Review of Partial Discharge Monitoring and Pattern Recognition for Stator Insulation of Rotating Hydrogenerator

Partial discharge (PD) detection and diagnosis represent the key techniques for evaluating the insulation condition of hydrogenerators. This article reviews recent significant research findings on PD in hydrogenerator stator insulation, focusing on insulation deterioration, PD detection, and pattern recognition issues. First, the stator insulation structure of hydrogenerators, typical insulation defects, and failure mechanisms are summarized. Second, the PD detection methods for hydrogenerators are discussed, including those based on the high-voltage terminal, neutral point, and electromagnetic wave detection, with a focus on analyzing the advantages and disadvantages of each detection technology. Subsequently, interference suppression approaches for stator insulation PD detection are systematically analyzed, identifying various sources, characteristics, and processing methods while comparing the advantages and disadvantages of different approaches. Furthermore, the PD pattern recognition methods are discussed in detail, examining both traditional machine learning and deep learning-based intelligent algorithms and emphasizing their respective advantages and limitations. Finally, this article indicates the future challenges and prospects of PD detection in hydrogenerators, aiming to provide a foundation and basis for the construction of large-scale hydropower engineering.