A fuzzy dissolved gas analysis method for the diagnosis of multipleincipient faults in a transformer

摘要

Dissolved gas analysis (DGA) of transformer oil has been one ofnthe most useful techniques to detect the incipient faults. Variousnmethods, such as the IEC codes, have been developed to interpret DGAnresults directly obtained from a chromatographer. Although these methodsnare widely used in the world, they sometimes fail to diagnose,nespecially when more than one fault exists in a transformer. This papernpresents a fuzzy logic technique which can diagnose multiple faults in antransformer and quantitatively indicates the likelihood/severity of eachnfault. Insulation deterioration at each fault location can then benmonitored closely according to its trend, which is important for antransformer in critical situation. Tests using this technique on annumber of transformers have given promising results