Fault diagnosis method of point machine based on vibration signal

Railway point machine is the key equipment to control the train path and change the operation direction. Considering the anti-interference characteristics of vibration signals, this paper innovatively proposes a fault diagnosis method for point machine based on vibration signal, which provides a new fault diagnosis method for point machine. Firstly, the vibration data of the point machine are preprocessed by using variational mode decomposition (VMD). Secondly, a hybrid feature extraction method is developed by combining various methods such as wavelet transform and frequency-domain feature extraction method, which can characterize the fault features more comprehensively. Finally, support vector machine (SVM) algorithm is used to judge fault conditions. The diagnosis accuracies of localization and inverse switching processes reach 100% and 98.9%, respectively. Comparative experiments indicate the superiority of the proposed method.