Summary

Description

Fault to Failure Progression (FFP) signature modeling and processing is a new method for applying condition-based signal data to detect degradation, to identify fault modes, and to produce system estimates for State of Health (SoH) and Remaining Useful Life (RUL). The base technology has been applied for prognostic purposes for various government-sponsored programs, but FFP signature modeling and processing has not been applied for the area of Fault Management, nor does it include such features as fault dictionaries, lookup tables, and management algorithms. The technology includes Ridgetop-designed and developed algorithms to do the following: (1) perform Kalman Filtering to reduce noise; (2) transform sensor signal data to reveal underlying (hidden) FFP signatures; (3) normalize units-of-measure dependent signal data into dimensionless FFP signatures to facilitate re-use and reduce the time to characterize and define new FFP signatures; (4) define and use model definitions that reduce memory requirements and support fast and accurate processing and calculations; (5) two forms of trajectory curve characterization, both straight-line and curvilinear; (6) a fast yet accurate, graphics-based mathematical routine to adapt an FFP model to received data; (7) amplitude and time updates similar to Extended Kalman Filtering to estimate how long it will take an adapted FFP model to reach a defined failure threshold; and (8) produce SoH and RUL estimates that rapidly converge to the estimated time-to-failure (TTF) solution. The FFP signature modeling and processing will include additional innovation to support FM to minimize application-specific programming, those include algorithms to simplify fault identification and isolation.