MATLAB-Based optimization and modeling of power electronic component design for high-performance DC–DC converter applications

This study introduces a multi-objective electro-thermal optimization method for a 500 W synchronous buck converter with a 48 V input, 12 V output, and 200 kHz switching frequency. The converter is modeled in MATLAB/Simulink using a detailed representation that includes practical parasitic elements, temperature-dependent characteristics, and the coupling between electrical losses and heat generation.  The design task is formulated as a constrained multi-objective optimization problem targeting higher efficiency, lower total loss, reduced junction temperature, and improved power density while satisfying voltage regulation and ripple limits. The NSGA-II algorithm is applied to obtain Pareto-optimal solutions. To reduce computational cost, Gaussian Process Regression is incorporated within the optimization loop. Compared with a conventional sequential design approach, the optimized converter increases full-load efficiency from 94.3% to 97.6%, reduces total loss by 60.5%, lowers peak junction temperature by 25.1%, and improves power density by 43.9%. The results show that simultaneous electrical and thermal evaluation inside the optimization process produces converter designs that better reflect practical operating conditions and realistic performance trade-offs.