The influence of uncertainty in usage and fatigue damage sensing on turbine engine prognosis

Hudak Jr. S.J., Lanning B.R., Light G.M., Major J.M., Moryl J.A., Enright M.P., McClung R.C., Millwater H.R.

Abstract:
Significant enhancements in the reliability and readiness of high-value assets are believed to be achievable by developing and implementing prognosis systems. This real time, or near-real time, decision making process is based on the acquisition and fusion of on-line sensor feedback, combined with physics-based analytical models for damage accumulation, and higher order reasoning for decision making. Since many elements of this process contain uncertainty it is necessary to treat this process probabilistically. The influence of two key sensor inputs on the probability of component failure is examined - uncertainty in mission usage, and uncertainty in the current material damage state. Example analyses for the case of fatigue cracking at a bolt-hole in a generic turbine disc serves to demonstrate that usage is highly variable, depends on mission type, and resulted in a 6 X variability in fatigue life and 10X to 100X variability in probability of failure (POF) at a given life. Using virtual sensors, the trade-off between sensor sensitivity and frequency of interrogation on the probability of disc failure is simulated, and target sensitivities for on-board sensor development are established. The concept of a novel thin-film, wireless sensor system for detection and monitoring fatigue cracking in the elevated temperature environment of turbine engines is then introduced. The general approach and current status of the development of this sensor system are also summarized.

回上页