Double Reliability Evaluation of Intelligent Diagnostic Methods baesd on DGA

Abstract

the classical DGA fault diagnosis methods, their diagnostic characteristics information from different angles, such as gas concentration, gas ratio, gas ratio, and the actual transformer faults is relatively complex. Due to various methods to diagnosis characteristics and criterion, the difference of the result also often appear inconsistent when using these methods in the diagnosis of the fault conclusion, their reliability in diagnosis of different fault types are not the same. Therefore, accurately assess the reliability of the diagnostic method difference, for the selection of diagnosis methods in practical, reasonable use of the characteristics of diagnosis and new diagnostic criterion of the proposed has the vital role.