Title :
Development of diagnostic and prognostic technologies for aerospace health management applications
Author :
Roemer, M.J. ; Nwadiogbu, E.O. ; Bloor, G.
Author_Institution :
Impact Technol. LLC, Rochester, NY
Abstract :
Effective aerospace health management integrates component, subsystem and system level health monitoring strategies, consisting of anomaly/diagnostic/prognostic technologies, with an integrated modeling architecture that addresses failure mode mitigation and life cycle costs. Included within such health management systems will be various failure mode diagnostic and prognostic (D/P) approaches ranging from generic signal processing and experience-based algorithms to the more complex knowledge and model-based techniques. While signal processing and experienced-based approaches to D/P have proven effective in many applications, knowledge and model-based strategies can provide further improvements and are not necessarily more costly to develop or maintain. This paper describes some generic prognostic and health management technical approaches to confidently diagnose the presence of failure modes or prognose a distribution on remaining time to failure. Specific examples of D/P strategies are presented that include Auxiliary Power Unit (APU) fuel system valves, APU performance degradation and hot section lifing, Power Take Off (PTO) shaft and AMAD snout bearing
Keywords :
Weibull distribution; aerospace expert systems; aircraft computers; aircraft maintenance; aircraft power systems; aircraft testing; backpropagation; computerised monitoring; condition monitoring; data mining; diagnostic expert systems; diagnostic reasoning; failure analysis; fault diagnosis; model-based reasoning; sensor fusion; state estimation; AI prognostics; Weibull distribution; aerospace health management; anomaly technology; auxiliary power unit; backpropagation; component level health monitoring; data mining; diagnostic technology; evolutionary prognostics; experience-based algorithms; failure mode mitigation; failure mode model; fuel system valves; generic signal processing algorithms; hot section lifing; integrated modeling architecture; life cycle costs; model-based approach; neural-fuzzy classifier; performance degradation; power take off shaft; prognostic technology; remaining time to failure; snout bearing; state estimation; subsystem level health monitoring; system level health monitoring; Condition monitoring; Costs; Degradation; Fuels; Knowledge management; Prognostics and health management; Shafts; Signal processing algorithms; Technology management; Valves;
Conference_Titel :
Aerospace Conference, 2001, IEEE Proceedings.
Conference_Location :
Big Sky, MT
Print_ISBN :
0-7803-6599-2
DOI :
10.1109/AERO.2001.931331