Speaker
Description
Non-light water reactor (non-LWR) concepts under development throughout the industry are significantly different from traditional LWR designs with respect to geometry, materials, and operating conditions. These differences result in significant differences with respect to the reactor physics behavior of non-LWRs as compared to LWRs. Given the limited operating experience with non-LWRs, the accurate simulation of reactor physics and the quantification of associated uncertainties are important for ensuring that the nuclear design for advanced reactor concepts include appropriate margins. Nuclear data are a major source of input uncertainties in reactor physics analysis.
As part of a project sponsored by the US Nuclear Regulatory Commission, the impact of nuclear data uncertainties for reactor physics analysis was quantified based on sensitivity and uncertainty analyses performed for six selected benchmarks: a graphite moderated high temperature gas-cooled reactor (HTGR), a fluoride salt-cooled high temperature reactor (FHR), a graphite-moderated molten salt reactor (MSR), a fast-spectrum heat pipe reactor (HPR), and a sodium-cooled fast reactor (SFR). For these benchmarks, comparisons of uncertainties were performed for several safety-relevant reactivity effects using the latest three Evaluated Nuclear Data File (ENDF)/B library releases: ENDF/B-VII.0, ENDF/B-VII.1, and ENDF/B-VIII.0. The analysis of the top contributing nuclide reactions to the observed uncertainties was used to identify the causes of the different magnitudes of observed output uncertainties.
In addition to nuclear data uncertainties that are often found relevant in LWR uncertainty analysis (e.g., $^{235}$U neutron multiplicity, $^{235}$U and $^{238}$U neutron capture), several large cross section uncertainties that are not commonly found to be relevant in LWR studies were found to be significant for several non-LWR systems. In particular, three nuclide reactions stood out: (1) The uncertainty of $^{238}$U inelastic scattering in the fast energy range contributes significantly to large reactivity uncertainties of the control rod worth in reactor designs with fast neutron spectrum. For the studied HPR and the SFR, large uncertainties of the multiplication factor, control rod worth, reactivity changes from grid expansion and fuel temperature increase are caused by this reaction. (2) The uncertainty of $^{235}$U (n,$\gamma$) in the fast energy range causes significant reactivity uncertainties in fast neutron spectrum systems that use $^{235}$U-enriched fuel. In the studied $^{235}$U-enriched HPR and SFR, this reaction causes large uncertainties of the multiplication factor, the control rod worth, and the reactivity changes from fuel temperature and coolant density increase, grid radial expansion, and fuel axial expansion. (3) The uncertainty of $^7$Li (n,$\gamma$) causes a large fraction of uncertainty in the reactivities investigated for MSR systems in which lithium is part of the salt. In the studied MSR and the FHR, this reaction causes large uncertainties of the multiplication factor and reactivity changes from fuel temperature and salt density increase.
Special attention should be paid to differences in cross section and uncertainties of different evaluated nuclear data library releases. Significant differences were found in nuclear data that can lead to major differences in reactivity calculations, even for well-known nuclides. In particular, differences in $^{235}$U, $^{238}$U, and $^{239}$Pu nominal and uncertainty data between the ENDF/B-VII.1 and ENDF/B-VIII.0 nuclear data releases are the major causes of differences in calculations when using these libraries.
A significant gap for the performed uncertainty analyses of the selected moderated systems was identified in terms of the unavailability of uncertainty for graphite and FLiBe thermal scattering data. For the analysis of fast spectrum systems, the unavailability of angular scattering data uncertainties was noted.
The presentation will cover an overview of the major outcome of the performed uncertainty analyses. Major differences between analyses with the different ENDF/B nuclear data libraries will be highlighted, and the magnitude of reactivity uncertainties including their major contributors will be discussed. Uncertainty analyses as presented in this work, in particular the ranking of contributions to the output uncertainties, can guide future measurement and evaluation efforts to reduce the identified large uncertainties. Reduced nuclear data uncertainties will contribute to the improved prediction of key figures of merit for reactor safety.