Intelligent aerodynamic design for airfoil based on Artificial Neural Network Method

Abstract

The Artificial Neural Network (ANN) method is applied to intelligent aerodynamic design for airfoils. A Self-Organizing Map (SOM) network is demonstrated selecting referenced airfoils which mostly meet or close to the design requirement from airfoil database. Then a Back-Propagation (BP) network automatically learns the relationship between referenced airfoil geometry and aerodynamic performance by means of supervised learning approach. After a set of training, the BP network is able to estimate airfoil aerodynamic characteristics using knowledge and criteria learned before. Design results indicate that trained network can give effective prediction and excellent aerodynamic efficiency for airfoil.