The Gallery of Turbulent Flows

We compute turbulent combustion and reacting flows in complex engineering environments including scramjets, jet engines and pool fires. We develop methods to reduce jet noise and predict aeroacoustics of high-speed flows. We develop methods to compute two-phase flows in realistic engineering geometries. We develop methods to predict near-wall turbulence fluctuations to enable large-eddy simulation of complex engineering geometries. Breakup of a liquid jet into droplets showing the formation of ligaments (by D. Kim, 2010). The rectangular nozzle is shown in gray with an isosurface of temperature (gold) cut along the center plane of the nozzle showing temperature contours (red/yellow). The acoustic field is visualized by (blue/cyan) contours of the pressure field taken along the same plane. The chevrons enhance turbulent mixing just downstream of the nozzle exit shaping how the jet spreads downstream. This significantly reduces the noise produced by the supersonic jet compared to a rectangular nozzle without chevrons (not shown). Joseph Nichols, Center for Turbulence Research DNS of zero-pressure-gradient flat-plate boundary layer (ZPGFPBL) by Xiaohua Wu and Parviz Moin. Image taken from DNS of the ZPGFPBL, which develops spatially from Re_theta = 80 at x=0.1 to Re_theta=1000 at x=3.5. Grid size is 4096 points in x, 400 in y, and 128 in z. Rayleigh-Taylor instabilities on the surface of a blast wave. See http://shocks.stanford.edu for more information. Large-Eddy simulation of an oblique shock/turbulent boundary layer interaction (by B.E. Morgan, 2012). Near-wall structures in boundary layer transition (isosurfaces of Q-criterion colored by velocity magnitude). See also Multi-element airfoil aerodynamics and noise reduction by J. Bodart, 2011. Direct numerical simulation of turbulent combustion of a swirl burner in Proceedings of the 2010 CTR Summer Program. Fires and buoyant plumes. Aeroacoustics of rectangle chevron nozzles by J. Nichols, 2010. Integrated hybrid RANS/LES of a realistic gas turbine engine. Hydrodynamics and nutrient transport in coral reefs. Color shades of the skin friction in the Direct Numerical Simulations of fully developed turbulent channel flow. The simulations were performed by John Kim, Parviz Moin and Robert Moser (Journal of Fluid Mechanics, v 177, pp 136-166, 1987). The plot is produced by Arthur Kravchenko. Direct numerical simulation of transition to turbulence. Large Eddy Simulations of Combustor Flows Knut Akselvoll, Charles Pierce and Parviz Moin Color shades of the instantaneous streamwise velocity in Large Eddy Simulations of turbulent flow in plane diffuser. The simulations were performed by Massimiliano Fatica, Rajat Mittal, Hans Kaltenbach and Parviz Moin. Computation of Flow in a Jet Engine Fuel Nozzle X. Wu, G. Iaccarino, P. Moin and F. Ham Flow over a Backward-Facing Step (Spanwise Vorticity) Xiaohua Wu, George Homsy and Parviz Moin Direct numerical simulation of a zero-pressure-gradient flat-plate boundary layer in Wu and Moin, Journal of Fluid Mechanics (2009). Droplet evaporation in a temporal shear layer (Re=200). Droplet position superimposed on mixture fraction contours. 5 million hexane drops are simulated in this direct numerical simulation by Olivier Desjardins. Simulation of evaporating isopropyl alcohol in a coaxial combustor corresponding to the experimental setup of Sommerfeld and Qiu (1998) is performed. Hot air enters the chamber through the annulus, while isopropyl alcohol droplets are injected through the central region. Inflow conditions for spray are based on measured size-velocity correlations at the inlet section. The predictions of mass-flow rates, droplet-size distributions, mean and rms velocities for both phases are in good agreement with the experimental data. See Mahesh, K., Constantinescu, G., Apte, S.V., Moin, P., 'LES of gas-turbine combustors,' Annual Research Briefs, Center for Turbulence Research, 2001. Langrangian particle paths as they enter a standing shock in the isotropic shock-turbulence problem of Larsson et al., CTR Annual Research Briefs 2007. Here, we extended the point-particle approximation by accounting for the finite-size of the dispersed phase. We validate this model by simulating capture Poisuille flow with solid particles arranged at the bottom of the channel (DNS by Choi & Joseph, JFM 2000). The standard point-particle approach does not predict any lift force on the particles and the particle layer moves in laminar layers. The finite-size model leads to Kelvin-Helmholtz type instability in the gas-phase velocity and lift of particles similar to the DNS. See Apte, S.V., Mahesh, K., & Lundgren, T., 'A Eulerian-Lagrangian model to simulate two-phase/particulate flows,' Annual Research Briefs, Center for Turbulence Research, Stanford, 2003. Snapshot of shock/turbulence interaction at M=2, Mt=0.15, Re_l = 40. The flow is from left to right, with the shock visualized by transparent isosurfaces of compression. Vortex cores are visualized by isosurfaces of the second invariant of the velocity gradient tensor, colored by the vorticity magnitude. See Larsson et al., CTR Annual Research Briefs 2007. The picture shows a Direct Numerical Simulation of a Mach 1.92 jet. The jet turbulence is visualized with contours of vorticity magnitude and the near acoustic field is visualized with gray levels indicating divergence of velocity, which highlights the weak shocks (dark) and relatively broad expansions (light). See J. B. Freund, S. K. Lele, and P. Moin. 