Simultaneous Detection of Radial Deformation Location and Axial Displacement Extent in a Transformer Winding Using Cross-correlation Technique and ANN

Abstract — In this paper, a new method to detect and locate mechanical defects, as one of the major faults in power transformer windings, is proposed. Axial displacement and radial deformation are considered as two major types of mechanical defects in this study. In the first step, a real transformer winding is modelled based on detailed model considering geometrical specifications and the frequency response characteristics are obtained for intact and defected modes. This technique is called Frequency Response Analysis (FRA). Thereafter, some features based on cross-correlation and other mathematical patterns are selected from the obtained signals. These features are then used to train an ANN classifier. The proposed method is able to discriminate between radial deformation and axial displacement defects precisely and determine the axial displacement extent and radial deformation location with a good accuracy. This accuracy rate reaches 81.25% for axial displacement extent and 75% for radial deformation location.