Hybrid Model Fusion for Gas Turbine Engine Diagnostics and Prognostics

In Phase I IAC, teamed with Pratt & Whitney and Luppold & Associates, developed and demonstrated a hybrid modeling approach to enhance diagnostics and prognostics performance on F-117 engines. The hybrid model fuses a physics-based model developed by Pratt & Whitney called STORM with an empirical model that uses neural networks to monitor and quantify unmodeled and/or mismodeled engine phenomena that corrupt STORM?s diagnostic outputs. The resulting hybrid model is called the enhanced STORM or eSTORM. In Phase I eSTORM was demonstrated to work extremely well for limited conditions when processing simulated engine data. In Phase II IAC proposes to develop a standalone system for real time implementation of a full F-117 engine eSTORM. The system will cover the full C-17 flight envelope. The system will include interfaces to accept data from the C-17 aircraft bus and will be capable of on-wing operation. The system will be demonstrated using real F-117 inputs running in a test bench environment. Though developed explicitly for an F-117 application, the hybrid modeling concept developed on this Phase II is generic and can be used for eSTORM development on all Pratt & Whitney military and commercial engines.