(Georgia Institute of Technology, 2018-06)
Rajaram, Dushhyanth; Yu, Cai; Chakraborty, Imon; Mavris, Dimitri N.
The aerospace industry’s current trend towards novel or More Electric architectures results in some unique challenges for designers due to both the scarcity or absence of historical data and a potentially large combinatorial space of possible architectures. These add to the already existing challenges of attempting to optimize an aircraft design in the presence of multiple possible objective functions while avoiding an overly compartmentalized approach. This paper uses the Integrated Subsystem Sizing and Architecture Assessment Capability to pursue a multi-objective optimization for a large twin-aisle aircraft and a small single-aisle aircraft using the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) algorithm with parallel function evaluations. One novelty of the optimization setup is that it explicitly considers the impacts of subsystem architectures in addition to those of traditional aircraft-level design variables. The optimization yields generations of nondominated designs in which substantially electrified subsystem architectures are found to predominate. As a first assessment of the impact of epistemic uncertainty on the results obtained, the optimization is rerun with altered sensitivities for the thrust-specific fuel consumption penalties due to shaft-power and bleed air extraction. This analysis demonstrated that the composition of architectures on the Pareto frontier is sensitive to the secondary power extraction penalties, but more so for the small single-aisle aircraft than the large twin-aisle aircraft.