Study on Identification Method of GIS Insulation Defects Based on ARMA

Abstract

GIS is an important part in power system. The operating status of GIS is closely related to the operation of stability and security for grid. For existing problems in GIS insulation defects of mainstream identification method, a new identification method of GIS insulation defects based on ARMA is put forward in this paper. The paper proposes a method of multi-cycle time-domain average data preprocessing, a construction method of identifying reference data, a diagnosis method of regarding the coefficient of a certain order ARMA model as comparative quantity. At the same time, the author of the paper conduct a test on new identification method using laboratory devices, and confirms data validation of new identification method. The result suggests that the new identification method is effective and reliable.