Bulletin of the American Physical Society
75th Annual Meeting of the Division of Fluid Dynamics
Volume 67, Number 19
Sunday–Tuesday, November 20–22, 2022; Indiana Convention Center, Indianapolis, Indiana.
Session J26: Turbulence: Wall-Bounded I |
Hide Abstracts |
Chair: Bjoern Hof, Institute of Science and Technology Aust Room: 234 |
Sunday, November 20, 2022 4:35PM - 4:48PM |
J26.00001: Friction scaling and the onset of large scale motions in pipe flow Bjoern Hof, Jose M Lopez, Davide Scarselli, Bowen Yang, Gregory Falkovich Fully turbulent flow is stable in pipes once the bulk Reynolds number |
Sunday, November 20, 2022 4:48PM - 5:01PM |
J26.00002: Adjoint-variational estimation of near-wall turbulence from outer observations Mengze Wang, Tamer A Zaki Turbulence in the vicinity of the wall and the associated surface stresses are challenging to probe experimentally. A recent study demonstrated that, given fully resolved outer measurements, the wall-attached horizontal layer can synchronize to the true flow state and evolution, as long as the thickness of the wall layer is less than thirty viscous units (M. Wang and T. Zaki, 2022, J. Fluid Mech. 943, A4). Beyond this critical thickness, synchronization is not possible, and the accuracy of estimating the wall layer has never been examined. This problem is studied using adjoint-variational data assimilation, where we seek a Navier-Stokes solution that reproduces the outer flow observations and predicts the unknown near-wall turbulence. The reconstructed wall layer remains almost identical to the true flow when the measurement height is within fifty wall units, and the estimation accuracy deteriorates as the observations are further separated from the wall. We also demonstrate that the accuracy of the optimally estimated near-wall flow is robust when filtered and subsampled outer observations are adopted. |
Sunday, November 20, 2022 5:01PM - 5:14PM |
J26.00003: Towards Real-time Reconstruction of Velocity Fluctuations in Turbulent Channel Flow Rahul Arun, H. Jane Bae, Beverley J McKeon We develop a streaming framework for efficient reconstruction of turbulent velocity fluctuations with the goal of enabling real-time prediction of turbulent flow features. The approach consists of training and reconstruction phases, with a minimal data requirement during reconstruction. During training, we efficiently and robustly compute the resolvent operator for channel flow via blockwise inversion. During subsequent testing, we process the incoming stream of measurements using a (temporal) sliding discrete Fourier transform to allow for continuous updates. We apply the technique to reconstruction of the flow in a minimal channel at Reτ ≈ 186 from sparse, planar measurements and evaluate the errors incurred relative to the input data required. The combination of data-driven and equation-based approaches bolsters reconstruction efficacy beyond using either in isolation. |
Sunday, November 20, 2022 5:14PM - 5:27PM |
J26.00004: Reconstructing wall-bounded turbulent flows from limited PIV measurements using Physics Informed Neural Networks Vamsi Krishna Chinta, Deep Ray, Assad Oberai, Mitul Luhar Turbulent flows are characterized by multiple length and time scales. Laboratory techniques used to acquire velocity field measurements such as Particle Image Velocimetry (PIV) are typically limited in their ability to acquire time-resolved velocity fields. In this work we train a Physics Informed Neural Network (PINN) to estimate the velocity field snapshots between consecutive 2D PIV snapshots. The PINN is defined to take the spatial and temporal coordinates as the input and the two velocity field components (streamwise and wall-normal) are the output. The PDE constraint imposed on the reconstructed flow field is consistent with the Taylor's frozen turbulence hypothesis, which has been used successfully for temporal reconstruction in previous studies. A composite loss function is defined to include the error from the PDE and the deviation from the PIV snapshots. The PINN is trained to minimize this composite loss function. The velocity field can now be reconstructed between the two snapshots at arbitrary spatio-temporal resolution. The errors from these reconstructions are evaluated using PIV-like snapshots from numerical simulations of turbulent channel flow from the Johns Hopkins Turbulence Database (JHTDB). |
Sunday, November 20, 2022 5:27PM - 5:40PM |
