Bulletin of the American Physical Society
2005 58th Annual Meeting of the Division of Fluid Dynamics
Sunday–Tuesday, November 20–22, 2005; Chicago, IL
Session NP: Turbulence Simulations: DNS and LES |
Hide Abstracts |
Chair: Philip Marcus, University of California, Berkeley Room: Hilton Chicago Stevens 1 |
Tuesday, November 22, 2005 11:01AM - 11:14AM |
NP.00001: Vortex sheet identification method and fourth-order velocity derivative invariants Kiyosi Horiuti, Yohei Takagi, Takeharu Fujisawa We considered a new identification method for vortex sheet structures in turbulent flows using the eigenvalue $[A_{ij}]_{+} $ of the tensor $A_{ij}\equiv(S_{ik}\Omega_{kj}+S_{jk}\Omega_ {ki})$, where $S_{ij}$ is the strain-rate tensor, $\Omega_{ij}$ the vorticity tensor. Effectiveness of the proposed method was verified in the assessment using DNS data for homogeneous isotropic turbulence. The region identified using the proposed method exhibited a characteristic feature consistent with that of the vortex sheet, i.e., comparable dominance of both strain rate and vorticity. As a result, intense dissipation took place in the educed region. The fractal dimension, $D$, of the individual structures with large amplitudes of $[A_{ij}]_{+}$ was $D \simeq 1.7$, with very little dependence on the threshold. This value indeed suggests structures in the form of vortex sheet, and was close to the fractal dimension of intense dissipation structures. The vortex sheets were spanned approximately by the two eigenvectors corresponding to other eigenvalues of $A_{ij}$, $[A_{ij}]_z$ and $[A_{ij}]_-$. Relationship between the eigenvalue $[A_{ij}]_{+}$ and the four invariants of fourth-order moments of velocity derivatives, $I_i$ (Siggia 1981), was examined. It was shown that $[A_{ij}]_ {+} \simeq \sqrt{\frac{1}{2}A_{ij}A_{ji}}=\sqrt{\frac{1}{2}(I_2- \frac{3}{2}I_3)}$. DNS data showed that the invariant, $A_{ij}A_ {ji}$, deviated markedly from the Gaussian value, indicating that both the strain rate and vorticity in the region educed using $[A_{ij}]_{+}$ were strongly correlated, and both were highly intermittent. [Preview Abstract] |
Tuesday, November 22, 2005 11:14AM - 11:27AM |
NP.00002: The effect of the Reynolds number on mass transfer at a free surface in a fully developed turbulence Ryuichi Nagaosa This study deals with mass transfer mechanism into a turbulent liquid at a free surface in an open channel. Both mass flux and subsurface hydrodynamics measured in laboratory measurements and found that the normalized mass transfer coefficient is proportional to the Reynolds number ${\rm Re}_m$ which is defined by water depth and the bulk mean velocity [S. Komori, R. Nagaosa and Y. Murakami, AIChE J. {\bf36}, 957, 1991]. Direct numerical simulations (DNS) of mass transport at the free surface in a fully developed turbulence have been carried out in this study to discuss suitability of the results of the previous laboratory experiments. The results of this study show that the predicted mass transfer velocities by the DNS technique agree well with our previous laboratory measurements. The mass transfer velocities predicted in the present DNS are, however, proportional to 3/4 power of ${\rm Re}_m$, rather than 1 as found in the laboratory experiments. The difference of the exponent could be a reason of underestimation of mass flux in the numerical predictions in a larger Reynolds number turbulence of about ${\rm Re}_m>10,000$. [Preview Abstract] |
Tuesday, November 22, 2005 11:27AM - 11:40AM |
NP.00003: Turbulence suppression in a stably stratified boundary layer Byung-Gu Kim, Changhoon Lee Suppression of turbulence in stably stratified flow is investigated by using large eddy simulation of channel flow. Stably stratified turbulent flow owing to negative buoyancy or adverse density gradient is frequently observed phenomenon in geophysical flow as well as in many engineering flows. Total suppression or relaminarization was reported from many experiments, but it has not been reproduced in numerical simulations typically performed for low Reynolds number flow with low Richardson number. Here, suppression means decrease of turbulent stresses relative to the wall-shear velocity. In this study we by using large eddy simulation at relatively high Reynolds number investigate modification of near-wall turbulence due to strong stratification and relevant dynamics. Particularly, we focus on the role of the near-wall vortical structures against baroclinic torque generated by the stratification. Detailed statistics will be reported in the meeting. [Preview Abstract] |
