Railway rolling noise emerges from combination of wheel and rail surfaces roughness. Previous researches evaluated roughness of running surface as the main cause of railway vehicles noise from 60 km/h up to 250 km/h running speed. As an example, rail welds or local irregularities on rail running surface may cause an increase of noise up to 4 dB(A) for the same running speed. Railhead roughness therefore became important parameter in order to conduct typical vehicle passing-by noise measurements. Within European Union’s strategy for harmonization of internationally running train services where developed new standards, as EN ISO 3095 or TSI-Noise, to set limits on the noise emitted by individual rail vehicles. Their requirement is “roughness level test” performed on a 'reference track' to confirm its influence on external passing-by vehicle noise.
This paper describes standard rail roughness measurement procedure on 'test sections' selected in order to determine passing-by noise of a new low floor EMV train constructed by Končar – Electric Vehicle for Bosnia and Herzegovina Federal Railways. It is mainly concentrated on measured data analysis according to European standards. Rail roughness with a broad spectrum of wavelengths is always present on the running surfaces of the rails and it has been shown that there is a high dependence between the amplitude and the wavelength of the roughness. Therefore, according to the standards above, measurements have been preformed with device that measures wavelengths from 0.1 up to 0.63 m. The track quality had to satisfy a rail roughness limit set by EN ISO 3095 based on one-third octave bands. Collected data needed to be filtered and grouped for further evaluation. The result of that spectral analysis is presented as a spectrogram which determined tested section as valid for further typical pass-by noise measurements.

rail roughness, measurement, spectral analysis, rail vehicles