The whirling helical structure obtained when pouring honey onto toast may seem like an easy enough problem to solve at breakfast. Specifically, one would hope that a quick back-of-the-envelope scaling argument would help rationalize the observed behaviour and predict the coiling frequency. Not quite: multiple forces come into play, both in the part of the flow stretched by gravity and in the coil itself, which buckles and bends like a rope. In fact, the resulting abundance of regimes requires the careful numerical continuation method reported by Ribe (J. Fluid Mech., vol. 812, 2017, R2) to build a complete phase diagram of the problem and untangle this sticky situation.

Liquid film flow along an inclined plane featuring a slit, normal to the main direction of flow, creates a second gas–liquid interface connecting the two side walls of the slit. This inner interface forms two three-phase contact lines and supports a widely varying amount of liquid under different physical and geometrical conditions. The exact liquid configuration is determined by employing the Galerkin/finite element method to solve the two-dimensional Navier–Stokes equations at steady state. The interplay of inertia, viscous, gravity and capillary forces along with the substrate wettability and orientation with respect to gravity and the width of the slit determine the extent of liquid penetration and free-surface deformation. Finite wetting lengths are predicted in hydrophilic and hydrophobic substrates for inclination angles more or less than the vertical, respectively. Multiple steady solutions, connected by turning points forming a hysteresis loop, are revealed by pseudo-arclength continuation. Under these conditions, small changes in certain parameter values leads to an abrupt change in the wetting length and the deformation amplitude of the outer film surface. In hydrophilic substrates the wetting lengths exhibit a local minimum for small values of the Reynolds number and a very small range of Bond numbers; when inertia increases, they exhibit the hysteresis loop with the second limit point in a very short range of Weber numbers. Simple force balances determine the proper rescaling in each case, so that critical points in families of solutions for different liquids or contact angles collapse. The flow inside the slit is quite slow in general because of viscous dissipation and includes counter-rotating vortices often resembling those reported by Moffatt (J. Fluid Mech., vol. 18, 1964, pp. 1–18). In hydrophobic substrates, the wetting lengths decrease monotonically until the first limit point of the hysteresis loop, which occurs in a limited range of Bond numbers when the Kapitza number is less than 300 and in a limited range of Weber numbers otherwise. Here additional solution families are possible as well, where one or both contact points (Cassie state) coincide with the slit corners.

For a liquid film falling down along a vertical fibre, classical theory (Kalliadasis & Chang J. Fluid Mech., vol. 261, 1994, pp. 135–168; Yu & Hinch J. Fluid Mech., vol. 737, 2013, pp. 232–248) showed that drop formation can occur due to capillary instability when the Bond number $G=\unicode[STIX]{x1D70C}ga^{3}/\unicode[STIX]{x1D6FE}h_{0}$ is below the critical value $G_{c}\approx 0.60$, where $\unicode[STIX]{x1D70C}$ is the fluid density, $g$ is the gravitational acceleration, $a$ is the fibre radius, $\unicode[STIX]{x1D6FE}$ is the surface tension and $h_{0}$ is the unperturbed film thickness. However, the experiment by Quéré (Europhys. Lett., vol. 13 (8), 1990, pp. 721–726) found $G_{c}\approx 0.71$, which is slightly greater than the above theoretical value. Here we offer a plausible way to resolve this discrepancy by including additional wall slip whose amount can be measured by the slip parameter $\unicode[STIX]{x1D6EC}=3\unicode[STIX]{x1D706}/h_{0}$, where $\unicode[STIX]{x1D706}$ is the slip length. Using lubrication theory, we find that wall slip promotes capillary instability and, hence, enhances drop formation. In particular, when slip effects are strong ($\unicode[STIX]{x1D6EC}\gg 1$), the transition films and the drop height scale as $(c/\unicode[STIX]{x1D6EC})^{-1/3}$ and $(c/\unicode[STIX]{x1D6EC})^{2/3}$, respectively, distinct from those found by Yu & Hinch for the no-slip case where $c$ is the travelling speed. In addition, for $\unicode[STIX]{x1D6EC}>1$, $G_{c}$ is found to increase with $\unicode[STIX]{x1D6EC}$ according to $G_{c}\approx 0.7\unicode[STIX]{x1D6EC}^{1/3}$, offering a possible explanation why the $G_{c}$ found by Quéré is slightly greater than that predicted by the no-slip model. Using the above expression, the estimated slip length in Quéré’s experiment is found to be of the order of several micrometres, consistent with the typical slip length range 1–$10~\unicode[STIX]{x03BC}\text{m}$ for polymeric liquids such as silicone oil used in his experiment. The existence of wall slip in Quéré’s experiment is further supported by the observation that the film thinning kinetics exhibits the no-slip result $h\propto t^{-1/2}$ for early times and changes to the strong slip result $h\propto t^{-1}$, where $h$ is the film thickness. We also show that when the film is ultrathin, although capillary instability can become further amplified by strong slip effects, the instability can be arrested by the equally intensified gravity draining in the weakly nonlinear regime whose dynamics is governed by the Kuramoto–Sivashinsky equation.

