Turbophoresis of inertial particles in inhomogeneous turbulence generated by oscillating grids

arXiv:2605.03646 · physics.flu-dyn, physics.ao-ph, physics.geo-ph · Submitted 2026-05-05 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.

Kai: We're looking at the paper: "Turbophoresis of inertial particles in inhomogeneous turbulence generated by oscillating grids".

Mira: Here's the gist of the paper. Turbophoresis in inhomogeneous turbulent flows leads to the formation of large-scale nonuniform particle number density distributions of inertial particles This effect is associated with an effective drift velocity directed toward regions of lower turbulence intensity It depends on the Stokes and Reynolds numbers, as well as on the gradient of the turbulence intensity In this study, turbophoretic transport is experimentally investigated in air turbulent flows generated by onegrid and two-grid oscillating systems The measurements show preferential accumulation of inertial particles in regions of minimum mean-square turbulent velocity.

Kai: Experimental Setup and Measurement The experiments were performed in isothermal inhomogeneous turbulence generated by one and two oscillating grids in air flows The chamber dimensions were Lx = Lz = 26 cm and Ly = 53 cm Each oscillating grid had dimensions of 25 × 25 cm2 and consisted of a square array of bars with a mesh size of 5 cm and a bar width of 1 cm The grids oscillated along the Y direction with frequency f = 10.5 Hz and stroke amplitude Ag = 6 cm The laser light sheet was located in the central vertical Y Z plane at X = 0 The inertial particles used were borosilicate hollow glass particles with an approximately spherical shape.

Mira: Flow Characterization and Turbulence Metrics The velocity field was measured using a Particle Image Velocimetry (PIV) system The mean velocity field and turbulent characteristics in the Y Z plane were analyzed Key turbulent characteristics obtained include the turbulent velocity u(rms) = ⟨u2y⟩ + ⟨u2z⟩1/2, the ratio between turbulent and mean velocities u(rms)/U, and the anisotropy parameter Au = u(rms)z /u(rms)y − 1.

Kai: The integral length scales ly and lz were determined from two-point velocity correlation functions.

Paper summary: Kai: We've just covered the general idea and what this paper is testing with these oscillating grids. The focus here is on how they set up the experiment and what they are measuring to prove the accumulative effect.

Mira: Right, so before we get into how they got those results, it helps to know their setup. They used air flows created by one and two oscillating grids in a chamber that was twenty-six centimeters by fifty-three centimeters by twenty-six centimeters.

Lev: That gives us a sense of the scale of the turbulence they are generating. The dimensions—twenty six by fifty three by twenty six—that sets the boundary conditions for what kind of flow we're looking at here.

Kai: Each grid was twenty five times twenty five centimeters, and they had a mesh size of five centimeters with a bar width of one centimeter, and these grids oscillated along the Y direction at a frequency of ten point five hertz with an amplitude of six centimeters.

Mira: And they placed the laser light sheet in the central vertical Y Z plane at an X coordinate of zero to get their measurements. They used borosilicate hollow glass particles that are approximately spherical in shape for their inertial particles.

Lev: That's good detail on the particle itself, because the size and shape of those particles really matter when you think about how they interact with the fluid motion described by this paper.

Kai: Next, they characterized the flow using a Particle Image Velocimetry system to measure both the mean velocity field and key turbulent characteristics in that Y Z plane.

Mira: They calculated several things like the turbulent velocity rms, which is defined as u squared y plus u squared z to the power of half, and then they look at that ratio of turbulent to mean velocities.

Lev: And they also found an anisotropy parameter, Au, which measures how different the turbulence is in the Y direction versus the Z direction. That tells you if it's isotropic or if there's a preferred direction in the turbulence structure.

Kai: They also determined integral length scales, l y and l z, by looking at two point velocity correlation functions to get an idea of those large-scale structures.

Conclusion: Mira: So, to wrap up the summary of this paper titled "Turbophoresis of inertial particles in inhomogeneous turbulence generated by oscillating grids," the main finding is that these particles cluster where the turbulence intensity is lowest.

