A small eddy viscosity or mass diffusivity that varies with height has been found to have unexpected effects on the Kelvin–Helmholtz (KH) instability of a stably stratified shear layer near the neutral stability boundary. In particular, varying viscosity can increase the growth rate of the instability in contrast to the effect of uniform viscosity. Here, these results are extended to parameter ranges relevant in many geophysical and engineering contexts. We find that linearization of the viscous terms based on the assumption of weak viscosity/diffusivity is valid for non-dimensional values (inverse Reynolds number) up to ${\sim}10^{-2}$. Decreasing the Richardson number far below its critical value $1/4$ can change, or even reverse, the effects of eddy viscosity and diffusivity. A primary goal is to explain the unexpected destabilization by viscosity. Varying viscosity affects vorticity (and other fluid properties) in a manner identical to advection with an advecting velocity equal to minus the gradient of viscosity. Destabilization occurs when this viscous ‘advection’ reinforces the vorticity distribution of a growing mode.

The three-dimensional effects of turbulence cannot be neglected when the spanwise wavelength of the incident turbulence is not effectively infinite with respect to the chord, which may invalidate the strip assumption. Based on three-dimensional theory, a general approach, expressed in terms of a two-dimensional Fourier transform of the correlation of the buffeting force, is proposed to identify the two-wavenumber spectrum and aerodynamic admittance of the lift force on an aerofoil. It is essential that the approach presented can be validated in wind tunnel experiments. The coherence of the lift force on an aerofoil in grid-generated turbulence is obtained by simultaneous measurements of unsteady surface pressures around several chordwise strips on a stiff sectional model, which controls the accuracy of results. For the purpose of the Fourier transform, three empirical coherence models of the lift force are presented to fit the experimental results. Compared with the linearized theory, the two-wavenumber aerodynamic admittance can describe well the pressure distribution and the pattern of energy transition in an isotropic turbulence field. Thus, the failure mechanism of the traditional strip assumption can be demonstrated explicitly. In addition, the results obtained also validate the theory proposed by Graham (Aeronaut. Q., vol. 21, 1970, pp. 182–198; Aeronaut. Q., vol. 22, 1971, pp. 83–100). The present approach can be extended to study the three-dimensionality of the buffeting force on line-like structures with arbitrary cross-configurations, such as long-span bridges and high-rise buildings.

We consider convection in an internally heated (IH) layer of fluid that is bounded below by a perfect insulator and above by a poor conductor. The poorly conducting boundary is modelled by a fixed heat flux. Using solely analytical methods, we find linear and energy stability thresholds for the static state, and we construct a lower bound on the mean temperature that applies to all flows. The linear stability analysis yields a Rayleigh number above which the static state is linearly unstable ($R_{L}$), and the energy analysis yields a Rayleigh number below which it is globally stable ($R_{E}$). For various boundary conditions on the velocity, exact expressions for $R_{L}$ and $R_{E}$ are found using long-wavelength asymptotics. Each $R_{E}$ is strictly smaller than the corresponding $R_{L}$ but is within 1 %. The lower bound on the mean temperature is proven for no-slip velocity boundary conditions using the background method. The bound guarantees that the mean temperature of the fluid, relative to that of the top boundary, grows with the heating rate ($H$) no slower than $H^{2/3}$.

We have performed numerical simulations of the flow in a large-aspect-ratio Couette–Taylor system with rotating inner cylinder and with a radial temperature gradient. The aspect ratio was chosen in such a way that the base state is in the conduction regime. Away from the endplates, the base flow is a superposition of an azimuthal flow induced by rotation and an axial flow (large convective cell) induced by the temperature gradient. For a fixed rotation rate of the inner cylinder in the subcritical laminar regime, the increase of the temperature difference imposed on the annulus destabilizes the convective cell to give rise to co-rotating vortices as primary instability modes and to counter-rotating vortices as secondary instability modes. The space–time properties of these vortices have been computed, together with the momentum and heat transfer coefficients. The temperature gradient enhances the momentum and heat transfer in the flow independently of its sign.

