This paper explores integrating artificial intelligence (AI) segmentation models, particularly the Segment Anything Model (SAM), into fluid mechanics experiments. SAM’s architecture, comprising an image encoder, prompt encoder, and mask decoder, is investigated for its application in detecting and segmenting objects and flow structures. Additionally, we explore the integration of natural language prompts, such as BERT, to enhance SAM’s performance in segmenting specific objects. Through case studies, we found that SAM is robust in object detection in fluid experiments. However, segmentations related to flow properties, such as scalar turbulence and bubbly flows, require fine-tuning. To facilitate the application, we have established a repository (https://github.com/AliRKhojasteh/Flow_segmentation) where models and usage examples can be accessed.

@en

This study investigates the turbulent/non-turbulent interface (TNTI) in self-similar turbulent axisymmetric jet flows, focusing on a novel approach named ’move with the flow’ where the image acquisition is space-based rather than time-based. Experiments were conducted at three different Reynolds numbers (9000, 12000, and 31000), utilizing particle image velocimetry (PIV) and laser-induced fluorescence (LIF) techniques. The core of this research was developing a traverse system specifically designed to follow the evolving flow structures at the TNTI synchronously. We succeeded in tracking large-scale events in TNTI from creation to dissolution.@en

Molecular tagging is used to study the dispersion and deformation of patterns written in turbulent air. The writing is done by fusing O2 and N2 molecules into NO in the focus of a strong ultraviolet laser beam. By crossing several of these laser beams, patterns that have both small and large scales can be painted. The patterns are visualized a while later by inducing fluorescence of the NO molecules with a second UV laser and registering the image. The width of the lines that make the pattern is approximately 50μm, a few times the Kolmogorov length η, the smallest length scale in turbulence, while the largest size of the patterns (≈4mm) is inside the inertial range of the used turbulent jet flow. At small scales molecular clouds disperse under the joint action of molecular diffusion and turbulence. The experiments reveal this highly nontrivial interaction. At inertial-range scales (≈200η) we verify the Batchelor dispersion of objects whose size is inside the inertial range. Patterns are compressible objects and spontaneously develop concentration fluctuations. We show for the first time the nontrivial statistical properties of these fluctuations. Finally, we use the information in written and deformed lines to quantify turbulent intermittency, obtaining results that agree with the established scaling anomaly of velocity structure functions.

@en

Spectral line shape models can successfully reproduce experimental Rayleigh-Brillouin spectra, but they need knowledge about the bulk viscosity ηb. Light scattering involves GHz frequencies, but since ηb is only documented at low frequencies, ηb is usually left as a free parameter, which is determined by a fit of the model to an experimental spectrum. The question is whether models work so well because of this freedom. Moreover, for light scattering in air, spectral models view "air"as an effective molecule. We critically evaluate the use of ηb as a fit parameter by comparing ηb obtained from fits of the Tenti S6 model to the result of Direct Simulation Monte Carlo (DSMC) for a mixture of Nitrogen and Oxygen. These simulations are used to compute light scattering spectra, which are then compared to experiments. The DSMC simulation parameters are cross-checked with a molecular dynamics simulation based on intermolecular potentials. At large values of the uniformity parameter y, y ≈ 4, where the Brillouin contribution to spectra is large, fitted ηb are 20% larger than the ones from DSMC, while the quality of the simulated spectra is comparable to that of the Tenti S6 line shape model. At smaller y, the difference between fitted and simulated ηb can be as large as 100%. We hypothesize the breakdown of the bulk viscosity concept to be the cause of this fallacy.

@en

In our experiment a vortical flow behind a traveling plate turns into turbulence. By exactly repeating this experiment 42 times with a robot, we study the statistics of this transition. In each realization the fate of the flow is followed over 1.7 s when the plate travels with a constant velocity. It suddenly turns turbulent at a scaled traveled distance of x∗≈5.5. We register the vorticity in a plane that divides the plate perpendicularly. We introduce an original Lagrangian measure of variability between the experiment realizations. The finite-time Lyapunov exponent field of a single experiment predicts this variability; thus we confirm ergodicity. Apart from pointwise measures, yielding a distribution over the field of view, we study the statistics of the circulation computed over the upper and lower half of the domain. The almost perfect symmetry both of the mean and of the fluctuations points to their origin as the fluctuating vortex ring trailing behind the plate. During the initial phase long-time correlations exist in the flow, but they cease once the flow turns turbulent. By ordering our repeated experiments we find that extreme circulations are preceded by circulations that are larger than the median.

