A Deep Neural Network-based Estimation of EMI Reduction by an Intermediate Coil in Automotive Wireless Power Transfer System

Abstract

In this paper, we proposed a deep neural network (DNN)-based estimation of EMI reduction by an intermediate (Int) coil in automotive wireless power transfer (WPT) system. The Int coil can reduce electromagnetic interference (EMI) level in the WPT system with the proper resonant frequency of the Int coil. The previous study has explained the resonant frequency of the Int coil should be higher than the operating frequency. According to the resonant frequency of the Int coil, we can achieve the amount of EMI reduction. Therefore, the design of the Int coil is essential for optimization of the EMI reduction of the automotive WPT system. However, it is impossible to get the optimized results of EMI reduction by simulations. The proposed DNN-based estimation method can predict the amount of reduced EMI level in real cases consisted of ferrites and shielding structures.