Bulletin of the American Physical Society
69th Annual Meeting of the APS Division of Fluid Dynamics
Volume 61, Number 20
Sunday–Tuesday, November 20–22, 2016; Portland, Oregon
Session R34: Turbulence: DNS Simulations |
Hide Abstracts |
Chair: Daniel Livescu, Los Alamos National Laboratory Room: Oregon Ballroom 203 |
Tuesday, November 22, 2016 1:30PM - 1:43PM |
R34.00001: Time tracking and interaction of energy-eddies at different scales Jose I. Cardesa, Alberto Vela-Martin, Javier Jimenez We study the energy cascade through coherent structures obtained in time-resolved simulations of incompressible, statistically steady isotropic turbulence. The structures are defined as geometrically connected regions of the flow with high kinetic energy. We compute the latter by band-pass filtering the velocity field around a scale $r$. We analyse the dynamics of structures extracted with different $r$, which are a proxy for eddies containing energy at those $r$. We find that the size of these ``energy-eddies'' scales with $r$, while their lifetime scales with the local eddy-turnover $r^{2/3}\epsilon^{-1/3}$, where $\epsilon$ is the energy dissipation averaged over all space and time. Furthermore, a statistical analysis over the lives of the eddies shows a slight predominance of the splitting over the merging process. When we isolate the eddies which do not interact with other eddies of the same scale, we observe a parent-child dependence by which, on average, structures are born at scale $r$ during the decaying part of the life of a structure at scale $r'>r$. The energy-eddy at $r'$ lives in the same region of space as that at $r$. Finally, we investigate how interactions between eddies at the same scale are echoed across other scales. [Preview Abstract] |
Tuesday, November 22, 2016 1:43PM - 1:56PM |
R34.00002: Dynamic Mode Decomposition of Jet in Channel Crossflow Zhao Wu, Dominique Laurence In this paper, the authors present a comparative analysis of Koopman modes computed from snap-shots of direct numerical simulations of a jet in channel crossflow (channel flow Re number =3333, jet-to-crossflow velocity ratio =1/6). The flow is complex due to interactions between the jet and the cross-flow, and contains geometry-dependent large-scale coherent structures; thus, the Koopman mode analysis provides a powerful tool for studying the spatial and spectral information of the flow. The Koopman modes are approximated by the DMD modes restricted to Krylov subspace, and the Koopman modes isolate structures associated with single frequency only. In this work, we address issues related to the physical interpretation of the DMD modes. The results show that the computed Koopman modes identify the relevant frequencies and the corresponding three-dimensional flow structures automatically. We present the selected DMD modes, which show big differences in the spatial structures and frequency. The shear layer vortices are separated from the horseshoe vortex. These modes have large amplitudes among all modes obtained. [Preview Abstract] |
Tuesday, November 22, 2016 1:56PM - 2:09PM |
R34.00003: Lagrangian statistics of turbulent dispersion from $8192^3$ direct numerical simulation of isotropic turbulence Dhawal Buaria, P.K. Yeung, B.L. Sawford An efficient massively parallel algorithm has allowed us to obtain the trajectories of 300 million fluid particles in an $8192^3$ simulation of isotropic turbulence at Taylor-scale Reynolds number 1300. Conditional single-particle statistics are used to investigate the effect of extreme events in dissipation and enstrophy on turbulent dispersion. The statistics of pairs and tetrads, both forward and backward in time, are obtained via post-processing of single-particle trajectories. For tetrads, since memory of shape is known to be short, we focus, for convenience, on samples which are initially regular, with all sides of comparable length. The statistics of tetrad size show similar behavior as the two-particle relative dispersion, i.e., stronger backward dispersion at intermediate times with larger backward Richardson constant. In contrast, the statistics of tetrad shape show more robust inertial range scaling, in both forward and backward frames. However, the distortion of shape is stronger for backward dispersion. Our results suggest that the Reynolds number reached in this work is sufficient to settle some long-standing questions concerning Lagrangian scale similarity. [Preview Abstract] |
Tuesday, November 22, 2016 2:09PM - 2:22PM |
R34.00004: DNS of turbulent Couette flow with transpiration - spectra and symmetry induced scaling laws. Sergio Hoyas, Stefanie Kraheberger, Martin Oberlack We present DNS results of turbulent plane Couette flow with constant wall-normal transpiration for Reynolds numbers of $Re_\tau=250,500,1000$ and several transpiration Reynolds numbers $Re_{tr}=V_0/U_w$. To obtain the DNS data, a pseudo-spectral code, which originally was developed at UP Madrid, see \emph{(Hoyas and Jim{\'e}nez 2006)}, is used for the simulations. Due to the lack of experimental and DNS data, the convergence of every simulation has been validated using the total shear stress equation and the relation between the friction velocities at the lower and upper wall.\\ Examining the spectra we found that the large and wide structures, which appear in pure Couette flow, see \emph{(Avsarkisov et al. 2014)}, are destroyed as soon as transpiration velocity is different from zero. This and the presence of anomalous spectra near the blowing wall indicates the strong influence of suction on the whole flow, which was observed in \emph{(Antonia et al. 1988)} as well. As classical scaling laws are not valid due to transpiration, new scaling laws of the mean velocity are derived using Lie symmetry methods. Additionally, suction creates a comparably larger $u_\tau$ which, in turn, causes a flat and long region in the indicator function for the largest transpiration rate. [Preview Abstract] |
