A TRAFFIC-DEPENDENT ACOUSTICAL GRINDING CRITERION

Abstract

On most lines of the Dutch railway network, where a substantial amount of block-braked trains have rough wheels, the average wheel roughness dominates over the rail roughness. Therefore, reducing wheel roughness is top priority in the Netherlands. However, for the situations where rail roughness exceeds wheel roughness, this roughness can be lowered at acceptable cost. The high rail roughness is often due to rail corrugation which can be removed by grinding. A method has been developed to assess periodically the rail roughness on each railway line of the network, to compare it with the average wheel roughness for that line and to determine whether a noise reduction can be achieved by grinding the rail. Roughness measurements can be carried out with an instrumented coach. The two axle-boxes of a measurement wheelset are equipped with accelerometers. Together with the train speed and the right frequency filter, the accelerometer signal is used to produce a wavelength spectrum of the rail roughness. To determine the average wheel roughness on a given line, the so-called Acoustical Timetable can be used. This database comprises train types, train intensities and train speeds for each track section in the Netherlands. An average wheel roughness spectrum is known for each type of braking system. The number of trains of each type passing by on a certain track section determine the average roughness. Analysis of the data shows on which track sections the rail roughness exceeds the wheel roughness by a specified level difference. If this track section lies in a residential area, the decision can be made to grind this piece of track to reduce the noise production locally. Using this methodology, the noise production can be kept to a minimum, determined by the local average wheel roughness.