2023/04/25 by Ruh, Marius L., Sarojini, Darshan, Fletcher, Andrew +2
#Computational Engineering #FOS: Computer and information sciences #FOS: Mathematics #Finance #J.2 #Optimization and Control (math.OC) #and Science (cs.CE)
paper · doi:10.48550/arxiv.2304.14889
The conceptual design of eVTOL aircraft is a high-dimensional optimization problem that involves large numbers of continuous design parameters. Therefore, eVTOL design method would benefit from numerical optimization algorithms capable of systematically searching these high-dimensional parameters spaces, using comprehensive and multidisciplinary models of the aircraft. By leveraging recent progress in sensitivity analysis methods, a computational framework called the Comprehensive Aircraft high-Dimensional DEsign Environment (CADDEE) has been developed for large-scale multidisciplinary design optimization (MDO) of electric air taxis. CADDEE uses a geometry-centric approach that propagates geometry changes in a differentiable manner to meshes for physics-based models of arbitrary fidelity level. The paper demonstrates the capabilities of this new aircraft design tool, by presenting large-scale MDO results for NASA's Lift+Cruise eVTOL concept. MDO with over 100 design variables, 17 constraints, and low-fidelity predictive models for key disciplines is demonstrated with an optimization time of less than one hour with a desktop computer. The results show a reduction in gross weight of 11.4% and suggest that CADDEE can be valuable in the conceptual design and optimization of eVTOL aircraft.