Discharge Fault Simulation System for High Voltage SF6 Gas Insulated Switch-gear and Its Intelligent Pattern Recognition

In this paper, the defect simulator for high voltage sulfur hexafluoride gas insulated composite electrical apparatus is developed. The device consists of four parts: sulfur hexafluoride gas chamber, solid insulator, defect simulator, observation and measurement device. The defect simulator can effectively simulate free metal particle discharge, tip discharge, suspension discharge and air gap discharge. A real-type integrated defect simulator based on GIS is developed, and the partial discharge signal is tested on the simulator, the change trend of decomposed gas with time is detected. Based on this, an artificial intelligence classification method combining fuzzy ISODATA algorithm and ant colony algorithm is proposed, and the structure parameters of the two algorithms are optimized by PSO algorithm. The field application results of HV combined electrical appliances show that the proposed method is effective. The fault type diagnosis method can effectively judge the fault mode intelligently according to time series of SF6 micro-decomposition gas and typical micro-decomposition gas. This paper not only collects the original classification data from the hardware platform of the defect simulator, but also develops an artificial intelligence classification algorithm software system which is easy to be programmed. It can be directly and effectively used to diagnose and evaluate the type of insulation defect in the field practical engineering of GIS. It has certain theoretical guidance value for GIS equipment fault diagnosis and pattern recognition.