We study the dynamics of a fixed wing at stall in a wind tunnel by measuring the aerodynamic forces. We report experimental evidence of a critical Reynolds number from which low-frequency oscillations in the force are replaced by random bistable dynamics. In this new regime, the flow explores each state intermittently with long residence times. This stochastic process can be modeled as a continuous Markov chain, and equivalently, it shows a superexponential scaling for the mean residence times. Furthermore, the probability density function of the lift coefficient exhibits the characteristic heavy tail of extreme events. Extreme minima and maxima are at the origin of the transitions. We analyzed the evolution of these tails using extreme-value theory to identify the bifurcation points of the associated dynamical system. The results are in good agreement with a user-defined threshold method, the advantage being the unambiguity in the computation.

Tomographic particle image velocimetry measurements of turbulent swirling jets open new explorations of mean and pulse components and pattern dynamics. We present direct experimental evidence of turbulent helicity cascade from large scales to small scales. We map the distributions of mean and fluctuating helicity components—the revealing anisotropic structure of helical jet flow. We show that a scenario of helicity generation and decay along the jet depends dramatically on the inflow swirl. The obtained results provide valuable materials to benchmark the modern numerical simulations with a turbulent closure technique.

Supervised super-resolution deep convolutional neural networks (CNNs) have gained significant attention for their potential in reconstructing velocity and scalar fields in turbulent flows. Despite their popularity, CNNs currently lack the ability to accurately produce high-frequency and small-scale features, and tests of their generalizability to out-of-sample flows are not widespread. Generative adversarial networks (GANs), which consist of two distinct neural networks (NNs), a generator and discriminator, are a promising alternative, allowing for both semisupervised and unsupervised training. The difference in the flow fields produced by these two NN architectures has not been thoroughly investigated, and a comprehensive understanding of the discriminator's role has yet to be developed. This study assesses the effectiveness of the unsupervised adversarial training in GANs for turbulence reconstruction in forced homogeneous isotropic turbulence. GAN-based architectures are found to outperform supervised CNNs for turbulent flow reconstruction for in-sample cases. The reconstruction accuracy of both architectures diminishes for out-of-sample cases, though the GAN's discriminator network significantly improves the generator's out-of-sample robustness using either an additional unsupervised training step with large eddy simulation input fields or a dynamic selection of the most suitable upsampling factor. These enhance the generator's ability to reconstruct small-scale gradients, turbulence intermittency, and velocity-gradient probability density functions. Conversely, the supervised super-resolution CNN network lacks the capability to reconstruct these statistics. The extrapolation capability of the GAN-based model is demonstrated for out-of-sample flows at higher Reynolds numbers. Based on these findings, incorporating discriminator-based training is recommended to enhance the reconstruction capability of super-resolution CNNs.

We are carrying out the experimental study of the emptying of a hermetically sealed cylindrical tank in a rotating frame. We determine the optimal rotational velocity for the tank, which minimizes the emptying time. Moreover, there exists a critical rotation velocity at which the emptying time exceeds that of the nonrotating case, depending on the relative duration of each of the three emptying regimes identified in this flow, namely, (i) a bubble regime, (ii) a vortical air jet regime, and (iii) a bathtub vortex regime. The three regimes, easily distinguished by the shape of the air-water interface, have different draining velocities. The lifetime of each regime, as well as the transitions between them, depend on both the water level and the rotating velocity of the platform.

