The collapse dynamics of an axisymmetric fluid cavity that wets the bottom of a rotating bucket bound by vertical sidewalls are studied. Lubrication theory is applied to the governing field equations for the thin film to yield an evolution equation that captures the effect of capillary, gravitational, and centrifugal forces on this converging flow. The focus is on the quasistatic spreading regime, whereby contact-line motion is governed by a constitutive law relating the contact-angle to the contact-line speed. Surface tension forces dominate the collapse dynamics for small holes with the collapse time appearing as a power law whose exponent compares favorably to experiments in the literature. Gravity accelerates the collapse process. Volume dependence is predicted and compared with experiment. Centrifugal forces slow the collapse process and lead to complex dynamics characterized by stalled spreading behavior that separates the large and small hole asymptotic regimes.

We study the fluid drift due to a time-dependent dumbbell model of a microswimmer. The model captures important aspects of real microswimmers such as a time-dependent flagellar motion and a no-slip body. The model consists of a rigid sphere for the body and a time-dependent moving Stokeslet representing the flagella. We analyze the paths of idealized fluid particles displaced by the swimmer. The simplicity of the model allows some asymptotic calculations very near and far away from the swimmer. The displacements of particles near the swimmer diverge in a manner similar to an isolated no-slip sphere, but with a smaller coefficient due to the action of the flagellum. Far from the swimmer, the time dependence becomes negligible due to both being very fast and decaying with distance. Finally, we compute the probability distribution of particle displacements and find that our model has fatter tails than previous steady models, due to the presence of a no-slip surface that drags particles along.

The free rise of isolated, deformable, finite-size bubbles in otherwise homogeneous isotropic turbulence is investigated by direct numerical simulation. The Navier-Stokes equations are solved in both phases subject to the pertinent velocity and stress conditions at the deformable gas-liquid interface. The bubble rise velocity is found to be drastically reduced by turbulence, as is widely known for microbubbles. The probability distribution functions of the horizontal bubble acceleration component are well fitted by a log-normal distribution. The distributions of the vertical components are negatively skewed, a property related to the fact that bubbles experience on average stronger decelerations than accelerations. An assessment of the correlations of bubble acceleration with properties of the surrounding flow is used to define estimates of the liquid velocity and vorticity entering in liquid acceleration and lift forces. Finally, fast rising bubbles are found to preferentially sample downflow regions of the flow, whereas those subjected to a higher turbulence level have an increased residence time in swirling regions, some features similar to those of small bubbles.

Unsteady motion of a rarefied gas between two parallel plates caused when one of the plates starts a harmonic oscillation in its normal direction is investigated under a slightly rarefied condition, i.e., for small Knudsen numbers. The compressible Navier-Stokes equations are employed and their appropriate temperature jump condition is derived systematically. The equations with the correct boundary conditions are solved numerically to give the unsteady flow field. In particular, the time-periodic solution established at later times is investigated in detail and it is shown that the one-period average of the oscillating part of the momentum and that of the energy transferred from the oscillating plate to the resting one take nonzero values in contrast to the linear theory. This confirms the numerical result based on the Bhatnagar-Gross-Krook model of the Boltzmann equation for intermediate Knudsen numbers [T. Tsuji and K. Aoki, Microfluid. Nanofluid. 16, 1033 (2014)]. It is also shown that the gas approaches the time-periodic motion exponentially fast in time.

A mixed suspension of red blood cells (RBCs) and microparticles flows through a cylindrical channel with a constriction mimicking a stenosed blood vessel. Our three-dimensional Lattice-Boltzmann simulations show that the RBCs are depleted right ahead of and after the constriction. Although the RBC mean concentration (hematocrit) is 16.5% or 23.7%, their axial concentration profile is very similar to that of isolated tracer particles flowing along the central axis. Most importantly, however, we find that the stiff microparticles exhibit the opposite behavior. Arriving on a marginated position near the channel wall, they can pass through the constriction only if they find a suitable gap to dip into the dense plug of RBCs occupying the channel center. This leads to a prolonged dwell time and, as a consequence, to a pronounced increase in microparticle concentration right in front of the constriction. For biochemically active particles such as drug delivery agents or activated platelets this clustering may have important physiological consequences, e.g., for the formation of microthrombi.

Understanding the gravity-driven motion of droplets on inclined substrates in the presence of external shear and substrate permeability is important for applications such as spray coating and filtration. In this paper, we use a lubrication-theory-based model to study how external shear and substrate permeability affect droplet wetting transitions. A nonlinear evolution equation for the droplet height as a function of time and two spatial variables is derived and numerically solved. The contact-line region is described using a precursor film and disjoining pressure. Depending on its direction, external shear can either suppress or drive wetting transitions, but does not appear to significantly change the critical droplet speeds associated with these transitions. Substrate permeability generally suppresses wetting transitions due to liquid absorption and does appear to significantly affect these critical droplet speeds. The strong influence of substrate permeability and external shear on droplet wetting transitions indicates that it will be important to account for these effects when developing accurate models for industrial applications.

