Session F08: Compressible Flow: Shock-Boundary Layer Interactions (3:55pm - 4:40pm CST)

Effect of Temperature on Turbulent Shear-Layer and Shockwave Interaction

The boundary-layer separation and subsequent reattachment due to the free shear-layer and shockwave interaction have a large impact on the aerothermal design of supersonic aerospace systems. This problem is prevalent in high-speed flights and can significantly affect the skin friction, aerodynamics loads, and heat transfer. In recent years considerable progress has been achieved in the prediction of reattaching free shear-layer for compressible turbulent flows using DNS and LES. However, not considerable DNS or LES results for surface heat transfer in shockwave-turbulent boundary layer interaction are available. This is particularly important since prior RANS simulations of strongly separated shockwave-turbulent boundary layer interactions have failed to predict heat transfer accurately. In this study, the effect of heat transfer on the turbulent shear-layer and shockwave interactions in a scramjet has been investigated. To this end, Large Eddy Simulations are performed to explore the effect of wall thermal conditions on the behavior of a reattaching free shear layer interacting with and oblique shock in compressible turbulent flows. Different cases of wall to recovery temperature ratios are performed, and results are compared to the adiabatic wall.   

Large eddy simulation of intersection shock-wave/ turbulent boundary layer interactions in hypersonic flow regimes

For high-speed vehicles, accurate surface heat flux prediction is paramount for designing effective thermal protection systems. In this work, hypersonic flow over the double-fin geometry (Kussoy and Horstman, 1992) characterized by three-dimensional intersecting shock-waves/turbulent boundary-layer interaction at Mach 8.3 is numerically simulated using wall modeled large eddy simulation (WMLES) in the charLES code. This geometry is meant to be representative of an inlet for a hypersonic, air-breathing flight vehicle. The flow is characterized by aggressive pressure gradients, intersecting shock waves and strong separation. We use a standard equilibrium eddy viscosity wall model modified with semi-local scaling (Huang, JFM 305), which we demonstrate to yield more accurate comparisons with the experimental measurements. Both wall pressure and heat flux profiles have strong non-monotonic behaviors that challenge RANS turbulence models. However, the agreement of WMLES with the experimental data is remarkably good quantitatively, and qualitatively in terms of mean flow structures.   

Numerical investigation of shock-turbulent boundary layer interaction on flexible panels using wall modelled large eddy simulations

The effect of wall elasticity on shock-turbulent boundary layer interaction (STBLI) is investigated through the use of a coupled fluid structure interaction (FSI) solver. The FSI solver incorporates a wall-modeled large eddy simulation (WMLES) finite-volume flow solver, a geometrically non linear finite-element solid mechanics solver with damping, and a finite element flow mesh deformation solver. An Arbitrary Lagrangian-Eulerian (ALE) approach is employed to account for mesh motion and deformation in the flow domain. For sufficiently strong flow separation, unsteady low frequency motions of the shock separation bubble are present. Numerical simulations are compared to three different experimental flow configurations that vary the Mach number and flow deflection angle. Wall pressure, panel displacement, spanwise normal center-plane snapshots, and three dimensional shock bubble geometry are recorded over time. Notable differences from the rigid flow configuration in the wall pressure power spectral density and shock bubble statistics are observed for cases in which the maximum panel deflection is roughly equal in magnitude to the boundary layer thickness.   

Investigation of Dynamic Shock-Vortex Interactions in Compressible Low Reynolds Number Flows

As the Martian environment combines very low atmospheric density with a lower than Earth speed of sound, rotorcrafts such as NASA's \textit{Ingenuity }craft operate in a unique compressible low Reynolds Number (Re) flow regime. A preliminary study of dynamic stall using forced oscillation tests of a NACA 5605 airfoil in the aforementioned flow regime has showcased a new phenomenon called dynamic shock-vortex interactions. At a Re number, Mach number, and reduced frequency range of \textasciitilde 16000, .55, and .05 - .1 respectively, it was observed that during the pitch up motion, shock and vortex formations at the leading edge occur simultaneously. The resulting flow field showcases a shock-vortex interaction where the leading edge shocks are displaced from the airfoil surface as a result of the leading edge vortex formations. Additionally, these leading edge shocks are dynamically decomposed across the leading edge vortices and advected downstream. Based on these findings a complete evaluation of this novel shock-vortex interaction will be studied using higher fidelity CFD modeling on a 3D propeller mesh. Characterization of these shock-vortex interactions in terms of flow field parameters and aerodynamic load dynamics will be studied.   

Linear Instability Mechanisms of Supersonic Flow Over a Hollow Cylinder Flare Model

Instability of compressible laminar and transitional flow over an axisymmetric hollow cylinder flare configuration is addressed within the framework of linear BiGlobal analysis. The laminar basic state is topologically identical to spanwise homogeneous planar compression ramp flow and is obtained using high-resolution axisymmetric laminar direct numerical simulations. The inflow base flow profiles are provided by solution of the compressible axisymmetric boundary layer equations, such that the leading-edge shock is naturally excluded from the simulation domain. The recompression shock-system is fully resolved in the basic flow and included in the stability analysis. The global (modal) eigenvalue problem and the (non-modal/transient growth) singular value decomposition are solved using the LiGHT solver. Results obtained point to instability mechanisms analogous to those discovered in the planar compression ramp flow, namely two- and three-dimensional self-excitation of the laminar recirculation bubble, modified by the essential inclusion of the recompression shock. Comparisons with experimental results will be presented during the talk.   

