Transformer Fault Recognition and Interpretation Using Kohonen Feature Mapping

코호넨 특징 대응을 이용한 변압기 고장 인식 및 해석

This paper presents fault recognition and interpretation in power transformers using dissolved gas analysis embedded Kohonen feature mapping. The imprecision of gas ratio analysis in dissolved gas analysis are managed by mapping in accordance with learning of Kohonen neural network. To verify the effectiveness of the proposed system, it has been tested by the historical gas records to power transformers of Korea Electric Power Corporation. More appropriate fault types can support the maintenance personnels to increase the disgnostic performance for fault of power transformers.