Uncertainty Quantification In Neutron Multiplicity Measurements: The Impacts Of Nuclear Data

Confidence in nuclear material balance evaluations rests on the quality of measurements. The premise behind paired operator/declaration - inspector/measured difference analysis is that measurement data is reliable and accompanied by a well-quantified total measurement uncertainty. In the framework of neutron multiplicity counting, an established neutron-based technique for assay of Pu-item mass, neutron multiplicity equations (so called point-model equations), are solved to obtain Pu mass from three measurement rates (Singles, S, Double, D, and Triples, T). The general neutron multiplicity counting approach therefore involves two main steps: calibration using 252Cf followed by measurement of unknown Pu items by solution of the point model equations. The 252Cf data is used to estimate detection efficiency, the doubles gate factor fd, and the triples gate factor, ft, each with associated uncertainties. Recently we developed a Bayesian framework for uncertainty quantification of the full process from neutron detector calibration parameters to final estimation of Pu mass that provides needed uncertainty distributions and correlations in all of the estimated assay-item parameters. This paper reports on the impact of the nuclear data and their uncertainties on the full measurement process. We discuss our recently-developed method for high fidelity certification of the 252Cf source neutron yield and the related nuclear data topics, including the effect of the nuclear data in the final estimation of Pu mass and in the definition of the 240Pueffective mass. Uncertainty associated with nuclear data in all measurement process steps from Cf calibration to Pu mass determination is exposed and discussed. Driven by the results, conclusions are presented regarding the impact on nuclear material verification.Acknowledgments. This work was sponsored by the U.S. Department of Energy (DOE), National Nuclear Security Administration (NNSA).