Standard Guide for Total Measurement Uncertainty Estimation and Reporting for Nondestructive Assay Measurements

Abstract

5.1 Adopting a consistent UQ framework based on commonly accepted scientific methods across all NDA techniques promotes a better understanding and implementation of both the NDA techniques and the UQ process. This supports measurement and reporting that is consistent with good practices. 5.2 To be meaningful, a physical measurement or calculation shall be accompanied by a justifiable and scientifically based statement of its associated TMU. The TMU is the fundamental statement of the measurement’s quality. 5.3 Constructing and verifying the magnitude of each TMU component suitable for the end user is often a complex and specialized task. This process is crucial for understanding the limitations and acceptability of the results. 5.4 This guide is intended to serve as a resource and reference for subject matter experts, those selecting an NDA approach, instrument operators, instrument designers, facility managers, instructors, regulators, and end users of NDA data. 5.5 TMU estimates are necessary for the regulatory acceptance of measurement results. 5.6 TMU estimation is also important for various other purposes, such as: 5.6.1 Evaluating the applicability of a given measurement technique; 5.6.2 Comparing results from different techniques, or from different devices using the same technique; 5.6.3 Planning and optimizing the measurements for the intended purpose; 5.6.4 Informing the process used to generate measurement items; 5.6.5 Identifying and iteratively improving measurement strategies and procedures; 5.6.6 Setting measurement control and operational priorities; and 5.6.7 Determining whether NDA results meet the end user’s DQO. 5.7 Inhomogeneous items can be assayed by NDA techniques (for example, calorimetry) in contrast to destructive analysis (DA) techniques that require homogeneity or have large sampling uncertainty. 5.8 NDA can be used in situ while preserving sample integrity. 5.9 NDA often relies much more heavily on process knowledge compared to DA. 5.10 Sampling errors, such as those in holdup measurements and isotopic sampling combined with bulk analysis, differ from those in DA and are typically less significant because the entire item is assayed as a whole.