A Wear Detection Parameter for the Wheel–Rail Contact Based on Emitted Noise

Abstract

This article describes how a wheel–rail wear detection parameter based on noise was developed by running a rapid transit train instrumented with microphones in a depot with a small curve radius. This full-scale test made use of previous knowledge gained from lab-scale tests. The lab- and full-scale tests showed that noise could be used as an indicator of wear transition, when normal wear turns into severe or catastrophic wear. This transition was accompanied by a significant increase in sound pressure and a broader sound pressure amplitude spectrum; that is, a narrow initial amplitude distribution in the normal wear regime and a broader distribution in the severe wear regime. The measured noise from the inner/low and outer/high rails was analyzed in conjunction with the wear from the outer wheel and comparisons were made between cleaned (exposed to severe wear) and lubricated (no wear) rail. Based on those analyses, a wear detection parameter criterion was developed for this specific train. This wear parameter has been implemented in a real-time condition monitoring system so that warnings of the risk for severe wear could be sent to the maintenance department. Validation of the wear parameter and challenges linked to open system conditions are discussed in this article.