Transformer State Assessment And Fault Diagnosis Based On Information Fusion

Transformer state assessment and fault diagnosis has always been a research area of great concern.Transformer state assessment and fault diagnosis both belong to state maintenance.Transformer state assessment is the important basis of transformer state maintenance.Transformer fault diagnosis is the key step of transformer state maintenance,and transformer state assessment is the premise of transformer fault prediction and diagnosis.This paper first introduces the transformer fault division method,in the same time it illustrates several common types of transformer fault.It briefly explains the method of transformer state assessment and fault diagnosis,mainly studies DGA(dissolved gas analysis in oil)diagnosis technology,and analyzes the shortcomings of this traditional diagnosis method.Then,in view of the limitations of traditional DGA analysis method,this paper established a prediction model of dissolved gas in transformer oil based on artificial neural network algorithm,conducted simulation and testing,and proved the validity and feasibility of the model through comparison and analysis with actual measurement data.This model can be used to predict the dissolved gas content in the oil of the transformer,so as to conduct a more accurate analysis of the future trend of the transformer state in combination with DGA,achieving the purpose of accurate assessment of the transformer state,which is conducive to the following fault diagnosis and the realization of transformer state maintenance.Finally,in order to analysis data more effectively using the DGA method,combining neural network algorithm and evidence reasoningtechnology,and using the DGA data as data sources,this paper proposes a model of transformer fault diagnosis based on information fusion,and proved the accuracy and practicality of the model combining with living example.