NEREIDA: A computational framework for the design and dosimetric characterization of fast-neutron facilities

Abstract

Reliable neutron flux characterization and dosimetry are critical to the design, licensing, and safe operation of nuclear facilities. However, obtaining these quantities from as-built geometries usually involves translating computer-aided design (CAD) models into radiation-transport inputs. This is a fragmented process that lacks an end-to-end, auditable pipeline. Widely used tools address transport, but not the seamless integration of CAD-to-dose, and often require restrictive licences. We present NEREIDA: a modular, scalable framework that automates the process of creating high-resolution, three-dimensional neutron dose maps from native CAD models. NEREIDA unifies aspects such as geometry voxelization and activation analysis in common structural materials, producing dose equivalents and across complex facilities. The NEREIDA framework integrates Geant4 with external components to support realistic sources, phantoms, and background fields. Validation against 252Cf and 241Am-Be reference spectra demonstrates spectrometric accuracy, and large-scale benchmarks on CAD-derived geometries show near-linear scaling up to voxels and 15 000 particles per run. This scalability enables precise, facility-specific radiation assessments within a single, reproducible toolchain for design iteration, licensing, and routine safety in fast-neutron environments.