Distributed Adaptive Control of Engine Systems

Aurora and Georgia Tech’s Phase I efforts demonstrated the feasibility of a partially distributed control scheme with separate controllers on the engine core and fan, where the controllers are linked by a supervisory controller. This scheme is representative of the situation encountered in VTOL UAV design and the design of new turbo-props and variable pitch turbofans by the large commercial gas turbine manufacturers. In Phase II Aurora proposes to develop the partially distributed controller from Phase I further to cover safe performance during non standard operations (including sensor failure etc.), culminating in a static engine test of a small turbo-prop engine running the developed distributed adaptive controller. BENEFIT: Moving to a distributed architecture will increase flexibility through common standards, improve redundancy properties by improving the overall system topology, and allow for self-diagnosing components and other benefits of ‘smart’ actuators and sensors, such as reduced harness weight. Distributed computing in the smart components allows for localization of A/D conversion and signal processing, supports open standards and modularity, and provides an opportunity for self-diagnosis. Beyond this, our approach will tap into the full potential of a distributed architecture by allowing the control algorithms themselves to be distributed. This reduces the system’s dependence on the FADEC, reducing the number of redundant components and interconnections required to insure reliability. Furthermore, the benefits of adaptation, robustness, and self-repair at the component level are envisioned through feedback control at the component level, improving overall system reliability and performance.