Many important practical applications involving porous media, cosmetics, biological systems, and food processing involve the transport of non-Newtonian fluids, which possess nonlinear material properties. Elastoviscoplastic fluids are indeed a complex example, simultaneously involving viscous, elastic, and plastic properties. In this study, we conduct numerical simulations of elastoviscoplastic fluids using a hybrid lattice Boltzmann solver in order to investigate the impact of plasticity, characterized by the Bingham number, on inertialess viscoelastic instabilities at high Weissenberg numbers. Results obtained using the four-roll mill and cellular-forcing scheme benchmark cases, which produce a strong elongational flow regime, reveal the emergence of three distinct flow states over time, namely a gradual shift from a laminar steady state to periodic orbits and ultimately evolving into aperiodic flow fluctuations during the late stages. The transition and behavior between these different flow states are found to strongly depend on the interplay between elasticity and plasticity. We demonstrate that the general effect of the Bingham number is to increase the unyielded regions in the fluid, which although appears initially contained in the vortical regions, naturally emerges in the vicinity of the polymer birefringent strands over time. The eventual effect is to laminarize and suppress the flow fluctuations of the viscoelastic instability in the late stages. Ultimately, this work demonstrates the impact of plasticity on the already complex inertialess viscoelastic instabilities that develop in the presence of strong elongational flow regimes, wherein the results provide an avenue for controlling the instability mechanism.

The flow of yield stress fluids in porous media presents interesting complexity due to the interplay between the nonlinear rheology and the heterogeneity of the medium. An important consequence is that the number of flow paths increases with the applied pressure difference and is responsible for a nonlinear Darcy law. Previous studies have focused on the protocol where the pressure difference is imposed. Here we consider instead the case of imposed flow rate, $Q$. In contrast to Newtonian fluids, the two types of boundary conditions have an important influence on the flow field. Using a two-dimensional pore network model we observe a boundary length of merging flow paths of size $\ell \left(Q\right)\sim Q^{-\mu /\delta }$ where $\mu =0.42\pm 0.02$ and $\delta \simeq 0.63\pm 0.05$. Beyond this length the density of the flow paths is homogeneous and grows as $Q^{\mu }$. Using a mapping to the directed polymer model we identify $\delta$ with the roughness exponent of the polymer. We also characterize the statistics of nonflowing surfaces in terms of avalanches pulled at one end.

Understanding the receptivity of hypersonic flows to free-stream disturbances is crucial for predicting laminar to turbulent boundary layer transition. Input-output analysis as a receptivity tool considers which free-stream disturbances lead to the largest response from the boundary layer using the global linear dynamics. Two technical challenges are addressed. First, we restrict the allowable forcing to physically realizable inputs via a free-stream boundary modification to the classic input-output formulation. Second, we develop a hierarchical input-output (H-IO) analysis which allows us to solve the three-dimensional problem at a fraction of the computational cost otherwise associated with directly inverting the fully three-dimensional resolvent operator. Next, we consider Mach 5.8 flows over a sharp cone and two blunt cones with $3.6\mathrm{mm}$ and $7.2\mathrm{mm}$ spherically blunt tips. H-IO correctly predicts that the sharp cone boundary layer is most receptive to slow acoustic waves at an optimal incidence angle of ${10}^{\circ }$, validating the method. We then investigate the effect of free-stream disturbances on the blunt cone boundary layer and identify two distinct vorticity-dominated receptivity mechanisms for the oblique first-mode instability at $10\mathrm{kHz}$ and an entropy layer instability at 40 and $70\mathrm{kHz}$. Our results reveal these receptivity processes to be highly three-dimensional in nature, involving both the nose tip and excitation along narrow bands at certain azimuthal angles along the oblique shock downstream. We interpret these processes in terms of critical angles from linear shock/perturbation interaction theory. Finally, we show how these receptivity processes vary with frequency and nose tip bluntness, and demonstrate how this methodology might be applied to transition prediction from first principles.

We examine the nonequilibrium nature of two-phase fluid displacements in a quasi-two-dimensional medium (a model open fracture) in the presence of localized constrictions (“defects”) from a theoretical and numerical standpoint. Our analysis predicts the capillary energy dissipated in abrupt interfacial displacements (jumps) across defects, and relates it to the corresponding hysteresis cycle, e.g., in pressure-saturation. We distinguish between “weak” (reversible interface displacement, exhibiting no hysteresis and dissipation) and “strong” (irreversible) defects. We expose the emergence of dissipation and irreversibility caused by spatial interactions, mediated by interfacial tension, among otherwise weak defects. We exemplify this cooperative behavior for a pair of weak defects and establish a critical separation distance, analytically and numerically, verified by a proof-of-concept experiment.

