Analysis of Navier–Stokes Codes Applied to Supersonic Retropropulsion Wind-Tunnel Test

Abstract

Advancement of supersonic retropropulsion as a technology will rely heavily on the ability of computational methods to accurately predict vehicle aerodynamics during atmospheric descent, where supersonic retropropulsion will be employed. A wind-tunnel test at the NASA Langley Unitary Plan Wind Tunnel was specifically designed to aid in the support of Navier-Stokes codes for supersonic retropropulsion applications. Three computational fluid dynamics codes [data parallel line relaxation, fully unstructured Navier-Stokes three-dimensional, and overset grid flow solver] were exercised for multiple nozzle configurations for a range of freestream Mach numbers and nozzle thrust coefficients. The computational fluid dynamics pretest analysis of this wind-tunnel test aided in the test model design process by identifying the potential for tunnel blockage or unstart, of liquefaction within the plume, and of separation occurring at the internal fingers of the nozzles. This analysis led to a reduced model diameter, heating of the plenum, and reducing the nozzle area ratio, and the requirement to radius the corners at the fingers, to counter these potentials, respectively. Comparisons to test data were used to determine the existing capability of the codes to accurately model this complex flow, identify modeling shortcomings, and gain insight into the computational requirements necessary for correctly computing these flows. All three codes predict similar surface pressure coefficients and flowfield structures, such as jet termination shock, interface, bow shocks, and recirculation regions. However, the codes differ on the level of unsteadiness predicted.

Similarity and Key Parameters of Retropropulsion Assisted Deceleration in Hypersonic Wind Tunnels

“…An implicit backward Euler time-stepping scheme was employed to increase convergence efficiency. The two-equation turbulence model Menter's Shear Stress Transport (SST) was selected, as it was widely used in previous SRP simulations for other CFD codes [8,11,[17][18][19]. All presented CFD simulations were performed as steady-state computations for calorically perfect gas, as steady RANS computations compared favorably to experimental data in a study of SRP flows [19].…”