Techniques and Models to Enhance RUL Prognostics and Fault Detection in Mechanical Systems

"Impact Technologies proposes to develop a suite of Statistical Influence Models (SIMs) and fusion techniques for enhancing physics-based prognostic models and calibrating them in the presence of various forms of fault detection or state awareness inaircraft mechanical systems. While the robustness and accuracy of physics-based Remaining Useful Life (RUL) Prognostic models and incipient fault detection tools have been improving, the improvements have thus far been largely independent of each other.Through state-of-the-art knowledge fusion of various Statistical Influence Models (SIMs) focused on usage profiles, manufacturing defects, random damage events, build tolerances, material condition and inspection capability, the integration of stateawareness and predictive prognostics promises to be significantly improved.The enhancement capabilities of the statistical models and fusion techniques will be demonstrated with simulations in a Prognostics Testbench focused on the STOVL lift fan transmission on the F-35 aircraft. The Prognostics Testbench architecture will besuch that generic bearing, shaft and clutch prognostic models will be simulated with pre-defined usage profiles. Statistical Influence Models (SIMs) addressing fault detection updates, damage and defect likelihoods, and manufacturing and maintenanceinduced conditions will be "plugged" into the Testbench to investigate their influence on the RUL predictions and associated c