Data Mining of Pareto-Optimal Transonic Airfoil Shapes Using Proper Orthogonal Decomposition

摘要

A new approach to extract useful design information from the shape data of Pareto-optimal solutions of annoptimization problem is proposed and applied to the optimization of airfoil shapes for good aerodynamicnperformance at transonic speed. The proposed approach decomposes shape data into principal modes andncorresponding base vectors, using proper orthogonal decomposition.The advantage of the proposed approach is thatnthe knowledge one can obtain does not depend on howthe shape is parameterized for design optimization.Analysis ofnthe airfoil shapes obtained as the Pareto-optimal solutions for aerodynamic performance at transonic speeds showsnthat the optimized airfoils can be categorized into three families (low-drag designs, high-lift-to-drag designs, andnhigh-lift designs), where the lift is increased by changing the camber near the trailing edge among the low-dragndesigns, while the lift is increased by moving the lower surface upward among the high-lift designs.