Most of the literature on flow through nanoporous two-dimensional membranes has focused on static membranes under quasiequilibrium conditions. However, various studies have shown the relevance of fluid-structure interactions—particularly dynamic coupling—on flow through nanopores and nanotubes. A fundamental understanding of the effects of membrane oscillations could provide the missing link to the development of ultraefficient nanofluidic systems for water desalination, gas separation, and other applications. Herein, we use molecular dynamics simulations to study the effects of rigid in-plane harmonic pore oscillations on water flow through nanoporous graphene. First, we repurpose a used technique as a framework to avoid injecting arbitrary amounts of heat into the interfacial water, thereby isolating the physical mechanisms caused by the pore's movement. Then we demonstrate that the considered pore oscillations enhance flow by increasing the average axial velocity and decreasing the average density of water inside the pore due to the dynamic opening and closing of flow routes inside the pore. Finally, we show how these effects manifest themselves in relevant fluid properties for both single-file or collective-diffusion and frictional flows.

We perform fully coupled numerical simulations using immersed boundary methods of finite-size spheres and fibers suspended in a turbulent flow for a range of Taylor-Reynolds numbers $12.8<{\mathrm{Re}}_{\lambda }<442$ and solid mass fractions $0\le M\le 1$. Both spheres and fibers reduce the turbulence intensity with respect to the single-phase flow at all Reynolds numbers, with fibers causing a more significant reduction than the spheres. The particles' effect on the anomalous dissipation tends to vanish as $\mathrm{Re}\to \infty$. A scale-by-scale analysis shows that both particle shapes provide a “spectral shortcut” to the flow, but the shortcut extends further into the dissipative range in the case of fibers. Multifractal spectra of the near-particle dissipation show that spheres enhance dissipation in two-dimensional sheets, and fibers enhance the dissipation in structures with a dimension greater than one and less than two. In addition, we show that spheres suppress vortical flow structures, whereas fibers produce structures which completely overcome the turbulent vortex stretching behavior in their vicinity.

Volumetric velocimetry is used to investigate the evolution of tip vortices in the wake of a model wind turbine. The power spectral density of streamwise velocity fluctuations is estimated at several downstream locations, and changes in the magnitude of the frequency peaks corresponding to the blade-pass and turbine rotational frequency are observed. Full volumetric visualization of the coherent structures depicts the vortex merging process responsible for the change in detected frequency peaks, capturing the physical mechanism likely responsible for many similar previous observations.

We investigate the behavior of a high-Reynolds-number polydisperse gravity current propagating along a channel of general cross-section form given by a width function $f\left(z\right)$ into a linearly stratified ambient fluid of density ${\rho }_a\left(z\right)$. The current of density ${\rho }_c$ is formed by $n$ types of particles of various densities and settling velocities suspended in an interstitial fluid of density ${\rho }_i$. We formulate shallow-water equations of motion and then solve the partial differential equation system of hyperbolic type by the Lax-Wendroff two-step finite-difference method. We present typical profiles for height ($h$) and velocity ($u$) of the current and mass concentration ($\mathrm{\Phi }$) of the particles. We show that initially the front of the current propagates with an almost constant speed. During the next stage the height and the speed of the nose decrease, which leads to the pseudosimilar final stage of propagation. The solutions are illustrated for flow in typical power-law geometry. The problem introduces two dimensionless parameters: (1) the stratification parameter $S$ ($0\le S\le 1$), which represents the magnitude of the stratification in the ambient fluid, and (2) the particle buoyancy parameter $\mathrm{\Pi }$ ($\mathrm{\Pi }\ge 0$). We show that increasing $S$ decreases the velocity of the propagation of the current. The effect of $\mathrm{\Pi }$ is the opposite: as $\mathrm{\Pi }$ increases, the current propagates faster. For a specific dependence between $S$ and $\mathrm{\Pi }$, an equilibrium is reached for a significant time and the system behaves like a system without particles propagating into the ambient of constant density.

