Evolutionary algorithm shape optimization of a hypersonic flight experiment nose cone

Hypersonic vehicle development, particularly for hypersonic airbreathers, will require robust-design optimization to achieve performance targets. Vehicle shape optimization will play an important role. This paper presents the first application to hypersonic vehicle shape optimization of a reduced parametric section shape representation coupled with a surrogate-assisted evolutionary-algorithm optimization approach. The particular case considered is the minimization of total drag of the nose cone of a hypersonic flight experiment. The computational fluid dynamics solver used here is ANSYS CFX. Single-point optimization at Mach 3 and Mach 8 is performed at altitudes relevant to those Mach numbers for a typical hypersonic flight experiment ascent trajectory. Without surrogate assistance, the optimized shapes for each Mach number were both found to result in significant drag-force reductions (1.39% at Mach 3 and 1.96% at Mach 8) when compared with the baseline blunted standard ogive nose-cone shape. When the surrogate assistance (which is a radial-basis-function network approximation to the drag dependence on the shape parameters) was introduced, the optimized shapes yielded nearly the same drag reduction as without surrogate assistance, but also resulted in significant savings in the computational cost. Finally, the performance of the optimum shape derived at Mach 3 is evaluated at Mach 8 and vice versa to illustrate the robustness of the nose-cone shapes derived using such an approach.