Design Optimization and Uncertainty Quantification of a Dual Airfoil System

We optimize the design of a dual airfoil system. The system consists of two NACA 4412 airfoils whose leading-edge positions are fixed in space in a staggered arrangement. The geometry is a three-dimensional one. The optimization aims to improve the lift to drag ratio. To that end, we resort to Reynolds-Averaged Navier Stokes (RANS) and Bayesian optimization. Firstly, RANS are conducted using the one-equation SA model, several two-equation models, as well as the machine learning model by Bin et al. Secondly, we optimize the two airfoils’ attack angles following a Bayesian strategy. The results provide an uncertainty measure of the RANS-based design optimization. The process leads to a steep angle of attack for the trailing wing. Lastly, we verify the optimal design via large-eddy simulation.