USE OF AN EXPERT SYSTEM MODEL TO RANK RELATIVE ATTRACTIVENESS FOR THEFT OF ALTERNATIVE WASTE FORMS

Year
2003
Author(s)
N. Slater Thompson - U.S. Department of Energy
John G. Vlahakis - U.S. Department of Energy
Chris Einberg - U.S. Department of Energy
Abstract
A novel approach has been developed to evaluate potential attractiveness from a theft standpoint of the Department of Energy’s (DOE) spent nuclear fuel (SNF) and other candidate nuclear materials relative to commercial SNF and vitrified high-level waste (HLW). This approach may eliminate the need for performing relatively expensive analytical measurements to characterize a wide variety of waste forms based on attractiveness criteria. Also, this tool may aid in determining the adequacy of safeguards and security at the monitored geologic repository. The method considers the general concept that a fuel form is attractive for theft if: (1) it is easy to acquire, (2) it contains large amounts of fissile material, and (3) that material can be readily separated from the fuel matrix. In the initial phase, a ranking system was developed to measure the difficulty of separation, or chemical separability, of special nuclear material from DOE-owned SNF. Separability was ranked relative to typical commercial SNF using a decision support tool that employs the Analytic Hierarchy Process (AHP), a pairwise comparison technique that relies on the elicitation of expert judgement. Fuels were grouped according to their similarities for chemical recovery of fissile material. After selecting representative fuels for each group, four operations were identified for separation of fissile material from the DOE SNF: (1) Mechanical operations, (2) Dissolution and conditioning, (3) Separation, and (4) Conversion. By grading a representative fuel with respect to each of the four operations relative to a reference process for a benchmark commercial SNF, a relative separability score was calculated. In the second phase, experts determined the overall weights of importance for the three characteristics of container mass, amount of fissile material, and chemical separability with reference to theft of a container. Those results were then combined with the separability scores from the initial phase to calculate overall relative attractiveness rankings for the entire spectrum of fuels studied. Of the hundreds of fuel records in the DOE database, none of the fuels were ranked significantly more attractive for theft than commercial SNF.