The three-dimensional development of instabilities and the subsequent spray formation in a gas-liquid mixing layer are important fundamental problems in the area of multiphase flows. It is highly desirable to visualize this detailed atomization process and to analyze the instabilities and mechanisms involved, and massive numerical simulations are required, in addition to experiment. Rapid development of numerical methods and computer technology in the past decade now allows large-scale three-dimensional direct numerical simulations of atomization to be performed. Nevertheless, the fundamental question, whether all the physical scales involved in the primary breakup process are faithfully resolved, has eluded researchers until now. In the present study, we conduct direct numerical simulations of spray formation in a gas-liquid mixing layer with state-of-the-art computational resources (using up to $4\times {10}^9$ cells and $16384$ cores), in order to obtain a high-fidelity numerical closeup of the detailed mechanisms of spray formation. We also aim to examine whether present computational resources are sufficient for a fully resolved direct numerical simulation of atomization.

We study the dynamic coarsening of wrinkles in an elastic sheet that is compressed while lying on a thin layer of viscous liquid. When the ends of the sheet are instantaneously brought together by a small distance, viscous resistance initially prevents the sheet from adopting a globally buckled shape. Instead, the sheet accommodates the compression by wrinkling. Previous scaling arguments suggested that a balance between the sheet's bending stiffness and viscous effects lead to a wrinkle wavelength $\lambda$ that increases with time $t$ according to $\lambda \propto t^{1/6}$. We show that taking proper account of the compression constraint leads to a logarithmic correction of this result, $\lambda \propto {(t/logt)}^{1/6}$. This correction is significant over experimentally observable time spans and leads us to reassess previously published experimental data.

The flow driven by a rapidly expanding and collapsing cavitation bubble in a narrow cylindrical gap is studied with the volume of fluid method. The simulations reveal a developing plug flow during the early expansion followed by flow reversal at later stages. An adverse pressure gradient leads to boundary layer separation and flow reversal, causing large shear stress near the boundaries. Analytical solution to a planar pulsating flow shows qualitative agreement with the CFD results. The shear stress close to boundaries has implications to deformable objects located near the bubble: Experiments reveal that thin, flat biological cells entrained in the boundary layer become stretched, while cells with a larger cross section are mainly transported with the flow.

This work uses the Mori-Zwanzig (M-Z) formalism, a concept originating from nonequilibrium statistical mechanics, as a basis for the development of coarse-grained models of turbulence. The mechanics of the generalized Langevin equation (GLE) are considered, and insight gained from the orthogonal dynamics equation is used as a starting point for model development. A class of subgrid models is considered which represent nonlocal behavior via a finite memory approximation [Stinis, arXiv:1211.4285 (2012)], the length of which is determined using a heuristic that is related to the spectral radius of the Jacobian of the resolved variables. The resulting models are intimately tied to the underlying numerical resolution and are capable of approximating non-Markovian effects. Numerical experiments on the Burgers equation demonstrate that the M-Z-based models can accurately predict the temporal evolution of the total kinetic energy and the total dissipation rate at varying mesh resolutions. The trajectory of each resolved mode in phase space is accurately predicted for cases where the coarse graining is moderate. Large eddy simulations (LESs) of homogeneous isotropic turbulence and the Taylor-Green Vortex show that the M-Z-based models are able to provide excellent predictions, accurately capturing the subgrid contribution to energy transfer. Last, LESs of fully developed channel flow demonstrate the applicability of M-Z-based models to nondecaying problems. It is notable that the form of the closure is not imposed by the modeler, but is rather derived from the mathematics of the coarse graining, highlighting the potential of M-Z-based techniques to define LES closures.

An intrinsic property of turbulent flows is the exponential deformation of fluid elements along Lagrangian paths. The production of enstrophy by vorticity stretching follows from a similar mechanism in the Lagrangian view, though the alignment statistics differ and viscosity prevents unbounded growth. In this paper, the stretching properties of fluid elements and vorticity along Lagrangian paths are studied in a channel flow at ${\text{Re}}_{\tau }=1000$ and compared with prior known results from isotropic turbulence. To track Lagrangian paths in a public database containing direct numerical simulation results, the task-parallel algorithm previously employed in the isotropic database is extended to the case of flow in a bounded domain. It is shown that above 100 viscous units from the wall, stretching statistics are equal to their isotropic values, in support of the local isotropy hypothesis. In the viscous sublayer, these stretching statistics approach values more consistent with an unsteady two-dimensional shear flow, in which exponential stretching no longer occurs. Normalized by dissipation rate, the stretching in the buffer layer and below is less efficient due to less favorable alignment statistics. The Cramér function characterizing cumulative Lagrangian stretching statistics shows that overall the channel flow has about half of the stretching per unit dissipation compared with isotropic turbulence.

