Adaptive Turbine Engine Control for Stall Threat Identification and Avoidance Navy STTR FY2010.A Sol No.: Navy STTR FY2010.A Topic No.: N10A-T008 Topic Title: Adaptive Turbine Engine Control for Stall Threat Identification and Avoidance Proposal No.: N10A-008-0358 Firm: Aurora Flight Sciences Corporation 9950 Wakeman Drive Manassas, Virginia 20110 Contact: Nathan Fitzgerald Phone: (617) 500-5045 Web Site: www.aurora.aero Abstract: Aurora Flight Sciences and MIT propose to develop a model-based adaptive health estimation and real-time proactive control to identify gas turbine engine stability risks and avoid them through control action. In this concept, the engine control system actively monitors sensors and actuators, compares them against physical models, and infers which components may be performing poorly and may need to be replaced. This estimation procedure will be based on Hidden Markov Model of the deterioration process. In addition, once a threat is identified, the control is given the ability to modify its behavior to avoid operations that increase the chance of compression system instability. Aurora intends to accomplish this through the development of model-based control based on the concept of Rapidly-Expanding Random Trees (RRTs). Using these techniques, maintenance personnel will have the ability to upload health tracking parameters from the control system, directly identifying component performance issues needing corrective action. The engine control system will also be able to adapt to changing deterioration conditions to ensure stall-free operation. Benefits: The proposed model-based health estimation and control system will allow engine maintenance personnel to proactively assess engine component health and replace components before an engine begins experiencing in-service stalls. The model-based estimation technology required to accomplish this task could be used in a variety of applications that may benefit from model-based control, anywhere from autonomous aircraft systems to chemical processing plants. The Rapidly-Expanding Random Tree control concept proposed is anticipated to be an economic method of performing constrained optimization in a model-based control for an aerospace subsystem. This technology is applicable to a variety of systems requiring path planning amid constraints, and ranging from gas turbine engine control to multi-vehicle coordination. Return