Constraint-driven multi-objective design and experimental validation of a high-frequency transformer for dual active bridge converters in solid-state transformer applications

The rapid growth of renewable energy systems has increased the demand for efficient and compact power electronic converters. Solid-state transformers (SSTs) provide high-frequency isolation, bidirectional power flow, and enhanced controllability, with the dual active bridge (DAB) converter being widely considered because of its modular structure and soft-switching capability. However, conventional high-frequency transformer (HFT) design approaches often treat magnetic sizing, losses, thermal performance, and converter-required leakage inductance separately, which can limit the practical suitability of the resulting design. This paper proposes a constraint-driven multi-objective HFT design framework that simultaneously considers transformer geometry, copper and core losses, thermal limits, efficiency, conductor sizing, manufacturability, and the leakage-inductance requirement imposed by DAB operation. In particular, converter-required leakage inductance is incorporated directly into the HFT optimization rather than being treated only as a parasitic parameter. GA, PSO, NSGA-II, MOGWO, SA, and WOA are comparatively evaluated to investigate their design trade-offs and select a suitable solution for practical implementation. The optimized HFT is verified using finite-element electromagnetic and thermal analysis and experimentally evaluated using a 150 W, 120/12 V, 20 kHz prototype. The analytical primary-referred leakage inductance of approximately 5 6 2 μ ⁢ H agrees with the FEA value of approximately 5 7 7 μ ⁢ H and the measured value of approximately 5 8 5 μ ⁢ H . The analytical and FEA peak flux densities are approximately 0.280 T and 0.288 T, respectively. The selected GA-optimized design achieves a predicted analytical HFT efficiency of 99.63%, while the fabricated transformer exhibits an experimental efficiency of approximately 98.75% at the rated operating condition. The analytical value is obtained from the estimated transformer losses, whereas the experimental value represents the transformer-side efficiency evaluated using oscilloscope-based power analysis. Experimental switching waveforms at a representative operating point show behaviour consistent with ZVS operation facilitated by the designed leakage inductance. Infrared thermography further shows an HFT surface temperature of approximately 35 °C under steady-state operation, supporting the thermal feasibility of the optimized design. The combined optimization, FEA, simulation, and experimental results demonstrate the feasibility of integrating converter-level leakage-inductance requirements with practical HFT design constraints for low-power DAB-based SST applications.