A Transformer Intelligent Fault Diagnosis System Design

Transformer is a vital and costly equipment in electric power system. Once some faults happen on it, not only does it have deep influence on security and stability of the power system, but also can make some economical damage badly. So it is very necessary to enhance the research on the fault diagnosis of the transformer. The thesis mainly does deep research on two contents; the one is to estimate whether there exists some faults in the transformer, the other is to distinguish which kind of the faults happens.After analyzing the shortcoming of the current fault diagnosis methods, the thesis designs a transformer fault diagnosis system model, which includes a transformer condition evaluation module based on information fusion and a fuzzy cluster diagnosis module based on characteristic gas.Firstly, the condition evaluation model is designed based on information fusion. And the conception of relative health index is introduced to describe the transformer's health degree quantificationally. The approaches to calculate the weight of the different test data and to evaluate the transformer's condition are discussed' in detail. Secondly, aiming at the shortcoming of the current methods to distinguish which kind of the faults happens, a new fuzzy cluster diagnosis method based on characteristic gas is discussed in the thesis, and compared with the conventional method. Additionally, a fuzzy cluster diagnosis model is designed by 80 groups of transformer oil gas data. And the validity of the model is checked by examples. Finally, the condition evaluation module and the fuzzy cluster diagnosis module are made into COM components respectively. The transformer fault diagnosis system is accomplished primarily by C++ Builder and SQL Server.