The lifting Hele-Shaw cell problem is a variant of the classical constant-gap Hele-Shaw situation in which the cell gap width is time dependent. Experiments on Newtonian fluid flows in lifting Hele-Shaw cells with large lifting velocities reveal the development of dendriticlike fingered structures that compete among themselves. The usual Darcy law description of the problem, where inertial effects are neglected, has not been able to explain the origin of these competing sidebranched patterns. In this work we use a generalized Darcy law and a perturbative mode-coupling theory to investigate the impact of inertia on the pattern-forming dynamics of the system. Two main results are deduced: Inertia induces a mechanism for finger sidebranching formation and favors the intensification of finger competition events.

We present an extended thermodynamics (ET) theory of dissipative dense gases. In particular, we study the ET theory with six fields, where we neglect shear viscosity and heat conductivity. We postulate a simple principle of duality between rarefied and dense gases. This principle is based on the microscopic analysis of the energy exchange between different modes of the molecular motion. The basic system of equations satisfies all principles of ET, that is, Galilean invariance, entropy principle, and thermodynamic stability (entropy convexity), and, as in the ET theory of rarefied gases, the constitutive equations are completely determined by the thermal and caloric equations of state. The system is simplest after the Euler system, but, in contrast to the Euler system, we may have a global smooth solution due to the fact that the system is dissipative symmetric hyperbolic and satisfies the so-called K condition. There emerge two nonequilibrium temperatures; one is due to the translational modes, and the other is due to the internal modes such as rotation and vibration of a molecule. This viewpoint allows us to understand the origin of the dynamic pressure in a more clear way. Furthermore we evaluate the characteristic velocities associated with the hyperbolic system and address the fluctuation-dissipation relation of the bulk viscosity. As a typical example, we analyze van der Waals fluids based on the present theory.

Direct numerical simulation data for channel flow at a friction Reynolds number of 4200, generated by Lozano-Durán and Jiménez [J. Fluid Mech. 759, 432 (2014)], are used to examine the properties of near-wall turbulence within subranges of eddy-length scale. Attention is primarily focused on the intermediate layer (mesolayer) covering the logarithmic velocity region within the range of wall-scaled wall-normal distance of 80–1500. The examination is based on a number of statistical properties, including premultiplied and compensated spectra, the premultiplied derivative of the second-order structure function, and three scalar parameters that characterize the anisotropic or isotropic state of the various length-scale subranges. This analysis leads to the delineation of three regions within the map of wall-normal-wise premultiplied spectra, each characterized by distinct turbulence properties. A question of particular interest is whether the Townsend-Perry attached-eddy hypothesis (AEH) can be shown to be valid across the entire mesolayer, in contrast to the usual focus on the outer portion of the logarithmic-velocity layer at high Reynolds numbers, which is populated with very-large-scale motions. This question is addressed by reference to properties in the premultiplied scalewise derivative of the second-order structure function (PMDS2) and joint probability density functions of streamwise-velocity fluctuations and their streamwise and spanwise derivatives. This examination provides evidence, based primarily on the existence of a plateau region in the PMDS2, for the qualified validity of the AEH right down the lower limit of the logarithmic velocity range.

We present results from the simultaneous application of planar laser-induced fluorescence (PLIF), particle image velocimetry (PIV) and particle tracking velocimetry (PTV), complemented by direct numerical simulations, aimed at the detailed hydrodynamic characterization of harmonically excited liquid-film flows falling under the action of gravity. The experimental campaign comprises four different aqueous-glycerol solutions corresponding to four Kapitza numbers ($\text{Ka}=14$, 85, 350, 1800), spanning the Reynolds number range $\text{Re}=2.3–320$, and with forcing frequencies $f_{\mathrm{w}}=7$ and $10\mathrm{Hz}$. PLIF was employed to generate spatiotemporally resolved film-height measurements, and PIV and PTV to generate two-dimensional velocity-vector maps of the flow field underneath the wavy film interface. The latter allows for instantaneous, highly localized velocity-profile, bulk-velocity, and flow-rate data to be retrieved, based on which the effect of local film topology on the flow field underneath the waves is studied in detail. Temporal sequences of instantaneous and local film height and bulk velocity are generated and combined into bulk flow-rate time series. The time-mean flow rates are then decomposed into steady and unsteady components, the former represented by the product of the mean film height and mean bulk velocity and the latter by the covariance of the film-height and bulk-velocity fluctuations. The steady terms are found to vary linearly with the flow $\text{Re}$, with the best-fit gradients approximated closely by the kinematic viscosities of the three examined liquids. The unsteady terms, typically amounting to $5\%–10\%$ of the mean and peaking at approximately $20\%$, are found to scale linearly with the film-height variance. And, interestingly, the instantaneous flow rate is found to vary linearly with the instantaneous film height. Both experimental and numerical flow-rate data are closely approximated by a simple analytical relationship with only minor deviations. This relationship includes terms such as the wave speed $c$ and mean flow rate $\overline{Q}$, which can be obtained by relatively simple and inexpensive methods, thus allowing for spatiotemporally resolved flow-rate predictions to be made without requiring explicit knowledge of the full flow-field information underneath the wavy interface.

