Vibration Signal Analysis For Rail Flaw Detection

Abstract

Rails are the foundation of rail transport, and any defects of the rails may directly affect the running state of the train or even lead to major safety incidents. However, the existing algorithms for rail crack detection are too expensive and difficult to perform on-line real-time monitoring. In this paper, we propose a method based on vibration signal for rail crack detection, which fits the vibration signal in the healthy rail and the cracked rail by least squares method. The transmission mode of vibration signal in the healthy rail and the cracked rail can be constructed, and the transmission mode can be used to distinguish the difference between the two types of rail on the higher harmonics. On this basis, the crack type of the cracked rail can be further distinguished. Using this method, we have established a rail crack detection system, which achieves a good on-line detection of cracks, and we discuss the reliability and safety of its on-line use in the future.