Fast Airfoil Design Based on Multi-output Gaussian Process Regression

Wind tunnel experiments and CFD technology are the traditional airfoil design methods. However, they are usually costly and time-consuming. In order to reduce costs and improve the airfoil design efficiency, this paper applies multiple-output Gaussian process (MGP) regression model to fast evaluation of aerodynamic performance and fast inverse design of airfoil. In this paper, a group of NACA series airfoils is selected, and the corresponding lift coefficient, drag coefficient and surface pressure distribution are calculated. Based on the experimental data, we have established a MGP regression model to perform aerodynamic performance evaluation and inverse design of airfoil. The experimental results show that the MGP regression model can improve the airfoil design accuracy compared with single-output Gaussian process (SMP) regression model and other multiple-output models, such as BP and RBF artificial neural networks.