Dispersed two-phase flows at air–water interfaces are ubiquitous in environmentally relevant flows such as in the dispersion of floating microplastics or transport processes across the air–sea interface. In the current study, we propose a method to study such flows through the study of a relatively flat turbulent free surface laden with spherical floating particles (“floaters”). The free surface is perturbed by a relatively low-mean nearly homogeneous subsurface turbulent flow that is produced in a turbulence box actuated by a 10×10 synthetic jet array. The free surface flow field is characterized using planar particle image velocimetry (PIV) simultaneously with Lagrangian tracking of floaters allowing insight into the floater dynamics and the surface flow coupling. This is enabled by a relatively simple setup of LED panels and a single camera. Distinction between the continuous (flow tracers) and the dispersed (floaters) phase is carried out by exploiting their size disparity and number density. The proposed method is employed to characterize the single-phase flow field and the clustering statistics of floaters for different turbulence levels, the latter achieved by varying the distance of the free surface from the jet array. Specifically, we study the effect of different turbulence levels on the floater clustering behavior. We observe that the time required for floaters to reach a clustered quasi-steady state decreases with increasing vorticity and surface divergence amplitude. In addition, the growth rate of the mean cluster size is observed to increase with increasing vorticity and surface divergence amplitude, with its temporal evolution exhibiting two distinct phases: an agglomeration phase and an equilibrium phase. In contrast, in the absence of a subsurface flow, floaters are observed to cluster at a relatively slower rate characterized by a prolonged agglomeration phase. Finally, to highlight the potential of this technique in studying floater-laden turbulent free surfaces, preliminary results of flow–floater interactions are discussed.

Segmenting gas bubbles in multiphase flows is a critical yet unsolved challenge in numerous industrial settings, from metallurgical processing to maritime drag reduction. Traditional approaches — and most recent learning-based methods — assume near-spherical shapes, limiting their effectiveness in regimes where bubbles undergo deformation, coalescence, or breakup. This complexity is particularly evident in air lubrication systems, where coalesced bubbles form amorphous and topologically diverse patches. In this work, we revisit the problem through the lens of modern vision foundation models. We cast the task as a transfer learning problem and demonstrate, for the first time, that a fine-tuned Segment Anything Model (SAM v2.1) can accurately segment highly non-convex, irregular bubble structures using as few as 100 annotated images.

The turbulent boundary layer (TBL) development over an air cavity is experimentally studied using planar particle image velocimetry. The present flow, representative of those typically encountered in ship air lubrication, resembles the geometrical characteristics of flows over solid bumps studied in the literature. However, unlike solid bumps, the cavity has a variable geometry inherent to its dynamic nature. An identification technique based on thresholding of correlation values from particle image correlations is employed to detect the cavity. The TBL does not separate at the leeward side of the cavity owing to a high boundary layer thickness to maximum cavity thickness ratio (δ/t_max = 12). As a consequence of the cavity geometry, the TBL is subjected to alternating streamwise pressure gradients: from an adverse pressure gradient (APG) to a favourable pressure gradient and back to an APG. The mean streamwise velocity and turbulence stresses over the cavity show that the streamwise pressure gradients and air injection are the dominant perturbations to the flow, with streamline curvature concluded to be marginal. Two-point correlations of the wall-normal velocity reveal an increased coherent extent over the cavity and a local anisotropy in regions under an APG, distinct from traditional APG TBLs, suggesting possible history effects.

Increasing utilization of ocean space and a global push for renewable energy solutions has spurred interest in wave behavior around Very Large Floating Structures, like floating photovoltaic (PV) systems. Flexible PV modules may be more suitable for the varying wave conditions found in offshore environments. However, while viscoelastic models are commonly used for wave prediction, they show notable discrepancies with experiments, likely due to untested assumptions of inviscid flow. This experimental study aims to fill that gap by investigating both the wave characteristics and velocity fields underneath flexible and rigid structures using simultaneous Particle Image Velocimetry (PIV) and wave elevation measurements. Wave attenuation is observed for short wavelengths over the flexible structure length. The 2nd order Stokes wave theory provides a good approximation of the wave-induced horizontal velocity profiles under the flexible structure but underestimates the velocities under the rigid one which further lacks the typical exponential decay with water depth. The presence of a wave boundary layer is showcased and compared to an adaptation of the Stokes 2nd problem.

Cambridge University Press apologise for an error with the supplementary material of the above article. Additional materials from an unrelated article were erroneously published alongside the intended supplementary material. This has been corrected.

