The influence of real-world rail head roughness on railway noise prediction

Harmonised methods for the prediction of noise from roads, railways, aviation and industry will eventually be available as a result of the HARMONOISE and IMAGINE EC projects, but not in the immediate future. As the first round of noise mapping under Directive 2002/49/EC is required to be completed in 2007, alternative models are required. The options for railways under the Directive are either to use the Dutch national procedure as the approved EC “interim method” or for Member States to use their own method if it can be adapted to produce results in terms of the indicators required by the Directive. In the UK it is therefore likely that the national method “Calculation of Railway Noise 1995” (CRN) will be applied. This paper discusses a shortcoming in CRN that results from its assumption that the rail head is comparatively smooth. In reality this is often not the case, leading to increased rolling noise levels. A statistically based study has been carried out for the UK Department for Environment, Food and Rural Affairs into the effects of rail head roughness on rolling noise prediction. Correction factors have been devised to account for the true roughness of rails in the UK, either at a local level or across the entire network. The acoustic effectiveness of rail grinding strategies has also been examined.