ENDF/B-VIII.0 Covariance Data Development and Testing for Advanced Reactors

Abstract

This report documents the ENDF/B-VIII.0 covariance data development and testing performed at Oak Ridge National Laboratory performed under the auspices of the US Department of Energy Office of Nuclear Energy. This work supports modeling and simulation of advanced reactor designs in the US by analyzing, validating and improving the nuclear data that provide underlying physics for calculations. This report assesses nuclear data uncertainty information for neutron cross sections, provided in the form of covariance matrices, in the 2018 ENDF/B-VIII.0 library. A review and interpretation of some of the issues with the current covariance library is provided. A new approach to augment the ENDF/B-VIII.0 covariance matrix to accurately reflect uncertainties in the results of modeling and simulation reliant on the current nuclear data library is documented. Rigorous testing of the ENDF/B-VIII.0 covariance data is documented. Finally, the performance of the ENDF/B-VIII.0 covariance data for advanced reactor applications is assessed.

Key Nuclear Data Impacting Reactivity in Advanced Reactors

“…In fact, the current ENDF/B format cannot even store correlations between these reactions. In a recent ORNL study, the existing covariance library was augmented with such correlations (Sobes et al 2018). It was demonstrated that the consideration of these additional correlations is relevant and that they have a visible impact on uncertainty analysis.…”

Hypothetical Sodium-Cooled Fast Reactor Fuel Transport Using the Existing ES-3100

“…The plots in Figure 5 and Figure 6 illustrate the difference in main nuclides of interest, 1 H, 235 U, and 238 U between the two wet conditions, demonstrating the influence of the additional spacing and moderation. 16 O has a significant sensitivity coefficient in both wet configurations, with an integrated sensitivity value of 0.11, with little energy dependent differences between the two. With increased moderation, the sensitivity to the primary moderator 1 H increases as expected.…”

“…The number of experiments is likely sufficient given the subcriticality of the system; in a more reactive subcritical system nearing criticality, such a validation basis may be insufficient. Additionally, the variation in c k due to the choice of covariance library has the potential to significantly affect the number of applicable experiments, though such a study on this system, in particular, is considered out of scope for this simple framing analysis [16]. Conversely, with such a significant margin to criticality even in the package's heavily compromised state, extensive validation for determining the most accurate bias and bias uncertainty to apply may be superfluous.…”

Nuclear Data Assessment for Advanced Reactors

“…In fact, the current ENDF/B format cannot even store correlations between these reactions. In a recent ORNL study, the existing covariance library was augmented with such correlations (Sobes et al, 2019). It was demonstrated that the consideration of these additional correlations is relevant and that they have a visible impact on uncertainty analysis.…”