Smart Monitoring of Nuclear Facilities: Implementation Concepts and Development Status

LA-UR-19-204042019 INMM AbstractTitle: Smart Monitoring of Nuclear Facilities: Implementation Concepts and Development StatusAuthors: Karen Miller, Emily Casleton, Janette Frigo, Paul Mendoza, Rosalyn Rael, Kendra Van Buren, Jonathan Woodring, Vlad HenzlAbstract:The world is at an inflection point where our ability to collect data far outpaces our ability to make sense of it, and the proliferation of sensors and so-called Internet of Things (IoT) is changing the way society uses data in profound ways. The wealth of information contained within large and heterogeneous datasets is widely recognized, yet the methodologies required to analyze them efficiently are not sufficiently mature. Building off the idea that the whole is greater than the sum of its parts, the goal of this work is to develop methods for combining disparate data representative of current or potential future safeguards data streams to improve detection of misuse or diversion in declared facilities and deter would-be proliferators through the threat of early detection. This project pulls inspiration from modern smart home and commercial sensing technologies and applies it to the challenge of monitoring and verification of nuclear facilities. The team is using machine learning techniques to integrate persistent, multisource data streams collected in one of the unique radiological facilities at Los Alamos National Laboratory (LANL) for patterns-of-life analysis. The intent is that this work will provide the foundational building blocks that can then be applied to data collected in other facilities. The paper provides an overview of the near- and longer-term implementation concepts envisioned for this work along with preliminary analysis on data streams collected in the testbed at Los Alamos.