The heat and mass transfer of deformable droplets in turbulent flows is crucial to a wide range of applications, such as cloud dynamics and internal combustion engines. This study investigates a single droplet undergoing phase change in isotropic turbulence using numerical simulations with a hybrid lattice Boltzmann scheme. Phase separation is controlled by a non-ideal equation of state and density contrast is taken into consideration. Droplet deformation is caused by pressure and shear stress at the droplet interface. The statistics of thermodynamic variables are quantified and averaged over both the liquid and vapour phases. The occurrence of evaporation and condensation is correlated to temperature fluctuations, surface tension variation and turbulence intensity. The temporal spectra of droplet deformations are analysed and related to the droplet surface area. Different modes of oscillation are clearly identified from the deformation power spectrum for low Taylor Reynolds number $Re_{\unicode[STIX]{x1D706}}$, whereas nonlinearities are produced with the increase of $Re_{\unicode[STIX]{x1D706}}$, as intermediate frequencies are seen to overlap. As an outcome, a continuous spectrum is observed, which shows a decrease in the power spectrum that scales as ${\sim}f^{-3}$. Correlations between the droplet Weber number, deformation parameter, fluctuations of the droplet volume and thermodynamic variables are also developed.

We investigate fully developed turbulence in stratified plane Couette flows using direct numerical simulations similar to those reported by Deusebio et al. (J. Fluid Mech., vol. 781, 2015, pp. 298–329) expanding the range of Prandtl number $Pr$ examined by two orders of magnitude from 0.7 up to 70. Significant effects of $Pr$ on the heat and momentum fluxes across the channel gap and on the mean temperature and velocity profile are observed. These effects can be described through a mixing length model coupling Monin–Obukhov (M–O) similarity theory and van Driest damping functions. We then employ M–O theory to formulate similarity scalings for various flow diagnostics for the stratified turbulence in the gap interior. The midchannel gap gradient Richardson number $Ri_{g}$ is determined by the length scale ratio $h/L$, where $h$ is the half-channel gap depth and $L$ is the Obukhov length scale. As $h/L$ approaches very large values, $Ri_{g}$ asymptotes to a maximum characteristic value of approximately 0.2. The buoyancy Reynolds number $Re_{b}\equiv \unicode[STIX]{x1D700}/(\unicode[STIX]{x1D708}N^{2})$, where $\unicode[STIX]{x1D700}$ is the dissipation, $\unicode[STIX]{x1D708}$ is the kinematic viscosity and $N$ is the buoyancy frequency defined in terms of the local mean density gradient, scales linearly with the length scale ratio $L^{+}\equiv L/\unicode[STIX]{x1D6FF}_{\unicode[STIX]{x1D708}}$, where $\unicode[STIX]{x1D6FF}_{\unicode[STIX]{x1D708}}$ is the near-wall viscous scale. The flux Richardson number $Ri_{f}\equiv -B/P$, where $B$ is the buoyancy flux and $P$ is the shear production, is found to be proportional to $Ri_{g}$. This then leads to a turbulent Prandtl number $Pr_{t}\equiv \unicode[STIX]{x1D708}_{t}/\unicode[STIX]{x1D705}_{t}$ of order unity, where $\unicode[STIX]{x1D708}_{t}$ and $\unicode[STIX]{x1D705}_{t}$ are the turbulent viscosity and diffusivity respectively, which is consistent with Reynolds analogy. The turbulent Froude number $Fr_{h}\equiv \unicode[STIX]{x1D700}/(NU^{\prime 2})$, where $U^{\prime }$ is a turbulent horizontal velocity scale, is found to vary like $Ri_{g}^{-1/2}$. All these scalings are consistent with our numerical data and appear to be independent of $Pr$. The classical Osborn model based on turbulent kinetic energy balance in statistically stationary stratified sheared turbulence (Osborn, J. Phys. Oceanogr., vol. 10, 1980, pp. 83–89), together with M–O scalings, results in a parameterization of $\unicode[STIX]{x1D705}_{t}/\unicode[STIX]{x1D708}\sim \unicode[STIX]{x1D708}_{t}/\unicode[STIX]{x1D708}\sim Re_{b}Ri_{g}/(1-Ri_{g})$. With this parameterization validated through direct numerical simulation data, we provide physical interpretations of these results in the context of M–O similarity theory. These results are also discussed and rationalized with respect to other parameterizations in the literature. This paper demonstrates the role of M–O similarity in setting the mixing efficiency of equilibrated constant-flux layers, and the effects of Prandtl number on mixing in wall-bounded stratified turbulent flows.

