Maximum likelihood estimation of transformer high frequency parameters from test data

A methodology for the development of transformer high-frequency models is presented. Time constants of the transformer transfer function are estimated from frequency response measurements by using the nonlinear least-squares method in the frequency domain. Then, the time response of transfer functions is used in conjunction with the maximum-likelihood method to estimate the transformer model parameters. As a case study, a high-frequency model of a 15 kVA, 7620/240 V single-phase distribution methodology.>