Application Of Information Fusion Theory On Fault Diagnosis Of Power Transformer

As the power transformer is one of the important equipments in power system, the technology of fault diagnosis for transformer is always taken into account by savants all over the world. As the fault information of power transformers has characteristics such as, complementarities, redundancy and uncertainty, the fault diagnosis has much uncertainty. The basic ideas of information fusion are introduced in this paper which can improve reliability and lower uncertainty in fault diagnosis.Information fusion can define to three levels which contain data level, feature level and decision level. It can merge information and data from multi-sources synthetically, then educe conclusion which can be more exact and more reliable. We select fusion level and algorithm in respond to application backgrounds. The thesis research on information fusion technique based on Dissolved Gas Analysis (DGA), the DGA is combined tightly with other information such as the results of conventional electrical test of power transformer. The power transformer fault diagnosis model based on information fusion is built. The models can diagnose both fault property and fault spot.Present general diagnosis model based on D-S Evidence theory. It can reduce uncertainty in fault diagnosis. The effective information from electrical test and analysis result of the dissolved gas in the oil can be fully utilized. The method proves high precision, high reliability and low uncertainty.Present general fault diagnosis fusion model based on Rough Set (RS) theory. The effective information from electrical test and analysis result of the dissolved gas in the oil can be fully utilized. First the decision table can be constructed by applying the RS theory. Then, the reduction of the decision table will be obtained, and an improved new fault diagnosis decision table for transformer can be acquired. Finally, two decision tables can be acquired which can diagnose fault character and fault spot of transformer. The realization of the software has been realized by MATLAB. The diagnosis results indicate the decision table method inRS theory can deal with imperfect information through avoiding omitted and wrong information. It can be used conveniently in practical project.