Extenics and rough set theory are brought into transformer fault diagnosing procedure in this paper to get rid of abundant information data and to obtain more precise diagnosing result. Using the dissolved gas data as fault diagnosing attribution set, attributions which are needed for transformer fault diagnosis are predigested and preliminarily grouped by means of rough set method, and then matter element model for transformer's fault diagnosing is built. With the transformer's standard fault

more »

... odes as the transformer's fault diagnosing decision set, utilizing extenics association function to calculate each fault degree, acceptance and rejection rule is defined to diagnose transformer's fault. 76 dissolved gas information data have been collected to verify the method proposed in this paper, the diagnosing results show that the correctness of diagnosing results got by this method is better than frequently used IEC three ratio methods. 67 scopes proposed by published papers, classical domain of element model is determined. Using extennics corresponding function, the diagnosing results about the fault type of transformer and fault degree can be determined. 73 3) The needing of large number of transformer faults samples is avoided by applying extenics method. By analyzing the diagnosing results of 76 transformer fault samples, the correctness of this method is verified.