Decomposition Characteristics Of SF₆ And Fault Identification Under Negative DC Partial Discharge

At present,a lot of studies have proved that using the SF6 decomposition component analysis(Decomposition Component Analysis,DCA)to diagnose the internal insulation fault of GIS is scientific and effective.But the current research mainly focuses on the AC GIS equipment,and has not yet obtained the DC gas insulation equipment represented by the DC wall casing in the typical insulation failure state.Whether decomposition characteristics in AC condition can be directly applied to fault diagnosis of DC GIS,it is an urgent problem that the vast majority of electric power technicians need to solve at present.In this paper,the decomposition characteristics of SF6 under different PD strengths of three typical insulation defects,such as metal protrusion,free metal particles and insulator gap,are studied and compared by a series of experiments.On this basis,the correlation characteristic between CRMS,CRRMS,RRMS and the PD intensity has been studied.Taking the metal protrusion defect as an example,the decomposition characteristic of SF6 under different AC and DC is compared.At the same time,in order to identify the three defects,the relationship between the CTRMS?CRTRMS and the PD fault type under a certain PD time is established,and the decision tree is built to identify the three types of PD faults based on the the three group of CRTRMS.Through the above research work,the innovative conclusions obtained in this paper are as follows::1)The negative DC PD under three kinds of insulation defect can all make SF6 decomposition,and it will produce five kinds of stable components,SO2F2?SOF2?CF4?CO2?SO2,which the SO2F2 and SOF2 are the main decomposition products,and the generation of each component are different with the extension of PD time.2)There are obvious differences in the SF6 decomposition characteristics under different PD strength caused by three typical insulation defects.The decomposition components show a specific change rule with the increase of PD quantity.Further research has found that the CRMS?CRRMs have a very obvious mapping with the Qsec.The correlation coefficient calculation further proves tha CRMS?CRRMS have a linear positive correlation with Qsec.On the basis of component content,the correlation between the three groups of effective content ratio CRRMS and PD strength was studied.The results showed that CRRMS[CF4/CO2]?CRRMS[SO2F2/(SOF2+SO2)]can be used to evaluate the severity of PD faults.3)SF6 will decompose under the different PD strength caused by metal protrusion defect,SO2F2?SOF2?CF4?CO2 these four components are generated under both the AC-PD and DC-PD And each component increases with the increase of Qsec,the growth trend of these components are similar except SOF2.Comparing the three sets of CRRMS under the AC-PD and DC-PD,it is found that the variation of CRRMS varies with Qsec are the different,but they all have obvious correlation with the discharge capacity.4)According to the decomposition of SF6 under DC-PD caused by three different defects,the correlation characteristic between theCTRMS?CRTRMS and PD fault types is studied.The results show that the CTRMS(SOF2)?CTRMS(SO2F2)?CTRMS(CO2)?CTRMS(SO2)and CRTRMS[CF4/CO2]?CRTRMS[(CF4+CO2)/(SO2F2+SOF2+SO2)]? CRTRMS[SO2F2/(SOF2+SO2)]are suitable to be associated with the fault,which can be used for defect type identification.5)Building fault identification decision tree using for PD recognition based on CRTRMS[CF4/CO2]?CRTRMS[(CF4+CO2)/(SO2F2+SOF2+SO2)]? CRTRMS[SO2F2/(SOF2+SO2)],the result shows that the decision tree has a good identification effect on the type of defect fault.