Enhancing Field Homogeneity of Dipolar Ultra-Low-Field MRI Magnets via Bézier Radial Parameterization and Dual-Material Optimization

Dipolar ultra-low-field (ULF) magnetic resonance imaging (MRI) systems, with compact size and open geometry, offer promise for portable MRI. However, the limited spatial distribution of the magnet constrains optimization toward high field homogeneity, thereby restricting imaging quality. This work presents a geometric optimization framework for dipolar magnets using a novel third-order Bézier radial parameterization (TBRP), which extends conventional single-radius designs to a multi-degree-of-freedom configuration, enhancing geometric flexibility. Optimizations of samarium–cobalt (SmCo) and neodymium–iron–boron (NdFeB) magnets are performed via a coupled Genetic Algorithm–Finite Element approach. A dual-material (DM) composite magnet is proposed to combine the advantages of both materials for improved field uniformity. Simulations show that, with the average main magnetic field maintained within the target range, the optimized SmCo- and NdFeB-based magnets achieve inhomogeneities of 1292 ppm and 3164 ppm, corresponding to improvements of approximately 20–30% over conventional disk designs, while the proposed DM magnet further reduces the inhomogeneity to 556 ppm, representing an improvement of nearly 60%. The analyses of magnetic thermal stability and engineering tolerances indicate that, although all magnet systems exhibit certain sensitivities to temperature variations and assembly deviations, the DM configuration demonstrates relatively better stability in maintaining field homogeneity. This study provides a scalable and cost-effective approach for the design of ULF MRI magnet systems.