power transformers in the power transmission system are one of a very important plantequipment. In the process of the operation of the transformer, achieve an accurate evaluationof the status, the correct diagnosis to fault happened, relates to the safety operation of the gridsystem. Condition assessment of transformer is based on the working state of the transformerfor recording, analysis and evaluation. Then the possibility of fault prediction and analysis oftransformer provided a basis for its processing and the suggestion in all for transformercondition assessment work. Fault diagnosis is an important part of the one, and on this basisto oil chromatographic analysis is put forward as the main method of transformer faultdiagnose is method.A combination diagnostic model is proposed in this paper, which makesthe original data secondary diagnosis,applys bayesian network classifier transformer in the fault diagnosis field,proposed bayesian classifier model for transformer fault diagnosis, established the bayesian rough model of the classifier with parameter self-learning functionthe resultas the secondary sample support vector machine (SVM) of secondary diagnosis improves thediagnostic accuracy greatly. |