Tuesday, November 22, 2005 11:40AM - 11:53AM |
NP.00004: Towards A Large Eddy Simulation of Supercritical CO$_2$ Pipe Flow: High-Reynolds-Flow Simulation Xiaohang Wang Studies on the numerical simulation of high-Reynolds-number flows encounter difficulties due to the wide range of characteristic length and time scales existing in the flow field, which are often smaller than the computational grid size. Besides, the near-wall region contains small vortical structures dynamically important to the flow, which have dimensions that scale with the viscous scale, making it impractical to resolve them in numerical simulations at very high Reynolds numbers. A thorough numerical investigation of high Reynolds number (Re$_{D}$=28666) supercritical CO$_{2}$ pipe flow was performed by using the large eddy simulation (LES), where a dynamic subgrid-scale model was used to account for the subgrid scale effects. The objective was to investigate the influence of subgrid scale modeling, filter size, domain size, grid resolution and initial condition on the quality of the predicted results. Certain resolution requirements in LES proposed by J. S. Baggett and J. Jimenez were studied. The simulation indicates that grid refinements in spanwise direction may cause the divergence, as the mean streak spacing is much smaller than 100 wall units, which is assumed to be consistent with near-wall streamwise vortices optimally configured to gain the most energy over an appropriate turbulent eddy turnover time. [Preview Abstract] |
Tuesday, November 22, 2005 11:53AM - 12:06PM |
NP.00005: Dispersion of a passive scalar around Gaussian hills in the atmospheric boundary layer simulated with LES and stochastic modeling. Serge Simoens, Ivana Vinkovic, Cesar Aguirre Because of environmental concerns, considerable attention has been given to the prediction of concentration levels downwind of polluting sources. A large Eddy-Simulation (LES) with a Smagorinski/Germano subgrid scale (SGS) model was used to simulate the velocity field of an atmospheric boundary layer (ABL) with Gaussian hills at the surface. The LES was combined with a Lagrangian stochastic model in order to simulate dispersion of a passive scalar released at various locations with respect to the hills. The first case studied is used to validate both the computation of the velocity and concentration fields. The second case shows the behavior of fluid particles carrying the scalar around the relief of the hill. Vertical and horizontal profiles of velocity and concentration are shown upwind and downwind of the hills. The recirculating zone downwind of the hill is of particular interest. The potential of such an approach for more complex simulations of real pollution sites will be discussed. [Preview Abstract] |
Tuesday, November 22, 2005 12:06PM - 12:19PM |
NP.00006: Profound Effects of Boundaries and Weak 3D Motions on Nearly 2D Turbulence Philip Marcus, Chung-Hsiaing Jiang Over the last 30 years, there have been many numerical simulations of 2D turbulent flows, most of which have used doubly-periodic spatial boundary conditions. One of the common goals of these simulations has been to reproduce and understand the inverse energy cascade to large scales. The flows are either run-down or forced (often at the small scales and dissipated at the large). Often the calculations include a beta plane (to mimic geophysical flows) which breaks symmetry and allows for the formation of large-scale zonal flows. Our interest is the understanding of the atmospheres of the large planets, which have both large, long-lived zonal flows {\it and} large, long- lived vortices. Flows in rotating laboratory tanks designed to simulate the atmospheres also produce vortices and zonal flows. However, 2D, forced simulations on a beta plane that are initialized with weak noise produce long-lived large zonal flows {\it or} long-lived, large vortices but not both simultaneously. We find that by including boundaries and/or by including weak 3D motions we can create both large, long-lived zonal flows and vortices. Here we examine why this is so and what the implications are for the modeling and simulating of real 3D flows with 2D equations of motion. [Preview Abstract] |
Tuesday, November 22, 2005 12:19PM - 12:32PM |