@en

I comment on the recent article Grüneisen approach for universal scaling of the Brillouin shift in gases by Liang et al (2022 New J. Phys. 24 103005). While the result of this article indeed provides a parametrization of measured Rayleigh Brillouin scattering spectra, I argue that the association with the Grüneisen effect is fortuitous. The Grüneisen effect is the change of the specific heat with compression of a gas. For the nontrivial case of the van der Waals gas, it predicts the decrease of the velocity of sound with increasing pressure. In solids, which is the context considered in the article by Liang et al, the effect is opposite: an increase of the velocity of sound with pressure.

@en

We study the dispersion of tiny molecular clouds in turbulence by writing patterns in turbulent air and following their deformation in time. The writing is done by fusing O2 and N2 molecules into NO in the focus of a strong ultraviolet laser beam. By crossing several of these laser beams, patterns that have both small and large scales can be painted. The patterns are visualized a while later by inducing fluorescence of the NO molecules with a second UV laser and registering the image. The width of the lines that make the pattern is approximately 50 μm, a few times the Kolmogorov length η, the smallest length scale in turbulence, while the overall size of the patterns (≈4 mm) is inside the inertial range of the used turbulent jet flow. At small scales molecular clouds disperse under the joint action of molecular diffusion and turbulence. The experiments reveal for the first time this subtle, yet very important interaction. At macroscales (≈200 η) we verify the Batchelor dispersion of objects whose size is inside the inertial range; however, the expected influence of molecular diffusion is smaller than the accuracy of the experiments.

@en

Abstract: Our quest is for the thumb and finger positions that maximize drag in front crawl swimming and thus maximize propulsion efficiency. We focus on drag in a stationary flow. Swimming is in water, but using Reynolds similarity the drag experiments are done in a wind tunnel. We measure the forces on real-life models of a forearm with hands, flexing the thumb and fingers in various positions. We study the influence on drag of cupping the hand and flexing the thumb. We find that cupping the hand is detrimental for drag. Swimming is most efficient with a flat hand. Flexing the thumb has a small effect on the drag, such that the drag is largest for the opened (abducted) thumb. Flow structures around the hand are visualized using robotic volumetric particle image velocimetry. From the time-averaged velocity fields we reconstruct the pressure distribution on the hand. These pressures are compared to the result of a direct measurement. The reached accuracy of ≈ 10% does not yet suffice to reproduce the small drag differences between the hand postures. Graphical Abstract: [Figure not available: see fulltext.].

@en

A scalar emanating from a point source in a turbulent boundary layer does not mix homogeneously, but is organized in large regions with little variation of the concentration: uniform concentration zones. We measure scalar concentration using laser-induced fluorescence and, simultaneously, the three-dimensional velocity field using tomographic particle image velocimetry in a water tunnel boundary layer. We identify uniform concentration zones using both a simple histogram technique, and more advanced cluster analysis. From the complete information on the turbulent velocity field, we compute two candidate velocity structures that may form the boundaries between two uniform concentration zones. One of these structures is related to the rate of point separation along Lagrangian trajectories and the other one involves the magnitude of strong shear in snapshots of the velocity field. Therefore, the first method allows for the history of the flow field to be monitored, while the second method only looks at a snapshot. The separation of fluid parcels in time was measured in two ways: The exponential growth of the separation as time progresses (related to finite-Time Lyapunov exponents and unstable manifolds in the theory of dynamical systems), and the exponential growth as time moves backward (stable manifolds). Of these two, a correlation with the edges of uniform concentration zones was found for the past Lyapunov field but not with the time-forward future field. The magnitude of the correlation is comparable to that of the regions of strong shear in the instantaneous velocity field.

