Airfoil optimization based on multi-objective bayesian

This paper proposes a novel aerodynamic optimization framework for airfoils, which utilizes OpenFOAM, an open-source computational fluid dynamics software, and a Bayesian network to achieve efficient optimization of airfoil aerodynamic performance. Aerodynamic analysis of the NACA 4-digit airfoil was performed by adopting the Spalart-Allmaras turbulence model to solve the Reynolds-averaged Navier-Stokes equations. With the use of this framework, the optimal lift-to-drag ratio can be found by using a small number of objective evaluations. The optimal angle of attack and aerodynamic shape can be obtained under different thicknesses. Finally, after various aerodynamic objectives are arranged and combined, the Pareto fronts were obtained by the multi-objective Bayesian algorithm. Compared with the original NACA four-digit airfoil, the lift-to-drag ratio of the airfoil after single-objective optimization is greatly improved as the thickness increases, and the airfoil after multi-objective optimization achieves different Pareto sets according to different sailing phases.