The effect of the streamwise velocity structuring on air lubrication is experimentally assessed in a flat plate turbulent boundary layer. High-speed, planar particle image velocimetry is performed on a wall-parallel plane within the logarithmic region upstream of the air injection, concurrently with high-speed shadowgraphy of the air phase immediately downstream of it. Three different air phase regimes are evaluated (bubbly, transitional, air layer), all deeply embedded in the log region and for a constant liquid freestream velocity. Air bubbles formed downstream of low-speed streaks are found to be larger and to convect much slower than those formed downstream of high-speed events. The spanwise variation of air coverage and its time evolution follow the spanwise structuring and evolution of log-region residing streaks for both the bubbly and transitional regimes, with low-speed streaks promoting air coverage and high-speed ones inhibiting it, a correlation stronger closer to the injector but remaining significant for large downstream distances (x>4δ). High-speed streaks are also shown to be largely responsible for liquid pockets developing close to the injector in the air layer regime, locally breaking its spanwise continuity.

Different air phase regimes are formed by controlled air injection in a spatially developing flat plate turbulent boundary layer (TBL). The air is introduced via a slot type injector without the use of a backward-facing step or cavitator upstream of the air injection position. The effect of different incoming liquid flow characteristics on the different regimes is investigated by varying both the liquid freestream velocity and the incoming TBL thickness. The latter is realized through changing the position of the air injection along the length of the water tunnel facility. That resulted in a downstream distance based Reynolds number from 1 to 5 million. Three different air phase regimes are identified under different air flow rates and freestream velocities: the bubbly regime, the transitional, and the air layer regime. The morphological differences of each one are described and quantitative analysis is performed based on the non-wetted area in each condition. The incoming TBL as well as the flow around the air layer are measured with planar particle image velocimetry. The latter enabled the determination of the air layer thickness. In addition, the ratio of the air layer to the incoming boundary layer thickness t_air/δ is also calculated (≈ 0.04 – 0.5). This is a significant dimensionless parameter for scaling, which indicates the extent to which the air layer is embedded within the incoming TBL. Depending on the incoming flow conditions, a two or three branch air layer is formed. The length of the air layer is found to increase with increasing liquid freestream velocities. A good agreement between the air layer length and a half gravity wave predicted by the dispersion relation is found. An increase of the air layer length is observed with a decreasing incoming TBL thickness. This is attributed to a decrease in the local mean velocity at the air–water interface due to the TBL growth. Finally, increasing the incoming TBL thickness delays the onset of the air layer regime.

The probability density function (PDF) of the instantaneous streamwise velocity has consistently been used to extract information on the formation of uniform momentum zones (UMZs) in wall-bounded flows. Its temporal evolution has previously revealed patterns associated with the geometry and amplitude of the underlying velocity fluctuations [Laskari and McKeon, J. Fluid Mech. 913, A6 (2021)0022-112010.1017/jfm.2020.1163]. In this paper, we examine the robustness of these patterns in a variety of data sets including experiments and wall-bounded flow models. Experimental data sets spanning a range of Reynolds numbers, with very long temporal and spatial domains, suggest that the rate of the observed temporal variations scales in inner units. The use of a convection velocity, uniform across heights, to transform space into time has a marginal effect on these features. Similarly, negligible effects are observed between internal and external geometries. Synthetic databases generated following the resolvent framework and the attached eddy model are employed to draw comparisons to the experimental databases. Our findings highlight the distinctive strengths of each: The broadband frequency input of the attached eddy model allows for a better statistical description as opposed to a narrow frequency input in the resolvent data sets; instantaneously, however, representative eddies are seen to lack some structural details leading to the observed temporal behavior, which is better replicated by resolvent modes. Overall, given the considerable variety of the input data tested, the agreement between the data sets highlights the robustness of the spatiotemporal characteristics of the examined UMZs. It also underpins the need for their proper inclusion in UMZ modeling from a statistical as well as an instantaneous viewpoint; the current analysis accentuates important performance indicators for both.

