A new optimization method for gapped core magnetics in LLC resonance converter

The LLC resonant converter design is a very challenging design problem for electric vehicle charging applications (Figure A), since Lm, Lr, Cr and the quality factor are complex functions of voltage gain and it is very hard to find a choice of switching frequency range that can provide Zero Voltage Switching serving the voltage gains depending on the application. Figure A. Primary series secondary parallel connected LLC multi-core structurePurpose:A systematic design method is proposed for inductor and transformers in LLC resonant converter, based on a specific magnetic core database, under efficiency, cost and volume constraints, as illustrated in Figure A. The proposed method can be used to determine the magnetic elements in a high-level LLC resonance converter optimization for all Lr and Lm combinations that meet the required voltage conversion gain in the entire load profile to which the converter is connected.Theory and Methods:The recommended method is mainly for applications with high output current and low output voltage. In this study, the transformer of an LLC resonant converter with operating values of 3700W, 400V/48V is analytically designed thanks to the proposed algorithm. Results are proposed for single and multi-core structures. The obtained results were simulated based on the finite element method, and their magnetic properties was evaluated. In addition, the effect of different winding types was also investigated. The particle swarm algorithm is used instead of sweeping all the points in the solution space.Results:The presented method was tested for a 3700W, 48V LLC resonant converter installed in a lightweight EV charger. In the algorithm, a design study was carried out considering the multi-core structures in which serial and parallel multi-couplings are applied in the primary and secondary windings, respectively. The designed magnetic components were validated in the co-simulation environment of Ansys Electronic Desktop and Simplorer, and the most convenient core structures were selected from the core database.Conclusion:This study presents a new nature-inspired mathematical method-based design methodology that can evaluate multiple design components and constraints together for air gap magnetic components. The proposed method selects the most convenient design among numerous cores in the created core database by using the Particle Swarm Algorithm. Unlike the traditional design algorithms, the proposed novel algorithm determines the optimal magnetic flux density by minimizing a penalty function that includes the loss, cost, and volume of magnetic components. The presented method was tested for a 3700W, 48V LLC resonant converter installed in a lightweight EV charger. It was shown that the proposed method can be used as part of a system-level optimization algorithm where multiple combinations can be evaluated together and rapidly to find the most suitable LLC resonant converter design.