We study the problem of drag reduction (DR) in a lubricated conduit, in which a thin layer of low-viscosity (e.g., water) fluid is injected in the near-wall region and facilitates the transport of a core of high-viscosity fluid (e.g., oil). In the present investigation, the flow instance is a channel flow, and consequently we have one thin layer of low-viscosity fluid lubricating each wall. We run direct numerical simulations of this flow instance, respecting the protocol of the constant power input approach. This approach prescribes that the flow rate is adjusted according to current pressure gradient, so to keep constant the power injected into the flow, it mimics closely real transport pipelines. A phase-field method is used to describe the dynamics of the liquid-liquid interface. As this technique is tailored toward the transport of very viscous fluids like oils, we study the drag reduction performance of the system by keeping fixed the lubricating fluid properties (e.g., water) and by considering two different types of oil characterized by different viscosities, 10 and 100 times more viscous than water, respectively. As in real instances the presence of impurities and surfactants—which act by locally reducing the local value of the surface tension—is inevitable, we consider, for each type of transported oil, a clean and a surfactant-laden interface. For all four tested configurations, we unambiguously show that significant DR can be achieved. Reportedly, compared to the single-phase case, we observe a reduction of the mean pressure gradient down to $p_x/p_{x,\mathrm{sp}}=0.25$ for the largest viscosity oil. By analyzing the features of turbulence in the lubricating layer, and the close interaction with the perturbations induced by the oil-water interface deformation, we elucidate the physical mechanisms leading to DR and we underline the effects of viscosity ratios and of surfactants.

Bubbles have always captivated our curiosity with their aesthetics and complexities alike. While the act of blowing bubbles is familiar to everyone, the underlying physics of these fleeting spheres often eludes reasoning. In this Letter, we discuss the dynamics of inflating a soap bubble using controlled airflow through a film-coated nozzle. We assess and predict the rate of inflation by varying the source pressure. Visualizing the previously unexplored internal flow reveals that air enters the bubble as a round jet, emerging from the nozzle opening and impinges on the expanding concave bubble interface to form a toroidal vortex. Several scaling laws of the associated vortical flow spanning the entire bubble and the vortex core are reported. The observed dynamics of this bubble-confined vortex ring formation indicate universality in certain aspects when compared to the free laminar vortex rings.

This Letter explores the influence of colloids at liquid-liquid interfaces on droplet pinch-off dynamics in microfluidic devices, examining hard polystyrene particles and soft pNIPAM microgels. We uncover a significant deviation in droplet formation time compared to pure systems, similarly to surfactant-laden systems, yet notably, colloids exert minimal impact on droplet size, indicating potential nonlinear effects. The dynamics of neck thinning without colloids agree with the classic pendant drop scaling laws, while particle presence replaces traditional viscous and inertial-viscous regimes with a single power law, suggesting an elastic behavior driven by soft particle interactions.

The Enskog equation with respect to orthonormal vielbein fields is derived in this paper, enabling the utilization of arbitrary coordinate systems to characterize spatial geometry. Additionally, an adapted coordinate system in the momentum space is employed, connected to the physical space through vielbeins. Within this framework, the momentum component perpendicular to a curved boundary can be treated as an independent one, facilitating the application of half-range Gauss-Hermite quadratures. An appropriate finite-difference lattice Boltzmann model is developed and validated against a direct simulation Monte Carlo particle-based method for solving the Enskog equation in curvilinear geometries. The test scenarios include cylindrical Couette flow, cylindrical Fourier flow between coaxial cylinders, and spherical Fourier flow between concentric spheres. Reasonable agreement between the two approaches is observed throughout the parameter range and curvature-specific effects are well captured.

