Study On Fault Diagnosis Approaches For Oil-immersed Transformer Based On Information Fusion Technology

Electric power transmission is closely related to people's production and life,and Oil-immersed transformer is the key hub in the electric power transmission and distribution system.In order to not affect production and life,ensure transformer's normal and stable operation,it is important to detect oil-immersed transformer's latent fault accurately as early as possible.So this paper proposed a method for how to improve the oil-immersed transformer fault diagnosis accuracy,by studying the existing methods.Firstly,in view of the extreme learning machine lack of output ability for probability,this paper introduces the classification probability limited machine learning methods,which decomposes classification problem into multiple binary classification problems then using the Sigmoid function to map binary classification extreme learning machine's output to probability output.Then by solving a quadratic programming problem,fuses multiple binary extreme learning machine's probability output to classification probability output.When the D-S evidence theory has conflicts,the fusion results will not in conformity with th e actual result,and need to be promoted.The paper introduced the concept of similarity,calculate the weighted average of the evidence collection,and after determine the primary element of the evidence set to correct the evidence set ensure its strong p rocessing ability in the case of conflict evidence fusion.Secondly,this article combined classification probability limited learning machine and improving D-S evidence theory,established a model of fault diagnosis based on information fusion technology,the whole diagnosis process can be divided into primary diagnosis layer and fusion diagnosis layer.In this paper,the original data of oil-immersed transformer DGA gas made of gas contents,gas ratio,and three ratio of input feature space as the input o f preliminary diagnosis layer,could get three probability out puts.Then use the improved D-S evidence theory for fusion diagnosis,and the final diagnosis results are obtained.Finally,on the basis of the above research contents,this paper designed and implemented an oil-immersed transformer fault diagnosis system based on information fusion technology.