A Stochastic Method for Estimating the Effect of Isotopic Uncertainties in Spent Nuclear Fuel

This report describes a novel approach developed at the Oak Ridge National Laboratory
(ORNL) for the estimation of the uncertainty in the prediction of the neutron multiplication factor
for spent nuclear fuel. This technique focuses on burnup credit, where credit is taken in criticality
safety analysis for the reduced reactivity of fuel irradiated in and discharged from a reactor.
Validation methods for burnup credit have attempted to separate the uncertainty associated with
isotopic prediction methods from that of criticality eigenvalue calculations. Biases and
uncertainties obtained in each step are combined additively. This approach, while conservative,
can be excessive because of aphysical assumptions employed. This report describes a statistical
approach based on Monte Carlo sampling to directly estimate the total uncertainty in eigenvalue
calculations resulting from uncertainties in isotopic predictions. The results can also be used to
demonstrate the relative conservatism and statistical confidence associated with the method of
additively combining uncertainties.
This report does not make definitive conclusions on the magnitude of biases and
uncertainties associated with isotopic predictions in a burnup credit analysis. These terms will
vary depending on system design and the set of isotopic measurements used as a basis for
estimating isotopic variances. Instead, the report describes a method that can be applied with a
given design and set of isotopic data for estimating design-specific biases and uncertainties.