3D approach for fault identification within power transformers using frequency response analysis

• Published in: IET science, measurement & technology measurement & technology, 2019-08, Vol.13 (6), p.903-911
• Main Authors: Abu-Siada, Ahmed, Radwan, Ibrahim, Abdou, Ahmed F
• Format: Article
• Language: English
• Subjects: 3D‐SFRA signature; aged power transformers; apparent power 40.0 MVA; apparent power 45.0 MVA; current interpretation technique; current SFRA technique; deformation; digital image processing‐based code; fault diagnosis; fault identification; feature extraction; finite element analysis; finite element simulation analysis; frequency response; frequency response analysis quantification; incipient faults; incipient mechanical deformation; interpretation process; mechanical integrity; power transformer testing; Research Article; sweep frequency response analysis technique; three‐dimensional SFRA signature; transformer winding deformations; transformer windings
• ISSN: 1751-8822; 1751-8830; 1751-8830
• DOI: 10.1049/iet-smt.2018.5573

With the growing pool of aged power transformers, application of the sweep frequency response analysis (SFRA) to assess power transformers’ mechanical integrity has been given much attention. One of the research gaps in this field is the lack of reliable and automated techniques to interpret SFRA signatures. Conventional interpretation technique relies on visual inspection and personnel level of expertise, which may lead to inconsistent interpretation for the same signature. Furthermore, current SFRA technique fails in detecting transformer incipient mechanical deformations of low levels. To overcome these limitations, this paper presents a new three-dimensional (3D)-SFRA signature that comprises frequency, magnitude and phase angle in one plot. In contrary to the current interpretation practice that relies only on the magnitude plot, the proposed 3D signature exhibits more features, which can improve the SFRA identification accuracy. To automate and standardise the fault identification process, a digital image processing code is developed to extract some unique features from the proposed signature. The proposed technique is validated through finite element simulation analysis to detect short-circuit turns, axial displacement and radial deformation of a three-phase 40 MVA transformer and practical feasibility is assessed through its application to detect short-circuit turns of a three-phase 45 MVA transformer.