Analysis of the BN-600 fast-spectrum core mock-up at BFS-2 zero-power facility using MCNPX

A 3D full-core heterogeneous model of the BFS-62-3A critical benchmark experiment was developed and validated using the Monte Carlo MCNPX-2.4.0 code.The BFS-2 critical facility at the Institute of Physics and Power Engineering (IPPE) was designed for simulation of fast reactor core neutronics, and for the validation of codes and nuclear data. The BFS-62-3A critical benchmark experiment was set-up as a mock-up of the BN-600 reactor core with (U, Pu) O2 fuel of 17% Pu content and stainless-steel reflectors. It was operated to measure the effective multiplication factor, spectral indices, radial fission rate distributions, control rod worths and sodium void effects.In the present study, special care was taken to run the MCNPX model to make Monte-Carlo confidence intervals comparable with uncertainties reported in the experiments; such as in material dimensions, number densities and isotopic compositions. In addition to the effective multiplication factor, sodium void effect, fission rate distributions and control rod worth were calculated. Simulations were carried out with four different modern nuclear data libraries; the primary aim being to estimate sensitivity of the results to the nuclear data. This task, besides being a library comparison, is also meant as a first step towards a code-to-code verification with deterministic methods.Results agree well with experimental values on most of the nuclear characteristics, even though a discrepancy up to more than 20% was found on the flux distribution in the stainless-steel reflector region, evaluated by measurements of the U235 and Pu239 fission rates; the sensitivity of this discrepancy to the stainless steel reflector density was estimated.

► We model the BFS-62-3A experiment with the MCNPX code and four nuclear libraries.
► We show the impact on reactivity of heterogeneous structures in the reactor.
► We model experimental uncertainties, e.g. in materials dimension and density.
► The model agrees with experiments on k-eff, CR worth, Na voids and fission rates.
► The analysis questions experimental data measured in the reflector region.