Airfoil Shape Optimization Using Variable-Fidelity Modeling and Shape-Preserving Response Prediction

Abstract

Shape optimization of airfoils is of primary importance in the design of aircraft and turbomachinery with computational fluid dynamic (CFD) being the major design tool. However, as CFD simulation of the fluid flow past airfoils is computationally expensive, and numerical optimization often requires a large number of simulations with several design variables, direct optimization may not be practical. This chapter describes a computationally efficient and robust methodology for airfoil design. The presented approach replaces the direct optimization of an accurate but computationally expensive high-fidelity airfoil model by an iterative re-optimization of a corrected low-fidelity model. The shape-preserving response prediction technique is utilized to correct the low-fidelity model by aligning the pressure and skin friction distributions of the low-fidelity model with the corresponding distributions of the high-fidelity model. The algorithm requires one evaluation of the high-fidelity CFD model per design iteration. The algorithm is applied to several example case studies at both transonic and high-lift flow conditions.