Hypersonic viscous flow simulations

Abstract

Computational work in continuum hypersonics and, in particular, the difficulties encountered during the development of the methods at the DLR Institute for Theoretical Fluid Mechanics are presented. Finite-difference schemes with implicit-explicit central discretisation and shock fitting and with upwind-TVD formulation and shock capturing are discussed. c.g., for strong shock--shock interactions. 3D flows of perfect gas at large angle of attack are considered, including those with radiation-adiabatic surface boundary conditions. 2D or axisymmetric as well as 3D tlows in thermochemical equilibrium are simulated using either the Tannehill curve-fitting approach for air or the minimization procedure for Gibbs' free energy which is more general at the expense of larger cost. Nonequilibrium tlow simulations are discussed for axisymmetric nitrogen and air model tlows. With the help of an appropriate model equation for finite-rate chemistry it is shown that the usual Harten-Yee approach does not allow to propagate shock waves in a time-accurate fashion if large source terms are present.

Separation length in high-enthalpy shock/boundary-layer interaction

“…45 When there is negligible dissociation upstream of reattachment ͑see Furumoto et al 45 or Grasso and Leone 17 ͒, the decrease in L sep may be due to lower pressure in region 3 caused by dissociation behind the reattachment shock. An explanation is not so obvious for the results of Oswald et al, 43 Brenner et al, 42 or Kordulla et al, 44 all of whom compared frozen flow to equilibrium flow on the hyperboloid flare. To complicate matters, the comparison by Oswald et al 43 was between an equilibrium flight condition at M ϱ ϭ25 and a perfect-gas wind tunnel condition at M ϱ ϭ10.…”

“…The result that separation length decreases for reacting flows relative to frozen flows was found in a number of computational studies on compression-corner flows, 17 axisymmetric hyperboloid-flare flows, [42][43][44] and shockimpingement flows. 45 When there is negligible dissociation upstream of reattachment ͑see Furumoto et al 45 or Grasso and Leone 17 ͒, the decrease in L sep may be due to lower pressure in region 3 caused by dissociation behind the reattachment shock.…”