An Alternative Method for GIS Insulation Performance Evaluation

Abstract

The continuous improvement of system capacity and voltage level is somewhat accompanied by insulation fault as one of the main faults of GIS device. In this paper, an alternative method for identifying the GIS insulation performance based on fuzzy evidence theory is proposed by way of judging the four main features of GIS insulation level. Firstly, the fuzzy statistic test was used to obtain the normalized values of the four feature quantities of GIS insulation level. Four methods were used to calculate their respective expectations and deviations. The membership degree value was determined by the membership function, and the difference between the membership degrees was also considered. The degree of fusion was used to determine the weight of each feature quantity, so that the membership degree of the same kind of feature information fusion was obtained. Finally, the merged membership degree was transformed into the basic trust distribution function. The evidence theory then took the job to realize the pairwise fusion of the trust distribution functions between different feature quantities. The experimental results showed that this alternative way was proven to make effective judgment on the GIS insulation performance.