Statistical Methods in Nuclear Material Safeguards

Abstract

The nuclear industry must account for nuclear material (NM), such as items that contain plutonium‐239, uranium‐233, or uranium enriched in the 235 or 233 isotopes. Such materials are expensive, plus play a role in weapons production, so there are national and international security considerations. There are two fields of NM safeguards. First, the field of NM domestic safeguards protects NM and monitors for possible NM loss, and does not include the possibility that a facility might alter the accounting data to mask NM diversion. Second, the field of NM international safeguards does not control NM, but monitors for possible NM loss or diversion where inspectors take measurements to determine whether a facility diverted NM and could have altered the accounting records to try to mask the diversion. Domestic safeguards has two main components: physical security to monitor and control access to NM and accounting procedures to monitor the flow and location of NM quantities. Analogously, the banking industry also uses physical security (guards, cameras, locked vaults, etc.) together with accounting procedures to monitor money flow. However, the nuclear industry is more complicated owing to the multitude of physical forms and chemical compositions of NM in bulk quantities, resulting in measurement errors and statistical methods to monitor NM flow.

Essential Statistical Concepts

“…Many of the published relative standard deviation estimates are based on top-down UQ, using specialized analysis of variance. Bonner et al (2016) and Burr (Burr, Croft, et al 2016) provide further discussion on how statistical methods are used to verify nuclear material inventories (Bonner 2016;Burr, Croft, et al 2016;Burr, Krieger, et al 2016). The American Society for Testing and Materials maintains a number of useful standards and guides in NDA instruments and methods and the National Institute of Standards and Technology has a very good on-line handbook on statistical methods (National Institute of Standards and Technology 2012).…”