Material Balance Areas and Frequencies For Large Reprocessing Plants*

It has long been recognized that facilities with a large nuclear material throughput will probably not meet the International Atomic Energy Agency (IAEA) goal for detecting trickle diversion of plutonium over periods of about one year. The reason is that measurement errors for plutonium concentration and for liquid volume are often approximately relative over a fairly wide range of true values. Therefore, large throughput facilities will tend to have large uncertainties assigned to their annual throughput. By the same argument, if frequent balances are performed over small material balance areas, then the uncertainty associated with each balance period for each balance area will be small. However, trickle diversion would still be difficult to detect statistically. Because the IAEA will soon be faced with safeguarding a new large-scale reprocessing plant in Japan, it is timely to reconsider the advantages and disadvantages of performing frequent material balances over small balance areas (individual tanks where feasible). Therefore, in this paper we present some simulation results to study the effect of balance frequency on loss detection probability, and further simulation results to study possibilities introduced by choosing small balance areas. We conclude by recommending frequent balances over small areas.