Characterization of transformer FRA signature under various winding faults

Abstract

Frequency response analysis (FRA) is gaining global popularity in detecting power transformer winding and core deformations due to the development of FRA test equipment. However, because FRA relies on graphical analysis, interpretation of its signatures is still a very specialized area that calls for skillful personnel to detect the sort and likely place of the fault as so far, there is no reliable standard code for FRA signature classification and quantification. This paper aims to initiate the establishment of standard codes for FRA signature interpretation through comprehensive simulation analysis on a detailed transformer distributed parameters-based model. Various mechanical faults such as axial displacement, buckling stress, disk space variation and bushing fault are simulated on the model to study its impact on the FRA signature. The main contribution of this paper is the comprehensive table it presents for FRA signature sensitivity to winding and core deformations that can be used for classification and quantification of the transformer FRA signature.