Application of Digital Image Processing to Detect Short-Circuit Turns in Power Transformers Using Frequency Response Analysis

Although a frequency response analysis (FRA) technique has been extensively used to detect mechanical deformation within power transformers, interpretation of FRA signature still needs a high level of expertise to identify and quantify faults as there is no FRA interpretation code widely accepted yet. All commercial frequency response analyzers can measure the magnitude and phase angle of the impedance, admittance, or transfer function of each phase in a wide frequency range; however, only magnitude is currently used for fault identification and quantification. This paper presents a novel approach for FRA signature interpretation by incorporating the FRA magnitude and phase plots into one polar plot that captures more features of the measured signal than the magnitude plot. Digital image processing-based techniques are employed to automate the fault identification and quantification process. To investigate the impact of transformer rating and size on the proposed technique, two transformers of different ratings and physical geometrical dimensions are simulated using 3-D finite-element analysis to emulate transformer real operation. Short-circuit of various fault levels is simulated at various locations within the high-voltage and low-voltage windings of the two transformer models and the obtained FRA polar plot signature for each case study is analyzed and compared with the healthy signature. Also, practical measurement is conducted to validate the simulation results. Results show that fault level along with fault location can be easily identified using the proposed polar plot signature along with the developed digital image processing technique.