Modeling Efforts to Aid in the Prediction of Process Enrichment Levels with the Intent of Identifying Potential Material Diversion

As part of an ongoing effort at Lawrence Livermore National Laboratory (LLNL) to enhance analytical models that simulate enrichment and conversion facilities, efforts are underway to develop routines to estimate the total gamma-ray flux and that of specific lines around process piping containing UF6. The intent of the simulation modeling effort is to aid in the identification of possible areas where material diversion could occur, as input to an overall safeguards strategy. The operation of an enrichment facility for the production of low enriched uranium (LEU) presents certain proliferation concerns, including both the possibility of diversion of LEU and the potential for producing material enriched to higher-than-declared, weapons-usable levels. Safeguards applied by the International Atomic Energy Agency (IAEA) are designed to provide assurance against diversion or misuse. Among the measures being considered for use is the measurement of radiation fields at various locations in the cascade hall. Our prior efforts in this area have focused on developing a model to predict neutron fields and how they would change during diversion of misuse. The neutron models indicated that while neutron detection useful in monitoring feed and product containers, it was not useful for monitoring process lines. Our current effort is aimed at developing algorithms that provide estimates of the gamma radiation field outside any process line for the purpose of determining the most effective locations for placing in-plant gamma-monitoring equipment. These algorithms could also be modified to provide both dose and spectral information and, ultimately, detector responses that could be physically measured at various points on the process line. Such information could be used to optimize detector locations in support of real-time on-site monitoring to determine the enrichment levels within a process stream. The results of parametric analyses to establish expected variations for several different process streams and configurations are presented. The benefits and issues associated with both passive and active interrogation measurement techniques are also being explored.