Test on the Correlation between Partial Discharge and Gas Decomposed Products in the GIS

Abstract: Thanks to its anti-interference against electromagnetic noises and vibration as well as its applicability for the diagnosis of overheating fault, the SF6 decomposition method demonstrates a broader application prospect in on-line fault detection. The present study attempts to fill the research gap by exploring the correlation between SF6 decomposed products and PD in GIS. Based on the simulation of three major PD patterns including protrusion discharge, floating discharge and surface discharge in GIS, long-term tests are carried out to study the impact of partial discharge pattern, PD severity and SF6 gas pressure over the volume fraction of SF6 decomposition products as well as the variation in volume fraction over time. It is found that three kinds of discharge can generate SOF2 + SO2 and HF. Thus, the two decomposed gas can be used as feature gas in detecting PD in the GIS; as PD severity increases, the volume fraction of both SOF2 + SO2 and HF increase as well. The SF6 decomposition method is more suitable for the diagnosis of severe PD; a negative correlation is identified between SF6 gas pressure and the generation rates of SOF2 + SO2 and HF.

Key words: GIS, partial discharge, gas decomposed product, discharge type, severity, gas pressure