Search space narrowing using the GWO algorithm to estimate the enhanced coupled model parameters via FRA measurements for transformer winding

Purpose This paper aims to present a novel analytical search space-narrowing optimization algorithm for accurately estimating the enhanced lumped mutually coupled (ELMC) equivalent model parameters of transformer windings. The ELMC model is an extension to the lumped mutually coupled (LMC) model, whose performance has been improved by adding stray mutual capacitances. Design/methodology/approach The ELMC model has been developed using a new proposed analytical formula of the total equivalent capacitance. Using this formula, a novel search space narrowing of the series and stray mutual capacitances for the grey wolf optimization (GWO) algorithm has been proposed. Findings The proposed approach proved its effectiveness and robustness in estimating the ELMC model parameters using frequency response analysis data measurements on two different cases of transformer windings. Moreover, the parameters obtained using the ELMC model via the GWO algorithm with a physics-informed constrained search space (GWO-PICSS) showed more accurate results than the LMC model despite the complexities introduced due to the increased number of parameters. Research limitations/implications The true internal behavior of the transformer winding is displayed as accurately as possible through the ELMC model and the proposed analytical relationship that illustrates the overall capacitive coupling to overcome the parameters estimation task. Originality/value A new methodology to construct the ELMC equivalent circuit of transformer winding using an improved search space narrowing GWO algorithm is presented to obtain a more accurate model, which may contribute to power transformer faults diagnosis.