Effective Diagnostics for Dynamic Operating Environments Navy SBIR FY2005.2 Sol No.: Navy SBIR FY2005.2 Topic No.: N05-104 Topic Title: Effective Diagnostics for Dynamic Operating Environments Proposal No.: N052-104-0082 Firm: Sentient Corporation 850 Energy Drive Idaho Falls, Idaho 83401 Contact: Sean Marble Phone: (208) 524-4865 Abstract: Fault detection and isolation algorithms are often developed with data from steady-state tests, and many algorithms perform well under these controlled conditions. However, when diagnostics are used in highly dynamic environments, sensitivities to changing loads or speeds become apparent. Vibration metrics tend to be particularly susceptible to operating condition variations. This sensitivity interferes with effective diagnosis and requires fault thresholds to be increased significantly to avoid false alarms, making the system less sensitive to real problems. Sentient Corporation, in cooperation with Pratt & Whitney, will develop new vibration diagnostics for dynamic environments that leverage existing physics-based models for fault induced vibration. These new algorithms will greatly enhance diagnostic performance in dynamic operating environments and will also improve fault severity indication for steady state applications. These new techniques will be developed for and demonstrated on a highly dynamic F-135 subsystem such as the lift fan drive. During Phase I, a functional proof-of-concept system will be implemented and tested using bearing seeded fault data collected under dynamic conditions. Benefits: Vibration is the most widely used diagnostic parameter for mechanical systems. However, unlike industrial machinery and commercial aircraft that operate at steady state for long periods of time, military aircraft are operated under constantly changing conditions. Enhanced capability to handle changing operating conditions without loss of accuracy would greatly improve the performance of PHM systems in military aircraft of all types. In addition, the proposed approach will allow for better indication of fault severity than is currently possible even for steady state environments. This enhanced state awareness will help enable true predictive prognostics, with the associated safety, maintenance and logistics benefits. Return