Data-driven airfoil shape optimization framework for enhanced flutter performance

This paper presents a machine learning-based airfoil shape optimization framework designed to increase flutter resistance and reduce drag. Using the National Advisory Committee for Aeronautics airfoil as the base design and a Hicks–Henne bump function, we employ multi-objective Bayesian optimization and harmonic balance-based flutter prediction. The optimization process yields a Pareto front revealing trade-off relationships between the flutter speed index and drag coefficient. The optimized airfoils, resembling those of evolved marine animals, outperform the base design in terms of flutter resistance and drag. These results demonstrate the framework's potential to enhance aircraft performance and safety by addressing aeroelastic factors.