Abstract

OpenMC is an open-source Monte Carlo code with increasing relevance in criticality safety and reactor physics applications. While its validation has covered a broad range of systems, its performance in spent nuclear fuel storage scenarios remains limited in the literature. This work benchmarks OpenMC against MCNP for eleven configurations based on the KBS-3 disposal concept, involving variations in geometry, fuel composition (fresh vs. spent), and environmental conditions (e.g., air, argon, flooding scenarios). Effective multiplication factors (keff) and leakage fractions were evaluated for both codes. Across all cases, OpenMC systematically yielded slightly lower keff values than MCNP, with absolute code-to-code differences bounded by 136(10) pcm. In the normal-operation configurations, differences were-32(3) pcm (spent) and-38(3) pcm (fresh), with negligible leakage. Under periodic boundary conditions, both codes showed the expected boundary-induced rise in keff; for both OpenMC and MCNP, the increase relative to vacuum was +237(3) pcm. By providing reproducible input decks and quantifying agreement under equivalent modeling assumptions, this study represents the first systematic OpenMC-MCNP benchmark for KBS-3-type spent fuel canisters. These findings support extending the validation domain of OpenMC toward spent fuel transport and geological disposal applications.