Data-driven Cable Terminal State Detection and Fault Diagnosis Methods

Cable terminals are critical for reliable operation of both traditional and traction power supply systems. During operation, cable terminals are subjected to a complex array of environmental influences including humidity, temperature fluctuations, electric fields, and mechanical stress. This complexity renders cable terminals prone to defects under extreme conditions, potentially leading to severe and catastrophic outcomes. Therefore, the condition monitoring and fault diagnosis of cable terminals are of paramount importance. In this study the most advanced diagnostic technologies for cable terminals are systematically reviewed, focusing on advances in signal processing, machine learning, and deep learning (DL). In addition, diagnostic techniques for cable terminals in scenarios with limited sample sizes are explored. This study identifies the challenges currently faced by fault diagnosis technologies for cable terminals and outlines future research directions. The findings suggest that emerging technologies such as DL, data augmentation, and semi-supervised learning have broad application prospects in the fault diagnosis of cable terminals.