@en

We study the relation between large-scale structures in the concentration field with those in the velocity field in a dye-seeded turbulent jet. The scalar concentration in a plane is measured using laser-induced fluorescence. Uniform concentration zones of an advected scalar are indentified using cluster analysis. We simultaneously measure the two-dimensional velocity field using particle image velocimetry. The structures in the velocity field are characterized by finite-time Lyapunov exponents. The measurement of the scalar- and velocity fields moves with the mean flow. In this moving frame, turbulent structures remain in focus long enough to observe well-defined ridges of the finite-time Lyapunov field. This field gauges the rate of point separation along Lagrangian trajectories; it was measured both for future and past times since the instant of observation. The edges of uniform concentration zones are correlated with the ridges of the past-time Lyapunov field, but not with those of the future-time Lyapunov field. We quantify this relation using both conditional averages and the ordinary correlation function.

@en

We use Faraday waves to measure interfacial tension σ between two immiscible fluids, with an interest in (ultra)low values of σ. The waves are excited by vertically oscillating the container in which the fluids reside. Using linear stability theory, we map out the accessible range of interfacial tensions. The smallest value (σmin ≈ 5 × 10-4 N/m) is limited by the joint influence of gravity and viscous dissipation. A further limitation is posed by the greatest accelerations that can be realized in a laboratory. We perform experiments on a water-dodecane interface with an increasing concentration of a surfactant in the water layer that decreases the interfacial tension into the ultralow domain [σ = [Formula: see text](10-6 N/m)]. Surprisingly, the smallest measured wavelength is larger by a factor of 2 than that predicted for vanishing σ. We hypothesize the effect of transport of the surfactant in the fluid flow associated with the waves.

@en

We present a new technique to study preferential concentration of droplets in a turbulent air flow. Preferential concentration is the tendency of droplets to cluster in regions of strain, while avoiding regions of rotation. We study the properties of the droplet concentration field in zero mean flow turbulence that was created using an array of synthetic jets. The droplets are made of a phosphorescent solution of Europium chelate. They are excited by a laser sheet from a pulsed UV laser, after which the glowing droplets are followed using a high-speed intensified camera. We quantify preferential concentration through measurement of moments of the coarse-grained local droplet density. At the Stokes numbers studied (St≈2) the fractal dimension, a scaling property of this coarse-grained density field, points to clustering. Clustering is a consequence of the compressibility of the droplet velocity field. We also quantify the dynamical behavior of clustering by moving with this velocity field. We find a preference for clustering in the Lagrangian frame during the time interval set by the decay of the phosphorescence.

@en

Abstract: We discuss the application of synthetic aperture particle image velocimetry for measuring the flow around human swimmers using small bubbles as tracer. We quantify the two-dimensional projection of the velocity field in planes perpendicular to the viewing direction of an array of six cameras. With help of simulations, modelled after the experiment, we address questions about depth selectivity and occlusion in dense bubble fields. Using vortex rings in the swimming pool, we provide a proof of principle of the method. It is further illustrated by the vorticity field produced by a human swimmer. Graphic abstract: [Figure not available: see fulltext.].

@en

Rayleigh-Brillouin scattering spectra of CO ₂ were measured at pressures ranging from 0.5 to 4 bars and temperatures from 257 to 355 K using green laser light (wavelength 532 nm, scattering angle of 55.7°). These spectra were compared to two line shape models, which take the bulk viscosity as a parameter. One model applies to the kinetic regime, i.e., low pressures, while the second model uses the continuum, hydrodynamic approach and takes the rotational relaxation time as a parameter, which translates into the bulk viscosity. We do not find a significant dependence of the bulk viscosity with pressure or temperature. At pressures where both models apply, we find a consistent value of the ratio of bulk viscosity over shear viscosity η _b /η _s = 0.41 ± 0.10. This value is four orders of magnitude smaller than the common value that is based on the damping of ultrasound and signifies that in light scattering only relaxation of rotational modes matters, while vibrational modes remain "frozen."

