Study On The Hybrid Of Artificial Neural Network And Tri-ratio Method And Its Application On Transformers Fault Diagnosis

With the rapid development of power industry, the number and rated capacity of the transformer are increasing. The role and status of Power transformer in the power system is becoming more and more important. The large power transformers usually run in an outdoor and its working, environment is relatively poor. Therefore, strengthening the view of transformer fault analysis and diagnosis is a work which is meaningful and necessary.Daqing oil field network is the most heavy enterprise network; the power supply of a year is more than 10 billion on average. The proportion of energy on the Oil field development costs is becoming bigger and bigger, nearly forty percent. At present, the development of Oil field is in the mid- Period, High and Stable output needs power supply smoothly and continuous. If an 110KV transformer station's power failure, it will result the collieries to thousands of oil well. If the electricity fails a long time, crude oil will harden in the oil pipeline, pipeline to the discharge and oil production loss, the economic loss is inestimable. And the transformer is the core equipment of a transformer station, once accident happened, you must find out the causes as soon as possible and to solve it, try your best to minimize the directly and indirectly economic losses which might bring to. When a Transformer has fault, it will bring about two aspects of harm, the one is own losses; and the other is that it will Influence the safe and quality run of the entire network, which economic losses might be more, with the improve of design and making level to the transformer, the capacity of the transformers is larger, and the cost is also increasing, the lose of serious consequences is more.In recent years, the main transformer substation data in Daqing oil field show that there are many similarities for why the failure happens, generally there are six kinds of situations can lead to the failure, product design flaws and quality problems caused by manufacturing defects; improper packaging and transport; Installation or operation errors, many times by the impact of improper operation and maintenance of overload; Earthquake, lightning, and other subject to problems which couldn't be predicted; Insulation aging and moisture; Man-made intentional damage (such as theft of transformer oil). Because of the core of the transformer in the substation equipment, the higher economic value, to ensure their safety and smooth operation is important. Users make their plans to maintenance, repair and test according to the actual situation, and then transformer failure can avoid the direct and indirect economic losses. In general, we could strengthen both oil from the transformer inspection and maintenance work.The status of oil field network equipment is different, and the advanced equipment monitoring and fault diagnosis of the depth and breadth of technology is not enough, there are still possible substantial increases that on-line monitoring technology in the accuracy and reliability, moreover, the scope of cover is relatively small, for the development of the on-line monitoring technology is well. As the on-line monitoring technology of Power groups in Zhangtiejiang first change and Xinghuo first change has operate a Period of time, it will be widely used in Oil field power 110kv level substation. Improve the diagnose technology of large transformer; to integrate the status surveillance and the diagnose technology, constantly improving the level of device detection, or quick, accurate and timely find out and remove the defects of the equipment provides powerful scientific, for scientific management to the equipment of the whole oil field network, for sustainable development in a state of repair work to provide technical and detailed evidence.Transformer fault diagnosis technology is concerned widely all the time. If the transformer internal fault does not develop to a certain extent, its electrical characteristics will not change, therefore, the ordinary electrical test often can not determine the status of the transformer accurately. It has been proposed in this article that make the neural networks and the three-ratio method used in transformer fault diagnosis at the same time, The application of theory about the trained neural network model and the three ratio method of transformer oil theory, for accurately determine the transformer internal fault is very effective. Combine the main transformer overheating analysis in Qingxin first change and Toutai first change, Make the neural network and the three ratio method applied in transformer internal fault diagnosis at the same time, the effect is obvious, and the judge is accurate. It is proved through the practice that the method proposed in this article is extremely accurate to transformer fault diagnosis, and it is suitable to promotion and application.