Kai: The authors found that the ratio of their mean particle number densities, n in divided by nn in, which they call nin over nnin, is approximated by u squared Y times the mean square velocity times a constant kappa turboph τ zero minus alpha kappa turboph.

Lev: That equation shows how strongly it depends on the turbulence itself—the u squared Y term being key to that dependence.

Mira: And what this means for us is that this accumulation isn't just some minor boundary effect; it’s a bulk effect caused by those large-scale gradients of turbulence intensity across the flow field.

Kai: They explicitly state the spatial distribution doesn't show localization near the boundaries; instead, it follows the large scale variations of turbulence intensity, which is consistent with what they see in direct numerical simulations.

Mira: The normalized inertial-particle concentration increases in regions of reduced turbulence intensity, which matches what we expect from the turbophoretic trend seen in other models.

Lev: And they give us a terminal fall velocity for particles with a diameter of ten micrometers, and that’s about zero point three three centimeters per second, which is significantly smaller than both the typical mean fluid velocities and the turbulent ones.

Kai: The limitation they mention is that they didn't use these measurements to calculate a specific quantitative value for the turbophoretic coefficient kappa turboph itself; they just compared their observed spatial trend to previous theory and direct numerical simulations.

Mira: So, while this provides strong experimental evidence of large-scale turbophoretic particle accumulation in inhomogeneous turbulence, it doesn't give us a number for how strong the force is driving it yet.

Lev: The method itself relies on measurements with relatively short acquisition times and ensemble averaging over many independent realizations to get those density distributions.

Kai: This experimental setup gives us a concrete picture of how these particle drifts manifest in real air turbulence generated by oscillating grids, setting a clear benchmark for future theoretical work trying to quantify this effect.

The Pearlstone Center for Aeronautical Engineering Studies · Department of Mechanical Engineering, The Ben-Gurion University of the Negev

physics.flu-dyn, physics.ao-ph, physics.geo-ph

Submitted: 2026-05-05

Updated: 2026-11-10

Comments: 27 pages, 16 figures, revtex4-2, revised paper

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 79/100

The gist: Turbophoresis in inhomogeneous turbulent flows leads to the formation of large-scale nonuniform particle number density distributions of inertial particles This effect is associated with an effective…

Key concepts

Turbophoresis
This is the process where inertial particles are pushed or pulled by gradients in the surrounding fluid's turbulence. In this case, it causes particles to drift toward regions of lower turbulence intensity.
Inhomogeneous Turbulence
This refers to turbulent flows where the intensity (how 'chaotic' or energetic) changes significantly across space. The experiment used oscillating grids to create such a flow with varying levels of turbulence.
Particle Number Density Distribution
This describes how the concentration of inertial particles is spread out throughout the flow area. The researchers found that this distribution shifts, concentrating particles where the turbulence intensity is minimal.
Mean-Square Turbulent Velocity ($u(rms)$)
This metric quantifies the overall turbulent motion in a specific plane (Y Z). It helps measure how energetic and chaotic the fluid movement is at different locations within the experimental chamber.

Terminology

Summary

Turbophoresis in inhomogeneous turbulent flows leads to the formation of large-scale nonuniform particle number density distributions of inertial particles This effect is associated with an effective drift velocity directed toward regions of lower turbulence intensity It depends on the Stokes and Reynolds numbers, as well as on the gradient of the turbulence intensity In this study, turbophoretic transport is experimentally investigated in air turbulent flows generated by onegrid and two-grid oscillating systems The measurements show preferential accumulation of inertial particles in regions of minimum mean-square turbulent velocity. p. 0

Where the paper says it

il; Abstract Turbophoresis in inhomogeneous turbulent flows leads to the formation of largescale nonuniform particle number density distributions of inertial particles. This effect is associated with an effective drift velocity directed toward regions of lower turbulence intensity.

