Diagnosis and prognosis of incipient faults and insulation status for asset management of power transformer using fuzzy logic controller & fuzzy clustering means

Power transformer is one of the most expensive and pivotal electrical equipment among all connected network apparatuses. Various electrical, mechanical, thermal, and chemical stresses can degrade the insulation of mineral oil transformers. This article uses novel combined approach of fuzzy logic controller and fuzzy clustering means. It uses expert system for diagnosis and prognosis of power transformer's incipient faults along with critical parameters like dissolved gas analysis, moisture content, furans, interfacial tension, degree of polymerization, etc., in decision making of asset management. Furthermore, surface viewer and contour plots are extensively used to examine various stresses and their impact on insulation structure. This proposed expert model uses 200 transformer test datasets for test and verification. For diagnostic accuracy analysis efficiency of dissolved gas analysis based techniques like Gouda's three ratio technique is 97.5%, Duval Triangle is 96%, Degree of poly-merization and furans based paper degradation is 93.5%, Moisture and IFT-based insulation degradation is 91.5%. Fuzzy clustering means is used for cluster analysis to investigate insulation status further. Four clusters are divided for insulation diagnosis, cluster 1 represents healthy insulation, cluster 2 depicts moderate defor-mation, cluster 3 shows heavy deformation, and cluster 4 shows nearer to the end of life.