NP.00007: Drag Reduction with Blowing and Suction from a Pair of Oblique Slots on the Wall Joon Ahn, Kaoru Iwamoto, Koji Fukagata, Nobuhide Kasagi We introduce a pair of oblique blowing and suction of zero net-mass-flux on the wall to reduce the drag by inducing negative Reynolds stress. It is applied to the channel flow at the bulk Reynolds numbers of 300 (laminar) and 2,800 (turbulent). Blowing and suction are made on one wall at an inclination angle set at 10 degrees. The magnitude of blowing and suction is the half of bulk velocity of the main flow. We also simulate the flows inside the blowing and suction slots by employing an immersed boundary method. The laminar flow results show that sub-laminar drag should be achievable. The pumping power to drive a constant mass flow rate is reduced by 9.8{\%} for the turbulent flow, although the blowing/suction is applied to only 10{\%} surface area of one wall. Detailed analysis including the energy balance will be discussed in the presentation. [Preview Abstract] |
Tuesday, November 22, 2005 12:32PM - 12:45PM |
NP.00008: Production of turbulent energy flux in inertial-range cascades Zuoli Xiao, Shiyi Chen, Minping Wan, Gregory Eyink We study the exact dynamical equation for the production of turbulent energy flux in the inertial-range, obtained by spatial filtering of the Navier-Stokes equation (Germano, 1992). We report here on the results from direct numerical simulations of both the 2D inverse-energy cascade and the 3D forward-energy cascade. Conditional averages of the various production and destruction terms in the equation are calculated, given the values of the energy flux at later times in a Lagrangian frame, using a backward particle-tracking method. In the 3D forward cascade, the dominant term is the production of subscale stress by the large-scale strain. However, in the 2D inverse cascade the dominant term is the production of subscale stress by the pressure-strain correlation. The physical mechanisms associated with these various terms will be discussed. [Preview Abstract] |
Tuesday, November 22, 2005 12:45PM - 12:58PM |
NP.00009: Towards the Study of Flow through Low-Pressure Turbine Cascade by Large-Eddy Simulation Technique Shirdish Poondru, Karman Ghia, Urmila Ghia This study aims to accurately predict and control the flow separation that occurs on the suction side of a LPT cascade, by application of higher-order compact-difference scheme to Large Eddy Simulation (LES) Technique. To achieve this goal, a MPI-based higher-order, Chimera version of the FDL3DI flow solver developed by the Air Force Research Laboratory at Wright Patterson Air Force Base is used. FDL3DI solves the full 3-D Navier-Stokes equations using the LES technique. To understand the LES module of the flow solver, two test cases are solved first: Flow through a channel, and flow past a circular cylinder. A multi-block structured grid is employed for all cases. The solutions obtained compare well with the available results for these flow problems. The study is proceeding to obtain a solution for flow through LPT cascade using the LES technique with dynamic sub-grid scale model, followed by implementation of a flow-control strategy. A parametric study of effect of grid density and location of upstream inflow boundary will also be studied. [Preview Abstract] |
Tuesday, November 22, 2005 12:58PM - 1:11PM |
NP.00010: Flexible Inhibitor Fluid-Structure Interaction Simulation in RSRM. Bono Wasistho, Ali Namazifard, Robert Fiedler We employ our tightly coupled fluid/structure/combustion simulation code 'Rocstar-3' for solid propellant rocket motors to study 3D flows past rigid and flexible inhibitors in the Reusable Solid Rocket Motor (RSRM). We perform high resolution simulations of a section of the rocket near the center joint slot at 100 seconds after ignition, using inflow conditions based on less detailed 3D simulations of the full RSRM. Our simulations include both inviscid and turbulent flows (using LES dynamic subgrid-scale model), and explore the interaction between the inhibitor and the resulting fluid flow. The response of the solid components is computed by an implicit finite element solver. The internal mesh motion scheme in our block-structured fluid solver enables our code to handle significant changes in geometry. We compute turbulent statistics and determine the compound instabilities originated from the natural hydrodynamic instabilities and the inhibitor motion. The ultimate goal is to studdy the effect of inhibitor flexing on the turbulent field. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700