Analysis of Heat Flux Prediction Capabilities Across Multi-Fidelity Simulations of Hypersonic Inflatable Aerodynamic Decelerators

We conduct Reynolds-averaged Navier-Stokes (RANS), wall-modeled large-eddy simulations (WMLES), and under-resolved large-eddy simulations (URLES) of a Mach 5.96 atmospheric reentry vehicle with a statically deformed isothermal aeroshell. The study focuses on comparing the ability of RANS (with a $k$-$\omega$-$\gamma$ model), WMLES (with an equilibrium wall-stress model), and URLES (with a no-slip, isothermal wall boundary condition) to predict wall heat flux. Mesh sensitivity analysis applied to the RANS and WMLES cases reveal that the wall heat flux predictions are noticeably sensitive not only to the mesh resolution but also to the mesh misalignment with the bow shock. The latter is particularly relevant in the wall heat flux prediction near the stagnation point and in laminar regions of the flow. In turbulent flow regions, downstream of boundary layer transition, both RANS and WMLES provide accurate prediction of wall heat flux peaks, but underpredict the heat flux in regions leading to those peak values. The effects of varying the wall-model exchange height on the wall heat flux prediction are explored, finding that decreasing the wall-model exchange height to the first cell off the wall provides accurate predictions for wall heat flux in separated flow regions, while maintaining the 10\% boundary layer exchange height provides a subpar separation wall peak heat flux prediction, but a superior reattachment wall heat flux prediction. The accuracy of the $k$-$\omega$-$\gamma$ RANS model in predicting the boundary layer transition location and wall heat flux is also assessed.