Research On GIS Partial Discharge Diagnosis Method Based On UHF Method And SF₆ Decomposition Product Concentration

Gas enclosed switchgear(GIS)has been widely used in current power systems due to its advantages such as small footprint and strong resistance to electromagnetic interference.During the long-term operation of GIS,its internal insulation structure will continue to age under the effect of various factors,which will lead to partial discharge,further causing surface flashover or air gap breakdown.Partial discharge is the main form of insulation deterioration of GIS.Accurate judgment of its discharge type and intensity is of great significance to the operation and maintenance of GIS.Therefore,based on the real and simulated GIS test platform,this paper uses ultra-high frequency signal PRPD pattern recognition and SF6 decomposition product component detection methods to realize the recognition of partial discharge type and size.(1)Based on the finite element method,the propagation characteristics of electromagnetic waves generated by partial discharge in GIS are simulated;The mathematical models of four typical defect discharge pulses are constructed,and the relationships between electromagnetic wave propagation and defect type,propagation distance,discharge direction,pulse amplitude and pulse gradient are studied.Research has shown that under four different discharge forms,electromagnetic waves will move along the metal surface towards each gas chamber,and the electromagnetic waves generated by different defects have different propagation processes,which are reflected in the propagation rate,time required to reach the peak,and reflection oscillation.The electromagnetic waves have a certain degree of attenuation and time delay with the increase of propagation distance,and the attenuation degree is greater in the T-shaped structure.During the discharge direction from the z-axis to the x-axis,the field strength inside the gas chamber gradually decreases.The electric field intensity detected at each detection point is positively proportional to the discharge pulse intensity.The steeper the pulse steepness,the more intense the oscillation of electromagnetic waves.(2)A GIS partial discharge test platform was designed and built.Based on this platform,ultra-high frequency signals under four typical partial discharge were obtained,and the PRPD spectrum was drawn based on Python.The improved WGAN model was used to expand the sample data set,and the pattern recognition of partial discharge was realized based on Mobilenet-V2 network.The results indicate that the method proposed in this article can effectively solve the problem of insufficient on-site experimental data,reduce the number of network parameters,and have high accuracy.(3)A test platform for SF6 decomposition products detection was designed and built.The concentration of SF6 decomposition products was detected based on electrochemical sensors,including SO2 and HF.Based on this platform,the relationship between partial discharge intensity and the concentration of decomposition products was studied.The results show that the growth rates of SO2 and HF are both related to the partial discharge intensity,and HF has a high generation concentration at the early stage of partial discharge.The detection of the concentration changes of SO2 and HF can be used as the basis for judging the partial discharge intensity.