Power Electronic Converters for Energy and e-Mobility: Topologies, Optimization, Modeling and Control

This comprehensive study dives deep into the realm of power electronics, addressing both the theoretical and practical aspects of advanced converter designs. A significant focus lies on the challenges posed by pulse-width modulation (PWM) inverters, particularly the distortions stemming from dead-times and switch voltage drops. Emphasizing the importance of a more nuanced approach, a novel compensation strategy grounded in a detailed physical model of power converters is introduced, followed by the assessment of innovative self-commissioning techniques employing Multiple Linear Regression. The discussion further expands into the field of electric mobility, presenting an innovative system architecture tailored for range extender systems. Harnessing the potential of an integrated multi three-phase PMSM and high-frequency converter modules employing silicon carbide (SiC) power devices, this work pushes the envelope in creating a modular and fault-tolerant solution for electric vehicles. In the second part of this thesis, DC/DC converters for On-Board charging application are studied, namely LLC and Dual Active Bridge (DAB). This work addresses the limitations of traditional methods (FHA and EDF) used to identify the small-signal model of the LLC converter. Instead of relying on these conventional techniques, which often necessitate a resistive load assumption, this study proposes approximating the converter's small-signal output current response using a second-order discrete-time transfer function. This transfer function's coefficients adapt based on the operating conditions, like output voltage and switching frequency. To optimize the coefficient fitting, a data-driven approach using simulations and the LASSO machine learning method is employed. The research seeks to provide an accurate, efficient approximation of a resonant converter's output current response using machine learning. Another commonly used converter in these kind of applications is the Dual Active Bridge. This converter has gained prominence for its features like bidirectional operation and galvanic isolation, making it ideal for interfacing with renewables, batteries, and smart grids. However, its control remains a challenge due to its intricate behavior. This work introduces a comprehensive model for the DAB converter, emphasizing its ability to adjust the average output current without external dynamics. This model offers insights into the design, operating point selection, and control of the DAB converter. A novel control loop and a Finite Control Set (FCS) that assures full ZVS method are proposed, both tested via simulations and experiments. Moreover, this research examines various DC-DC converter designs for solid oxide fuel cell (SOFC) systems, focusing on their efficiency and gravimetric power density in hydrogen storage and energy distribution systems. Key performance metrics, like rise/fall times and ripple current limits, are outlined. Multiple converter topologies, such as the three-level multi-channel buck and buck with active filter, are assessed. Detailed design factors, from inductor values to switching frequencies, are explored for each topology. The study culminates in a design optimization procedure comparing the efficiency and weight of the converters through a Pareto analysis. Lastly, the of hardware design of the previous converter is analyzed, focusing on the significance of PCB layout in the context of minimizing parasitic components. Silicon carbide's rising prominence is underscored, setting the stage for discussions on gate driver designs that can harness SiC's rapid switching capabilities. In totality, this work serves as a cornerstone in power electronics, bridging theoretical advancements with practical implementations, ensuring optimized performance across a spectrum of applications.