INVESTIGATION OF THE BENEFIT OF THE COUPLED AERO-ELASTIC ADJOINT APPROACH FOR FLEXIBLE WINGS

Kurzfassung

This master thesis presents a comprehensive investigation into the optimization of XRF1 aircraft configurations, specifically focusing on enhancing their aerodynamic efficiency. The primary objective is to identify the most effective approach for optimizing various wing configurations, taking into account their varying levels of elasticity. The study examines and compares three distinct adjoint methods: the aerodynamic adjoint, aero-elastic adjoint with adjustable convergence rates, and the hybrid adjoint technique. The research outcomes indicate that, for configurations characterized by rigid wings with deflection to span ratios up to 4%, the aerodynamic adjoint consistently delivers superior optimization improvements compared to other approaches. However, it is worth noting that the hybrid adjoint approach can also offer substantial benefits and should be considered as a valuable lternative to the aerodynamic adjoint. In particular, when the hybrid adjoint is predominantly reliant on the aerodynamic adjoint to evaluate a considerable number of design iterations, it demonstrates notable optimization gains while maintaining computational efficiency. Further investigation is recommended to explore the full potential of this approach in such cases. Conversely, for configurations featuring moderate to high levels of wing elasticity, deflection to span ratio between 4% and 8%, both the hybrid adjoint and aero-elastic adjoint exhibit significant advantages. The hybrid adjoint, which combines the aerodynamic adjoint with gradients obtained from the aero-elastic adjoint using low to moderate convergence rates, showcases excellent optimization values per design iteration and effectively reduces computational effort. Consequently, it emerges as the preferred method for optimizing moderate to highly elastic wing configurations. Although the aero-elastic adjoint necessitates considerable computational resources, it achieves ultimate optimization outcomes and can be employed with convergence rates adjusted to strike a balance between accuracy and computational costs. Moreover, for extreme cases involving either very rigid or highly flexible aircraft wings with deflection to span ratios of approximately 8% or higher, the hybrid adjoint algorithm should be designed to rely on the aerodynamic adjoint to a high extent to evaluate a larger number of design iterations before transitioning to the aero-elastic adjoint, like when a smaller improvement threshold ( 1%) is attained. This adjustment ensures that the computational costs associated with highly elastic wing configurations are mitigated without compromising optimization outcomes. In conclusion, this thesis highlights the significance of selecting an appropriate adjoint approach based on the specific characteristics of the wing configuration. The aerodynamic adjoint is recommended for rigid wings with some recommendations on hybrid adjoint, while the hybrid adjoint demonstrates clear advantages for elastic wings. The insights and recommendations provided here contribute to the advancement of optimization techniques for aircraft configurations, allowing for enhanced aerodynamic efficiency while managing computational costs effectively.