Automated Reasoner Technology for Managing Military Aircraft Navy SBIR FY2006.1 Sol No.: Navy SBIR FY2006.1 Topic No.: N06-023 Topic Title: Automated Reasoner Technology for Managing Military Aircraft Proposal No.: N061-023-0573 Firm: Sentient Corporation 850 Energy Drive Idaho Falls, Idaho 83401 Contact: Sean Marble Phone: (802) 876-3100 Abstract: The F-35 prognostic health management (PHM) system will generate a large volume of data on current and predicted component health. Tools are needed to automatically assimilate and analyze this data, and present findings and recommended actions to the user. These automated reasoner tools are necessary to ensure that PHM data is utilized for maximum benefit, to allow maintainers to focus on maintenance rather than data interpretation, and to present capability and readiness information to commanders and decision makers. Sentient Corporation, in cooperation with JSF OEM partners, will develop automated reasoner algorithms and software tools that fuse PHM data into current and predicted capability information, preserve and propagate uncertainty information, and present this information to the user in an intuitive yet powerful way. Phase I activities will include research on user information needs and a prototype demonstration which combines a novel model-based prognostic data fusion technique, a fuzzy expert system for interpretation, and a web-based user interface. This demonstration will be conducted for a selected F-35 subsystem, and will include multiple realistic maintenance scenarios. A large existing fault-to-failure progression database for 48 bearings will be used to provide raw data to feed the Phase I demo. Phase II will include further development and implementation for multiple subsystems, and a demonstration using actual F-35 PHM data. Benefits: Automated reasoner and data presentation tools for PHM would free maintenance personnel to focus on maintenance, rather than data analysis and interpretation. This will help insure consistent diagnoses, faster turn-around, and higher sortie generation rates by taking full advantage of the tremendous knowledge embedded in PHM data. Return