A feature pseudo-fusion method for intelligent fault diagnosis of electro-hydraulic switch machine inspired by contrastive learning

Abstract

The railway system is an important part of the transportation system. Its scheduling process is carried out by the switch machines. The accuracy of determining the health status of the switch machines is related to the operational efficiency and reliability of the whole system. However, manual fault diagnosis for these machines is always unstable and expensive. The intelligent fault diagnosis (IFD) method can perform accurate fault diagnosis at low cost and high efficiency, but requires a large amount of labeled data. In this case, this study realizes the Feature Pseudo-Fusion (FPF) of left and right oil pressure signals of the electro-hydraulic switch machine. It uses contrastive learning to regularize the feature representation of original signals. Based on FPF, a fault diagnosis method applicable to electro- hydraulic switch machines is constructed. This method reduces the need for labeled data laterally without introducing additional measurement content to the field signal acquisition system. The effectiveness of FPF and the superiority of the fault diagnosis method have been verified through experiments.