Multiwinding Transformer Leakage Inductance Optimization for Power Flow Decoupling in Multiport DC-DC Converters

Isolated multiport DC-DC converters manifest some prominent advantages over usual multiple two-port DC-DC converters such as smaller holistic magnetics and higher power density. However, the operation of such a converter is tied to the power flow decoupling capability in the magnetic medium frequency transformer (MFT). This paper targets to optimize the decoupling between the multiwinding MFT ports by means of multi-objective optimization of the leakage inductance network. Three different multi-objective cost functions are proposed and solved by Genetic Algorithm (GA). The obtained results show that a winding topology where primary winding is sandwiched by the two secondaries and are tightly wound to the core, provides the minimum possible leakage inductance without interleaving the windings. The obtained topologies from solving different objectives can be used as a benchmark in design and manufacturing of mutliwinding transformers. Experimental results are provided to verify the obtained optimum design.