RASA Prognostics from State of Health (SOH) Modeling

75406-Radionuclide Aerosol Sampler/Analyzer (RASA) units are deployed in many parts of the world by the United States and other countries for the monitoring of nuclear explosions. Although raw State of Health (SOH) information is available for many of these units, no methodology exists to exploit this SOH data for the diagnosis and detection of subtle hardware faults. Predicting the failure of a critical component of a RASA unit would enable a schedule of service or repairs prior to component failure, thus reducing costs, reducing down time, and increasing network capability. This project will develop a system to enable users at the United States National Data Center to better monitor and predict component failure of deployed RASA units, using information from transmitted SOH data. The system will include a predictive failure model, or prognostic model, which will either remove or parameterize environmental factors that may skew the data model. Phase I will demonstrate the ability of the prognostic model to predict component failure from monitored SOH data, develop database tables and attributes, and develop a software system to visualize data and results. The prototype system will enable end users at the United States National Data Center to evaluate its capability early in the development cycle. Commercial Applications and Other Benefits as described by the awardee: In addition to the national security application, the prognostic models for component failure should easily translate into the commercial sector for equipment of the same specific nature as that utilized within the RASA units.