Intelligent Detection System for Partial Discharge of Transmission Cable

With the improvement of power system's requirements for safety and stability, the research of partial discharge detection technology becomes more and more important. In this paper, an intelligent detection system of partial discharge based on adaptive multi-scale decomposition and Hilbert-Huang transform (HHT) is proposed. Combined with You Only Look Once-v5 (YOLOv5) model, it can realize efficient monitoring of transmission cables. In the paper, high-frequency digital oscilloscope and high-precision sensor are used to collect partial discharge signals of different types of cables, and their time-frequency characteristics are extracted. Compared with the traditional machine learning model and other deep learning models, the results show that the proposed system has significantly improved in accuracy, recall and F1 value. The detection accuracy, recall and F1 value of YOLOv5 model reach 95.8%, 93.8% and 94.8% respectively, which is higher than other models. In addition, the actual performance verification of the system shows that it exceeds the expected target in data acquisition accuracy, signal processing delay and classification accuracy. This paper not only provides a new solution for partial discharge monitoring, but also provides strong support for the safe operation of power equipment.