In this paper we investigate the Tayler instability in an incompressible, viscous and resistive liquid metal column and in a model of a liquid metal battery (LMB). Detailed comparisons between theory and numerics, both in linear and nonlinear regimes, are performed. We identify the timescale that is well adapted to the quasi-static (QS) regime and find the range of Hartmann numbers where this approximation applies. The scaling law $\mathit{Re}\sim \mathit{Ha}^{2}$ for the amplitude of the Tayler destabilized flow is explained using a weakly nonlinear argument. We calculate a critical electrolyte height above which the Tayler instability is too weak to disrupt the electrolyte layer in a LMB. Applied to present day Mg-based batteries, this criterion shows that short circuits can occur only in very large batteries. Finally, preliminary results demonstrate the feasibility of direct numerical multiphase simulations of the Tayler instability in a model battery.

The rotational motion of a prolate spheroidal particle suspended in shear flow is studied by a lattice Boltzmann method with external boundary forcing (LB-EBF). It has previously been shown that the case of a single neutrally buoyant particle is a surprisingly rich dynamical system that exhibits several bifurcations between rotational states due to inertial effects. It was observed that the rotational states were associated with either fluid inertia effects or particle inertia effects, which are always in competition. The effects of fluid inertia are characterized by the particle Reynolds number $\mathit{Re}_{p}=4Ga^{2}/{\it\nu}$, where $G$ is the shear rate, $a$ is the length of the particle major semi-axis and ${\it\nu}$ is the kinematic viscosity. Particle inertia is associated with the Stokes number $\mathit{St}={\it\alpha}\,\mathit{Re}_{p}$, where ${\it\alpha}$ is the solid-to-fluid density ratio. Previously, the neutrally buoyant case ($\mathit{St}=\mathit{Re}_{p}$) was studied extensively. However, little is known about how these results are affected when $\mathit{St}\neq \mathit{Re}_{p}$, and how the aspect ratio $r_{p}$ (major axis/minor axis) influences the competition between fluid and particle inertia in the absence of gravity. This work gives a full description of how prolate spheroidal particles in the range $2\leqslant r_{p}\leqslant 6$ behave depending on the chosen $\mathit{St}$ and $\mathit{Re}_{p}$. Furthermore, consequences for the rheology of a dilute suspension containing such particles are discussed. Finally, grid resolution close to the particle is shown to affect the quantitative results considerably. It is suggested that this resolution is a major cause of quantitative discrepancies between different studies. Thus, the results of this work and previous direct numerical simulations of this problem should be regarded as qualitative descriptions of the physics involved, and more refined methods must be used to quantitatively pinpoint the transitions between rotational states.

Longwave Marangoni convection in two-layer films under the action of heating modulation is considered. The analysis is carried out in the lubrication approximation. The capillary forces are assumed to be sufficiently strong, and they are taken into account. Periodic or symmetric boundary conditions are applied on the boundaries of the computational region. Numerical simulations are performed by means of a finite-difference method. Two regions of parametric instabilities have been found. In the first region, one observes the competition or coexistence of standing waves parallel to the boundaries of the computational region. The multistability of the flow regimes is revealed. In the second region, the regimes found in the case of periodic boundary conditions are more diverse than in the case of symmetric boundary conditions.

Motivated by understanding the liquid core dynamics of tidally deformed planets and moons, we present a study of incompressible flow driven by latitudinal libration within rigid triaxial ellipsoids. We first derive a laminar solution for the inviscid equations of motion under the assumption of uniform vorticity flow. This solution exhibits a resonance if the libration frequency matches the frequency of the spin-over inertial mode. Furthermore, we extend our model by introducing a reduced model of the effect of viscous Ekman layers in the limit of low Ekman number (Noir & Cébron, J. Fluid Mech., vol. 737, 2013, pp. 412–439). This theoretical approach is consistent with the results of Chan et al. (Phys. Earth Planet. Inter., vol. 187, 2011, pp. 404–415) and Zhang et al. (J. Fluid Mech., vol. 692, 2012, pp. 420–445) for spheroidal geometries. Our results are validated against systematic three-dimensional numerical simulations. In the second part of the paper, we present the first linear stability analysis of this uniform vorticity flow. To this end, we adopt different methods (Lifschitz & Hameiri, Phys. Fluids A, vol. 3, 1991, p. 2644; Gledzer & Ponomarev, Acad. Sci., USSR, Izv., Atmos. Ocean. Phys., vol. 13, 1977, pp. 565–569) that allow us to deduce upper and lower bounds for the growth rate of an instability. Our analysis shows that the uniform vorticity base flow is prone to inertial instabilities caused by a parametric resonance mechanism. This is confirmed by a set of direct numerical simulations. Applying our results to planetary settings, we find that neither a spin-over resonance nor an inertial instability can exist within the liquid core of the Moon, Io and Mercury.