An air layer within a liquid turbulent boundary layer (TBL) is formed by controlled air injection underneath a flat plate. The incoming boundary layer as well as the flow around the air layer were measured with planar particle image velocimetry (PIV). The effect of different incoming liquid flow characteristics on the air layer geometry is investigated by varying both the freestream velocity and the streamwise development length of the TBL. The latter was realized through changing the position of the air injection along the length of the water tunnel facility. Increasing the freestream velocity resulted in an increase of the air layer length, while its maximum thickness remained relatively unaltered. An increase in the TBL development length, had a similarly marginal effect on the resulting maximum air layer thickness but led to a shorter air layer length. The latter could be attributed to a decrease in local mean velocity due to the TBL growth, reflected in a decrease of the air layer to boundary layer thickness ratio (from 0.27 to 0.17). The results of this study are expected to provide insight on the design conditions of an air layer drag reduction system installed in the hull of a ship.

Small spatial perturbations grow into fingers along the unstable interface of a fluid displacing a more viscous fluid in a porous medium or a Hele-Shaw cell. Mitigating this Saffman-Taylor instability increases the efficiency of fluid displacement applications (e.g., oil recovery), whereas amplifying these perturbations is desirable in, e.g., mixing applications. In this work, we investigate the Saffman-Taylor instability through analysis and experiments in which air injected with an oscillatory flow rate outwardly displaces silicone oil in a radial Hele-Shaw cell. A solution for linear instability growth that shows the competing effects of radial growth and surface tension, including wetting effects, is defined given an arbitrary reference condition. We use this solution to define a condition for stability relative to the constant flow rate case and make initial numerical predictions of instability growth by wave number for a variety of oscillations. These solutions are then modified by incorporating reference conditions from experimental data. The morphological evolution of the interface is tracked as the air bubble expands and displaces oil between the plates. Using the resulting images, we analyze and compare the linear growth of perturbations about the mean interfacial radius for constant injection rates with and without superimposed oscillations. Three distinct types of flow rate oscillations are found to modulate experimental linear growth over a constant phase-averaged rate of fluid displacement. In particular, instability growth at the interface is mildly mitigated by adding to the base flow rate provided by a peristaltic pump a second flow with low-frequency oscillations of small magnitude and, to a lesser extent, high-frequency oscillations of large amplitude. In both cases, the increased stability results from the selective suppression of the growth of large wave numbers in the linear regime. Contrarily, intermediate oscillations consistently destabilize the interface and significantly amplify the growth of the most unstable wave numbers of the constant flow rate case. Numerical predictions of low-frequency oscillations of opposite sign (initially decreasing) show promise of even greater mitigation of linear instability growth than that observed in this investigation. Looking forward, proper characterization of the unsteady, wetting, and nonlinear dynamics of instability growth will give further insight into the efficacy of oscillatory injection rates.

Velocity structures associated with large streamwise density gradients in an incompressible turbulent boundary layer (with air as the working fluid, ) are analysed experimentally using planar image velocimetry and aero-optic measurements. The resulting flow topologies for the velocity fluctuations associated with large negative and positive density gradients are in excellent agreement with results for coolings and heatings in time, respectively (Antonia & Fulachier, J. Fluid Mech., vol. 198, 1989, pp. 429-451). The current results are complimentary to those from Saxton-Fox et al. (AIAA J., vol. 57 (7), 2019, pp. 2828-2839), on the signature of the vertical velocity structures associated with large density gradients. In the present work, these structures are shown to exhibit a sign change, consistent with the scalar gradient, and are localised in the wall-normal direction with an average height of approximately , almost constant for increasing distance from the wall. The corresponding small-scale streamwise fluctuations also exhibit a consistent sign change, which is found to originate, on average, from upstream leaning structures. The emerging picture for the velocity field is then that of a multiscale phenomenon, where small-scale structures, responsible for large optical aberrations, are superimposed on the back of large-scale bulge-like structures that are known to populate the outer layers. The proposed conceptual model is consistent with early ideas of 'typical' eddies (Falco, Phys. Fluids, vol. 20 (10), 1977, pp. S124-S132), the hairpin vortex model and associated shear layers (Adrian et al., J. Fluid Mech., vol. 422, 2000, pp.1-54), as well as with notions of multiscale velocity organisation in shear layers (Klewicki & Hirschi, Phys. Fluids, vol. 16 (11), 2004, pp. 4163-4176; Saxton-Fox et al. 2019), and it provides new insight into the geometry of the small-scale velocity structures.

