Study On ±320KV Flexible DC Cable Insulation Defect Criteria Based On Partial Discharge

With the gradual increase of electricity demand as well as the construction and reformation of urban electric grid,high-voltage power cable has gradually become the main body of urban transmission project.XLPE cable was invented since the beginning of 1960 s,after more than 40 years' development,the oil-impregnated paper insulation cable has been totally replaced by XLPE cable in various voltage levels and applications.With the development of flexible DC transmission technology,high-voltage DC XLPE cable has been developing rapidly.As for the voltage level,Xiamen ± 320 kV flexible DC transmission project,is the highest DC transmission line in China,accordingly it has brought unprecedented challenges to the operation and maintenance personnel.XLPE cable may suffer a lot of damage during the production,transportation,laying process,in the long-term high-voltage DC operation environment may eventually cause cable insulation breakdown accident,endangering the safe operation of power grid.It is necessary to study the cable insulation degradation.In this paper,± 320 kV DC XLPE cable laid in ± 320 kV flexible HVDC project in Xiamen is used as the research object,and the fault criterion of main insulation fault of cable based on partial discharge signal is studied.Firstly,the partial discharge detection method of XLPE cable is reviewed.On the basis of studying PD mechanism,HFCT(high frequency current sensor)is selected.The hardware circuit and FIR digital filter of partial discharge signal are designed,which laid the foundation for the fault simulation experiment.Secondly,the main insulation defect simulation experiment of XLPE DC cable was carried out,in the high-voltage shielding laboratory.The partial discharge signal was detected during the step-up process by artificially fabricating excessive bending defects and air gap defects.The q-t graph and the n-q statistical curve under different voltage levels are analyzed.The n-q statistic curve can be used for the fault alarm criterion of the main insulation defect.Finally,a feature extraction method based on EMD decomposition is proposed,which is used to characterize the main insulation fault of the cable and to build the knowledge base of partial discharge signal feature.In addition,the use of ELM limit learning machine for fault diagnosis,verifies the validity of the criteria.