limited; only straightforward one-third octave band measurements of pass-by sound pressure and vertical railhead vibration are needed.
This paper describes the method, giving the derivation of the calculation by which the roughness levels, transfer function spectra levels and the vibration decay rates in the track are determined from the measured quantities. Typical results are shown.
Among other applications, the method allows fast assessment of wheel roughness for whole vehicles or trains, speed-independent assessment of the effectiveness of track and vehicle noise control measures and the separation of rolling noise from noise due to other sources.