Tuesday, November 22, 2016 2:22PM - 2:35PM |
R34.00005: Multiscale modeling of turbulent channel flow over porous walls Sudhakar Yogaraj, Ugis Lacis, Shervin Bagheri We perform direct numerical simulations of fully developed turbulent flow through a channel coated with a porous material. The Navier-stokes equations governing the fluid domain and the Darcy equations of the porous medium are coupled using an iterative partitioned scheme. At the interface between the two media, boundary conditions derived using a multiscale homogenization approach are enforced. The main feature of this approach is that the anisotropic micro-structural pore features are directly taken into consideration to derive the constitutive coefficients of the porous media as well as of the interface. The focus of the present work is to study the influence of micro-structure pore geometry on the dynamics of turbulent flows. Detailed turbulence statistics and instantaneous flow field are presented. For comparison, flow through impermeable channel flows are included. [Preview Abstract] |
Tuesday, November 22, 2016 2:35PM - 2:48PM |
R34.00006: Lagrangian and Eulerian statistics in homogeneous, anisotropic flows Kartik Iyer, Fabio Bonaccorso, Federico Toschi, Luca Biferale We report results from highly resolved direct numerical simulations of anisotropic homogeneous flows using up to $2048^3$ collocations points. We examine a turbulent Kolmogorov flow with randomly correlated phases in order to recover space homogeneity on average. We present Eulerian and Lagrangian measurements concerning the universality of isotropic and anisotropic contributions using a systematic decomposition based on the eigenfunctions of the SO$(3)$ group of rotations in three dimensions. Additionally, we discuss absolute dispersion statistics of particles in flows subjected to different large-scale anisotropies. [Preview Abstract] |
Tuesday, November 22, 2016 2:48PM - 3:01PM |
R34.00007: Lagrangian statistics in turbulent channel flow: implications for Lagrangian stochastic models Nickolas Stelzenmuller, Juan Igancio Polanco, Ivana Vinkovic, Nicolas Mordant Lagrangian acceleration and velocity correlations in statistically one-dimesional turbulence are presented in the context of the development of Lagrangian stochastic models of inhomogeneous turbulent flows. These correlations are measured experimentally by 3D PTV in a high aspect ratio water channel at $Re_\tau=1450$, and numerically from DNS performed at the same Reynolds number. Lagrangian timescales, key components of Lagrangian stochastic models, are extracted from acceleration and velocity autocorrelations. The evolution of these timescales as a function of distance to the wall is presented, and compared to similar quantities measured in homogeneous isotropic turbulence. A strong dependance of all Lagrangian timescales on wall distance is present across the width of the channel. Significant cross-correlations are observed between the streamwise and wall-normal components of both acceleration and velocity. Lagrangian stochastic models of this flow must therefore retain dependance on the wall-normal coordinate and the components of acceleration and velocity, resulting in significantly more complex models than those used for homogeneous isotropic turbulence. [Preview Abstract] |
Tuesday, November 22, 2016 3:01PM - 3:14PM |
R34.00008: Statistics of Vortical Structures in Variable-Density Turbulent Mixing Layers Jon Baltzer, Daniel Livescu Direct Numerical Simulations are performed of temporal incompressible shear-driven planar mixing layers between two miscible streams of fluids with different densities. The simulations begin from thin disturbed interfaces and develop into self-similar states. We use very large domain sizes, corresponding to grids of up to 6144 x 2048 x 1536 points, to produce high-quality statistics and allow natural growth of turbulent structures. A wide range of Atwood numbers are explored, ranging from nearly constant density to A=0.87 (or a density ratio of 14). At high Atwood numbers, a variety of statistics show that variable-density effects produce significant asymmetries. Here we focus on the differences in vortical structure of the light and heavy fluid streams and the importance of non-Boussinesq effects as Atwood number increases. Detailed budgets of vorticity moments are examined in conjunction with the alignments of vorticity relative to other flow quantities. The results display the variable density effects due to compositional variations, a distinctly different mechanism from the density variations associated with compressibility in high-speed flows. [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