The Sounds of Partial Discharge
* Presenting author
Abstract:
Partial discharge (PD) is a critical phenomenon that often occurs in medium and high voltage electrical equipment due to insulation faults and generates various types of acoustic emissions. The characteristics of the PD signal depend on the type of equipment, physical condition, operating voltage, and insulating substances used (e.g., gases, oils, nonmetallic solids). Corona, arcing, and surface discharges are the most common PD faults. Usually, these faults are quantified by electrical and acoustic detection methods. Acoustic methods gained popularity over electrical methods due to their insensitivity to electrical noise and localization capabilities. PD signals have different levels of acoustic emission, represented by time-frequency characteristics. Currently, the analysis of PD data, detection and classification of faults is done manually or semi-automatically by experts. This paper provides the current state of the literature on acoustic condition monitoring of partial electrical discharges in various high voltage equipment (e.g., transformers, cables, switchgears). The typical approaches used in the acoustic measurement and feature representation of signals for PD defect recognition are presented in detail. Expanding the possibilities for automated detection of PDs using signal processing and neural networks is also discussed for future research.