Bulletin of the American Physical Society
71st Annual Meeting of the APS Division of Fluid Dynamics
Volume 63, Number 13
Sunday–Tuesday, November 18–20, 2018; Atlanta, Georgia
Session F38: DNS and LES |
Hide Abstracts |
|
Chair: Johan Larsson, University of Maryland, College Park Room: Georgia World Congress Center Ballroom 1/2 |
|
Monday, November 19, 2018 8:00AM - 8:13AM |
F38.00001: Scaling and structure of extreme velocity gradients in turbulence Dhawal Buaria, Alain Pumir, Eberhard Bodenschatz, Pui-Kuen Yeung The formation of very localized and intense velocity gradients underlies intermittency in fully turbulent flows. While the typical value of velocity gradients scales with the Kolmogorov time scale τK, their fluctuations can be orders of magnitude larger and become increasingly intense as Reynolds number increases. Such extreme events play a crucial role in numerous applications, e.g. cloud physics, turbulent combustion, but a complete description still remains a major challenge. Using direct numerical simulations of isotropic turbulence with an unprecedented small-scale resolution, we characterize such extreme events over a wide range of Taylor-scale Reynolds numbers (Rλ). Specifically, by studying the PDFs of dissipation and enstrophy, we find that the extreme events scale as τK-1Rλβ, with β=0.775±0.025, weaker than β=1 predicted by existing theories. The observation that the velocity differences across very small distances can be as large as urms leads to the conclusion that the smallest length scale in the flow scales as ηRλ-α, with α=β-0.5, where η is the Kolmogorov length scale. Comparisons with the multifractal theory are also drawn. We further relate the exponent β<1 to the nonlocal stretching acting on a given vortex structure.
|
|
Monday, November 19, 2018 8:13AM - 8:26AM |
F38.00002: Sensitivity analysis of large eddy simulation data Shervin Sammak, Ling Miao, Peyman Givi, Cyrus K Madnia The discontinuous Galerkin (DG) methodology has proven very effective for large eddy simulation (LES), as a larger portion of the resolved energy is captured as the order of spectral approximation is increased. In this work, a filtered density function (FDF) based method is used for DG-LES of turbulent shear flows. The effects of the discretization order (p), the grid resolution (h), and the filter width (Δ) are examined, and comparative assessments are made against direct numerical simulation (DNS) data. It is shown that p-refinement yields a better convergence rate and requires relatively less computational times compared to h-refinement with the filter width of the same order as the mesh spacing. |
|
Monday, November 19, 2018 8:26AM - 8:39AM |
F38.00003: How turbulence is generated from zero—and the interscale dynamics in stationary state Sualeh Khurshid, Diego A. Donzis, Katepalli R. Sreenivasan A large database of direct numerical simulations of isotropic turbulence, with Taylor-Reynolds numbers ranging from $R_\lambda \approx 1$ to $300$ and grid sizes of up to $2048^3$, is analyzed to understand how broadband turbulence develops from the forcing at a few low wavenumbers. Different forcing schemes at large scales are used. The energy growth in various wavenumbers follows power laws to very good accuracy. The energy in the first few wavenumbers close to the forcing band grows simultaneously with forcing; but that in higher wavenumbers occurs after a time lag of the order of a large-eddy time scale, qualitatively consistent with the cascade picture. However, it appears that part of the energy leaks to dissipative wavenumbers as an instantaneous reaction to forcing. Once in a steady sate, while Gaussian statistics are observed for temporal fluctuations in the inertial range consistent with the fluctuations in forcing, increasingly skewed probability density functions emerge at higher wavenumbers. In particular, fluctuations from the mean in the far-dissipation range, defined here as wavenumbers larger than twice the mean Kolmogorov wavenumber, are very large. Different transfer models to predict the observed behavior are discussed. |
|
Monday, November 19, 2018 8:39AM - 8:52AM |
F38.00004: Persistence of reflectional non-symmetry in freely-decaying homogeneous turbulence Katsunori Yoshimatsu, Yukio Kaneda We consider freely-decaying incompressible homogeneous helical turbulence whose energy spectrum E(k) and helicity spectrum H(k) are respectively given by E(k)=C k^2 + o(k^2) and H(k)= C_h k^3 +o(k^3) at k -> 0. Here, k is the wavenumber, C is a dynamical invariant, and C_h is a k-independent constant. It is shown that C_h is another dynamical invariant, the O(k^0)-tem of the velocity spectral correlation tensor is time-independent, and the term may be non-reflectional symmetric. A theoretical analysis based on a flow self-similarity and the time-independence suggests the persistence of the reflectional non-symmetry for fully-developed helical turbulence. The decay rates of energy and helicity are obtained by a simple dimensional analysis. We examine the theoretical results by the use of direct numerical simulation of incompressible helical turbulence in a periodic box. |
|
Monday, November 19, 2018 8:52AM - 9:05AM |
F38.00005: Missing large scales: Development and implementation of a physics-based turbulence forcing scheme for numerical simulations of compressible flows Guillaume Beardsell, Guillaume Blanquart When performing direct numerical simulations of highly-turbulent flows, it is often prohibitively expensive to simulate complete flow geometries. A well-selected portion of the domain is then chosen. However, by doing so one usually misses Turbulent Kinetic Energy (TKE) injection due to shear by the large scales. There are many techniques available in the literature to inject TKE, however none of these follow directly from the Navier-Stokes equations. This is the goal of the present work. We decompose the velocity field into small-scale and large-scale components. The latter is assumed to be known beforehand, and we solve for the small-scale component only. We have already applied this strategy to incompressible flows, but not to compressible ones, where special care must be taken regarding the energy equation. Implementation of this scheme in the finite-difference solver NGA is discussed and preliminary results are presented. In particular, we investigate the impact of periodic boundary conditions, which can cause some dilatational velocity modes to grow boundlessly. |
|
Monday, November 19, 2018 9:05AM - 9:18AM |