In the present study, we report a Mach number invariant function (${\varphi }_S$) for the mean temperature field in compressible wall turbulence. We demonstrate its validation by comparing it with the invariant functions derived in the previous studies—the semilocal-type (${\varphi }_{SL}$) and van Driest-type (${\varphi }_{VD}$) scalings—case by case. To be specific, ${\varphi }_{SL}$ works well in the inner layer of compressible channel flows with isothermal walls; ${\varphi }_S$ works well in the inner layer of compressible channel flows with adiabatic walls, and supersonic/hypersonic turbulent boundary layers with cold walls; and ${\varphi }_{VD}$ does not work the best among all three functions in the flows under consideration. The proposed temperature transformations based on ${\varphi }_S$ show an improvement in channel flows over adiabatic walls and supersonic/hypersonic turbulent boundary layers with cold walls. The effects of the generated high-order terms during derivation are also clarified. These findings may be revealing for the development of the near-wall model in high-speed aerodynamics.

We discuss the mechanism of electrokinetic regularization of the shortwave catastrophe in electrodeposition. Without suitable regularizations, both morphological and nonequilibrium electroosmotic instabilities occurring in electrodeposition are singular for shortwave perturbations in the following sense. In morphological instability the infinitesimal wavelength perturbation grows the fastest. In the electroosmotic instability in terms of the local electroneutrality approximation, the quiescent state is least stable for infinitesimal wavelength perturbations. Regularization of electroosmotic instability requires considering the finite width of the electric double layer. It was recently shown that in electrodeposition too, taking into account the finite width of the electric double layer combined with the finite electrode reaction rate constant, yields selection of a finite critical perturbation wavelength for instability, thus providing a regularization mechanism alternative to surface tension. We discuss the physical mechanism of this length selection, along with the effect exerted by the nonequilibrium electroosmotic flow which emerges in the limiting-overlimiting current regime.

Employing the discrete element method, we study the rheology of dense granular media, varying in size, density, and frictional properties of particles, across a spectrum from quasistatic to inertial regimes. By accounting for the volumetric contribution of each solid phase, we find that the stress ratio, $\mu$, and concentration, $\varphi$, scale with the inertial number when using volume averaging to calculate mean particle density, friction, and size. Moreover, the critical packing fraction correlates with skewness, polydispersity, and particle friction, irrespective of the size distribution. Notably, following the work of Kim and Kamrin Phys. Rev. Lett. 125, 088002 (2020), we introduce a rheological power-law scaling to collapse all our monodisperse and polydisperse data, reliant on concentration, dimensionless granular temperature, and the inertial number. This model seamlessly merges the $\mu \left(I\right)$-rheology and kinetic theory, enabling the unification of all local and nonlocal rheology data onto a single master curve.

This paper presents a quadrant analysis of turbulent rotor wakes, placing a particular emphasis on the self-similarity in single-point turbulent statistics. The self-similarity analysis reveals that within the wake self-similar region, the distribution of the Reynolds shear stress $-\overline{u^{\prime }{v}^{\prime }}$ and the turbulent kinetic energy $k$ in different quadrants can also attain a state of self-similarity. The length scaling is the same for $-\overline{u^{\prime }{v}^{\prime }}, k$, and their different quadrant contributions. However, the velocity scaling of the ejection contribution differs from that of $-\overline{u^{\prime }{v}^{\prime }}$ and $k$. Additionally, the analysis of hole filtering and spectrum filtering effects on the self-similarity and scaling of the ejection contribution to $-\overline{u^{\prime }{v}^{\prime }}$ reveals a strong connection between ejection events and large-scale coherent structures, as well as deceleration extreme events. This finding is further confirmed through quadrant analysis of the time evolution of velocity fluctuations. The present study can be helpful in improving our understanding of the turbulent wake self-similarity.