Bulletin of the American Physical Society
67th Annual Meeting of the APS Division of Fluid Dynamics
Volume 59, Number 20
Sunday–Tuesday, November 23–25, 2014; San Francisco, California
Session D28: Turbulence: LES Modeling |
Hide Abstracts |
Chair: Scott Murman, NASA Room: 2011 |
Sunday, November 23, 2014 2:15PM - 2:28PM |
D28.00001: New subgrid-scale model for large-eddy simulation of turbulent flows Changping Yu, Xinliang Li, Zuoli Xiao, Shiyi Chen Based on the theory of energy and helicity cascades, a new subgrid-scale (SGS) stress model is proposed for large eddy simulation of turbulent flows. The new SGS model is a type of eddy-viscosity model, and the eddy-viscosity is proportional to the magnitude of the mean product of the large-scale strain rate tensor and symmetric vorticity gradient tensor. The new SGS model is first tested \textit{a priori} and \textit{a posteriori} in homogeneous and isotropic helical turbulence, and the statistical results show that the new model can predict most of the results better than Smagorinsky model and the mixed helical model. Then, we apply the present model to simulate the channel flows, and also our model can support satisfied simulating results of mean velocity, turbulent stress and skin-friction coefficients, etc. The surprising findings is that the new model can describe much more realistic flow structures than DES-SA model and reproduce the skin friction force much more accurately than the Smagorinsky model. The new SGS model is proved to be universal model in large eddy simulation of turbulent flows. [Preview Abstract] |
Sunday, November 23, 2014 2:28PM - 2:41PM |
D28.00002: A dynamic sub-grid model for variational multiscale methods Scott Murman, Laslo Diosady, Anirban Garai The variational multiscale method uses an explicit {\em{a priori}} separation of scales, along with Galerkin projection as the filter operator, to develop an alternative to the classical LES approach.\footnote{Hughes {\em et al.} Comput. Visual Sci. Sci. \textbf{3}, 47 (2000)} A dynamic, parameter-free, multiscale extension of this approach is developed from the variational implementation of Germano's identity.\footnote{Oberai and Wanderer, J. of Turbulence \textbf{6}, 7 (2005)} The method is implemented in an entropy-stable Discontinuous-Galerkin spectral-element solver.\footnote{Diosady and Murman, AIAA 2014-2784} We outline the relevant details of the method using {\em a priori} testing, before demonstrating the performance in {\em a posteriori} testing on several canonical flows, including homogeneous isotropic turbulence, channel flow, and separated flow over an array of periodic hills. These computed results are compared against DNS, classical LES, and experimental data. [Preview Abstract] |
Sunday, November 23, 2014 2:41PM - 2:54PM |
D28.00003: Autonomic Closure for Large Eddy Simulations Ryan King, Werner Dahm, Peter Hamlington Motivated by the application of adjoint techniques for rapidly solving large optimization problems, a fundamentally new autonomic closure is presented that allows an essentially model-free, dynamic subgrid-scale closure for large eddy simulations (LES). The autonomic closure addresses nonlinear, nonlocal, and nonequilibrium turbulence effects and, in its most general form, is based on all possible tensorally-invariant, dimensionally-consistent relations between the local subgrid-stress tensor and resolved scale primitive variables. This introduces a large matrix of spatially and temporally varying coefficients that are optimized using a test filter approach and then applied at the LES filter scale by invoking scale similarity. The autonomic closure avoids the need to specify a model for the subgrid stresses, and instead allows the simulation by itself to determine the best local relation between the subgrid stresses and resolved state variables. A priori tests of this new autonomic closure approach are presented using data from direct numerical simulations of homogeneous isotropic and sheared turbulence, and application of the closure to practical simulations is discussed. [Preview Abstract] |
Sunday, November 23, 2014 2:54PM - 3:07PM |
