Assessment of nuclear data libraries for fast reactor systems with MCNP and NDaST

Abstract

The paper presents results for multiplication factors of reactor physics experiments with fast spectrum systems, representative of liquid metal fast reactors, documented in the IRPhEP Handbook. The calculations were performed with the Monte Carlo code MCNP 6.1, using continuous-energy data based on ENDF/B-VII.1, ENDF/B-VIII.0, JEFF-3.1.2, JEFF-3.2, and JEFF-3.3. Considerable discrepancies were observed between the results obtained with the different data libraries. In comparison with the benchmark values of the experiments, the ENDF/B-VII.1 results show the best agreement, while for the JEFF-3.3 results, the deviations are largest. In addition, the sensitivity tool NDaST, using linear perturbation theory, was applied to identify the most important microscopic cross sections responsible for the differences of the latest library revisions ENDF/B-VIII.0 and JEFF-3.3 as compared to ENDF/B-VII.1. For JEFF-3.3 vs ENDF/B-VII.1, where the deviations are overall largest, the main contributors are U-238n, γ and, for systems containing plutonium, Pu-239 fission and total neutron multiplicity.