F38.00006: Scale averaged trends of dissipation and enstrophy in fluid turbulence Kartik Iyer, Joerg Schumacher, Katepalli Raju Sreenivasan, Pui-Kuen Yeung Fluid turbulence is typically characterized as a tangle of high enstrophy (vorticity squared), low pressure vortices, embedded in regions of straining motions which possess high kinetic energy dissipation. The intermittent statistics of enstrophy and dissipation, which quantify rotation and strain respectively, considered over inertial scales, are expected to approach one another, in the traditional paradigm of small-scale universality. Using a scale-based analysis of the Poisson equation that relates pressure, energy dissipation and enstrophy, obtained from the incompressible Navier-Stokes equations, we show that regions of high enstrophy and low dissipation are more prevalent than regions of low enstrophy and high dissipation. Consistent with this discrepancy is the finding that the intermittency exponents of local averages of enstrophy and dissipation differ by a finite amount at least up to Taylor micro-scale Reynolds number, Rλ = 1300, and that these differences can be attributed to the non-trivial pressure laplacian contributions. |
|
Monday, November 19, 2018 9:18AM - 9:31AM |
F38.00007: Reducing the computational demand of direct numerical simulation via assimilation of experimental data Callum Atkinson, Vassili Kitsios, Julio Soria Direct numerical simulations (DNS) rapidly establish small scales of turbulence, however large scales require large domains and long simulation times. In contrast, experimental measurements capture large scales yet struggle to resolve smaller scales due to measurement noise. This work seeks to reduce the required computational domain of DNS by replacing the simulation of large scales with data from experimental measurements. To account for measurement uncertainty and disparity in measured and resolved scales, data assimilation is performed by an ensemble Kalman filter, via which the limited domain DNSs are nudged towards the experimental measurements. To explore the extent to which this approach can reduce the computational domain without sacrificing the accuracy of the resolved scales, an incompressible homogeneous isotropic code was modified to simultaneously simulate a fully resolved DNS and an ensemble of smaller simulations from which the Kalman gain is calculated and applied. The fully resolved DNS is subsampled and filtered to represent experimental data of varying resolution and accuracy. |
|
Monday, November 19, 2018 9:31AM - 9:44AM |
F38.00008: The impact of boundary conditions on spectral condensation of turbulence: Experiments and numerics Chayanon Wichitrnithed, Jordon Tyler Campbell, Logan Kageorge, Roman O Grigoriev, Michael F Schatz Flow patterns that are coherent over the entire domain can emerge in 2D and anisotropic 3D turbulent flows. In a combined experimental and numerical study of flow in a quasi-2D electrolyte layer, we observe, for electromagnetic driving that is sufficiently strong, the emergence of well-ordered large scale patterns, whose structure strongly depends on boundary conditions. For rectangular lateral boundaries of high aspect ratio, the patterns exhibit a well-defined length scale that is comparable to the smaller lateral dimension. For periodic lateral boundaries, an array of jets are observed; by contrast, with no-slip lateral boundaries, a pattern of vortices are found. The patterns are found to persist, even with forcing that is spatially random. The spectral behavior of the patterns will be described in this talk.
|
|
Monday, November 19, 2018 9:44AM - 9:57AM |
F38.00009: Effects of Low Mach-Number Correction in Turbulent Mixing Transition Simulations Fernando F Grinstein, Juan A Saenz, Josh Dolence, Tom Masser, Marianne M Francois Transition and turbulence decay with the Taylor-Green vortex have been effectively used to demonstrate emulation of high Reynolds-number (Re) physical dissipation through numerical convective effects of various non-oscillatory finite-volume algorithms for implicit large eddy simulation (ILES), e.g. using the Godunov-based Eulerian adaptive mesh refinement code xRAGE. The inverse-chevron shock-tube simulations have been also used to assess xRAGE based ILES for shock driven turbulent mixing, compared with available simulation and laboratory data. The previous assessments are extended to evaluate new directionally-unsplit high-order algorithms in xRAGE, including a correction to address the well-known issue of excessive numerical diffusion of shock-capturing (e.g., Godunov-type) schemes for low Mach numbers. Basic issues of transition to turbulence and turbulent mixing are discussed, and results of simulations of high-Re turbulent flow and mixing in canonical test cases are reported. Compared to the directional-split cases, and for each grid resolution considered, unsplit results exhibit transition to turbulence with much higher effective Re -- and significantly more so with the low Mach number correction. |
|
Monday, November 19, 2018 9:57AM - 10:10AM |
F38.00010: Grid-adaptation in large eddy simulation: finding the optimal filter-width as a function of space and direction Siavash Toosi, Johan Larsson The grid is the single most important factor determining the accuracy of a large eddy simulation, controlling both the numerical errors and the errors due to the modeling of the unresolved scales. The simultaneous dependence of both numerical and modeling errors on the grid resolution makes finding the optimal grid more complicated than in DNS or RANS, where the grid only controls the numerical errors. The purpose of the present study is to introduce an efficient technique that eliminates or reduces the need for input from user's knowledge of turbulence in grid generation, is capable of producing optimal unstructured grids with varying resolution in space and direction, and that is still practical for generating LES grids in realistic settings. The proposed method is tested on turbulent channel flow (where the optimal grid is relatively well known) and on the flow over a backward-facing step, which has a more complex flow that requires a grid with varying resolution in space and different optimal aspect ratios in each location. The proposed method performs well in both test cases and produces nearly optimal grids. |
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. |
© 2023 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
1 Research Road, Ridge, NY 11961-2701
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700