Search

The main challenge of medium-frequency transformers is the number of design parameters, constraints and objectives, and the difficulty of handling them on a particular design. This paper presents a novel computer-aided optimal design for MF transformers using a multiobjective genetic algorithm, in particular the nondominated sorting genetic algorithm II. The proposed methodology has the aim of reaching the best MF transformer for a given power converter topology, by optimizing transformer efficiency, weight, and also, transformer leakage and magnetizing inductances at the same time. The proposed methodology and the optimal solutions are validated with the design and the development of two 10-kVA/500-V transformers considering two different topologies. Finally, some experimental measurements are presented so as to demonstrate the proposed models and the performance of built transformers. [ABSTRACT FROM AUTHOR]

This paper presents new analytical equations, parametrized with respect to the number of converter phases, for multiphase isolated dc–dc topologies. This paper describes five multiphase isolated converter models, as well as their level of accuracy. Based on these new parametric models, a thorough comparative among various common high-power converter topologies is carried out, highlighting the best topology regarding efficiency, cost, and power density. The analytical results are firstly compared and verified using PLECS simulation software, and validated finally by experimental measurements with a 10-kW/20-kHz/500-V configurable laboratory prototype. Due to the low computation cost and accuracy, the developed analytical models provide important insights for optimization-based design methodologies. Furthermore, this paper highlights, depending on the targeted application and objectives, the benefits of each analyzed converter configuration. [ABSTRACT FROM AUTHOR]