We present wall-resolved large-eddy simulations (LES) of flow over a smooth-wall circular cylinder up to $Re_{D}=8.5\times 10^{5}$, where $Re_{D}$ is Reynolds number based on the cylinder diameter $D$ and the free-stream speed $U_{\infty }$. The stretched-vortex subgrid-scale (SGS) model is used in the entire simulation domain. For the sub-critical regime, six cases are implemented with $3.9\times 10^{3}\leqslant Re_{D}\leqslant 10^{5}$. Results are compared with experimental data for both the wall-pressure-coefficient distribution on the cylinder surface, which dominates the drag coefficient, and the skin-friction coefficient, which clearly correlates with the separation behaviour. In the super-critical regime, LES for three values of $Re_{D}$ are carried out at different resolutions. The drag-crisis phenomenon is well captured. For lower resolution, numerical discretization fluctuations are sufficient to stimulate transition, while for higher resolution, an applied boundary-layer perturbation is found to be necessary to stimulate transition. Large-eddy simulation results at $Re_{D}=8.5\times 10^{5}$, with a mesh of $8192\times 1024\times 256$, agree well with the classic experimental measurements of Achenbach (J. Fluid Mech., vol. 34, 1968, pp. 625–639) especially for the skin-friction coefficient, where a spike is produced by the laminar–turbulent transition on the top of a prior separation bubble. We document the properties of the attached-flow boundary layer on the cylinder surface as these vary with $Re_{D}$. Within the separated portion of the flow, mean-flow separation–reattachment bubbles are observed at some values of $Re_{D}$, with separation characteristics that are consistent with experimental observations. Time sequences of instantaneous surface portraits of vector skin-friction trajectory fields indicate that the unsteady counterpart of a mean-flow separation–reattachment bubble corresponds to the formation of local flow-reattachment cells, visible as coherent bundles of diverging surface streamlines.

