The Research On Intelligent Fault Diagnosis Of Oil-immersed Power Transformer

Power transformer is one of the most significant equipment in power system, and its safe and reliable operation has great effect on ensuring the power energy economic transmission, flexible distribution and safety using. Once there is damage to transformers, the serious effect will range very broadly. Therefore, it is very important to research the fault diagnosis technology and discover the potential failure of transformer promptly and accurately. In this paper, in order to improve the methods of transformer fault diagnosis, we concluded the previous studies, selected the most widely used oil-immersed power transformers as the research object, and had a work on power transformer intelligent fault diagnosis. The main work of this paper is as follows:First, taking into account the capacity of neural networks with pattern recognition, combinating the topology and learning algorithm of neural networks, we selected some typical representative networks from feed-forward networks, competitive networks and feedback networks. Then, the BPNN, PNN, LVQNN, ElmanNN transformer fault diagnosis models were established. In the paper, these models were compared with each other from mathematical model, implementation principles and diagnostic performance for transformer fault diagnosis firstly and completely. By a large number of experimental simulations, we had good understanding to the theoretical knowledge of the neural networks and gave the advantages and disadvantages and applicability to transformer fault diagnosis using different neural networks.Second, for the development of the method for intelligent fault diagnosis of transformer, clear and widely used extension theory was introduced into the transformer fault diagnosis in this paper. According in-depth study of the mechanism for transformer oil dissolved gas analysis, this paper concluded the relationship between the type of transformer failures and the transformer oil dissolved gas analysis, and so gave an improved matter-element model for descripting the transformer fault diagnosis problem. Then by the correlation function in extension set, the quantitative diagnosis of transformer faults was achieved. The test result proved that this method can be applied to oil-immersed power transformer fault diagnosis.The above research results are summarized finally. The further investigative direction is put forward in the end.In this paper, we have made beneficial attempts for the transformer failure diagnosis research; this has certain practical significance to improve the safe and economic operation level of transformer and the reliability of power system operation.