Reduced-Order Modeling of a High-Fidelity Propulsion System Simulation

Ever stringent aircraft design requirements on simultaneous reduction in fuel consumption, emissions, and noise necessitate innovative, integrated airframe designs that require concurrent engine designs. To fulfill these design challenges, aerospace engineers have relied on a physics-based engine modeling environment, such as the numerical propulsion system simulation. To expedite the use of numerical propulsion system simulation in aircraft design, this research proposes a methodology for the reduced-order modeling of numerical propulsion system simulation by incorporating the following two techniques: probabilistic principal component analysis for basis extraction and neural networks for weighting coefficient prediction. To efficiently achieve an empirical orthogonal basis, this research capitalizes on an expectation-maximization algorithm for probabilistic principal component analysis to handle numerical propulsion system simulation engine decks that typically lack some data due to failed offdesign performance analyses; they result from the numerical instabilities present in the Newton-Raphson method used within numerical propulsion system simulation to find a converged solution at a given flight condition. In addition, to effectively explore a weighting coefficient space, this research uses neural networks to deal with six numerical propulsion system simulation engine modeling parameters. As a proof of concept, the proposed numerical propulsion system simulation reduced-order modeling method is applied to a numerical propulsion system simulation turbofan engine model usually employed for conventional civil transport aircraft. Comprehensive prediction quality investigations reveal that engine performance metrics estimated by the reduced-order numerical propulsion system simulation model show considerably good agreement with those directly obtained by numerical propulsion system simulation. Furthermore, the reduced numerical propulsion system simulation engine model is integrated with the flight optimization system in lieu of directly using numerical propulsion system simulation as an illustration of the utility of numerical propulsion system simulation reduced-order modeling for aircraft design research.