Abstract: The traditional power transformer fault diagnosis method was bettered by an improved gray clustering algorithm based on the theory of gray system. A concept of influence gene was introduced in the optimization of the power transformer fault standard model. Two ends of the trapezium whiting function were selected to analyze and judge the fault pattern and location of the transformer by the trapezium gray clustering algorithm. The results of the proposed diagnosis method were in good agreement with those of actual hanging hood detections, illustrating the new method is effeitive and practical, and characterized by the quickness and the realtime possibility.

Key words: electronics technology, influence gene, trapezium gray clustering, fault diagnosis, power transformer