부산대학교 도서관

This article provides an in-depth review of recent research into high-voltage circuit breaker (HVCB) fault diagnosis practices that use digital vibration signal analysis technologies. HVCB vibration signals provide a wealth of mechanical condition data, allowing for real-time, noninvasive, and comprehensive mechanical fault (MF) diagnosis. However, because HVCB vibration signals are typically nonlinear, nonperiodic, and transient, precisely extracting fault features and identifying fault types is difficult. The rapid development of digital analysis techniques has opened up new avenues for solving this problem. First, three crucial stages in fault diagnosis—vibration data acquisition, feature extraction, and fault identification—are introduced and analyzed. The data acquisition platform construction, signal acquisition process, and corresponding parameters are briefly introduced, including multisensor information fusion fault identification methods. Furthermore, the advantages and disadvantages, similarities, and differences of mechanistic, machine learning (ML), and deep learning (DL) approaches are examined. Next, existing difficulties in the field are described, and the solutions presented by recent studies are discussed, including imbalanced data training, fault degree identification, and noise immunity. Finally, we summarize and provide a research outlook; this work should serve as a useful guide for researchers developing HVCB diagnosis and online monitoring procedures based on vibration signals.