GAS CENTRIFUGE DIVERSION PATH ANALYSIS FOR NEW MATERIALS CONTROL AND ACCOUNTABILITY CONCEPTS, LA-UR-06-0630

Year
2006
Author(s)
Mark M. Pickrell - Los Alamos National Laboratory
S. Tobin - Los Alamos National Laboratory
Scott F. Demuth - Los Alamos National Laboratory
Justin E. Doak - Los Alamos National Laboratory
Kristen Klasky - Los Alamos National Laboratory
Abstract
New concepts have been, and continue to be, proposed for safeguarding uranium gas centrifuge enrichment plants (GCEPs)[1]. As new materials control and accountability (MC&A) concepts are proposed, standardized diversion path analyses should be utilized to estimate proliferation risk reduction. As part of the diversion path analysis, process modeling should be included to quantify the operations and/or reconfiguration complexity. For instance, reconfiguration concepts that consider only enrichment increases, but do not evaluate centrifuge hydrodynamic requirements, can seriously oversimplify the adversary’s effort. Consequently, an effort has been undertaken to couple conventional diversion path analysis with process modeling as applied to new MC&A concepts for GCEPs. In large scale low enriched uranium GCEPs, many possibilities exist for reconfiguring to achieve highly enriched uranium. The approaches to reconfiguring can include (1) maintaining an “ideal” cascade configuration in a continuous mode, (2) maintaining an “ideal” cascade configuration in a batch plus continuous mode, and (3) utilizing a “non-ideal” cascade configuration. As used here, “ideal” refers to conditions where the 235U fraction in the product from stage “n-1” and the tails from stage “n+1” are equal. If the centrifuge hydrodynamic requirements are considered, the above scenarios (1) through (3) can dictate significantly different levels of operational and reconfiguration effort. These scenarios can also require different MC&A concepts. For instance, a scenario dependent on batch processing will likely require de-sublimation capability; whereas a scenario utilizing a non-ideal cascade configuration will likely require significant testing prior to operation. It is for these reasons that process modeling will be coupled with diversion path analysis so that the adversary’s effort can be factored into proliferation risk reduction estimates. The paper will describe the concepts utilized for combining the diversion path analysis and process modeling. [1] M. Pickrell, Design of an Automated Safeguards System for an LEU Enrichment Plant, LA-UR-05-7195, Los Alamos National Laboratory, August 2005.