SYNTHETIC INSULATION FAULT DIAGNOSTIC MODEL OF OIL-IMMERSED POWER TRANSFORMERS UTILIZING INFORMATION FUSION
SHANG Yong1, YAN Chun-jiang1, YAN Zhang1, CAO Jun-ling2 (1.Xian Jiaotong University, Xian 710049, China; 2.Nanjing Automation Co. Ltd, Nanjing 210003, China)
As the fault information of large power transformers has characteristics such as, complementarity, redundancy and uncertainty , the basic ideas of information fusion are introduced in this paper. In accordance with the basic principles of information fusion, a new type of multi-level comprehensive fault decision model is proposed, using back-propagation artificial neural networks and the technique of evidence reasoning. Within the diagnostic model, the dissolved gas-in-oil analysis (DGA) and the results of conventional electrical tests of power transformers are combined tightly. Also, the on-site experiences in operation, diagnosis and maintenance are highly utilized in the model. It has shown that the model possesses satisfactory capacity of knowledge representation and strong solving ability to deal with uncertain facts.