Integrated fuzzy approach for incipient fault detection in power transformers

Present work describes the design of an integrated tool for detection and recognition of faults in transformer by the analysis of dissolved gases in the insulating oil. This approach is based on the combined use of well established standardized methods like the Duval's Triangle method, Roger's Ratio method, IEC599-Gas Ratio Code method, Denkyoken's method (Modified IEC-599) and the Improved IEC-599 method. In this paper, both the Boolean logic algorithm and fuzzy logic techniques are implemented to these methods individually and results are compared for specifying the internal faults of transformer. The results so obtained have a higher extent of reliability with enhanced success level.