An opportunity exists to develop an Intelligent Condition Based Maintenance System (ICBMS) to provide "early warning" of equipment maintenance needs. Adaptive process models will estimate changes in machine health from analysis of sensor inputs and machine usage. A troubleshooting and repair knowledge base will provide advice on maintenance scheduling and procedures, and thus support ongoing operations and training of new staff. ICBMS will minimize the cost and disruption of maintenance, repair and unscheduled downtime. Innovations include the use of advanced modeling technologies including neural nets to provide "virtual sensors" and to estimate critical but unmeasurable process and machine health parameters and other available information. Our Automated Knowledge Acquisition technology will extract structured rules by analyzing operational decisions and problem solving approaches provided by machine operators and maintenance staff. Objectives include preliminary process modeling, acquisition of troubleshooting and maintenance expertise suitable for automated knowledge extraction, and a preliminary conceptual design of ICBMS outlying system elements integration strategies and functionality. COMMERCIAL APPLICATIONS: Failure prediction and intelligent scheduling of maintenance and repair for: (1) rotating machinery, including power generation, machine tools, propulsion systems, (2) thermal processes, (3) electromechanical systems, (4) high performance/extreme conditions systems
