Online Monitoring Method for Transformer Winding Deformation Based on Leakage Magnetic Induction Coefficient Identification

In response to the shortcomings of existing transformer winding deformation detection methods, this paper proposes an online monitoring scheme for winding deformation based on changes in leakage magnetic induction coefficients. Using the Roth method, an analytical calculation model for leakage magnetic fields under normal and deformed conditions of the windings is established, which is then validated against finite element simulations, with an error margin of less than 2%. The Recursive Least Squares algorithm is employed to extract the leakage magnetic induction coefficients in real time, which are then compared with the coefficients obtained from the analytical model to derive characteristic parameters that represent the health status of the windings. Corresponding threshold calibration methods are established, resulting in a high-sensitivity and high-reliability online monitoring scheme capable of accurately identifying various winding deformations of 5% or more, as well as determining the location of the deformed windings. Finally, finite element simulation models for both normal and deformed transformer windings are constructed in ANSYS-Maxwell, confirming the effectiveness of the online monitoring scheme.