Effective interpretation of transformer SFRA signature using multiresolution decomposition technique

Power Transformers are one of the most important and critical devices of any power system network so the interpretation of the faults and failures within transformers are of utmost importance. In monitoring the condition of the transformers sweep frequency response analysis has gained huge interest in assessing the core or winding disintigrity and deformations, open or shorted turns, axial or radial shifts etc. Changes in core and winding can be detected from the knowledge of poles and zeroes of the transfer function by statistical and mathematical indicators. In spite of efforts of researchers on SFRA there is no universally accepted or systematic method to detect deformations of winding and core of transformers. This new method is based on multiresolution decomposition technique using discrete wavelet transform. This method is further supported by statistical and mathematical indicators using the frequency band where faults actually exist which adds robustness and accuracy to the diagnosis of the fault. This methodology is tested for several cases of transformers and validated using two of the cases which show the difference between normal and abnormal condition of the transformers.