Prediction of Breakdown Voltage Value in SF6-N2 Gases Mixture Under DC Voltage with the Influence of Conducting Particles by Analyzing Experimental Data using ANN

Abstract

Sulphur hexafluoride (SF6) is the most commonly used dielectric gas in electrical systems, DC-GIL is considered as a new type of transmission lines has the potential to replace the overhead lines and cables in the future for use in DC distribution networks. It deserves great concern that abundant amount of SF6 use in the earth’s atmosphere because this gas is known to be a kind of “greenhouse gas”. Also, metal particles inevitably appearing in the pipeline of DC GIL which may do damage to GIL and cause a serious threat to its insulation safe and reliable operation. This paper aims to focus on experimental study on the difference in insulation strength of SF6 gas and the mixture of two gases (SF6 and N2) with and without using conducting particle under DC voltage in the ball-bowl electrode arrangement. Moreover, spherical conducting particle is selected to simulate metal particle defects in GIL and they are set to the floating position. Breakdown characteristics of SF6 gas and (SF6-N2) gas mixtures have been measured in non-uniform electric field when the SF6 gas mixture ratio (70%, 50%, 30% and 20%) and the pressure is between 0.1MPa and 0.5MPa under the positive and negative polarity of DC voltage. The experiment results show that DC negative polarity is more severe as compared with positive polarity. In addition, it is found that there is no significant increase in breakdown voltage when the pressure is more than 0.3MPa in the presence of particle but the growth trend is clear without particle. The equivalencies between the breakdown voltage of SF6 and SF6-N2 gas mixtures were analyzed with and without the presence of conducting spherical particle and tried to predict the improved breakdown voltage SF6-N2 gases mixture using Artificial Neural Network technique.