Surface Discharge Characteristics And Risk Assessment Method Of Epoxy Resin Insulation Parts Under Long-term Constant AC Voltage

Partial discharge is the main manifestation of insulation degradation of electrical equipment,so partial discharge detection is an effective means to evaluate the insulation status of equipment.The on-site detection of GIS partial discharge detection mainly found metal particles and suspension discharge defects,and there were few cases of solid insulation defects.However,the grid's operation failure records involve sudden breakdown or flashover failures of GIS insulation pots.Occurred,in many cases,the equipment had an online monitoring system installed,but the fault signal characteristics were not captured before the incident.Therefore,effective detection of GIS solid insulation defects and timely warning of dangerous defects that cause serious deterioration of insulation are urgent problems to be solved by the power transportation inspection department.To this end,this paper builds a long-term partial discharge test platform for GIS solid insulation defects along the surface discharge.It uses the constant voltage method closer to the actual operating conditions and the UHF method widely used in GIS partial discharge engineering detection,which is harmful to GIS equipment insulators.Conduct the long-term constant voltage test research on the most serious creeping discharge,and collect the discharge signal data of the whole process from discharge to breakdown.The data of UHF partial discharge signals during the whole process from creeping discharge to breakdown caused by the attachment of metal foreign body defects on the surface of epoxy resin are analyzed.The statistics,such as amplitude,number of times,and pulse interval,as well as the trends and laws of PRPD patterns.Through research,found the intermittent characteristics of this type of discharge.The idea of dividing the two stages of stable development stage and high breakdown risk stage is proposed,and the state characteristics of this type of defects in these two stages are analyzed and summarized.The characteristics and development of the discharge stage at different voltage levels are studied separately The similarities and differences of the trend have obtained the influence of the voltage level on the development process of discharge,and summarized the change law of statistics during the whole process of discharge development under different voltages.Through the analysis of intermittent discharge,combined with the on-site detection strategy,pointed out the loopholes in the existing detection mode,and analyzed the reasons for the intermittent signal in the long-term discharge process.Based on the development trend graph of statistical parameters such as the number of discharges,amplitude parameters and discharge pulse interval parameters,three machine identification methods of SVM,BP neural network and DBN are used to perform cluster analysis on the discharge state.Compare the effectiveness of different methods for dividing the long-term solid insulation defects along the surface discharge stage.Considering the situation of on-site partial discharge detection,we explored the characteristic quantities that can effectively distinguish the two stages of stable development and high breakdown risk.A method that can effectively discriminate the development status of long-term partial discharge along the surface and provide early warning is proposed.The research results of this paper provide a test basis for the status monitoring and fault severity assessment of partial discharge of surface defects under long-term operating conditions of GIS.It has practical guiding significance for early warning of dangerous failures in engineering applications.