Multi-objective Optimization of Square Magnetic Coupler for Undersea Inductive Power Transfer

This paper proposes a multi-objective design and optimization methodology for square magnetic couplers concerning power density and efficiency for undersea inductive power transfer (IPT). To establish the objective functions, an analytical model of mutual inductance computation for square coils with ferrite core, and a loss prediction model via backpropagation (BP) neural network based on finite element models (FEM) are proposed. Objectives’ sensitivities analysis is conducted to select design and operating parameters. Based on analytical calculations and FEM-based data-driven models, valid design guidelines for a scaled 1 kW prototype magnetic coupler are presented. A magnetic coupler with an efficiency of 96.07% at a power density of 0.4045 kW/d$\text{m}^{2}$, of which the length is less than 170 mm with a 60-mm air gap, is designed to validate the methodology.