Bulletin of the American Physical Society
77th Annual Meeting of the Division of Fluid Dynamics
Sunday–Tuesday, November 24–26, 2024; Salt Lake City, Utah
Session R37: Turbulence: DNS |
Hide Abstracts |
Chair: Adrian Lozano-Duran, Caltech / MIT Room: 355 C |
Monday, November 25, 2024 1:50PM - 2:03PM |
R37.00001: Analyses of extreme-scale turbulence datasets in short simulations of forced isotropic turbulence enabled by exascale computing Pui-Kuen Yeung, Kiran Ravikumar, Rohini Uma-Vaideswaran, Daniel L Dotson, Charles Meneveau, K.R. Sreenivasan, Stephen B Pope, Stephen Nichols Access to the world's first exascale computer (named Frontier) combined with a protocol of multiple independent simulation segments at high resolution [Phys. Rev. Fluids 2020, 110517] have led to the creation of a new DNS data collection of isotropic turbulence at the scale of 32768-cubed grid points. Advances in GPU computing techniques [paper in review] have also greatly facilitated both on-the-fly processing on a time-resolved basis, and the memory-intensive post-processing of single-time snapshots of size at 1/4 of a petabyte or larger. We will discuss intermittency corrections for energy spectra and velocity structure functions, as well as the behavior of cumulative distribution functions that quantify the likelihood of extreme events present in both pointwise and locally-averaged fluctuations of the energy dissipation and enstrophy. Attention is given to both dissipation-range and inertial-range phenomena. |
Monday, November 25, 2024 2:03PM - 2:16PM |
R37.00002: Statistics of pressure Hessian quantities in well-resolved simulations at high Reynolds number. Rohini Uma-Vaideswaran, Pui-Kuen Yeung, Charles Meneveau, Daniel Livescu, Michael Wilczek The pressure Hessian, a second-order tensor that consists of second-derivatives of the pressure field, is known to have important effects on the evolution of velocity gradients following Lagrangian fluid particle trajectories in turbulence. Accurate characterization of the properties of this tensor at high Reynolds number is challenging due to strong intermittency, which also make accurate interpolation along particle trajectories a more demanding task. Nevertheless, access to almost the full power of the world's first exascale computer is expected to allow us to obtain reliable results at Reynolds numbers significantly higher than in previous results in the literature as well as recent work using machine learning. In particular, we will investigate both Eulerian and Lagrangian conditional averages of the pressure Hessian contracted with the velocity gradient tensor given the second and third-order invariants of the latter. |
Monday, November 25, 2024 2:16PM - 2:29PM |
R37.00003: A study of passive scalar turbulence at high Reynolds numbers enabled by exascale computing Daniel L Dotson, Pui-Kuen Yeung, K.R. Sreenivasan Many sources of data in the literature suggest, at least at Schmidt numbers of order unity or higher, that passive scalar fields transported by turbulence tend to be more intermittent than the velocity field, and also show significant departures from local isotropy when a production effect arising from a mean scalar gradient is present. In this talk we present direct numerical simulations at high resolution enabled by the Frontier GPU platform to investigate effects of Reynolds number and resolution to greater detail than in the past. For unity Schmidt number it appears that a grid spacing close to the Kolmogorov scale coupled with Courant number 0.3 for time-stepping can provide sufficiently accurate results for scalar gradient moments up to the fourth order, and for the peak value of normalized scalar dissipation rate, which is higher than that for the energy dissipation. It appears that inadequate resolution reduces the gradient skewness relevant to local anisotropy, in a manner related to ramp and cliff structures in the scalar field. |
Monday, November 25, 2024 2:29PM - 2:42PM |
R37.00004: Quantifying the entanglement complexity in turbulent flows Fathima Farheen Nambipunnilath Siddique, Yeonsu Jung, David Palmer, Kartik P Iyer, Lakshminarayanan Mahadevan Turbulence be regarded as a dynamically disordered whirl of vortical structures on multiple scales. Complementing statistical measures of the complex flows in terms of correlation functions, here we consider the use of a locally-averaged average crossing number (ACN) of filamentary vortical structures to quantify the dynamical entanglement of vorticity. Using data from direct numerical simulations of homogenous isotropic turbulence which show that vortex filaments can entangle with each other into higher-order structures of ever-increasing complexity, we calculate the ACN for different Reynolds numbers. We describe the evolution of the ACN and compare the topological complexity of such structures in Navier-Stokes turbulence with that of a synthetic Gaussian Random Field with the same two-point correlation in order to distinguish between dynamical and kinematic effects on the topology of the vortex filaments. |
Monday, November 25, 2024 2:42PM - 2:55PM |
R37.00005: ABSTRACT WITHDRAWN
|
Monday, November 25, 2024 2:55PM - 3:08PM |
R37.00006: Impact of low floating-point precision on high-fidelity simulations of turbulence Philipp Schlatter, Martin Karp, Ronith Stanly, Hang Song, Timofey Mukha, Luca Galimberti, Siavash Toosi, Manuel Münsch, Lisandro Dalcin, Saleh Rezaeiravesh, Niclas Jansson, Stefano Markidis, Matteo Parsani, Sanjeeb T Bose, Sanjiva K Lele Modern computing clusters offer specialized hardware with reduced-precision arithmetic that can speed-up the time to solution significantly, mainly due to less data movement and increased arithmetic performance. However, for high-fidelity simulations of turbulence, separation, and transition the impact of lower floating-point precision on the computed solution and the uncertainty it introduces has not been explored in sufficient detail. This limits the optimal utilization of new and upcoming exascale machines. In this work, the effect of reduced precision for numerical solution of the Navier-Stokes equations is considered across different spatial and temporal discretization approaches. We compare four solvers, two compressible and two incompressible, across three test cases: K-type transition in a channel, turbulent channel flow up to Ret=2000 and flow over a cylinder at ReD=3900. Different terms of the Navier-Stokes equations are perturbed to lower floating-point precision, ranging from conventional 64 bit IEEE double precision down to recent 8 bit formats highlighting the opportunities and the drawbacks of low-precision arithmetic in high-fidelity computational fluid dynamics. |
Monday, November 25, 2024 3:08PM - 3:21PM |
R37.00007: ABSTRACT WITHDRAWN
|
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. |
© 2025 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