Indirect Detection of Short-Pitch Rail Corrugation Using Acoustic Signal

Rail corrugation, particularly short-pitch corrugation, is prevalent on nearly all metro lines and is a primary cause of various issues within metro systems at frequencies corresponding to several hundred Hz. To ensure passenger comfort and safety, timely and effective rail grinding in response to the condition of rail corrugation is crucial. Traditional methods for directly detecting rail corrugation require regular measurement of each line section during the limited timeframe when the metro is not operational each night, which entails significant manpower and time costs. Therefore, the development of a rapid and intelligent indirect detection method for rail corrugation is essential. This study introduces a support vector machine (SVM) approach to facilitate the indirect detection of short-pitch rail corrugation in metro systems through the analysis of acoustic signals in the bogie area. Extensive test data from a line in China are utilized for intelligent learning, and this indirect detection method is subsequently applied to other metro lines for generalization performance verification, yielding promising results in detecting short-pitch rail corrugation.