New Approaches in Process Monitoring for Fuel Cycle Facilities

Process monitoring (PM) has been part of the safeguards approach for fuel cycle facilities for many years, but its use has been limited. For example, aqueous reprocessing plants may use bulk level measurements to generate a bulk material balance that can be correlated with traditional nuclear material accountancy measurements. However, advanced measurement technologies and modern data analytics approaches may provide new approaches for PM. New facility types may also drive the need for better approaches. Pyroprocessing plants have unique safeguards challenges and unique measurements that could be used to monitor operations. The purpose of the work presented here is to examine improved PM approaches for both aqueous and pyroprocessing facilities. Both unique measurements specific to those facility types as well as machine learning techniques to correlate various data types are being examined. The success metrics are either to improve detection probability or timeliness of detection or to reduce safeguards burden through increased use of unattended monitoring systems. This work relies on safeguards performance modeling to generate simulation data for training followed by diversion or misuse scenarios for testing the approaches. The approaches being considered and preliminary results are presented.