Uncertainty analysis for VVER-1000 Core simulations with MCNP/ATHLET

Based on a coupling scheme between the Monte Carlo neutron transport code MCNP6 and the thermo-hydraulics code ATHLET, developed earlier, steady-state calculations for the full-power state of a VVER-1000 reactor core specified within an OECD/NEA benchmark were performed. This was done with the multi-group option of MCNP6, with self-shielded cross sections for all materials generated beforehand with the SCALE 6.1.2 code system, thus enabling uncertainty and sensitivity analyses with respect to the underlying nuclear data using the XSUSA package. Nuclear data and covariance data are taken from SCALE 6.1.2. Analyses were also performed with stand-alone MCNP6 calculations, yielding results for radial power distribution not much different from results obtained with continuous-energy data in the framework of the benchmark, confirming the high quality of the multi-group approach. Uncertainties in the radial power distribution are significant, with values of up to 9% for the relative standard deviation in the core center. The obtained uncertainties in the radial power distribution from the coupled MCNP6/ATHLET simulation are very similar to those from the stand-alone MCNP6 calculation, meaning that the thermo-hydraulic feedback does not significantly change the uncertainties. Sensitivity analyses show that the main contributions to the uncertainties in the radial power distribution are due to uncertainties in the average number on neutrons per fission of ²³⁹Pu. (authors)