A dissolver diversion scenario illustrating the value of process monitoring

In large throughput spent nuclear fuel reprocessing plants such as the Rokkasho Reprocessing Plant (RRP) there is a low detection probability for material losses of interest to the IAEA (8 kg of Pu) using even the most optimistic near-real-time accounting (NTRA) methods currently employed. A particularly low detection probability is seen in the head end where “input” shipper declarations (via reactor burnup calculations) having relatively large uncertainties (5-10%) are compared to “output” measurements consisting of waste (leached hulls) measurements plus accountability tank measurements. Currently, a dissolver monitoring system applied by the IAEA utilizes semi-quantitative neutron assay of hull batches to detect changes in the neutron count rate that could indicate excess Pu in the leached hulls. The goal of the exercise reported in this paper is to provide an alternative dissolver process monitoring concept. The approach is to infer the completeness of spent fuel dissolution from easily- monitored process parameters. To provide a framework, a scenario was developed and evaluated where fuel and its contained Pu is purposely left undissolved, resulting in excess Pu in the hulls. The magnitude of the scenario was calculated based on the loss of 8 kg of Pu over the course of 90 working days. Based on the chemical models and material balance calculations presented here, relatively large changes in temperature, acid concentration or reaction time are needed for the stated material loss. Further, these process changes would be easily observable using current process monitoring technologies, but further work is needed to evaluate authentication strategies and performance under plant and long term conditions. Total uncertainties will depend upon the errors associated with model calculations and measurement errors. Estimation of these uncertainties is the next logical step for understanding the value of process monitoring in this scenario.