Amplifications of flow past a backward-facing step with respect to optimal inflow and initial perturbations are investigated at Reynolds number 500. Two mechanisms of receptivity to inflow noise are identified: the bubble-induced inflectional point instability and the misalignment effect downstream of the secondary bubble. Further development of the misalignment results in decay of perturbations from $x=28$ onwards (the step is located at $x=0$), as has been observed in previous non-normality studies (Blackburn et al., J. Fluid Mech., vol. 603, 2008, pp. 271–304), and eventually limits the receptivity. The receptivity is found to be maximized at an inflow perturbation frequency of ${\it\omega}=0.50$ and a spanwise wavenumber of ${\it\beta}=0$, where the inflow noise takes full advantage of both mechanisms and is amplified over two orders of magnitude in terms of the velocity magnitude. In direct numerical simulations (DNS) of the flow perturbed by optimal or random inflow noise, vortex shedding, flapping of bubbles, three-dimensionality and turbulence are observed in succession as the magnitude of the inflow noise increases. Similar features of linear and nonlinear receptivity are observed at higher Reynolds numbers. The Strouhal number of the bubble flapping is 0.08, at which the receptivity to inflow noise reaches a maximum. This Strouhal number is close to reported values extracted from DNS or large eddy simulations (LES) at larger Reynolds numbers (Le et al., J. Fluid Mech., vol. 330, 1997, pp. 349–374; Kaiktsis et al., J. Fluid Mech., vol. 321, 1996, pp. 157–187; Métais, New Trends in Turbulence, 2001, Springer; Wee et al., Phys. Fluids, vol. 16, 2004, pp. 3361–3373). Methods to further clarify the mechanisms of receptivity and to suppress the noise amplifications by modifying the base flow using a linearly optimal body force are proposed. It is observed that the mechanisms of optimal noise amplification are fully revealed by the distribution of the base flow modification, which weakens the bubble instabilities and misalignment effects and subsequently reduces the receptivity significantly. Comparing the base flow modifications with respect to amplifications of inflow and initial perturbations, it is found that the maximum receptivity to initial perturbations is highly correlated with the receptivity to inflow noise at the optimal frequency ${\it\omega}=0.50$, and the correlation reduces as the inflow frequency deviates from this optimal value.

The aim of this study is to provide a better insight into the heat transfer mechanisms involved in single bubble growth in forced convection. In a set-up with vertical upflow of demineralized water under saturation conditions special bubble generators (BGs) were embedded at various positions in the plane wall. Power to a BG, local mean wall temperature and high-speed camera recordings from two viewing angles were measured synchronously. An accurate contour analysis is applied to reconstruct the instantaneous three-dimensional bubble volume. Interface topology changes of a vapour bubble growing at a plane wall have been found to be dictated by the rapid growth and by fluctuations in pressure, velocity and temperature in the approaching fluid flow. The camera images have shown a clear dry spot under the bubbles on the heater surface. A micro-layer under the bubble is experimentally shown to exist when the bubble pins to the wall surface and is therefore dependent on roughness and homogeneity of the wall. The ratio of heat extracted from the wall to the total heat required for evaporation was found to be around 30 % at most and to be independent of the bulk liquid flow rate and heat provided by the wall. When the bulk liquid is locally superheated this ratio was found to decrease to 20 %. Heat transfer to the bubble is also initially controlled by diffusion and is unaffected by the convection of the bulk liquid.

