Fault Diagnosis Of Oil Immersed Transformer Based On Multi Source Information Fusion

The oil immersed transformer as the research object,through reading the relevant references,to understand the structure of oil immersed transformer,fault classification and fault causes,and to study and analyze the gas that may be generated when the internal fault occurs.Specifically understand the traditional detection methods and the latest intelligent fault detection methods.Combined with my project,a transformer fault diagnosis model based on multi-source information fusion is proposed.(1)The reason why the diagnosis model is difficult to obtain is that there is a complex nonlinear mapping relationship between fault types and fault characteristics.Artificial neural network uses its own advantages,such as self-adaptive and self-learning ability,nonlinear mapping and associative memory,and uses distributed processing to open up a new way to solve this problem.Therefore,this paper uses BP neural network for research.However,BP network has some defects,such as easily falling into local minimum,slow convergence speed and so on.(2)In view of the shortcomings of BP neural network,this paper uses genetic algorithm and gray wolf algorithm to improve the weight and threshold of BP network.The results show that the optimized BP network effectively overcomes the shortcomings of traditional methods and improves the diagnosis accuracy of traditional methods.(3)In the field of transformer fault diagnosis,the idea of information fusion is introduced,and a diagnosis model based on multi-source information fusion is constructed by combining neural network and evidence theory.The model uses dissolved gas data in oil as the main data,supplemented by other corresponding test data,to make comprehensive diagnosis,determine the cause and type of fault,and comprehensively improve the accuracy and reliability of the model.