AC loss analysis of magnetic gear system with superconducting component

Magnetic gears, which utilize specially arranged permanent magnets in rotating mechanisms, offer significant advantages over conventional mechanical gears. Despite their capability for high torque transmission, their industrial adoption remains limited due to torque density and loss constraints. To address this issue, recent studies have focused on improving flux modulation between the rotors through innovative magnetic and material configurations. Superconductors, with their unique electromagnetic properties, introduce new possibilities for enhancing magnetic gear performance. In this study, a superconducting magnetic gear system was analyzed using finite element simulations in COMSOL Multiphysics. A cylindrical coaxial magnetic gear with a 20/6 pole configuration was evaluated under three stator (pole piece) material arrangements: Steel & Air, Steel & Superconductor (SC), and Superconductor & Air. Torque optimization was performed using the derivative-free BOBYQA algorithm, and AC (iron) losses were assessed based on the Bertotti loss model. The results demonstrate that optimization enhances torque transmission by factors of 3.5–5.1, while losses increase only 2.6–2.7 times. Across all configurations, the torque growth consistently outpaces the rise in losses, confirming an overall improvement in energy efficiency and torque density. Among the examined configurations, the Steel & SC combination yielded the highest absolute torque, whereas the SC & Air configuration exhibited the greatest relative improvement due to the absence of iron losses. These outcomes indicate that superconductors can substantially improve torque performance while maintaining manageable loss levels, effectively balancing the torque–loss trade-off. The study also reveals that optimization alters the effective gear ratio by modifying material volume distributions, underscoring a critical design consideration for superconducting magnetic gears. Overall, the findings provide valuable insights for multi-objective optimization strategies and offer a solid foundation for future experimental implementations.