On the spectral characteristics of separated flows in a spiked body at supersonic flow

A sharp-tip hemispherical spiked body experiences a form of unsteadiness due to separation caused by a shockwave boundary layer interaction in a supersonic flow (M$=$2). A series of computational studies using the Ansys-Fluent: Detached Eddy Simulation sheds light on understanding the spectral characteristics of the unsteadiness. The fluctuations intensity of the unsteady events like the shock motions, shedding of large-scale coherent structures from the separated shear layer, and charging/ejecting of fluid mass from the recirculation region, is found to be dependent on the separation point from the spike's leading edge and the separation angle. As the spike length is increased beyond a critical length, the recirculation region's length remains almost constant, whereas, the upstream boundary layer thickness increases. The large scale structures traveling along the separated shear layer carry shocklets, and parts of their feet interact with the spike-forebody wall around the recirculation region. The resulting gas dynamics enable the disturbances to bounce back and forth between the forebody wall and the separation point. The unsteady spectra obtained from the x-t plot of static pressure fluctuations along the spike and forebody wall for different upstream boundary layer thicknesses supplement the observation.   

Causality in the shock-wave/turbulent boundary layer interaction

The present work tackles the study of the unsteady behaviour of an impinging oblique shock and its interaction with a Mach 2 turbulent boundary layer. The investigation is made through the large-eddy simulation presented in Jiang \textit{et al.}, 2017, which has been extensively validated against experimental data. The main objective is to track the causes of the low-frequency fluctuations within the interaction zone, leading to the unsteadiness of the shock foot position. Evaluation of the spectrum in this region indicates that it is dominated by two-dimensional fluctuations which can be isolated via averaging in the spanwise direction. Data-driven approaches such as empirically derived transfer functions and spectral proper orthogonal decomposition (SPOD) are then performed in the averaged flow field. The results indicate the existence of a feedback mechanism between downstream fluctuations and the shock motion. Furthermore the leading SPOD mode comprises upstream travelling waves and enables the reconstruction of a significant portion of the energy of the shock motion from downstream measurements only. The current results indicate that downstream fluctuations are the driving mechanism behind the unsteady shock fluctuations.   

Turbulence behavior in supersonic channel flows with two- and three-dimensional sinusoidal roughness

Direct numerical simulations were performed to study supersonic turbulent channel flows over isothermal rough walls. The effect of roughness was incorporated as a body force in the momentum equations and a heat source in the energy equation, using an immersed boundary method. The rough surfaces included four sinusoidal geometries---two two-dimensional (2D) geometries (sine waves in the streamwise direction) and two three-dimensional (3D) geometries (sine waves in both streamwise and the spanwise directions). The surfaces shared the same roughness height but differed in their wavelengths. Simulations were carried out at bulk Reynolds number of $\textup{Re}=3000$ and Mach number of $\textup{Ma}=1.5$. Comparison of flow statistics shows a strong dependence of mean flow properties, turbulence anisotropy, and Reynolds stress budgets on the roughness wavelengths. Results also revealed major differences in the shock patterns induced by 2D and 3D roughness geometries, the effects of which were propagated over the entire channel and modified coherent turbulent motions.   

Schlieren Imaging of a Shock-Boundary Layer Interaction Over a Porous Plate

Schlieren imaging was performed to obtain qualitative measurements of a shock wave-boundary layer interaction (SBLI) generated by a 2D impinging shock over a porous surface. The experiments were conducted for the SBLI over a flat plate with and without a porous insert. The compression ramp was designed to have a turn angle of 11.5\textdegree in Mach 2 flow to generate a 43\textdegree impinging shock which interacts with the boundary layer on the tunnel floor. The experiment serves as an initial investigation into how the presence of a porous surface influences the interaction strength and boundary layer separation in a qualitative sense. Past works have primarily considered the effects of porous surfaces in the form of ordered holes, slots, and devices such as meso-flaps, but a limited number of experiments have considered randomly oriented porous media.   

Impact of Freestream Tunnel Noise on the Mach 6 Flow Interacting with a 35 Degree Compression Ramp

Conventional blow down wind tunnels generate acoustic fields from their boundary layers that impinge on the test article. At supersonic speeds, the model’s bow shock distorts the sound and generates vortical and entropy waves that also impact the model. These disturbances affect the model’s boundary layers and alter their transition and separation, subsequently affecting global properties of the flowfield. We study this process for a Mach 6 flow approaching a 35 degree compression ramp mounted on a flat plate with a sharp leading edge using direct numerical simulation. The conditions and model geometry match experiments conducted in the NASA Langley 20-inch Mach 6 tunnel. The incoming sound field is constructed from tunnel characterization data and included in the DNS. Comparisons of the flat plate--compression ramp flows are made between the quiet and sound-laden freestreams.