Evaluation of Mechanical Fault’s Severity in Power Transformers by using Maximum Points of FRA Window Curves

Frequency Response Analysis has emerged as a crucial tool for diagnosing mechanical faults in power transformers, enabling early detection of defects and hence minimizing transformer downtime. This study presents an advanced approach combining Frequency response analysis and the window calculation technique to enhance the accuracy of fault’s severity estimation, a crucial aspect in maintaining the reliability of power transformers. In this method, the interpretation algorithm divides the transfer function into small frequency windows. Then, each window is interpreted by a numerical index and the interpretation results form window curves. Finally, the approach scrutinizes the monotonic behavior of maximum point amplitudes in the window curves of frequency response analysis with selected indices. The investigation aims to identify the numerical index that exhibits the best monotonicity correlated with fault severity across different mechanical fault types including axial displacement, radial deformation, and disk space variation. Experimental validation across various test transformer mechanical faults in inductive inter-winding measurement test mode underscores the effectiveness of this novel technique.