A Data-driven approach to predicting neutron penetration through media

Abstract

A novel methodology is developed to predict the neutron permeation through media, enabling easier selection of shields and attenuators. The methodology relies on generalized metrics parametrized from MCNP®6.2 simulations where the average neutron energy and relative intensity are tallied through various media. This work builds on previously developed neutron energy attenuation coefficients by discussing accompanying neutron permeation coefficients. Both metrics are used to accurately predict the intensity of neutrons penetrating any medium and the average penetrating neutron energy, allowing for a simplified and generalized approach to predicting neutron penetration. Benchmarks with increasing complexity are used to demonstrate the applicability of the metrics to any geometry and medium. The benchmarks revealed 1.29–3.25% deviation from results in higher fidelity simulations. The data-driven methodology enables streamlined approaches to analyze complex neutron-attenuating media and presents a computationally efficient alternative to iterative neutronics simulations for optimizing neutron shielding and moderation setups.