Aerodynamic design optimization using computational fluid dynamics

An aerodynamic design optimization technique is presented that couples direct optimization algorithms with the analysis capability provided by appropriate computational fluid dynamics (CFD) programs. This technique is intended to be an aid in designing the aerodynamic shapes and establishing test conditions required for the successful simulation of aircraft engine inlet conditions in a ground test environment. However, the method is also applicable to other aerodynamic design problems such as airfoil design, turbomachinery cascade design, and nozzle design.

The approach involves minimization of a nonlinear least-squares objective function which may be defined in a region remote to the geometric surface being optimized. In this study, finite-difference Euler and Navier-Stokes codes were applied to obtain the objective function evaluations, although the applied optimization method can be coupled with any appropriate CFD analysis technique. Using CFD to compute design space gradients within an optimization algorithm has received little prior attention in the literature. It is demonstrated, herein, that CFD can be used in this manner by applying the developed design technique to a variety of aerodynamic design problems. Results are presented for several typical aerospace examples, including algebraic test functions, inviscid and viscous flow over airfoils, flows in convergent/divergent nozzles in both two and three dimensions, and supersonic flow about a planar forebody simulator.