Towards a Local Model for Crossflow Transition in Hypersonic Boundary-Layers

The prediction of laminar-turbulent transition is of crucial importance for the design of hypersonic flight vehicles. Past research has shown that a multitude of transition scenarios are possible depending on the flow conditions and the vehicle geometry. For hypersonic, axisymmetric, slender vehicles and zero angle of attack, transition is often dominated by second-mode waves. Changing the angle of attack or the geometry cross-section (e.g. el-liptic shape) can, however, lead to dominance of crossflow instability. Therefore, both mechanisms have to be incorporated into newly deployed hypersonic transition models for Reynolds-averaged Navier-Stokes codes. This paper discusses the extension of a transi-tion model based on second-mode amplification factor transport equation to also account for crossflow instability. The transition model production term is modeled with a neural network. Linear stability theory investigations for 40 different Falkner-Skan-Cooke base flows with swept freestream provide amplification factor curves for network training. The freestream conditions were based on the Purdue Mach 6 quiet tunnel (BAM6QT) condi-tions and the pressure gradient and sweep angle considered are inspired by the flow-field on the HiFIRE-5 elliptical cone. To identify useful analytical relationships for transition model development, relevant terms in the amplification factor transport equation are plot-ted versus local quantities and non-local quantities such as the edge Mach number.