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 X40: Turbulent Wall Bounded Flow III |
Hide Abstracts |
Chair: George Park, University of Pennsylvania Room: 355 F |
Tuesday, November 26, 2024 8:00AM - 8:13AM |
X40.00001: Estimating near-wall turbulence using adjoint-variational data assimilation Sugan Durai Murugan Velazhagan, Mengze Wang, Tamer A Zaki Experimental measurements of near-wall turbulence are challenging to acquire. Using adjoint-variational data assimilation (4DVar), we augment incomplete measurements that do not sample the viscous or buffer layer. The time series is divided into a sequence of shorter assimilation horizons, and 4DVar is performed within each window to estimate the flow within the gap, between the first measurement point and the wall, at full spatial resolution. The accuracy of the estimation is first verified using synthetic measurements from an independent direct numerical simulation. When the full system is simulated and the first measurement plan is near y+= 50, the predictions of the wall shearstresses and pressure are accurate. Spectral analysis of the errors in the estimation is interpreted in terms of the influence, or lack thereof, of near-wall eddies on the observation in the core of the channel.We then proceed to examine the case where the assimilation is performed in truncated sub-domains of the channel, where the boundary conditions are unknown. This configuration is designed to model application to experimental measurements, where we wish to predict the instantaneous wall shear stresses and pressure at high Reynolds numbers within truncated simulation domains. |
Tuesday, November 26, 2024 8:13AM - 8:26AM |
X40.00002: A data-driven quasilinear approximation for supersonic turbulent channel flow Zecheng Zou, Yongyun Hwang This work models the compressibility effects in supersonic turbulent channel flow using the approach of data-driven quasi-linear approximation (DQLA) (Holford & Hwang, J. Fluid Mech., vol. 980, 2024, p.A12) proposed for incompressible channel flow. Full nonlinear equations are considered for time-averaged mean, and the fluctuations are linearised around the mean with the nonlinear term consisting of an eddy-viscosity model and stochastic forcing. The streamwise weight of the stochastic forcing is determined by matching the velocity spectra from DNS of incompressible flow and that produced by the eddy viscosity enhanced linearised compressible Navier-Stokes equations. The spanwise weights are determined self-consistently, so that mean density, momentum and temperature equations are satisfied with their prescribed properties. The proposed compressible data-driven quasi-linear approximation demonstrates the capability to produce turbulent intensity profiles and energy spectra that exhibit qualitatively similar behavior across the entire wall-normal domain as DNS data up to Ma=3, where Ma is the Mach number. |
Tuesday, November 26, 2024 8:26AM - 8:39AM |
X40.00003: Multi-sensors source inversion algorithm based on remote measurement and domain of dependence Zejian You, Qi Wang, Xiaowei Zhu Identification of passive scalar sources from remote measurements raised serious challenges due to the combined effect of molecular diffusion and turbulent dispersion. To enable fast identification, the adjoint operator is adopted to exploit the sensitivity of remote measurements to source locations. This sensitivity is also coined as the measurement's domain of dependence (DoD). Our work utilizes measurements and the DoDs of a multi-sensor system to identify the source location and provide its confidence interval swiftly. The algorithm is tested in a turbulent channel flow with Reynolds number $Re_\tau=180$. The forward-adjoint duality relation shows that the time-averaged measurement and mean adjoint field are related by a proportionality equal to the source strength, where uncertainty is embedded in the finite window of length $T$ used for the time average. We quantify this averaging uncertainty represented by Gaussian random variables and add to the duality relation, which is then used to estimate the confidence interval of the source location. The standard deviation of the Gaussian variable scales with $1/\sqrt{T}$, and depends on the wall-normal location of the source-sensor pair. These trends are presented and discussed in the current research. |
Tuesday, November 26, 2024 8:39AM - 8:52AM |
