Accurate detection of micro-discharge information induced by defects is crucial for the early identification of faults in gas-insulated equipment. Thus

this work proposes a defect-induced partial discharge pattern recognition model based on phase-resolved photon counting. First

photon generated by metallic tip defects during the evolution of partial discharge was analyzed using a needle plate electrode. Subsequently

phase-resolved photon counting (PRPC) and statistical analysis methods were employed to extract characteristic parameters under different discharge patterns

enabling quantitative characterization of the photon response induced by defects. Finally

dimensionality reduction and improved clustering analysis were utilized to establish a partial discharge pattern recognition model

and the classification criteria for discharge states were clearly defined. The current study aims to provide new insights into discharge pattern recognition and operational risk assessment for gas-insulated equipment.