Experimental Setup and Measurement

The experiments were performed in isothermal inhomogeneous turbulence generated by one and two oscillating grids in air flows The chamber dimensions were Lx = Lz = 26 cm and Ly = 53 cm Each oscillating grid had dimensions of 25 × 25 cm2 and consisted of a square array of bars with a mesh size of 5 cm and a bar width of 1 cm The grids oscillated along the Y direction with frequency f = 10.5 Hz and stroke amplitude Ag = 6 cm The laser light sheet was located in the central vertical Y Z plane at X = 0 The inertial particles used were borosilicate hollow glass particles with an approximately spherical shape. p. 6

Where the paper says it

In the second approach, the integral length scale is determined from an exponential fit of the form exp(−r 2/ℓ2 0 ) to the normalized correlation function. For experimental study of turbophoresis, borosilicate hollow glass particles with an approximately spherical shape are used as inertial particles.

Flow Characterization and Turbulence Metrics

The velocity field was measured using a Particle Image Velocimetry (PIV) system The mean velocity field and turbulent characteristics in the Y Z plane were analyzed Key turbulent characteristics obtained include the turbulent velocity u(rms) = ⟨u2y⟩ + ⟨u2z⟩1/2, the ratio between turbulent and mean velocities u(rms)/U, and the anisotropy parameter Au = u(rms)z /u(rms)y − 1. The integral length scales ly and lz were determined from two-point velocity correlation functions. p. 6

Where the paper says it

The coordinates Y and Z are normalized by Lz = 26 cm. The integral length scales of the turbulence ℓy and ℓz in the horizontal Y and vertical Z directions are determined from two-point velocity correlation functions.

Turbophoretic Transport Analysis

The spatial distribution of inertial particles was determined by equation for the particle number density n(t, x) The steady-state solution for the mean particle number density given by Eq. (6) was used to obtain the ratio of the mean number densities nin and nnin. p. 5

Where the paper says it

The mean velocity and rms velocity profiles are also recalculated using subsets of the recorded data, and they are compared with the final profiles obtained from all 530 velocity maps. The profiles obtained from the later subsets and from the full data set showed only small variations, indicating that the statistical estimates are stable with respect to the number of velocity maps used in the averaging.

This ratio is approximated by nin(Y) / nnin ≈ u2(Y)⟨u2⟩max! p. 2
Where the paper says it

The DNS results also show that the product κturboph and urms ≡ p ⟨u2⟩ increases linearly with St for St ≪ 1, reaches a maximum near St ∼ 10, and decreases as St−1/3 for large St. Here, St = τp/τ0 is the Stokes number defined as a ratio of the Stokes time τp and the characteristic turbulent time scale τ0 associated with the integral turbulence scale.

− a∗κturboph τ0 The results demonstrated that inertial particles preferentially accumulate in regions of lower turbulence intensity. p. 3
Where the paper says it

These configurations differ in the spatial structure of turbulence intensity, anisotropy, integral length scales, characteristic turbulent time, and Reynolds-number distribution. The experiments show that in both turbulent-flow configurations, inertial particles preferentially accumulate in regions of lower turbulence intensity.

Conclusions on Accumulation Mechanism

The observed particle accumulation is primarily a bulk effect due to turbophoresis caused by large-scale gradients of turbulence intensity rather than by boundary effects The spatial distribution of inertial particles does not exhibit localization near the boundaries but instead follow the large-scale variations of turbulence intensity The normalized inertial-particle concentration increases in regions of reduced turbulence intensity, consistent with the turbophoretic trend observed in DNS. The terminal fall velocity for particles with diameter 10 µm is approximately 0. p. 1

Where the paper says it

Turbophoresis originates from averaging of particle velocity. In particular, the mean particle velocity U (p) is given by U (p) = U + V turboph + Vg [17, 18], where U is the mean fluid velocity, Vg = τp g is the terminal fall velocity of inertial particles in the gravity field, and V turboph = −κturboph ∇ u 2 (2) is the turbophoretic velocity of inertial particles in an inhomogeneous turbulence, and κturboph is the turbophoretic coefficient, which depends on the Stokes and Reynolds numbers.

