Robust high-fidelity dataset-based shape optimization for aircraft design

This paper demonstrates the feasibility of high-fidelity aerodynamic dataset-based aircraft shape optimization for mission performance improvements within an automated framework.The framework was applied to a simple, yet realistic test case: the Optimization Test Interceptor with Fan (OTIFAN) - a fixed-wing unmanned aerial vehicle (UAV) featuring an internal electric ducted fan (EDF) propulsion system. To ensure viable aircraft designs, lower-fidelity methods are incorporated for the disciplines of mass properties, flight mechanics, and structural analysis. Three optimization strategies have been implemented and assessed: Grid search (GS) for setup and validation, gradient-based optimization (GO) for efficient local optimization and Bayesian optimization (BO) for global, gradient-free optimization. The strategies are applied across two objective functions, illustrating the applicability of the method to geometric, mission-profile and structural optimizations. This novel approach advances the state of the art in high-fidelity optimization for aircraft design and showcases the engineering potential for enhancing mission-based performance of industrially relevant configurations through automated optimization. While the application of high-fidelity RANS computations mitigates the limitations of simplified aerodynamics, robustness of the design is ensured by the aerodynamic dataset-based approach, in which the complete flight envelope is considered for each generated shape.