Application Of Information Fusion Technology In The Fault Diagnosis Of Voltage Transformer

Power transmission is closely related with people’s daily life and industrial production. In order to keep producing, it’s necessary to ensure the electric transmission equipments work in the right condition. According to statistics, voltage transformer is the weakest one. It’s hard to repair it, and it can cause significant losses. So it’s meaningful to do some research on how to diagnosis the fault of voltage transformer. In this article, a new method has been put forward to deal with the problem.Firstly, introduce the Rough set and D-S evidence theory, then do some changes to let them work better. Attribute reduction is an important part of the Rough Set, but it usually need a lot of time and space to get the final result, this paper presents a new attribute reduction method by using the core attributes, and the new way costs less time and space. The traditional evidence theory is inconsistent when it’s used to deal with evidences of high conflict in multi-source fault diagnosis system. This paper proposes a new method based on modifying the evidence source. Get the trust degree of evidences, by calculating the reliable of evidences and the support among evidences.Secondly, use the improve rough set theory and evidence theory in transformer fault diagnosis, and combine the BP neural network with D-S evidence theory. Let the amount of specific gas and the specific gas ratio as the input of BP neural network. Use the improved Rough set theory to speed up the neural network convergence speed, and get the final result by using the improved D-S evidence theory, then compare with the traditional D-S evidence theory and Yager method to prove that the method has certain advantages.Finally, in the view of the current transformer fault diagnosis of reality, discusses possible future development direction.