Health Management for Gas Turbine Engine Accessory Components

The overall objective of this project is to develop an effective and practical diagnostic/prognostic capability for turbine engine accessories. The Phase I effort demonstrated the effectiveness of combining high fidelity physics-based models with feature based algorithms to diagnose incipient faults in turbine engine accessories. The goals for Phase II are to fully implement these techniques and to develop complete, functional diagnostic software for an accessory subsystem and demonstrate its performance on high fidelity bench test data using the actual aircraft hardware. A logical choice would be a fluid pump such as the one used for the Phase I proof-of-concept; data from Tinker AFB indicates that fluid pumps are one of the more problematic accessory components in the F100 and F110 engines. Phase II work will include further development of a detailed physics-based model, and coupling that model with fault diagnosis algorithms. Sentient will leverage its extensive in-house test capability to perform seeded fault bench tests on actual engine accessory hardware. Using this data, the final diagnostic algorithms will be used to identify the presence, type and severity of the various types of fault present in the hardware. This will be performed over the full range of speeds and loads occurring in actual use, including typical transients, to demonstrate resistance to false alarms. The final algorithms will also be benchmarked for computational requirements. By the end of Phase II, Sentient will demonstrate diagnostic performance on real aircraft hardware (at TRL 6) with sufficient fidelity such that the Air Force can make a Phase III transition decision.