@en

The flow field around a full-scale swimmer’s hand model with varying thumb positions is investigated by robotic volumetric PIV. The experiment is conducted in the Open Jet Facility wind tunnel at TU Delft at 15m/s. Quantitative flow field information is constructed with 3D-PTV in a 120 liter volume, encompassing the full hand and arm. The effect of spatial resolution on the time-averaged flow field is investigated. A large-scale recirculating wake behind the hand is accurately identified at a linear bin size of 20mm whereas the accelerated flow between individual fingers can only be resolved at bin sizes below 10mm where 5mm results in a statistically unconverged velocity field. The influence of the thumb is limited to one side of the hand where its presence results in a larger stagnated region in front and larger wake behind the hand, depending on the thumb position. Closing the thumb strengthens the recirculation but results in a smaller velocity deficit downstream, suggesting a smaller propulsive force generation which is considered disadvantageous in competitive swimming. @en

Rayleigh-Brillouin light scattering spectra of CO2 at ultraviolet wavelengths are computed from molecular dynamics which depends on intermolecular potentials only. We find excellent agreement with state of the art experimental data. This agreement was reached in a minimal computational box with sides one scattering wavelength long and integrating the classical trajectories over 20 ns. We also find complete consistency with models based on kinetic theory, which take known values of the transport coefficients as input.@en

We present results on the analysis of measurements on the dispersion of a contaminant in a turbulent boundary layer. The contaminant is a fluorescent dye (Rh.-B) that was introduced in a turbulent boundary layer (푅푅푒푒# = 3050) in a water channel. The velocity field was measured by means of tomographic PIV. Simultaneously, the concentration field was measured by means of laser-induced fluorescence (LIF). The LIF data clearly reveal uniform concentration zones with well-defined boundaries. These zones are extracted from the probability density function of the local concentration. The key question is what flow structures make up their boundaries, and how these uniform concentration zones are related to uniform momentum zones. We evaluate the Lyapunov exponents of the rate of separation of two close points while they are advected by the flow. Local maxima of the Lyapunov exponents are known to create temporary barriers for transport in two-dimensional flows. Alternatively, the boundaries of uniform concentration zones may also be formed by shear layers, which are local maxima of the shear vorticity. There is a striking concentration of both the Lyapunov exponent and the shear vorticity on real boundaries. The correlation with the Lyapunov field is the weakest of the two. However, the Lyapunov exponent is a Lagrangian average, extending over episodes where these boundaries were formed, whereas the shear vorticity is based on a planar cross section of the instantaneous velocity field. @en

The effect of finger spread on overall drag on a swimmer’s hand is relatively small, but could be relevant for elite swimmers. There are many sensitivities in measuring this effect. A comparison between numerical simulations, experiments and theory is urgently required to observe whether the effect is significant. In this study, the beneficial effect of a small finger spread in swimming is confirmed using three different but complementary methods. For the first time numerical simulations and laboratory experiments are conducted on the exact same 3D model of the hand with attached forearm. The virtual version of the hand with forearm was implemented in a numerical code by means of an immersed boundary method and the 3D printed physical version was studied in a wind tunnel experiment. An enhancement of the drag coefficient of 2% and 5% compared to the case with closed fingers was found for the numerical simulation and experiment, respectively. A 5% and 8% favorable effect on the (dimensionless) force moment at an optimal finger spreading of 10° was found, which indicates that the difference is more outspoken in the force moment. Moreover, an analytical model is proposed, using scaling arguments similar to the Betz actuator disk model, to explain the drag coefficient as a function of finger spacing.@en

Understanding the power balance of a swimmer, who needs to overcome power losses to drag and to water set in motion, requires detailed insight into the hydrodynamics of the flow around the swimmer. This will be done from a hydrodynamic point of view with techniques familiar from fluid mechanics. One of our objectives is to develop a system to visualize the flow around a swimmer in practice. Currently, the visualisation system is designed and built and we will present the underlying ideas of the experimental setup.

@en