Vortex-induced vibration of a circular cylinder with low mass ratio ($0.05\leqslant m^{\ast }\leqslant 10.0$) is investigated, via a stabilized space–time finite element formulation, in the laminar flow regime where $m^{\ast }$ is defined as the ratio of the mass of the oscillating structure to the mass of the fluid displaced by it. Computations are carried out over a wide range of reduced speed, $U^{\ast }$, which is defined as $U/f_{N}D$, where $U$ is the free-stream speed, $f_{N}$ the natural frequency of the spring mass system in vacuum and $D$ the diameter of the cylinder. In particular, the situation where the lock-in regime extends up to infinite reduced speed is explored. Studies at large $Re$, in the past, have shown that the normalized amplitude of cylinder oscillation at infinite reduced speed, $A_{\infty }^{\ast }$, exhibits a sharp increase when $m^{\ast }$ is reduced below the critical mass ratio ($m_{crit}^{\ast }$). This jump signifies a shift from desynchronized response to lock-in state. In this work it is shown that in the laminar regime, a jump in $A_{\infty }^{\ast }$ occurs only beyond a certain $Re$ ($=Re_{j}\sim 108$). For $Re<Re_{j}$, the response increases smoothly with decrease in $m^{\ast }$ with no discernible jump. In this situation, therefore, the identification of $m_{crit}^{\ast }$ based on jump in response at $U^{\ast }=\infty$ is not possible. The difference in the $A^{\ast }-m^{\ast }$ variation on the two sides of $Re=Re_{j}$, is attributed to the difference in the transition between the lower branch of cylinder response and desynchronization regime. This transition is brought out more clearly by plotting $A^{\ast }$ with $f_{v_{o}}/f$, where $f_{v_{o}}$ is the vortex shedding frequency for the flow past a stationary cylinder and $f$ is the cylinder vibration frequency. In the $A^{\ast }-f_{v_{o}}/f$ plane, the response data as well as other quantities related to free vibrations, for different $m^{\ast }$, collapse on a curve. Unlike at high $Re$, the collapsed curves show a dependence on $Re$ in the laminar regime. The transition between the lock-in and desynchronized state, as seen from the collapsed curves, is qualitatively different for $Re$ on either side of $Re_{j}$. The collapsed curves, at a certain $Re$, are utilized to estimate $A^{\ast }$ for the limiting case of $(U^{\ast },m^{\ast })=(\infty ,0)$. Interestingly, unlike at large $Re$, this limit value is found to be lower than the peak amplitude of cylinder vibration at a given $Re$. Hysteresis in the cylinder response, near the higher-$U^{\ast }$ end of the lock-in regime, is explored. It is observed that the range of $U^{\ast }$ with hysteretic response increases with decrease in $m^{\ast }$. Interestingly, for a certain range of $m^{\ast }$, the response is hysteretic from a finite $U^{\ast }$ up to $U^{\ast }=\infty$. We refer to this phenomenon as hysteresis forever. It occurs because of the existence of multiple response states of the system at $U^{\ast }=\infty$, for a certain range of $m^{\ast }$. The study brings out the significant differences in the response of the fluid–structure system associated with the critical mass phenomenon between the low- and high-$Re$ regime.

The global stability of laminar axisymmetric low-density jets is investigated in the low Mach number approximation. The linear modal dynamics is found to be characterised by two features: a stable arc branch of eigenmodes and an isolated eigenmode. Both features are studied in detail, revealing that, whereas the former is highly sensitive to numerical domain size and its existence can be linked to spurious feedback from the outflow boundary, the latter is the physical eigenmode that is responsible for the appearance of self-sustained oscillations in low-density jets observed in experiments at low Mach numbers. In contrast to previous local spatio-temporal stability analyses, the present global analysis permits, for the first time, the determination of the critical conditions for the onset of global instability, as well the frequency of the associated oscillations, without additional hypotheses, yielding predictions in fair agreement with previous experimental observations. It is shown that under the conditions of those experiments, viscosity variation with composition, as well as buoyancy, only have a small effect on the onset of instability.

The formation of a singularity in a compressible gas, as described by the Euler equation, is characterized by the steepening and eventual overturning of a wave. Using self-similar variables in two space dimensions and a power series expansion based on powers of $|t_{0}-t|^{1/2}$, $t_{0}$ being the singularity time, we show that the spatial structure of this process, which starts at a point, is equivalent to the formation of a caustic, i.e. to a cusp catastrophe. The lines along which the profile has infinite slope correspond to the caustic lines, from which we construct the position of the shock. By solving the similarity equation, we obtain a complete local description of wave steepening and of the spreading of the shock from a point. The shock spreads in the transversal direction as $|t_{0}-t|^{1/2}$ and in the direction of propagation as $|t_{0}-t|^{3/2}$, as also found in a one-dimensional model problem.

Orderly, or natural, transition to turbulence in dilute polymeric channel flow is studied using direct numerical simulations of a FENE-P fluid. Three Weissenberg numbers are simulated and contrasted to a reference Newtonian configuration. The computations start from infinitesimally small Tollmien–Schlichting (TS) waves and track the development of the instability from the early linear stages through nonlinear amplification, secondary instability and full breakdown to turbulence. At the lowest elasticity, the primary TS wave is more unstable than the Newtonian counterpart, and its secondary instability involves the generation of $\unicode[STIX]{x1D6EC}$-structures which are narrower in the span. These subsequently lead to the formation of hairpin packets and ultimately breakdown to turbulence. Despite the destabilizing influence of weak elasticity, and the resulting early transition to turbulence, the final state is a drag-reduced turbulent flow. At the intermediate elasticity, the growth rate of the primary TS wave matches the Newtonian value. However, unlike the Newtonian instability mode which reaches a saturated equilibrium condition, the instability in the polymeric flow reaches a periodic state where its energy undergoes cyclical amplification and decay. The spanwise size of the secondary instability in this case is commensurate with the Newtonian $\unicode[STIX]{x1D6EC}$-structures, and the extent of drag reduction in the final turbulent state is enhanced relative to the lower elasticity condition. At the highest elasticity, the exponential growth rate of the TS wave is weaker than the Newtonian flow and, as a result, the early linear stage is prolonged. In addition, the magnitude of the saturated TS wave is appreciably lower than the other conditions. The secondary instability is also much wider in the span, with weaker ejection and without hairpin packets. Instead, streamwise-elongated streaks are formed and break down to turbulence via secondary instability. The final state is a high-drag-reduction flow, which approaches the Virk asymptote.