We study experimentally the occurrence of spontaneous spinning (Quincke rotation) of an ellipsoid in a uniform direct current (dc) electric field. For an ellipsoid suspended in an unbounded fluid, we find two stable states characterized by the orientation of the ellipsoid long axis relative to the applied electric field: spinless (parallel) and spinning (perpendicular). The phase diagram of ellipsoid behavior as a function of field strength and aspect ratio is in close agreement with the theory of Cēbers et al. [Phys. Rev. E 63, 016301 (2000)]. We also investigate the dynamics of the ellipsoidal Quincke rotor resting on a planar surface with normal perpendicular to the field direction. We find behaviors, such as swinging (long axis oscillating around the applied field direction) and tumbling, due to the confinement.

The relative role of convection and diffusion is characterized both numerically and experimentally for porous media flows due to a Rayleigh-Taylor instability of a horizontal interface between two miscible solutions in the gravity field. We show that, though globally convection dominates over diffusion during the nonlinear regime, diffusion can locally be as important as convection and even dominates over lateral convection far away from the fingertips. Our experimental and numerical computations of the temporal evolution of the mixing length, the width of the fingers, and their wavelength are in good agreement and show that the lateral evolution of fingers is governed by diffusion.

When a droplet is placed onto a vertically vibrated bath, it can bounce without coalescing. Upon an increase of the forcing acceleration, the droplet is propelled by the wave it generates and becomes a walker with a well-defined speed. We investigate the confinement of a walker in different rectangular cavities, used as waveguides for the Faraday waves emitted by successive droplet bounces. By studying the walker velocities, we discover that one-dimensional confinement is optimal for narrow channels of width of $D\simeq 1.5{\lambda }_F$. Thereby, the walker follows a quasilinear path. We also propose an analogy with waveguide models based on the observation of the Faraday instability within the channels.

The dynamics and statistical mechanics of $N$ chaotically evolving point vortices in the doubly periodic domain are revisited. The selection of the correct microcanonical ensemble for the system is first investigated. The numerical results of Weiss and McWilliams [Phys. Fluids A 3, 835 (1991)], who argued that the point vortex system with $N=6$ is nonergodic because of an apparent discrepancy between ensemble averages and dynamical time averages, are shown to be due to an incorrect ensemble definition. When the correct microcanonical ensemble is sampled, accounting for the vortex momentum constraint, time averages obtained from direct numerical simulation agree with ensemble averages within the sampling error of each calculation, i.e., there is no numerical evidence for nonergodicity. Further, in the $N\to \infty$ limit it is shown that the vortex momentum no longer constrains the long-time dynamics and therefore that the correct microcanonical ensemble for statistical mechanics is that associated with the entire constant energy hypersurface in phase space. Next, a recently developed technique is used to generate an explicit formula for the density of states function for the system, including for arbitrary distributions of vortex circulations. Exact formulas for the equilibrium energy spectrum, and for the probability density function of the energy in each Fourier mode, are then obtained. Results are compared with a series of direct numerical simulations with $N=50$ and excellent agreement is found, confirming the relevance of the results for interpretation of quantum and classical two-dimensional turbulence.

We investigate the flow over shallow cavities as a representative configuration for modeling small surface irregularities in wall-bounded shear flows. Due to the globally stable nature of the flow, we perform a frequency response analysis, which shows a significant potential for the amplification of disturbance kinetic energy by harmonic forcing within a certain frequency band. Shorter and more shallow cavities exhibit less amplified responses, while energy from the base flow can be extracted predominantly from forcing that impacts the cavity head on. A structural sensitivity analysis, combined with a componentwise decomposition of the sensitivity tensor, reveals the regions of the flow that act most effectively as amplifiers. We find that the flow inside the cavity plays a negligible role, whereas boundary layer modifications immediately upstream and downstream of the cavity edges contribute significantly to the frequency response. The same regions constitute preferred locations for implementing active or passive control strategies to manipulate the frequency response of the flow.