Study Of Intelligent Information Fusion In Transformer Fault Diagnosis

Power transformer is one of the most important electrical equipments in the electric system. It is a significant issue for electrical department to find potential faults of the transformer so as to keep it operating safely. Dissolved Gases Analysis (DGA) is an important method to diagnose the internal fault of transformer and it offers a main basis to find the general incipient fault of the transformer indirectly. Because fault symptom space and fault pace have complicated, it has shortcomings, for example, non-linear, uncertainty and complexity relations and so on, the mathematical model of diagnosis system is difficult to build up, in this paper, a new integrated diagnostic model based on the information fusion technique is brought forward. The study of this paper focuses on the following aspects.The dissertation systematically elaborates the basic features of the smart grid, the three stages development of transformer equipment maintenance and sort the transformer condition information. This paper has put forward the method of electric power equipment fault diagnostic applying information fusion.The paper introduces various failures of power transformer and the present study condition of electrical equipments’fault diagnosis systematically, it has been important subjected that how to apply transformer condition information efficiently and scientifically to judge the transformer’s state. This paper also introduces the development trend of on-line monitoring, analyzes (DGA) in different fault conditions, introduces the three-ratio method、the Rogers-ratio、the no-ratio method、the rate of growth in oil gas and so on. It has the advantages in the use of dissolved gas analysis in transformer fault diagnosis.On the basis of the analysis of the biological neural networks, the back-propagation networks are expounded. The standard BP algorithm is derived in detail and its shortcomings are analyzed qualitatively. Then in order to expand the application of BP network, a basic network design method is summarized. Focuses on the BP (back propagation) network combined with the term of adaptive learning rate and momentum term. The power transformer diagnostic model based on information fusion is built, within this diagnostic model, the dissolved gas-in-oil analysis (DGA) is combined tightly with the results of conventional electrical test of power transformers. In this paper,it has been used30groups as Sample data, through the BP neural network learning and training, come to the BP neural network weight and threshold, and the model fault diagnosis is formed, the diagnostic model is applied to another30sets of experimental data. The results show that he improved algorithms can overcome the neural network stacked into the minimal valve locally and achieve the goal of fast convergence to a certain extent.