A Collaborative Diagnosis Method for Abnormal Discharge of SF6/N2 Mixed Gas Insulation Based on Multi-parameter Sensing

At present, SF6/N2 gas mixture with 30% volume fraction has been gradually applied to gas-insulated power equipment. However, most of the existing operation and maintenance management experience comes from traditional GIS equipment. Consequently, there is an urgent need to propose a discharge fault diagnosis method suitable for SF6/N2 mixed gas insulation equipment based on the development characteristics of SF6/N2 gas discharge and the applicability of detection technology. In this study, a gas insulation experiment platform was initially established to collect pulse current, optical, and UHF signals during SF6/N2 defect discharge through the simulation of three typical insulation defects: metal spikes, surface anomalies of insulation components, and suspension potential, as well as sensor settings. Then, the phased changes of various physical characteristics with the development of discharge were analyzed. Finally, a partial discharge collaborative diagnosis model based on multivariate information perception was proposed. By comparing the recognition accuracy of SF6 and SF6/N2 databases, the differences in feature selection and pattern recognition criteria between the two were found. It has high guiding value for the establishment of differentiated operation and maintenance criteria for SF6/N2 mixed gas GIS equipment.