D28.00004: A dynamic subfilter-scale stress model for Large eddy simulations Amirreza Rouhi, Ugo Piomelli, Bernard Geurts In conventional large eddy simulation, the filter width is related to the grid size; this decouples the filter width from turbulence physics and results in unwanted dependence of the subfilter model on the grid arrangement. Relating the filter width to the integral length-scale is a potential solution. We proposed an approximation for the integral length-scale, in which a single model parameter was determined based on the global contribution of unresolved (subfilter) scales to the resolved ones denoted as subfilter activity (Piomelli \& Geurts, \textit{Direct and Large-Eddy Simulation VIII}, pp. 15-20, 2011). We have devolped a localized model in which we assign a target value to subfilter activity locally, requiring the model parameter to adapt itself to the local state of the flow. This dynamic modification is coupled with a local formulation for the integral scale. The modified model was applied on channel flow at $Re_\tau$ up to $2,000$, accelerating boundary layer and backward-facing step flow at high $Re$ with comparable accuracy as the Dynamic Smagorinsky model but with less computational expense. [Preview Abstract] |
Sunday, November 23, 2014 3:07PM - 3:20PM |
D28.00005: Development of subgrid-scale model using machine learning Masataka Gamahara, Yuji Hattori Neural network, which automatically finds patterns or rules from big data, is applied to construct an improved sub-grid scale (SGS) model used in LES. SGS stress tensors are obtained by filtering data of direct numerical simulation (DNS) of turbulent channel flow. We use velocity gradient tensors and distance from the wall as inputs of the neural network aiming at improving conventional SGS models which include the Smagorinsky model. The back-propagation method is used in the learning process of the neural network. The results show that the neural network is able to learn SGS stress tensors. High correlation coefficients between SGS stress tensors obtained from DNS data and those estimated by the neural network are obtained. The results do not depend very much on the training data used for learning. Furthermore we investigate dependence on the size of training data, the filter size and the number of neurons. In particular the learning of neural network depends on the filter size. We also obtain high correlation coefficients at all Reynolds numbers tested. In order to find an explicit form of the estimated SGS stress tensors we try to identify the minimum set of independent variables by reducing the number of inputs. Physics behind the obtained model will be also discussed. [Preview Abstract] |
Sunday, November 23, 2014 3:20PM - 3:33PM |
D28.00006: Slip wall modeling approaches for separated flows and heat transfer Sanjeeb Bose, Brian Pierce, Parviz Moin Resolution of near-wall turbulent structures is computational prohibitive, necessitating the need for wall-modeled large-eddy simulation approaches. Standard wall models are often formulated to represent the wall stress assuming an equilibrium, attached boundary layer. This assumption is invalid in complex flows that include transition to turbulence or boundary layer separation. A dynamic slip wall boundary condition has been recently proposed (Bose \& Moin, PoF, 2014) as an alternative for wall-modeled LES, where a slip wall boundary condition is derived from the differentially filtered LES governing equations with no assumption on the state of the local boundary layer. Results will be presented from the application of the dynamic slip wall model to flows with 3D separation (asymmetric stalled diffuser) and from the extension of the model to the prediction of wall heat transfer (turbine blade). The wall modeled LES predicts the primary quantity of interest in these flows: the pressure recovery in the diffuser and the heat transfer coefficient on the turbine blade. [Preview Abstract] |
Sunday, November 23, 2014 3:33PM - 3:46PM |
D28.00007: A Wall Model for Large-Eddy Simulation of Compressible Channel Flows Barrett McCann, Antonino Ferrante We have developed a new wall model for the large-eddy simulation (LES) of compressible channel flows with isothermal walls by extending the incompressible model of Chung and Pullin (\emph{J. Fluid Mech.