Validation of Different Power Loss Models for a Bidirectional Isolated Dual-Active-Bridge DC-DC Converter used in Ultra-fast, Modular Electric Vehicle Chargers

Abstract

Due to the rapid electrification of vehicles, EnergyVille is researching 350kW modular, Ultra-fast chargers consisting of multiple DAB converters. Thanks to these kinds of converters, the charger is able to output multiple voltage and current levels independently of each other. This allows the charger to adapt to the vehicles’ specifications to ensure optimal fast charging. To reduce the size of the converter, a high switching frequency is used, which causes large switching losses. These switching losses prove difficult to estimate in high power components due to their parasitic elements. Nowadays, these losses are always verified experimentally via a prototype. The goal of this masters’ thesis is to establish different models to estimate the switching losses based on datasheet parameters of the components. These models are set up in Python and are validated through an experimental setup. Specifically, different operating points of the converter are set and the switching losses are determined via the measured values. Since the models differ in complexity, it is interesting to see which model is the most accurate. Additionally, a model is created to estimate the other converter losses throughout this operating range. The experimental verification shows that the most complex model, model 3, is the most accurate compared to the experimental results. Therefore, it is used to determine the switching losses throughout the entire operating range.