J26.00005: Simulating flow over superhydrophobic surfaces at high $Re$ and high gas fractions using octree discretizations Frederic Gibou, Luis Á Larios-Cárdenas, Fernando Temprano-Coleto, Paolo Luzzatto-Fegiz, Julien R Landel, Oliver E Jensen, Samuel D Tomlinson, Francois Peaudecerf Superhydrophobic surfaces have demonstrated high drag reductions in turbulent flows. However, simulating the high gas fractions and high $Re$ typical of naval applications remains challenging, as very sharp gradients arise near the transitions between no-slip regions and gas-liquid interfaces. To achieve accurate yet computationally manageable simulations, we use adaptive octree discretizations in distributed environments, building on the parallel level-set schemes of Mirzadeh et al. (2016), using {\tt p4est} (2011), and extending the work of Egan et al. (2021). We thus leverage performant multicore systems and scalable linear solvers. We simulate SHS structures consisting of streamwise gratings and discretize the computational domain with 6-by-2-by-3 octrees, enforcing a wide band of the smallest cells layering the walls to capture local phenomena. Our simulation approach enables relatively rapid exploration of the parameter space up to gas fractions of 93.75\% and shows how flow properties, including Reynolds stresses, mean momentum advection, and vortical structures, are affected by gas fraction and Reynolds number. In this conference contribution, we will present flow results as well as the numerical tools and algorithms that have enabled our simulations. |
Sunday, November 20, 2022 5:40PM - 5:53PM |
J26.00006: Causality analysis of turbulent channel flow based on massive computation Kosuke Osawa, Javier Jimenez The causally important flow features of wall turbulence are studied by measuring the effect of perturbing local flow structures in direct numerical simulations of turbulent channel flow at Reτ=600. The flow is initially perturbed by substituting the velocity in a cubic cell with its mean velocity, and the perturbed and original flow fields are allowed to develop for a given time. The causal significance of the perturbation is quantified by the growth of the L2 norm of the velocity difference between the perturbed and original flows. To avoid pre-assumptions as much as possible, the experiment is repeated many times, changing the location, the size of the perturbation cell and the initial flow field. The causal flow features are then studied from the ensemble of these realizations, which in this study includes 76320 simulation experiments. The rate of growth of the perturbations and their propagation normal to the wall show that the causal effect amplifies fastest when the perturbation hits the wall, where the energy production is high due to the mean shear. Which properties determine whether flow structures are more or less causal depend on the distance from the wall of the initial perturbations. High and low shear are found to be more and less causal, respectively, for near-wall perturbations, whereas sweeps and ejections are causal and less causal for perturbations farther from the wall. The growth of causal and less causal perturbations scales well when time is normalized by the local mean shear. This suggests that the causal effect of the flow structures depend on the total energy production that the structures experience during their lifetime. |
Sunday, November 20, 2022 5:53PM - 6:06PM |
J26.00007: Onsager Theory of Momentum Cascade in Wall-Bounded Turbulence Hao Quan, Gregory L Eyink Onsager analyzed energy cascade exactly for individual flow realizations at infinite Reynolds number by spatial filtering. We study momentum cascade in wall-bounded turbulence by filtering out eddies of size h to the wall. We consider interior flows through channels and exterior flows past solid bodies, with surfaces either hydraulically rough or smooth. We show that skin friction τw and surface pressure pw exist at the wall in the infinite-Re limit. But both space-time fields are obtained also by taking Re?∞ first and then matching to spatial flux of momentum as h,l?0. This deterministic “momentum cascade” does not generally occur in the traditional inertial sublayer, e.g. in a rough-wall pipe it occurs for h < roughness height. When the inviscid-limit velocity satisfies no-penetration at the wall, momentum flux vanishes and τw=0. Then only form drag remains for Re?∞ and wall pressure is obtained by taking the limit to the surface of the pressure of inviscid Euler solutions in the fluid interior. As an application, we show that Lighthill’s theory of vorticity generation holds even for Re?∞.Our analysis motivates novel LES methods for wall-bounded turbulence and applies also at finite Reynolds numbers. |
Sunday, November 20, 2022 6:06PM - 6:19PM |
