Large-eddy Reynolds-averaged Navier-Stokes Research Articles

Large-eddy/Reynolds-averaged Navier–Stokes simulation of cavity-stabilized ethylene combustion

In this study, a hybrid large-eddy/Reynolds-averaged Navier–Stokes (LES/RANS) method is used to simulate ethylene combustion inside a cavity flameholder. The cavity flameholder considered is Configuration E of University of Virginia’s Scramjet Combustion Facility, which consists of a Mach 2 inlet nozzle, a constant-area isolator, a combustor, and an extender, through which the exhaust gases are vented to the atmosphere. To increase the fuel-residence time, a cavity is fitted along the upper wall inside the combustor section of the flameholder. The configuration has the capability of injecting ethylene through a series of ports located upstream of and inside the cavity along the upper wall the combustor. In the simulations, ethylene combustion is modeled using a 22-species ethylene oxidation mechanism. Also, a synthetic eddy method is used to introduce turbulence at the inflow plane of the flameholder. For an equivalence ratio of 0.15, a cavity stabilized flame is predicted. Predictions are compared with line-of-sight temperature, water column-density, water mole-fraction, CO column-density, and CO2 column-density measurements at three stations within and downstream of the cavity. Agreement with experiment is generally good within the cavity. Downstream of the cavity, the simulations predict higher temperatures near the wall. Analysis of the flame structure predicted by the LES/RANS method indicates that the flame propagates into a stoichiometric to fuel-rich mixture near the cavity. Flame angles captured in the simulation are in close agreement with those predicted through classical premixed turbulent flame-speed estimates. Further downstream, the flame structure is non-premixed in character, and near complete conversion of CO to CO2 is observed by the time the flame reaches the combustor exit.

Numerical Simulation of the Effects of Mesoflaps in Controlling Shock/Boundary-Layer Interactions

This work utilizes an immersed-boundary method to simulate the effects of an array of aeroelastic mesoflaps in controlling oblique shock/turbulent boundary-layer interactions. A loosely coupled approach is adopted for the fluid-structure interaction problem, with separate solvers used for the fluid and structural domains. The mesoflaps are rendered as immersed objects for the fluid solver and modeled as cantilevered Euler-Bernoulli beams for the structural solver. Determination of the aerodynamic loads acting on the mesoflap surfaces from the surrounding fluid is done using a nearest neighbor approach combined with a bi-linear interpolation scheme. Simulations are performed for a Mach 2.46 shock / boundary layer interaction with and without control, based on experiments conducted at University of Illinois at Urbana-Champaign. Both Reynolds-averaged Navier-Stokes (RANS) and hybrid large-eddy/Reynolds-averaged Navier-Stokes (LES/RANS) turbulence closures are used. For the computations of the flow with mesoflap control, both 2-D quasi-steady and 3-D dynamic simulations of the fluid-structure interaction problem are performed. Comparisons made with experimental laser Doppler anemometry data and wall pressure measurements for flows with and without control show reasonable agreement, with better predictions away from the separation region. An analysis of the flow indicates that the mesoflap control system does not eliminate axial flow separation. Also, analysis of the frequency content of the mesoflap deflections suggests that a correlation might exist between the dominant frequency of the flap deflection and the low-frequency shock motion observed in separated flows.

Large-eddy/Reynolds-averaged Navier–Stokes simulation of a supersonic reacting wall jet

This work presents results from large-eddy/Reynolds-averaged Navier–Stokes (LES/RANS) simulations of the well-known Burrows–Kurkov supersonic reacting wall-jet experiment. Generally good agreement with experimental mole fraction, stagnation temperature, and Pitot pressure profiles is obtained for non-reactive mixing of the hydrogen jet with a non-vitiated air stream. A lifted flame, stabilized between 15 and 20 cm downstream of the hydrogen jet, is formed for hydrogen injected into a vitiated air stream. Flame stabilization occurs closer to the hydrogen injection location when a three-dimensional combustor geometry (with boundary layer development resolved on all walls) is considered. Volumetric expansion of the reactive shear layer is accompanied by the formation of large eddies which interact strongly with the reaction zone. Time averaged predictions of the reaction zone structure show an under-prediction of the peak water concentration and stagnation temperature, relative to experimental data, but display generally good agreement with the extent of the reaction zone. Reactive scalar scatter plots indicate that the flame exhibits a transition from a partially-premixed flame structure, characterized by intermittent heat release, to a diffusion-flame structure that could probably be described by a strained laminar flamelet model.

Simulation of Shock/Boundary-Layer Interactions with Bleed Using Immersed-Boundary Methods

This work utilizes an immersed boundary (IB) method to simulate the effects of arrays of discrete bleed ports in controlling shock wave / turbulent boundary layer inter actions . Both Reynolds averaged Navier -Stokes (RANS) and hybrid large -eddy / Reynolds -averaged Navier -Stokes (LES/RANS) turbulence closures are used with the IB technique. The approach is validated by conducting simulations of Mach 2.5 flow over a perfo rated plate containing 18 individual bleed holes. Predictions of discharge coefficient as a function of bleed plenum pressure are compared with experimental data. Simulations of an impinging oblique shock / boundary layer interaction at Mach 2.45 with an d without active bleed control are also performed. The 68 -hole bleed plate is rendered as an immersed object in the computational domain. Wall pressure predictions show that, in general, the LES/RANS technique under -estimate s the upstream extent of axi al separation that occurs in the absence of bleed. Good agreement with P itot -pressure surveys throughout the interaction region is obtained, however. Active suction completely removes the separation region and induces local disturbances in the wall pres sure distributions that are associated with the expansion of the boundary layer fluid into the bleed port and its subsequ ent re -compression. Predicted Pitot -pressure distributions are in good agreement with experiment for the case with bleed. Swirl stre ngth probability -density distributions are used to estimate the evolution of turbulence length -sca les throughout the interaction, and the effects of bleed on the amplification of Reynolds stresses are highlighted. Finally, simple improvements to engineerin g-level bleed models are proposed based on the computational results.