Using data mining to dissolved gas analysis for power transformer fault diagnosis

The objective of this paper is to develop the fault diagnosis of power transformer by making decision tree according to the percentage of H₂, CH₄, C₂H₂, C₂H₄, C₂H₆ in sampling oil analysis for Discharge, Partial Discharge, Thermal from power transformer faults. This paper utilizes data mining decision tree technology based on IEC TC 10 database and 115 fault records from TPC. Meanwhile, comparing the analysis results of decision tree with five traditional criteria of the dissolved gases analysis published in different standards, we verify the more reliable approach by TPC historical transformers gas records and show its effectiveness in transformers diagnosis.