} 2009). The wall model computes the local, instantaneous wall shear stress and heat flux, which are then applied as wall boundary conditions, by solving two time-dependent, parameter-free ordinary differential equations (ODEs) at each time step. These ODEs are obtained by integrating the filtered momentum and energy equations in the wall-normal direction from the wall to the first grid point in the log layer. In contrast to so-called ``wall resolved'' LES, employment of this wall model allows use of relatively coarse computational meshes of fixed size, independent of Reynolds number. The wall model is first validated by comparing our LES results at $M=0.15$ and $Re_{\tau}=2003$ to the DNS results of Hoyas and Jim\'{e}nez (\emph{Phys. Fluids} 2006). We present LES results of channel flow simulations at $M=0.15$ and $M=0.75$, over a three-order of magnitude range of friction Reynolds numbers ($Re_{\tau}=2 \times 10^3$, $2 \times 10^4$, $2 \times 10^5$ and $2 \times 10^6$), on a mesh with $256 \times 32 \times 128$ grid points in the streamwise, wall-normal, and spanwise directions. [Preview Abstract] |
Sunday, November 23, 2014 3:46PM - 3:59PM |
D28.00008: An integral wall model for Large Eddy Simulation (iWMLES) and applications to developing boundary layers over smooth and rough plates Xiang Yang, Jasim Sadique, Rajat Mittal, Charles Meneveau A new wall model for Large-Eddy-Simulations is proposed. It is based on an integral boundary layer method that assumes a functional form for the local mean velocity profile. The method, iWMLES, evaluates required unsteady and advective terms in the vertically integrated boundary layer equations analytically. The assumed profile contains a viscous or roughness sublayer, and a logarithmic layer with an additional linear term accounting for inertial and pressure gradient effects. The iWMLES method is tested in the context of a finite difference LES code. Test cases include developing turbulent boundary layers on a smooth flat plate at various Reynolds numbers, over flat plates with unresolved roughness, and a sample application to boundary layer flow over a plate that includes resolved roughness elements. The elements are truncated cones acting as idealized barnacle-like roughness elements that often occur in biofouling of marine surfaces. Comparisons with data show that iWMLES provides accurate predictions of near-wall velocity profiles in LES while, similarly to equilibrium wall models, its cost remains independent of Reynolds number and is thus significantly lower compared to standard zonal or hybrid wall models. [Preview Abstract] |
Sunday, November 23, 2014 3:59PM - 4:12PM |
D28.00009: Implicit LES of Flow Over Wall-Mounted Hump Susheel Sekhar, Nagi Mansour Implicit LES of turbulent flow over wall-mounted hump is conducted to understand the physics of separated flows, and to provide data for RANS modeling and development. A modified version of the FDL3DI code\footnote{D.V.\ Gaitonde \& M.R.\ Visbal, AFRL-VA-WP TR-1998-3060 (1998)} that solves the compressible Navier-Stokes equations using high-order compact difference scheme and filter, and the standard recycling/rescaling method for generating a fully developed turbulent boundary layer at the inlet,\footnote{B.\ Morgan et al., AIAA J., \textbf{49} (3), 582-597 (2011)} is used. A mean velocity profile with Re$_\theta=1,400$ is imposed at the inlet. Qualitative assessment shows that the separation bubble is comparable in size with experiment. A detailed analysis, including comparisons of mean velocity profiles with experimental data before separation and after reattachment, is made. Quantitative comparisons of Reynolds stress profiles, as well as budgets of Reynolds stresses and turbulent kinetic energy are also presented. Physics of the flow post-reattachment is the focus of this study. Results from this effort will be used to further set up simulations at a higher Reynolds number (Re$_\theta=3,500$). [Preview Abstract] |
Sunday, November 23, 2014 4:12PM - 4:25PM |
D28.00010: Influence of vortices on turbulence statistics Koujiro Anayama, Katsunori Yoshimatsu, Yukio Kaneda We consider the importance or unimportance of the role of vortices at small scales in the determination of the turbulence statistics, on the basis of the method of the so-called ``Computational Surgery.'' Two fields, true and false fields, are generated. The true field obeys the Navier-Stokes (NS) equations for an incompressible fluid. In the false field, the NS dynamics are artificially modified so that the intense tube-like structures of the vortices are lost. Comparing the two fields, we may get some idea on the role of the vortices. The comparison so far made suggests that the statistics at larger scales are not sensitive to the exact vortex structure at small scales. [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