An ensemble learning estimation of the effect of magnetic coupling on switching frequency value in wireless power transfer system for electric vehicles

Abstract

Wireless power transmission (WPT) systems of small power levels used in the medical and communications sectors have been developed in recent years for tens of kW power levels for charging stations of electric vehicles. In wireless charging systems, power transfer is provided by magnetic coupling using coreless transformers, and in these systems, power electronics circuit design is the crucial point. The inductor behavior as a series resonance circuit element required in the power electronics circuit of WPT systems varies according to the magnetic coupling positioning errors between the primary and secondary sides of the coreless transformer. Therefore, the considering that the resonant capacitor value is constant in the resonance tank circuit, the switching frequency value in the power electronics circuit must be adaptively controlled so that the transferred power value can be carried out efficiently. In this study, the parametric simulations have been performed using Ansys-Electronics software to adaptively control the switching frequency value in the inverter circuit depending on the magnetic coupling coefficients in the WPT circuit designed at a power value of 25 kW. Based on the data obtained from these simulation studies according to different scenarios, the switching frequency value can be changed adaptively and thus the WPT efficiency can be kept at a certain level by providing resonance in each condition. Also, 105 efficiency data were obtained by using Ansys-Electronics parametric solver for many variables such as coreless transformer, resonant circuit parameters of power electronic circuit, switching frequency and magnetic coupling. The WPT efficiency is predicted by ensemble decision trees algorithm. The results show that the estimation with ensemble decision trees is quite successful.