One-dimensional energy spectra in flat plate zero pressure gradient boundary layers and pipe flows are examined over a wide range of Reynolds numbers ($2600\leqslant \mathit{Re}_{{\it\tau}}\leqslant 72\,500$). The spectra show excellent collapse with Kolmogorov scaling at high wavenumbers for both flows at all Reynolds numbers. The peaks associated with the large-scale motions (LSMs) and superstructures (SS) in boundary layers behave as they do in pipe flows, with some minor differences. The location of the outer spectral peak, associated with SS or very large-scale motions (VLSMs) in the turbulent wall region, displays only a weak dependence on Reynolds number, and it occurs at the same wall-normal distance where the variances establish a logarithmic behaviour and where the amplitude modulation coefficient has a zero value. The results suggest that with increasing Reynolds number the energy is largely confined to a thin wall layer that continues to diminish in physical extent. The outer-scaled wavelength of the outer spectral peak appears to decrease with increasing Reynolds number. However, there is still significant energy content in wavelengths associated with the SS and VLSMs. The location of the outer spectral peak appears to mark the start of a plateau that is consistent with a $k_{x}^{-1}$ slope in the spectrum and the logarithmic variation in the variances. This $k_{x}^{-1}$ region seems to occur when there is sufficient scale separation between the locations of the outer spectral peak and the outer edge of the log region. It does not require full similarity between outer and wall-normal scaling on the wavenumber. The extent of $k_{x}^{-1}$ region depends on the wavelength of the outer spectral peak (${\it\lambda}_{OSP}$), which appears to emerge as a new length scale for the log region. Finally, based on the observations from the spectra together with the statistics presented in Vallikivi et al. (J. Fluid Mech., 2015 (submitted)), five distinct wall-normal layers are identified in turbulent wall flows.

This work experimentally examines the detachment of liquid droplets from both oleophilic and oleophobic fibres, using an atomic force microscope. The droplet detachment force was found to increase with increasing fibre diameter and forces were higher for philic fibres than phobic fibres. We also considered the detachment of droplets situated on the intersection of two fibres and arrays of fibres (such as found in fibrous mats or filters) and found that the required detachment forces were higher than for similarly sized droplets on a single fibre, though not as high as expected based on theory. A model was developed to predict the detachment force, from single fibres, which agreed well with experimental results. It was found that the entire dataset (single and multiple fibres) could be best described by power law relationships.

Full three-dimensional numerical simulations are employed to investigate the flows over inclined trapezoidal low-aspect-ratio plates at low Reynolds numbers, aiming to understand the unsteadiness induced by the interaction between the trailing vortical wake structures originating from the swept edges, and those from the leading and trailing edges. The flows past eighteen different plate geometries in three broad sets are simulated to study the influence of aspect ratio, taper angle and angle of attack on the wake vortices and the force coefficients. Both taper ratio and angle of attack of plates with the same area are found to have a broadly predictable influence on the wake stability and asymptotic forces. Smaller taper ratios result in lower maximum lift, while an increase in the angle of attack results in a reduction in the differences in maximum lift. Two distinct modes of periodic unsteady flow with significant differences in frequency are observed. The corresponding vortex-shedding mechanisms are analysed with the aid of $Q$-criterion isosurfaces and streamlines. A low wake frequency is observed at small taper angles when there is relative independence between the von Kármán vortices originating from the leading and trailing edges, and weak swept-edge vortices. The dominant Strouhal number in this state is approximately 0.09. When the taper angle or angle of attack increases, the flows over the swept edges form stronger trailing vortex structures which interact strongly with the leading-edge vortices, combining to produce a regular stream of vortex loops shed into the wake. In this regime, the dominant Strouhal number increases to approximately 0.14–0.18. Higher Reynolds numbers and/or angles of attack result in a loss of centre plane reflection symmetry in the wake. The aerodynamic forces have been quantified as a function of the problem parameters and plate geometry.

