I . INTRODUCTIONHigh-frequency planar transformers are widely adopted for power converters . They are also sources of unwanted parasitics such as winding resistances and leakage inductances . Thus, a good electrical model of high frequency planar transformers is essential for accurate circuit simulation of switchedmode power converters . In such cases, it is necessary to represent a transformer with a precise equivalent electrical circuit that can show the behavior of the transformer under varying operating frequency, voltage and current . In this paper, a method of estimating the equivalent circuit of a planar transformer using transformer terminal voltages and currents is presented . An algorithm, based on differential evolution, was developed to estimate the parameters of the equivalent circuit of a transformer from experimental data . The electrical model, shown in Fig .1, consists of an ideal transformer T x with turns ratio n, magnetizing inductance, L m , leakage inductance L lk , core loss shunt resistance, R c , leakage inductances L 1 and L 2 , winding resistances, R 1 and R 2 , and stray capacitances C 1 , C 2 and C 12 . The performance of the algorithm was validated by comparing the estimated parameters of the model with that of 3D FEA, analytical calculation and LCR measurements . II . DIFFERENTIAL EVOLUTION ALGORITHM A differential evolution algorithm (DEA) from [5] was adopted and implemented in MATLAB to estimate the parameters of the model in Fig .1 . The main goal of the differential evolution algorithm is to minimize the two objective functions for every measured primary voltage and current . The objective functions arewhere N is the number of test samples . V s,M , I s,M are the measured secondary voltage and current, respectively while V s,DE and I s,DE are, respectively, the computed secondary voltage and current using the circuit parameters provided by the current iteration of the algorithm . III . EXPERIMENTAL SETUP A high-frequency low profile transformer with Ferrite core and PCB windings with turns ratio of 1:6, was used to obtain experimental data . Sinusoidal input voltage is supplied by a signal generator driving a power amplifier . Signal amplitude was varied at 1V pk increments until the transformer exhibited saturation characteristics . This was conducted for frequencies from 20kHz to 100kHz at 20kHz interval . The load, placed at the secondary side, is a 50Ω resistor . The primary and secondary voltages and currents were obtained using an oscilloscope with voltage and current probes and data were then processed in MATLAB . The input primary winding voltage and measured primary winding current were used as input to the algorithm while the measured secondary winding voltage and current were used in the objective functions . IV .ALGORITHM VALIDATION The accuracy of the proposed algorithm was validated by comparing the estimated parameters with numerical calculation, 3D FEA simulation and LCR measurement . A simplified version of Fig .1, which lumps the winding resistances into a sin...