Neutron wall loading in stellarator optimization

Abstract

Non-axisymmetric stellarator equilibria produce spatially varying neutron wall loading (NWL), with localized peaks that affect first-wall lifetime, shielding requirements, magnet protection, maintenance planning, and ultimately plant availability. Although high-fidelity Monte Carlo transport can resolve this inherently three-dimensional loading, its computational cost makes it impractical for direct use in equilibrium optimization. We therefore integrate a deterministic line-of-sight NWL model into the DESC stellarator optimization package, providing a fast, differentiable engineering metric for use in optimization objectives or constraints. The model represents the plasma as a volumetric neutron source, using the Bosch–Hale fusion reactivity and prescribed flux-surface profiles to determine the source distribution before mapping its contributions to first-wall points through a geometry-dependent kernel. The resulting NWL maps closely agree with ParaStell/OpenMC Monte Carlo calculations and capture the dominant spatial loading patterns. With this model incorporated into DESC, proof-of concept optimizations demonstrate that localized NWL peaks can be directly targeted and reduced, allowing neutron loading to inform stellarator design at the equilibrium optimization stage.