Influence of nuclear data on fast reactor calculations

The paper presents a study for evaluating the uncertainties with respect to nuclear data when performing neutron transport calculations for sodium-cooled fast reactor arrangements. The analyses have been performed on four representative fuel assemblies taken from the SFR-FT benchmark specification. The estimation of the uncertainties is done in two ways: firstly fuel assembly criticality and depletion calculations with the Monte Carlo code MCNP6 using different evaluated nuclear data libraries in continuous-energy representation, and secondly systematic sensitivity and uncertainty analyses based on evaluated covariance data files with the TSUNAMI and SAMPLER codes from the SCALE 6.2 beta 4 code package, with both cross sections and covariance data in multi-group representation. Uncertainties for multiplication factors and reactivity worths are evaluated. The application of different evaluated data libraries yields significant discrepancies; in particular, JEFF-3.1.2 multiplication factor results show large deviations from the results from ENDF/B-VII and JEFF-3.2. These differences are, however, covered by the uncertainties obtained from the S/U analyses with the SCALE covariance files, which lead to uncertainties in the multiplication factor of 1.5% or above. The main contributor, and likewise to the Doppler reactivity uncertainty, is the uncertainty in the inelastic scattering cross section of U-238; for the sodium void reactivity; additional important contributors are the scattering cross sections of Na-23. These findings are valid independently of the SCALE 6.2 covariance file used.