A novel multi-objective design tool for multilevel converters by an electro-thermal evolutionary approach

Multilevel converters will be one of the most important conversion systems in future power networks. These technologies can facilitate the integration of flexible DER in the grid also getting a significant role in microgrids. Designing high performances converters is not a trivial task due to the numerous different variables involved in the design process. This paper proposes a novel design approach and an easy-to-use executable program, employing a graphical user interface, dedicated to the design of multilevel converters. This software tool aims to obtain multilevel trade-off solutions in terms of efficiency, Mean Time Between Failure and price. The core of this tool consists in a multi-criterion optimization technique, based on the use of an evolutionary genetic algorithm, matched to accurate circuital models (i.e. electro-thermal reliable routine - TER and multilevel converter price routine). The solutions, plotted in two-dimensional and three-dimensional graphics, allow to easily choose the optimal solution for the specific context and visualize the corresponding bill of materials (BOM). The software also permits to consider innovative converter systems employing novel switching devices as the silicon carbide and the gallium nitride types.