IEEE Princeton/Philadelphia PELS Chapter Seminar

This talk is focused on composite switched-mode power converter architectures where device stresses and sizes of passive components can be substantially reduced compared to conventional converter configurations. The composite architectures are inspired by the concepts of direct and indirect power processing and the fundamental limits of dc-dc converter networks. In electric drivetrain architectures, including xEV and electrified aircraft applications, a boost dc-dc converter is often utilized to interface a battery system to variable-speed ac drives, thus enabling the system operation at increased dc bus voltages. Using standard boost converter realizations, it has been shown how wide bandgap SiC devices lead to improvements in efficiency and reductions in the size of magnetic components. We then show how further improvements can be achieved using the composite architectures where SiC device and passive component stresses and losses are substantially reduced. Optimization of composite converters involves complex design tradeoffs in terms of losses, size, and reliability. To address these challenges, a design methodology is presented, which involves selection of the converter architecture, magnetics design, control techniques, and electro-thermal co-design. The approach is illustrated by modeling, simulations, and experimental results on a 125 kW SiC composite converter prototype featuring 21.3 kW/L power density and 99% drive-cycle weighted efficiency.