The longitudinal periodic irregularity of rail top surface called “rail corrugation” poses a larger problem with noise and vibration. Some types of corrugation are formed on railway track in Japan. In particular, a type of corrugation in commonly caused in sharp curves by repeated train loads (Fig.1). Focusing on settling down such ecological problems, it is very important to check and improve the degree of corrugation. On the other hand, visual inspection is currently carried out at track patrol to identify rail corrugation, but practically visual inspection is not efficient to cover a whole track length. So that it is necessary to discover and monitor rail corrugation more efficiently than visual inspection.
In this study, we examined the detection technique of the rail corrugation using the on-train measured data. First of all, the data are analyzed in the frequency domain (Fig.2), and the signals that originate by the corrugation are extracted by the band-pass filter processing. Then in the next step, we have developed an evaluation method by which the processed wave forms are converted to the vibration level (Fig.3). In addition, a detailed field investigation was executed on the section where the corrugation was detected by this technique. In the field investigation, we measured the wave height and the wavelength of the corrugation (Fig.4). Finally, we have shown that the generation location and the degree of the corrugation can be estimated efficiently from the on-train measured data, such as axle-box acceleration, car body vibration or internal noise.