FEA-Free Multiparameter Optimization of Low-EMI Toroidal Air-Core Reactors via NSGA-II for Compact Power Conversion Systems

Compared with column-type air-core reactors, toroidal air-core reactors confine most magnetic fields within the coil structure, significantly reducing magnetic flux leakage and minimizing the footprint. Via employing toroidal air-core reactors, this work aims at minimizing electromagnetic interference and reducing footprint of power conversion systems. First, this article proposes a framework driven by the nondominated sorting genetic algorithm to optimize structural parameters for minimizing cost and temperature rise while satisfying inductance constraints. Second, an inductance calculation and a temperature rise empirical equation are proposed, enabling rapid parameter design without finite element analysis. Finally, a laboratory-fabricated prototype toroidal reactor achieves 30 A, 8.7 mH inductance, and a footprint of 0.65 m². Using the distance where the external field attenuates to 0.5 mT as the effective radius, the toroidal reactor reduces the actual footprint by 70% compared to its cylindrical counterpart.