J26.00008: Generalizing the Concept of a Surface Layer as Wall-Modulated Eddies Samantha J Sheppard, James G Brasseur, John A Farnsworth, J. Christos Vassilicos, Pierre Braganca, Christophe Cuvier The classical description of the surface layer (SL), often called the log layer, is the inertia-dominated wall-adjacent region within the high Reynolds number turbulent boundary layer (TBL) in which integral scales scale linearly on the distance from surface z and where law-of-the-wall scaling applies. We generalize the SL concept to the wall-adjacent inertia-dominated region in any wall-bounded turbulent flow with vertical turbulence fluctuations w' directly modified by surface impermeability and causing integral scales involving w' to scale on z. We further explore the hypothesis that coherent motions can be separated into those that are directly modulated by the surface and those that are not. To do so, we compare analysis of time-resolved PIV data from two sets of wind tunnel experiments: (1) grid turbulence from three different grids interacting with an impermeable flat plate and (2) the classical TBL. In both flows we successfully identify SL regions characterized by integral scales that grow linearly with z and we apply conditional wavelet filtering to extract coherent wall-modulated eddy structures and quantify their contributions to SL statistics. |
Sunday, November 20, 2022 6:19PM - 6:32PM |
J26.00009: Finite-Reynolds-number asymptotics of wall-turbulence fluctuations Xi Chen, Katepalli R Sreenivasan As a generalization of our earlier work (Chen and Sreenivasan, JFM 2021 and 2022---hereafter referred to as CS), we present a perspective on finite-Reynolds-number asymptotics for turbulence mean profiles (including the variance of fluctuations of streamwise and spanwise velocities as well as pressure) in turbulent channels, pipes and flat-plate boundary layers. When normalized by the peak values, the near-wall turbulence profiles are characterized by single point functions independent of the friction Reynolds number. Utilizing the defect power-law given in CS, it is shown that all the data share the same inner scaling form. Using a matching scheme between the above inner scaling and the celebrated outer flow similarity, it is shown that bounded decay is in very good agreement with all current data in the three wall flows, and more importantly, that it leads to a finite plateau in the bulk for asymptotically high friction Reynolds numbers. |
Sunday, November 20, 2022 6:32PM - 6:45PM |
J26.00010: Wall turbulence at $Re_\tau=10k$: Kinematics and Symmetry scaling laws Sergio Hoyas, Martin Oberlack A new direct numerical simulation of a Poiseuille channel flow has been conducted for a friction Reynolds number of 10 000, extending a recent work published in PRL/PRF. Therein, we have shown using symmetry theory that the logarithmic law of the mean streamwise velocity and a related power-law for arbitrarily high moments is a valid solution to the multi-point moment equations. The DNS confirmed a long region of validity extending from 400 to 2500 wall units. Similarly, symmetry theory also gave rise to a related power-law scaling of arbitrary moments of the streamwise velocity in the center/deficit region of the channel. The DNS data also confirmed it. The symmetry theory has now been extended to scaling the moments of the wall-normal and spanwise velocities both in the near-wall and center region. Our most recent DNS data confirm this. About the flow's kinematics, the maximum intensity of the streamwise velocity increases with the Reynolds number, as was thought. The collapse of the turbulent budgets is perfect above 25 wall units. About the well-known scaling failure of the dissipation, we have computed the value of the near-wall gradient of u' for several Reynolds numbers. Our data confirm this value's relationship to the |
Sunday, November 20, 2022 6:45PM - 6:58PM |
J26.00011: Lagrangian 3D particle tracking for turbulent flow in rapid contraction Abdullah A Alhareth, Kenneth Langley, Vivek Mugundhan, Nathan B Speirs, Sigurdur T Thoroddsen The present work discusses the experimental investigation of a turbulent flow through rapid contraction using the Lagrangian particle tracking approach “Shake-The-Box” method. The experimental setup was created based on the work of (Mugundhan et al., 2020), it comprises an active bi-planar grid with flaps that can be rotated in either synchronous or random mode with a max rotational speed of 210 rpm and the maximum Reynolds number based on the Taylor microscale Reλ ≈220. The characteristics of the turbulent flow inside the 2D and 3D contractions with contraction ratios of 4:1 and 16:1 respectively were investigated. To resolve the flow field, 50,000 time-resolved volumetric velocity and vorticity fields were used with 136,000 tracked particles. A total of 45 volumes we obtained at different streamwise locations within the contraction to understand the evolution of vortical structures along with the contraction. The progression of the RMS fluctuations was compared against existing literature to understand the influence of the contractions in the alignment and amplification of coherent vertical structures. Additionally, the contraction and stretching of vertical structures in the YZ plane, due to the extensional strain in the streamwise direction. |
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