X40.00004: Ensemble-averaged domain of dependence of wall stress in turbulent channel flow Qi Wang, Tamer A Zaki Estimating past events from wall measurements in turbulent channel flow, by aid of the governing equations, is a challenging and ill-posed problem. To quantify the extent to which we can determine the earlier flow state from wall data, the adjoint Navier-Stokes equations are solved backward in time, initialized by the measurements kernel. These adjoint fields define the spatio-temporal domain-of-dependence (DOD) of the measurement. In turbulent channel flow, the energy of the adjoint field amplifies exponentially in backward time at the Lyapunov exponent, while the ensemble average of adjoint realizations decays. Slow convergence of the ensemble motivates a Reynolds-averaged approach. We adopt a linear eddy-viscosity model to close the forward-adjoint correlation term in the adjoint RANS equations. As such, the averaged DOD can be computed efficiently, and compares favorably with the ensemble approach. We additionally compare the DOD to the domain of influence, the later computed from the linearized forward RANS equations. We discuss the implications of these findings for data assimilation and information propagation in wall turbulence. |
Tuesday, November 26, 2024 8:52AM - 9:05AM |
X40.00005: Local Eddy Structure Underlying the Linear Growth in Integral Scale in Two Different Wall-Bounded Turbulent Flows Samantha Janvier Sheppard, James Brasseur, John A. Farnsworth, John Christos Vassilicos Two classes of wall-bounded turbulent flows are considered, each of which produce surface layers characterized by linear scaling normal to the wall (z) of the horizontal correlation length of wall-normal (“vertical”) fluctuations, w': lww,x(z). Quantifications from two distinctly different sets of wind tunnel experiments are compared: 1) a canonical flat plate turbulent boundary layer with a shear-dominated surface layer, and 2) a class of shear-free wall-bounded surface layers created by advecting inertia-dominated grid turbulence at moderate Reynolds numbers over a flat plate. The surface layer exists both in the presence and absence of mean shear-rate and only integral length scales of w' increase linearly with distance from the surface. This indicates that the surface layer is generated from the blockage of vertical fluctuations by the impermeable surface. Analysis of instantaneous turbulence structures through conditional sampling was conducted to reveal the local structure and distribution of concentrations of vertical velocity fluctuations within the individual samples that underly the statistically identified linear increase in lww,x with z, both in the presence and absence of mean shear-rate. We contrast the quantified structure of the wall modulated w' “eddies” with the general form proposed by Townsend (1976) in context with “attached eddy” spectral representations (e.g., Perry, Henbest & Chong, 1986). |
Tuesday, November 26, 2024 9:05AM - 9:18AM |
X40.00006: Eulerian techniques for Lagrangian investigation of wall-bounded turbulent flows George I Park, Imran Hayat The Lagrangian viewpoint, following fluid particles, is effective for understanding turbulent diffusion, which is fundamental to mixing, dispersion, entrainment, and turbulence modeling. However, Lagrangian studies of turbulence are sparse due to the high cost and complexity of tracking exceedingly large number of space-filling tracer particles. In this talk, we showcase our recent work with novel Eulerian approaches that allow computation of key Lagrangian kinematics/statistics without explicit particle tracking. First, the reference map (RM) technique, originally developed for Eulerian treatment of solid mechanics, replaces particle tracking by returning initial tracer locations sampled at grid points and updating them through the Eulerian transport equation. The RM allows computation of Lagrangian coherent structures (finite-time Lyapunov Exponent, FTLE) and material surface tracking. Second, we compute Lagrangian mean fields (averaged along tracer trajectories) by solving their exact transport equations. We apply these techniques to turbulent channel flow at Reτ=180. Suprisingly, Largrangian mean fields computed over one eddy turnover time contain spatial features finer than the instantaneous counterparts. The Lagrangian covariance field is highly correlated with FTLE. We also demonstrate how the RM can be leveraged to compute ensemble particle statistics such as particle displacement variance. |
Tuesday, November 26, 2024 9:18AM - 9:31AM |
X40.00007: ABSTRACT WITHDRAWN
|
Tuesday, November 26, 2024 9:31AM - 9:44AM |
X40.00008: Traveling waves in soap-film turbulent channel flow Chien-chia Liu The periodic vorticity structures emerging from a phase-specific average of the vorticity fluctuation provide crucial information about the underlying flow dynamics. Here it reports the identification of periodic patterns, by averaging over a fixed phase, in wall-bounded quasi-two-dimensional soap-film random turbulence, revealing the traveling wave nature of turbulence. |
Tuesday, November 26, 2024 9:44AM - 9:57AM |
X40.00009: High Spatial Resolution Measurement of Turbulent Channel Flow Using 2C-2D PIV and 3C-3D Digital Holographic PIV Bihai Sun, Callum Atkinson, Julio Soria Experimental measurements of turbulent channel flow have historically lagged behind direct numerical simulations (DNS). However, they remain essential for verifying simulations that involve surface treatments, such as micro-texturing of the channel walls. This presentation reports on high-resolution measurements of turbulent channel flow using 2-component – 2-dimensional (2C-2D) particle image velocimetry (PIV) and 3-component – 3-dimensional (3C-3D) Digital Holographic PIV (DHPIV). |
