Classification of Partial Discharge in Vehicle-Mounted Cable Termination of High-Speed Electric Multiple Unit: A Machine Learning-Based Approach

• DOI: 10.3390/electronics13030495
• Publication Date: 2024-01-24T14:57:42Z
• Authors: Yanhua Yang; Jiali Li; Zhenbao Chen; Yong-Chao Liu; Kui Chen; Kai Liu; Dong-Li Xin; Guoqiang Gao; Guangning Wu
• Keywords: Feature (linguistics); Envelope (radar); SIGNAL (programming language)

This paper presents a machine learning-based approach to identify and separate partial discharge (PD) two typical pulse interference (PI) signals in the vehicle-mounted cable terminations of high-speed electric multiple units (EMUs). First, test platform was established capture PD PI these terminations. The acquired were then processed using square envelope method extract feature parameters, such as rise time proportion, left–right symmetry, upper–lower symmetry. signal classification carried out on signals, utilizing waveform parameters derived from hierarchical clustering algorithm. results validate that extracted components effectively classify EMUs.