Validating Geant4 with Monte Carlo simulations in the context of nuclear disarmament verification

Manuel Kreutle, Alessandro Borella, Simon Hebel, Gerald Kirchner

Research outputpeer-review


In this paper neutron transport simulation results of experimental configurations used during the 2019 IPNDV measurement campaign at the Belgian Nuclear Research Centre SCK CEN in Mol are presented. As these are a good basis for benchmarking simulations, results of the Monte-Carlo simulation codes Geant4, SCALE/KENO-VI, MCNP and openMC are compared with the aim of validating Geant4. With fissile material present as plutonium–uranium mixed oxide, (α,n) emissions are included in the simulations. A Geant4 extension to calculate neutron multiplication factors keff was developed and applied to the IPNDV configurations and a set of OECD/NEA benchmark experiments. Neutron fluxes of the IPNDV configurations through a reference volume are simulated. The Geant4 (α,n) simulation toolkit SaG4n is compared to the SCALE code ORIGEN and small deviations are observed. Geant4 calculations of keff for criticality benchmark experiments deviate in the mean by +0.4%, for MCNP and openMC deviations are ≤6%, for KENO-VI ≤8%. Geant4's total neutron fluxes through a reference volume next to the IPNDV configurations agree within a margin of ±5% with MCNP and openMC and −14 % with KENO-VI. Considerable differences between all codes were observed for thermal neutron scattering but Geant4 results are still within the other codes’ variability. Differences to Geant4 for epithermal energies can be addressed to a not yet included “Doppler Broadening Rejection Correction”. By comparing Geant4 simulations of a complex set-up with various reference codes and demonstrating its good performance as well as some discrepancies, this study contributes to the validation of Geant4 neutron physics in fissile material systems and for nuclear disarmament verification simulations.

ASJC Scopus subject areas

  • Nuclear and High Energy Physics
  • Instrumentation

Cite this