Optimization of Local Stiffness Properties for Improved Aeroelastic Stability of a Transport Aircraft

Abstract

This thesis investigates the optimization of local structural stiffness in a transport aircraft to improve aeroelastic stability, particularly flutter behavior. High-aspect-ratio wings, an approach for more efficient aircraft, are more flexible and thus susceptible to aeroelastic instabilities. It is therefore crucial for the aircraft manufacturers to minimize the risk of encountering instabilities late in the design process. The contribution towards this risk mitigation is to explore optimization capabilities within the MSC Nastran environment. The feasibility of integrating flutter constraints in the sizing optimization of aircraft design processes is then evaluated based on the results. The specific objective is to increase the flutter velocity with minimal mass increase, using gradient-based optimization. After outlining the fundamentals of flutter analysis and optimization algorithms, the D2AE reference aircraft model was adapted from the in-house cpacs2MONA process. Initial flutter analyses of the free aircraft model revealed error-prone mode tracking due to the flutter solver. As a result, the model was simplified to a fixed-wing configuration to improve robustness and allow for more reliable optimization. Parameter studies were conducted in an attempt to provide consistent flutter analysis. The design model was then modified to include skin fields and spars as design fields. In addition, the pylon was introduced as a separate optimization model. Several optimization algorithms were tested and different objective functions were compared. Although all solvers showed consistent global trends in wing thickness distribution, tuning the optimization parameters revealed a high sensitivity toward move limits and constraint scaling. The wing model achieved a targeted 5% increase in flutter velocity. The pylon model, on the other hand, failed to improve flutter stability. Gradients were obtained and it was concluded that the specific flutter mode was not sensitive to pylon stiffness. Applying the optimized wing stiffness distribution back to the full aircraft model provided a flutter velocity increase of 4.62% at the cost of approximately 130 kg of additional mass. This demonstrated that local structural modifications can effectively improve flutter stability with moderate weight penalties. In conclusion, this thesis presents a successful proof of concept for flutter-constrained optimization using Nastran on complex models. However, it also highlights the current difficulties of integrating this process into automated design workflows.