The importance of well-defined rail and wheel roughness measurement methods has increased considerably since roughness limits have become part of type testing specifications (2001) and Technical Specifications for Interoperability (TSIs, 2002). Also, the European interim model (RMR) for railway noise predictions (2002) provides a method to correct measured noise emission by using the difference between the reference roughness and the measured roughness. Looking at the future, it can be expected that accurate assessment of roughness will become a standard requirement, as the Harmonoise method, which is intended to become the standard noise prediction method in Europe, uses the roughness spectrum as main input data for rolling noise. With these developments in mind, it is time to review what a decade of roughness measurements has achieved in terms of instrumentation, measurement procedures and analysis methods. Ten years of experience using various roughness measurement systems lay at the basis of this paper. Directly measured (using a sensor) and indirectly measured (using a train) roughness results are compared. Over the years, the need to specify and approve measurement equipment has grown. A preliminary suggestion for standardisation is proposed and discussed. This will include a standard graphical representation and units for roughness. Average and spread of rail roughness and typical wheel roughness patterns are discussed. A separate section deals with rail roughness growth and grinding results. With regard to grinding as a noise reduction measure (so-called acoustical grinding), it is made clear that roughness monitoring is required to maintain the acoustical qualifications. (c) 2006 Elsevier Ltd. All rights reserved.
