Novel Partial Discharge Pattern Recognition Method Based on Deep Learning and SVM

Partial discharge (PD) is a critical issues in the high-voltage power systems. In the study, we raise a novel PD pattern recognition method based on both deep learning and Support Vector Machine (SVM). This method include a deep learning model to PD signals, capturing complex patterns and variations. These features are then inputted into an SVM classifier for precise classification and recognition of different types of PD. Its potential applications include condition monitoring, fault diagnosis, and predictive maintenance of power equipment. Further research can explore enhancements and optimizations of this method to achieve better performance and broader application.