33 cm/s, which is substantially smaller than both the typical mean and turbulent fluid velocities. p. 5
Where the paper says it

In particular, the mean velocity field, mean velocity shear, root-mean-square (r. ) velocities, turbulent anisotropy parameter, two-point velocity correlation functions and integral turbulence length scales are obtained.

Future Directions

The present results provide experimental evidence of large-scale turbophoretic particle accumulation in strongly inhomogeneous turbulence A broad parametric experimental study of turbophoresis is the subject of a separate future investigation. The authors did not use the present measurements to infer a quantitative value of the turbophoretic coefficient κturboph. p. 1

Where the paper says it

Turbophoresis originates from averaging of particle velocity. In particular, the mean particle velocity U (p) is given by U (p) = U + V turboph + Vg [17, 18], where U is the mean fluid velocity, Vg = τp g is the terminal fall velocity of inertial particles in the gravity field, and V turboph = −κturboph ∇ u 2 (2) is the turbophoretic velocity of inertial particles in an inhomogeneous turbulence, and κturboph is the turbophoretic coefficient, which depends on the Stokes and Reynolds numbers.

The comparison with previous theory and DNS is therefore made at the level of the observed spatial trend. The particle number density distributions are obtained from measurements with relatively short acquisition times and ensemble averaging over many independent realizations. p. 3
Where the paper says it

Oscillating-grid turbulence provides a controlled laboratory configuration with relatively weak mean flow and spatially varying turbulence intensity, allowing direct investigation of particle accumulation induced by turbophoresis. Simultaneous measurements of turbulent velocity fields and particle number density distributions allow direct comparison between the spatial distribution of inertial particles and the turbulence intensity.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Incorporate a model for turbophoretic transport based on Equation (7) into particle tracking simulations to predict large-scale accumulation patterns in inhomogeneous turbulence. This allows AI systems to simulate how inertial particles preferentially accumulate in regions of lower turbulence intensity, consistent with the theoretical estimate given by Eq. (7).

  2. Develop an image processing module that utilizes the normalized relative particle-density fields as a direct indicator of regions dominated by turbophoretic transport rather than mean fluid flow or boundary effects. This capability enables AI systems to distinguish between different particle transport mechanisms in experimental data, as noted by the authors' approach: "To isolate turbophoretic particle accumulation from other possible transport mechanisms, the scattered-light intensity Ein measured for inertial particles is normalized pointwise by the scattered-light intensity Enin obtained in separate experiments with noninertial particles under identical flow conditions."

  3. Enhance turbulence characterization capabilities to estimate key flow inhomogeneity parameters, such as the Reynolds number Re, directly from spatial distributions of turbulent velocity and integral length scales. This enables AI systems to rapidly quantify the spatial gradients of turbulence intensity responsible for turbophoresis, which is crucial for understanding the competition between the effective turbophoretic velocity and turbulent diffusion.

Abstract

Turbophoresis in inhomogeneous turbulent flows leads to the formation of large-scale nonuniform particle number density distributions of inertial particles. This effect is associated with an effective drift velocity directed toward regions of lower turbulence intensity. It depends on the Stokes and Reynolds numbers, as well as on the gradient of the turbulence intensity. In the present study, turbophoretic transport is experimentally investigated in air turbulent flows generated by one-grid and two-grid oscillating systems. The flow velocity field is measured using Particle Image Velocimetry, and the particle spatial distribution is obtained by applying image processing techniques. To isolate the effect of particle accumulation due to turbophoresis from that associated with mean fluid flow, the measured particle number density of inertial particles is normalized by the corresponding distribution obtained for noninertial tracer particles under identical flow conditions. The measurements show preferential accumulation of inertial particles in regions of minimum mean-square turbulent velocity, consistent with the expected behavior of turbophoretic transport.

Related papers