A hybrid aerodynamic shape optimization approach for axisymmetric body in hypersonic flow

Abstract

A unique hybrid Aerodynamic Shape Optimization (ASO) framework is devised for axisymmetric bodies with minimum drag coefficient in hypersonic inviscid flow at zero angle of attack. The hybrid ASO framework that has been developed in this paper makes use of a combination of low fidelity framework and high fidelity framework, with the intention to reduce the turn-around time from initial guess to final optimal solution with accurate estimation of cost function. This was achieved when low fidelity framework was able to generate near optimal solution with quick and reasonably accurate estimation of coefficient of drag (cost function for present study). The near optimal solution was then fed as the initial guess for high fidelity framework and it accelerated the convergence to global optimal solution. The low fidelity framework comprised of Modified Newtonian Theory as flow solver while ANSYS FLUENT (v 14.5) formed the flow solver for high fidelity framework. Optimization was effected using Steepest Decent in both the framework and Bezier curves were used for generic shape representation of bodies. ICEM-CFD was used to create the structured grid as required in Fluent. The result show significant reduction in computational cost from initial guess to final optimal solution. It was also observed that the optimal solution obtained from proposed multi-fidelity framework and optimal solution from CFD simulation only, are nearly comparable with marginal difference in the volume. The optimization also resulted in a reduction of coefficient of drag by more than 28%.