Tuesday, November 26, 2024 9:57AM - 10:10AM |
X40.00010: Understanding the dynamics of near-wall turbulence using network motif identification Emma Lenz, Ahmed Elnahhas, Giovanni Iacobello, H. Jane Bae A graph-theoretic approach is used to identify and quantify key dynamical processes in wall-bounded turbulence. Classically, wall-bounded turbulence is considered to have a universal near-wall inner region where all velocity statistics are dependent only on the distance from the wall. However, as the Reynolds number and domain size increase, dynamic interactions between structures of different wall-normal sizes (inter-scale interactions) or different instances of structures of the same size (intra-scale interactions), modify this view. In this work, we aim to extract the signature of these dynamics and quantify their significance. This analysis is performed using direct numerical simulation data of a minimal flow unit (MFU) at Reτ = 180, a larger-domain channel at Reτ = 180 to capture the effect of intra-scale interactions of near-wall structures, and a channel at Reτ = 550 to capture the effect of inter-scale interactions with flow further from the wall. A proper orthogonal decomposition of the Reτ = 180 MFU is performed to identify energetic structures, and the energy captured by the structures is tracked over time, forming a temporal trajectory. Temporal patterns of high significance, known as network motifs, are then extracted from the trajectories. Similar analyses are performed on MFU-sized boxes embedded in the larger-domain simulations at Reτ = 180 and Reτ = 550 to understand the difference in the dynamics. |
Tuesday, November 26, 2024 10:10AM - 10:23AM |
X40.00011: Effects of Off-axis Rotation on Energy Transfer Mechanism of the Isotropic Turbulence Yijie Wang, Jun Chen, Leonardo Chamorro Turbulent flows in rotational systems have significant applications across diverse fields, spanning from large scales such as climate and oceanic systems to small scales such as turbomachinery. Previous experimental studies have primarily focused on homogenous turbulence and the rotational effect on the mean flow, while computational research tried to modify the existing turbulent models developed in inertial coordinates. There is a lack of understanding of the rotational effects on the energy transfer mechanism of turbulence as well as turbulent models directly developed in non-inertial coordinates. With PIV measurements applied on a zero-mean isotropic turbulent flow, the anisotropy between radial, tangential, and axial direction, induced by the presence of centrifugal force and Coriolis force, is investigated with a variety of statistics of turbulence. A Coriolis transportation mechanism is identified as a TKE transfer between radial and tangential directions, which may break the isotropic state of turbulence. This Coriolis transportation is quantified with PIV data experimentally and compared with other TKE terms to evaluate the rotational effect as well as provide a new perspective for modeling rotating turbulence. |
Tuesday, November 26, 2024 10:23AM - 10:36AM |
X40.00012: Unsteady effects and relaminarization in compressible channel flows Mateo Landazuri, Diego A. Donzis Because of their complexity and bigger parameter space, compressible wall-bounded flows are less understood than their incompressible counterpart. This is even more so when flows also experience temporal changes. In this work, we study unsteady effects in compressible turbulent channels using well-resolved Direct Numerical Simulation (DNS) for a range of Reynolds and Mach numbers. In particular, we investigate relaminarization processes and time-dependent scaling laws. Starting from a steady flow sustained by a constant body force in the streamwise direction, the forcing is cut-off and the flow is allowed to decay. As the turbulent kinetic energy dissipates, different measures of compressibility are tracked and related to quantities of practical relevance such as wall-friction and widely used compressible scaling laws for mean velocity and Reynolds stresses. Relaminarization and unsteady effects are assessed by comparing snapshots of the decaying flow with dynamically equivalent steady simulations. Preliminary results show deviations of the flow properties in the turbulent-to-laminar transition region but good agreement with both laminar solutions and turbulent incompressible approximations. Detailed studies of turbulent properties unveil the rapid shrinking of turbulent structures leading to negligible turbulent stresses and relaminarization. The possibility and evidence of hysteresis as a consequence of unsteadiness is discussed in light of the DNS data. |
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