Volcanic eruptions commonly produce buoyant ash-laden plumes that rise through the stratified atmosphere. On reaching their level of neutral buoyancy, these plumes cease rising and transition to horizontally spreading intrusions. Such intrusions occur widely in density-stratified fluid environments, and in this paper we develop a shallow-layer model that governs their motion. We couple this dynamical model to a model for particle transport and sedimentation, to predict both the time-dependent distribution of ash within volcanic intrusions and the flux of ash that falls towards the ground. In an otherwise quiescent atmosphere, the intrusions spread axisymmetrically. We find that the buoyancy-inertial scalings previously identified for continuously supplied axisymmetric intrusions are not realised by solutions of the governing equations. By calculating asymptotic solutions to our model we show that the flow is not self-similar, but is instead time-dependent only in a narrow region at the front of the intrusion. This non-self-similar behaviour results in the radius of the intrusion growing with time $t$ as $t^{3/4}$, rather than $t^{2/3}$ as suggested previously. We also identify a transition to drag-dominated flow, which is described by a similarity solution with radial growth now proportional to $t^{5/9}$. In the presence of an ambient wind, intrusions are not axisymmetric. Instead, they are predominantly advected downstream, while at the same time spreading laterally and thinning vertically due to persistent buoyancy forces. We show that close to the source, this lateral spreading is in a buoyancy-inertial regime, whereas far downwind, the horizontal buoyancy forces that drive the spreading are balanced by drag. Our results emphasise the important role of buoyancy-driven spreading, even at large distances from the source, in the formation of the flowing thin horizontally extensive layers of ash that form in the atmosphere as a result of volcanic eruptions.

Direct numerical simulations of turbulent channel flows at friction Reynolds numbers (Re) of 550, 1000 and 1500 are used to analyse the turbulent production, transfer and dissipation mechanisms in the compound space of scales and wall distances by means of the Kolmogorov equation generalized to inhomogeneous anisotropic flows. Two distinct peaks of scale-energy source are identified. The first, stronger one, belongs to the near-wall cycle. Its location in the space of scales and physical space is found to scale in viscous units, while its intensity grows slowly with $\mathit{Re}$, indicating a near-wall modulation. The second source peak is found further away from the wall in the putative overlap layer, and it is separated from the near-wall source by a layer of significant scale-energy sink. The dynamics of the second outer source appears to be strongly dependent on the Reynolds number. The detailed scale-by-scale analysis of this source highlights well-defined features that are used to make the properties of the outer turbulent source independent of Reynolds number and wall distance by rescaling the problem. Overall, the present results suggest a strong connection of the observed outer scale-energy source with the presence of an outer region of turbulence production whose mechanisms are well separated from the near-wall region and whose statistical features agree with the hypothesis of an overlap layer dominated by attached eddies. Inner–outer interactions between the near-wall and outer source region in terms of scale-energy fluxes are also analysed. It is conjectured that the near-wall modulation of the statistics at increasing Reynolds number can be related to a confinement of the near-wall turbulence production due to the presence of increasingly large production scales in the outer scale-energy source region.

A model of skin-stabilized interfacial cavitation nuclei and their response to tensile and compressive stressing is presented. The model is evaluated in relation to experimental tensile strength results for water at rest at the bottom of an open water-filled container at atmospheric pressure and room temperature. These results are obtained by recording the initial growth of cavities generated by a short tensile pulse applied to the bottom of the container. It is found that the cavitation nuclei shift their tensile strength depending on their pressure history. Static pressurization for an extended period of time prior to testing is known to increase the tensile strength of water, but little information is available on how it is affected by compression pulses of short duration. This is addressed by imposing compression pulses of approximately 1 ms duration and a peak intensity of a few bar prior to the tension pulse. The observations are interpreted on the basis of the new model.

We studied the tank treading motion of an erythrocyte (red blood cell, or RBC) in linear shear flows by using a boundary-element fluid-dynamics model coupled with a multiscale structural model of the cell. The purpose was to investigate the correlation between the reference (stress-free) state of the cytoskeleton and the cell dynamics in shear flows with relatively high capillary numbers. We discovered that there exist two distinctive modes of tank treading, mode 1 and mode 2. In mode 1 the membrane elements originating from the dimple areas keep close to the central plane, whereas in mode 2 these elements remain near the farthermost locations from the central plane. Mode 1 is also characterized by significantly higher breathing and swinging oscillations. During tank treading one mode may become unstable and switch to the other. Their stability depends on the viscosity ratio and the capillary number. At a fixed viscosity ratio, when the capillary number is increased the cell experiences sequentially a region dominated by mode 2, a mode 1/mode 2 bistable region and a region dominated by mode 1. More profoundly, these regions are highly sensitive to the reference state of the cytoskeleton. For example, compared with a cell with a biconcave reference state, a cell with a spheroidal reference state features a much smaller region dominated by mode 2. This finding may guide experiments to identify the actual reference state of these cells.