Abstract: In this study, we quantify the accuracy of a simple pressure estimation method from 2D snapshot PIV in attached and separated flows. Particle image velocimetry (PIV) offers the possibility to acquire a field of pressure instead of point measurements. Multiple methods may be used to obtain pressure from PIV measurements, however, the current state-of-the-art requires expensive equipment and data processing. As an alternative, we aim to quantify the efficacy of estimating instantaneous pressure from snapshot (non-time resolved) two-dimensional planar PIV (the simplest type of PIV available). To make up for the loss of temporal information, we rely on Taylor’s hypothesis (TH) to replace temporal information with spatial gradients. Application of our approach to high-resolution 2D velocity data of a turbulent boundary layer flow over ribs shows moderate to good agreement with reference pressure measurements in average and fluctuations. To assess the performance of the 2D TH method beyond average and fluctuation statistics, we acquired a time-resolved measurement of the same flow and determined temporal correlation values of the pressure from our method with reference measurements. Overall, the correlation attains good values for all measured locations. For comparison, we also applied two time-resolved approaches, which attained values of correlation similar to our approach. The performance of the 2D TH method is further assessed on 3D time-resolved velocity data for a turbulent boundary layer and compared with 3D methods. The root-mean-square (RMS) pressure fluctuations of the 2D TH, 3D TH and 3D pseudo-Lagrangian methods closely follow the pressure fluctuation distribution from DNS. These observations on the RMS pressure estimates are further supported by similar analysis on synthetic PIV data (based on DNS) of a turbulent channel flow. The values of spatial correlation between the 2D TH method and the DNS pressure fields in this case, are similar to the temporal correlations achieved in the turbulent flow over the ribs. Finally, we discuss the accuracy of instantaneous pressure estimates and provide a rule of thumb to determine regions where the pressure fluctuation estimate from the 2D TH methods is likely to fail.

Planar particle image velocimetry (PIV) measurements were performed in a thermal boundary layer with simultaneous measurements of the streamwise density gradient, averaged across the boundary layer height. Previous work on the topic (Saxton-Fox et al., 2019) revealed that large density gradients were associated with tall wall-normal velocity structures extending across the entire boundary layer height. The authors suggested that these structures were a result of a wall-normal average of smaller scales, which reside at different locations from the wall, and contribute to the same density gradient change. Results from the present work support this model and show that, if a second condition on the vertical location of the identified features is imposed, one can extract wall-normal velocity structures that are localized in the vertical direction. Their average wall-normal extent is found to be around 0.1δ, with no significant variation across the boundary layer height. The corresponding conditionally averaged streamwise velocity fluctuations also show a change in sign from the wall up to edge of the boundary layer, consistent with the underlying density gradient condition. This sign change is found to originate on average from upstream leaning structures of opposite signs, while the inferred flow topology is shown to agree with earlier results on the topic (Antonia and Fulachier, 1989).

Time-resolved planar particle image velocimetry was used to analyse the structuring of a turbulent boundary layer into uniform momentum zones (UMZs). The instantaneous peak-detection method employed by Adrian et al. (J. Fluid Mech., vol. 422, 2000, pp. 1-54) and de Silva et al. (J. Fluid Mech., vol. 786, 2016, pp. 309-331) is extended to account for temporal coherence of UMZs. The resulting number of zones detected appears to follow a normal distribution at any given instant. However, the extreme cases in which the number of zones is either very high or very low, are shown to be linked with two distinct flow states. A higher than average number of zones is associated with a large-scale event in the log region which creates increased small-scale activity within that region. Conversely, a low number of zones corresponds to a large-scale event in the log region and decreased turbulent activity away from the wall. The residence times, within the measurement plane, of zones belonging to the latter scenario are shown to be on average four times larger than those of zones present during higher than average zone structuring states. For both cases, greater residence times are observed for zones of higher momentum that are generally closer to the free stream.

This paper deals with pressure estimation from snapshot and time-resolved three-component (3C) volumetric PIV data using Taylor’s hypothesis, an Eulerian and a pseudo-Lagrangian approach. The Taylor’s hypothesis approach has been shown to provide accurate results for pressure in the case of 3C planar PIV data with an appropriate choice of convection velocity (de Kat and Ganapathisubramani 2013), and here we extend its use on 3C volumetric velocity snapshots. Application of the techniques to synthetic data shows that the Taylor’s hypothesis approach performs best using the streamwise mean as the convection velocity and is affected the least by noise, while the Eulerian approach suffers the most. In terms of resolution, the pseudo-Lagrangian approach is the most sensitive. Its accuracy can be improved by increasing the frame time-separation when computing the material derivative, at the expense of volume loss from fluid parcels leaving the FOV. Comparison of the techniques on turbulent boundary layer data with DNS supports these observations and shows that the Taylor’s hypothesis approach is the only way we can get pressure when time information is not present.