Aerodynamic Optimization Using Analytic Descriptions of the Design Space

This paper discusses a strategy for performing aerodynamic single-point and multipoint optimization studies. The approach uses a combination of three methods: an Euler computational e uid dynamics (CFD) code that provides accurate information at one chosen point in the design space, a linear approximation method of the quasianalytic type that rapidly evaluates data at many design points, and a function-e tting algorithm that maps aerodynamic and geometric data across the whole design space. As the method proceeds, a good quality database of aerodynamic and geometric quantities is rapidly constructed at a fraction of the cost of using conventional methods. The availability of this database and the analytic representation of the quantities over the whole design space permits optimizations to be performed using constraints on specie c aerodynamic parameters, e.g., maximum surface Mach number and pressure gradients, over chosen regions of the aerofoil. The approach is also highly e exible, enabling optimization problems of increasing complexity to be built up quickly, and is easily integrated within modern engineering methods. Results are presented for a typical transonic aerofoil giving comparisons with individual CFD analyses.