Motivated by the work on stagnation-point-type exact solutions (with infinite energy) of 3D Euler fluid equations by Gibbon et al. (Physica D, vol. 132 (4), 1999, pp. 497–510) and the subsequent demonstration of finite-time blowup by Constantin (Int. Math. Res. Not. IMRN, vol. 9, 2000, pp. 455–465) we introduce a one-parameter family of models of the 3D Euler fluid equations on a 2D symmetry plane. Our models are seen as a deformation of the 3D Euler equations which respects the variational structure of the original equations so that explicit solutions can be found for the supremum norms of the basic fields: vorticity and stretching rate of vorticity. In particular, the value of the model’s parameter determines whether or not there is finite-time blowup, and the singularity time can be computed explicitly in terms of the initial conditions and the model’s parameter. We use a representative of this family of models, whose solution blows up at a finite time, as a benchmark for the systematic study of errors in numerical simulations. Using a high-order pseudospectral method, we compare the numerical integration of our ‘original’ model equations against a ‘mapped’ version of these equations. The mapped version is a globally regular (in time) system of equations, obtained via a bijective nonlinear mapping of time and fields from the original model equations. The mapping can be constructed explicitly whenever a Beale–Kato–Majda type of theorem is available therefore it is applicable to the 3D Euler equations (Bustamante, Physica D, vol. 240 (13), 2011, pp. 1092–1099). We show that the mapped system’s numerical solution leads to more accurate (by three orders of magnitude) estimates of supremum norms and singularity time compared with the original system. The numerical integration of the mapped equations is demonstrated to entail only a small extra computational cost. We study the Fourier spectrum of the model’s numerical solution and find that the analyticity strip width (a measure of the solution’s analyticity) tends to zero as a power law in a finite time. This is in agreement with the finite-time blowup of the fields’ supremum norms, in the light of rigorous bounds stemming from the bridge (Bustamante & Brachet, Phys. Rev. E, vol. 86 (6), 2012, 066302) between the analyticity-strip method and the Beale–Kato–Majda type of theorems. We conclude by discussing the implications of this research on the analysis of numerical solutions to the 3D Euler fluid equations.

A tiny air bubble can be entrapped at the bottom of a solid sphere that impacts onto a liquid pool. The bubble forms due to the deformation of the liquid surface by a local pressure buildup inside the surrounding gas, as also observed during the impact of a liquid drop on a solid wall. Here, we perform a perturbation analysis to quantitatively predict the initial deformations of the free surface of a liquid pool as it is approached by a solid sphere. We study the natural limits where the gas can be treated as a viscous fluid (Stokes flow) or as an inviscid fluid (potential flow). For both cases we derive the spatiotemporal evolution of the pool surface, and recover some of the recently proposed scaling laws for bubble entrapment. On inserting typical experimental values for the impact parameters, we find that the bubble volume is mainly determined by the effect of gas viscosity.

The capabilities of the triple-deck theory of receptivity for subsonic compressible boundary layers have been thoroughly investigated through comparisons with numerical simulations of the compressible Navier–Stokes equations. The analysis focused on the two Tollmien–Schlichting wave linear receptivity problems arising due to the interaction between a low-amplitude acoustic wave and a small isolated roughness element, and the low-amplitude time-periodic vibrations of a ribbon placed on the wall of a flat plate. A parametric study was carried out to look at the effects of roughness element and vibrating ribbon longitudinal dimensions, Reynolds number, Mach number and Tollmien–Schlichting wave frequency. The flat plate is considered isothermal, with a temperature equal to the laminar adiabatic-wall temperature. Numerical simulations of the full and the linearised compressible Navier–Stokes equations have been carried out using high-order finite differences to obtain, respectively, the steady basic flows and the unsteady disturbance fields for the different flow configurations analysed. The results show that the asymptotic theory and the Navier–Stokes simulations are in good agreement. The initial Tollmien–Schlichting wave amplitudes and, in particular, the trends indicated by the theory across the whole parameter space are in excellent agreement with the numerical results. An important finding of the present study is that the behaviour of the theoretical solutions obtained for $\mathit{Re}\rightarrow \infty$ holds at finite Reynolds numbers and the only conditions needed for the theoretical predictions to be accurate are that the receptivity process be linear and the free-stream Mach number be subsonic.