Estimating plutonium (Pu) mass in spent nuclear fuel assemblies (SFAs) helps inspectors ensure that no Pu is diverted. Therefore, nondestructive assay (NDA) methods are being developed to assay Pu mass in SFAs. Uncertainty quantification is an important task in most assay methods, and particularly for SFA assay. A computer model (MCNPX) is being used to predict isotope masses and the spatial distribution of masses in virtual SFAs for 64 combinations of initial fuel enrichment (IE), fuel utilization [burnup (BU)], and cooling time (CT) values. Additional MCNPX modeling for the same 64 virtual SFAs provided the expected detector responses (DRs) for several NDA techniques such as the passive neutron albedo reactivity method and the 252Cf interrogation with prompt neutrons method.

A previous paper describes one uncertainty quantification approach involving Monte Carlo (MC) simulation using individually any of six new NDA options together with IE, BU, and CT. This paper provides an interpretation of the MC approach that is suited for a numerical Bayesian alternative, separately assesses the impact of MCNPX interpolation error, and compares several options to use subsets of IE, BU, CT, and one DR.