Genetic Algorithm Sculpts Compact High-Frequency Transformer for

A study from A. D. Patel Institute of Technology in Vallabh Vidyanagar, India, shows how a genetic algorithm can design a compact high-frequency transformer for solid-state transformers. Instead of conventional line-frequency magnetics, solid-state systems combine power converters with a high-frequency isolation stage. The work targets the dual active bridge (DAB) converter, a topology commonly used for isolated DC-DC stages. In a DAB, two full-bridge circuits face each other and a phase shift controls both the magnitude and direction of power flow. Operating at 20 kHz (versus 50/60 Hz) shrinks magnetic components, but it also increases core losses, skin-effect losses in copper, and thermal stress. The algorithm optimizes multiple interacting constraints by minimizing area product, copper loss, and core loss, while the team validates the design on a laboratory prototype.