Optimal MOSFET Selection Method of Dual Active Bridge Converter for Electric Vehicle Fast Charger based on Deep Learning

This paper proposes a novel method based on deep learning to select the optimal MOSFET in dual active bridge converters used in electric vehicle fast chargers. Although various studies have applied artificial intelligence to dual active bridge converters, most of them focus on designing transformers or deriving optimal operating points. However, an optimal switch that minimizes power loss across the entire operating area must be selected to design a high power dual active bridge converter and achieve high efficiency. The proposed method uses the power loss results of simulation as training data for deep learning, allowing the model to generate a power loss profile data and select the optimal MOSFET. The validity of the proposed method is verified by comparing the power loss estimates from the deep learning model with the actual power loss from the simulations.