Partial Discharge Identification and Area Location in Transformer Based on Multi-Terminal Polarity

Partial discharge identification and location are critical for ensuring the safe operation of transformers. High frequency current monitoring, known for its ease of installation and high sensitivity, is widely used for online partial discharge monitoring in transformers. This technique typically identifies the initial waveform polarity to locate the partial discharge source. However, the presence of electromagnetic noise in field environments poses a significant challenge to accurately identify the initial waveform polarity, resulting in difficulties in both the identification and location of partial discharges. To address these challenges, this paper investigates the signal propagation characteristics and polarity features of partial discharges in transformers using the full-winding circuit model. A waveform polarity-based method for partial discharge identification and location is proposed, consisting of two components: a waveform polarity identification approach utilizing principal component analysis (PCA), and a multi-terminal waveform polarity criterion for partial discharge location. Finally, the proposed method is validated through partial discharge detection experiments conducted on a 500 kV transformer. It was demonstrated that the polarity identification accuracy reached 97.78%, with the detected multi-terminal polarity patterns conforming to the established criterion for locating partial discharges. This paper provides an effective technical basis for accurate partial discharge location in transformers.