Emerging Applications of Bottom-Up Uncertainty Quantification in Nondestructive Assay

The Guide to the Expression of Uncertainty in Measurement (GUM) provides guidance on expressing measurement uncertainty for calibration, laboratory accreditation, and metrology services. Nondestructive assay (NDA) of items containing nuclear material uses calibration and modelling to infer item characteristics on the basis of detected radiation such as neutron and gamma emissions. Uncertainty quantification (UQ) can be approached from a bottom-up or top-down analysis. Top-down UQ compares measurements of the same item(s) from multiple assay techniques and/or laboratories. Bottom-up UQ quantifies sources of input uncertainty to a process and the resulting output uncertainty. Bottom up UQ is the focus of this paper. UQ for NDA has always been important, but there is a need for application of better statistical techniques and for UQ to play more of a role in assay development and assessment. This paper describes NDA applications (the enrichment meter, distributed-source term analysis, Cf shuffler, Uranium neutron collar) that have emerging UQ topics that are not specifically addressed by the GUM, including: item-specific biases, errors in predictors, model error effects, and quantification of uncertainty in computer model errors and parameters. This paper also includes an initial UQ case study using the Uranium neutron collar assay method.