Detection of Arc Faults in Transformer Windings via Transient Signal Analysis

In power transformers, lightning strikes, switching operations, and short circuit faults can deform the winding insulation, resulting in an electric arc between the windings. If the arc is not detected in its initial phase, it may lead to a solid short circuit and damage the transformer, potentially causing an explosion due to overheating and high pressure. In this study, winding arcs in the transformer are identified from the terminal current and voltage signals. A 3D magnetic model of a 15 MVA power transformer is constructed in Ansys@Maxwell, and the nonlinear arc model is simulated in Matlab@Simulink. The transient voltage and current signals related to the arcing conditions at five different points in the high-voltage side winding are obtained by running Ansys and Matlab simultaneously using ANSYS@Simplorer (Twin Builder). These signals are transformed into the frequency domain using Fast Fourier Transform (FFT). The arcs are detected from the transient-generated frequency components of the transformer voltage and current signals.