Expert system for condition monitoring of power transformers using DGA

摘要

Dissolved Gas Analysis (DGA) of transformer Oil is one of the most effective tool for detecting incipient faults in an energized transformer. This paper presents the development of an Expert System (ES) for prediction of incipient faults in a large power transformer using DGA. In this developmental work integration of Total Combustible Gas, Roger's, and Key Gas methods have been utilized for interpretation of DGA results. The problem of uncertain inferences, that may arise in interpretation of DGA results, has been solved by fuzzy set theory. The ES has been implemented in Turbo C++. The developed software is versatile in nature, it judges the transformer from various angles giving the over-all assessment of its health, the incipient fault(s) within it and their severity level. The Expert System has been verified for various test data and satisfactory results has been obtained under all most all possible cases.