Strengthened Nuclear Safeguards: a Statistical View in the Context of Combining Process Monitoring and Nuclear Material Accounting Data

Year
2012
Author(s)
Kory Budlong-Sylvester - Los Alamos National Laboratory
John Howell - University of Glasgow
Mitsutoshi Suzuki - Japan Atomic Energy Agency
Tom Burr - Los Alamos National Laboratory
Brian Weaver - Los Alamos National Laboratory
Claire Longo - Los Alamos National Laboratory
Michael S. Hamada - Los Alamos National Laboratory
Abstract
“Integrated/strengthened safeguards” is the international atomic energy agency (IAEA) concept that effective safeguards must be maintained at monitored nuclear facilities, the mission of monitoring for undeclared facilities must be added, and the safeguards budget must remain approximately constant. In traditional safeguards, periodic nuclear materials accounting (NMA) measurements confirm the presence of special nuclear material (SNM) in accountability units to within relatively small measurement error. Process monitoring (PM) is used to confirm the absence of undeclared flows that could divert SNM for illicit use. Despite occasional attempts to quantify the diversion detection capability of PM, nearly all quantified statements regarding safeguards effectiveness involve NMA, with PM used as a qualitative added measure. To assess the extent to which PM can provide quantitative assessment in effectiveness evaluation is one of ten recognized technical challenges (discussed during the IAEA’s “Consultancy Meeting on Proliferation Resistance Aspects of Process Management and Process Monitoring/Operating Data” held in Vienna, 28-30 September 2011) in the anticipated increased used of PM data. Effective resource allocation requires effective safeguards system evaluation. This paper reviews safeguards system evaluation methods based on statistical and decision theory, proposes a new evaluation method that quantifies an overall system consisting of both PM and NMA via detection probabilities for specified scenarios that are prioritized using diversion path analysis, and gives three examples at a hypothetical aqueous reprocessing facility. We also describe two complications in using data-driven monitoring of combined PM and NMA subsystems.