Self-Evolving Maintenance and Operations Reasoning

Impact Technologies, in collaboration with Intelligent Automation Corporation (IAC), proposes to develop and demonstrate an integrated, agent-based, self-evolving maintenance reasoning system that can effectively utilize PHM system data/information and support an autonomic logistics concept. The Impact/IAC team will develop and implement this concept as a web-based, data management application with associated software agents that provide reinforcement learning capabilities that can act upon dynamic information sources. To accomplish this goal, an innovative approach to database management, self-learning and adaptation are required. It will include the development and integration of many core elements, ranging from model-based reasoning to OSA-CBM interface development, with various sources of data, information and knowledge being considered to obtain a truly intelligent and evolvable maintenance decision support system. One of the key components of the proposed system consists of various reasoning agents and Maintenance Integrated Models (MIM) containing PHM, FMECA, sensor, and maintenance information. In addition, the open systems architecture (OSA) and associated software interfaces to existing JSF asset management systems (i.e. ALIS) and PHM software containing pertinent maintenance and logistics management information will be addressed. Most importantly, the SEMOR system will employ a combination of case-based and Bayesian processes to evolve the underlying Maintenance Integrated Model (MIM) when given maintainer feedback. The SEMOR system will be demonstrated on a propulsion system application with the support of Pratt & Whitney.