'Direct numerical simulation of a Mach 1.92 turbulent jet and its sound field,' AIAA Journal, 38 (11), 2023--2031, (2000). A multi-scale, multi-physics simulation of turbulent reacting flow in a realistic Pratt & Whitney combustor is performed. This simulation includes all the complex models for spray breakup, evaporation, and turbulent combustion and represents the first study of the reacting multiphase flow in complex combustor geometry using CDP. The simulation is performed for a single injector, which represents one sector of the full combustor containing 18 injectors. Liquid fuel (Jet-A) enters the combustion chamber through an annular ring at the injector exit. This liquid film is approximated by large drops of the size of the annulus radius. The drops are convected by the surrounding hot air, they break, evaporate, and the fuel vapor thus formed mixes with the surrounding air giving a non-premixed spray flame. Overall comparisons of mass-flow splits, exit temperature profiles are in agreement with the experimental data. See Ham, F., Apte, S.V., Iaccarino, G., Constantinescu, G., Mahesh, K., Moin, P., 'Unstructured LES of reacting multiphase flows in realistic gas-turbine combustors,' Annual Research Briefs, Center for Turbulence Research, 2003. Multi-material mixing in the Stanford Scidac project, see http://shocks.stanford.edu for more information. DNS of zero-pressure-gradient flat-plate boundary layer (ZPGFPBL) by Xiaohua Wu and Parviz Moin. Image taken from DNS of the ZPGFPBL, which develops spatially from Re_theta = 80 at x=0.1 to Re_theta=1000 at x=3.5. Grid size is 4096 points in x, 400 in y, and 128 in z. DNS of zero-pressure-gradient flat-plate boundary layer (ZPGFPBL) by Xiaohua Wu and Parviz Moin. Image taken from DNS of the ZPGFPBL, which develops spatially from Re_theta = 80 at x=0.1 to Re_theta=1000 at x=3.5. Grid size is 4096 points in x, 400 in y, and 128 in z. DNS of zero-pressure-gradient flat-plate boundary layer (ZPGFPBL) by Xiaohua Wu and Parviz Moin. Image taken from DNS of the ZPGFPBL, which develops spatially from Re_theta = 80 at x=0.1 to Re_theta=1000 at x=3.5. Grid size is 4096 points in x, 400 in y, and 128 in z. DNS of zero-pressure-gradient flat-plate boundary layer (ZPGFPBL) by Xiaohua Wu and Parviz Moin. Image taken from DNS of the ZPGFPBL, which develops spatially from Re_theta = 80 at x=0.1 to Re_theta=1000 at x=3.5. Grid size is 4096 points in x, 400 in y, and 128 in z. DNS of zero-pressure-gradient flat-plate boundary layer (ZPGFPBL) by Xiaohua Wu and Parviz Moin. Image taken from DNS of the ZPGFPBL, which develops spatially from Re_theta = 80 at x=0.1 to Re_theta=1000 at x=3.5. Grid size is 4096 points in x, 400 in y, and 128 in z. DNS of zero-pressure-gradient flat-plate boundary layer (ZPGFPBL) by Xiaohua Wu and Parviz Moin. Image taken from DNS of the ZPGFPBL, which develops spatially from Re_theta = 80 at x=0.1 to Re_theta=1000 at x=3.5. Grid size is 4096 points in x, 400 in y, and 128 in z. We performed LES of particle-laden, swirling flow in a coaxial-jet combustor. A mixture of air and lightly loaded, spherical glass beads with a prescribed size-distribution enters the primary jet, while a swirling stream of air flows through the annulus. Particles are treated as point sources. The particle-dispersion characteristics are examined in detail; in particular the dependence of particle trajectories and residence times upon particle sizes is emphasized. The predictions of mean and RMS velocity components for the gaseous and dispersed phases are in good agreement with the experimental data of Sommerfeld and Qiu (1991). This was the first LES prediction of swirling, particle-laden flows in coaxial geometries using an unstructured grid solver (CDP). See Apte, S.V., Mahesh, K., Moin, P., & Oefelein, J.C., 'Large-eddy simulation of swirling particle-laden flow in a coaxial combustor,' International Journal of Multiphase Flow, Vol. 29, pp. 1311-1331, 2003. A stochastic subgrid model for LES of atomizing spray is developed. The model is based on Kolmogorov’s concept of viewing solid particle-breakup as a discrete random process. Atomization of liquid blobs at high relative liquid-to-gas velocity is considered in the framework of uncorrelated breakup events, independent of the initial droplet size. Kolmogorov’s discrete model is rewritten in the form a Fokker- Planck equation for the pdf of droplet radii. See Apte, S.V., Gorokhovski, M., & Moin, P., “LES of atomizing spray with stochastic modeling of secondary breakup, Integrated RANS-LES of the NASA stage 35 compressor and a diffuser: the compressor stage is computed with the RANS flow solver TFLO, the subsequent diffuser with the LES flow solver CDP. Here, we look at the vorticity magnitude distribution at the 50% clip plane of the stator. Again, we can identify the wakes of the stator passing the interface. The different description of turbulence in the two mathematical approaches is apparent. While on the RANS side the turbulence is modeled in a turbulence model and cannot be seen in the vorticity distribution, on the LES side the fine scale turbulence is regenerated and can be identified as small-scale structures in the LES solution. See Integrated RANS-LES Computations in Gas Turbines: Compressor-Diffuser, J. U. Schlüter, X. Wu, S. Kim, J. J. Alonso, and H. Pitsch, AIAA-2004-0369, 42nd Aerospace Sciences Meeting and Exhibit Conference, January 2004. Center for Integrated Turbulence Simulations at Stanford University. Comprehensive simulation of the flow through an entire jet engine, done in collaboration with Pratt & Whitney. Contours of entropy in the high-pressure compressor and in the first two stages of the turbine, as well as contours of temperature in the combustor of a Pratt & Whitney engine.