Hypersonic Transition to Turbulence Triggered by Isolated Roughness Elements

Abstract

Detached Eddy Simulations have been performed to study the complex physics of boundary layer transition triggered by isolated roughness elements in hypersonic flows. Both frozen and reacting flow solutions have been obtained for a hemisphere with a disk-like surface roughness element in a ballistic range at Mach 12 in air. The results show a significant effect of non-equilibrium thermo-chemistry on high-enthalpy roughness-induced transition. The effect of Reynolds numbers on the Space Shuttle trip transition in a Mach 10 wind tunnel has also been studied. I. Introduction The design of hypersonic vehicles is challenging in several critical technology areas. The severe heating environment encountered during hypersonic flight dictates the shape of the vehicle. Boundary-layer transition at hypersonic speeds poses an especially significant challenge. Prediction and control of boundary layer transition in hypersonic flows are of crucial importance for the design of planetary entry vehicles as well as two-stage-to-orbit airbreathing access-to-space systems. Since turbulent heat transfer rates can be higher than laminar heating rates, reductions in the weight of thermal protection systems can be realized with an improved understanding of the physics of transition from laminar to turbulent flow. The hypersonic heating environment, coupled with an emphasis on reusability, creates additional severe technology challenges for materials, material coatings, and structures that not only carry aerodynamic loads but also repeatedly sustain high thermal loads while requiring longlife and durability with minimum weight. The gap-filler incident of the Space Shuttle mission STS-114 in 2005 was a potent reminder of the importance of accurate prediction of roughness-induced boundary layer transition and subsequent increase in surface heating 1 . Direct Numerical Simulation (DNS) solves the Navier-Stokes equations by resolving a wide range of spatial and temporal scales of turbulence. Since DNS requires a grid fine enough to resolve the Kolmogorov dissipative scales, it is not feasible for high Reynolds number flow simulations even with today’s most powerful supercomputers. Large Eddy Simulation (LES) requires less computational resources than DNS by modeling small eddies using subgrid scale models while still resolving large eddies. However, even with this improvement, the grid requirements for high Reynolds number LES calculations are still impractical. Coarse-Grid Direct Numerical Simulation (CDNS) is a turbulent flow simulation method without a sub-grid scale model but which is not a fully resolved DNS. Since the high cost of computation for LES comes from the near-wall region, hybrid models like Detached Eddy Simulation (DES) have been developed which alleviate this difficulty by using a Reynolds-Averaged Navier-Stokes