Recent Progress in The Defense Nuclear Nonproliferation Research and Development Initiatives in Data Science for Safeguards

摘要

The Department of Energy's National Nuclear Security Administration (DOE/NNSA) Defense Nuclear Nonproliferation Research and Development (DNN R&D) program has invested in researching the use of data science methods to improve the detection of diverted material, facility misuse, and other undeclared activities of concern. Nuclear Safeguards is a data-rich field that is ideal for the application of modern data analytic techniques. Application of data analytics and machine learning to test data sets and controlled safeguard situations will yield insights that will mature these techniques for future deployment. This multi-lab effort uses actual data obtained from an operating nuclear facility at the Oak Ridge National Laboratory (ORNL) for functional evaluations. The projects are investigating the use of common and new data science methods, with the goal of developing safeguards-specific algorithms and methodologies that will have to ability to use non-traditional data for domestic or international safeguards programs. This paper will report on the development of these initial efforts. Areas of research include examining facility behavioral patterns, data surety, anomaly detection, rapid handling, and analysis of large datasets.