Diagnostic/prognostic modeling and reconfigurable control

Repairable, complex systems, whose components tend to degrade with use, require skilled professionals, supported by digital processing and decision aids to monitor and manage system maintenance. The complexity of present day aircraft systems and the increased demands on their reliability motivate the need for more capable diagnostic and control systems that not only detect component/system degradation but can estimate the capabilities of the degraded system and adapt system controls to maximize the overall performance. This paper discusses the bond graph modeling approach and its application to aircraft fuel systems and reconfigurable control that has demonstrated such capabilities. Our approach enables fault diagnosis, isolation, and system reconfiguration as well as system prognostics. The system described has the ability to dynamically update the system or component model and continue to track the system characteristics, providing important feedback information to reconfigurable control processes, and predictive performance estimators. Example aircraft fuel systems that were analyzed and evaluated with our technique are presented.