The topology of vortex breakdown in the confined flow generated by a rotating lid in a closed container with a polygonal cross-section geometry has been analysed experimentally and numerically for different height/radius aspect ratios $h$ from 0.5 to 3.0. The locations of stagnation points of the breakdown bubble emergence and corresponding Reynolds numbers were determined experimentally and numerically by STAR-CCM+ computational fluid dynamics software for square, pentagonal, hexagonal and octagonal cross-section configurations. The flow pattern and velocity were observed and measured by combining seeding particle visualization and laser Doppler anemometry. The vortex breakdown size and position on the container axis were identified for Reynolds numbers ranging from 500 to 2800 in steady flow conditions. The obtained results were compared with the flow structure in the closed cylindrical container. The results allowed revealing regularities of formation of the vortex breakdown bubble depending on $Re$ and $h$ and the cross-section geometry of the confined container. It was found in a diagram of $Re$ versus $h$ that reducing the number of cross-section angles from eight to four shifts the breakdown bubble location to higher Reynolds numbers and a smaller aspect ratio. The vortex breakdown bubble area for octagonal cross-section was detected to correspond to the one for the cylindrical container but these areas for square and cylindrical containers do not overlap in the entire range of aspect ratio.

When hovering over sandy terrain, the rotor of helicopters generates a downward jet that induces resuspension of dust and debris. We investigate the mechanisms that govern particle resuspension in such flow using an Eulerian–Lagrangian approach based on large-eddy simulation of turbulence. The wake generated by the helicopter is modelled as a vertical impinging jet, to which a sequence of periodically forced azimuthal vortices is superposed. The resulting flow field provides a unique range of flow scales with which the particles can interact. Downstream of the impingement region, layers of negative azimuthal vorticity (secondary vortices) form on the upwash side of the primary azimuthal (large-scale) vortices. These layers then detach from the surface together with the near-wall (small-scale) vortices populating the wall-jet region. We show how the dynamics of sediments is governed by its interaction with these structures. After initial lift off from the impingement surface, particles accumulate in regions where near-wall vortices roll around the impinging azimuthal vortex, forming rib-like structures that either propel particles away from the azimuthal vortex or entrap them in the shear layer between the azimuthal and secondary vortices. We demonstrate that these trapped particles are more likely to reach the outer flow region and generate a persistent cloud of airborne particles. We also show that, in a time-averaged sense, particle resuspension and deposition fluxes balance each other near the impingement surface.

The role of the Coriolis effect on the attachment of the leading edge vortex (LEV) is investigated. Toward that end, the Navier–Stokes equations are solved in the non-inertial reference frame of a high angle of attack $\unicode[STIX]{x1D6FC}$ rotating wing with the Coriolis term being artificially tuned. Reynolds numbers in the range $Re\in [100;750]$ are considered to identify the interplay between Coriolis and viscous effects. Similarly, artificial tuning of the centrifugal term is achieved to identify the interplay between Coriolis and centrifugal effects. It is shown that (i) the Coriolis effect is the key element in LEV stability for $Re>200$ , (ii) viscous effects are the key element for $Re<200$ and (iii) centrifugal effects have a marginal role. The Coriolis effect is found to promote spanwise flow in the core and behind the LEV, which is known to promote outboard vorticity transport and presumably contributes to stabilizing the aft boundary layer. These mechanisms of LEV stabilization have increased authority as $\unicode[STIX]{x1D6FC}$ decreases.

The effect of large-scale forcing on the second- and third-order longitudinal velocity structure functions, evaluated at the Taylor microscale $r=\unicode[STIX]{x1D706}$, is assessed in various turbulent flows at small to moderate values of the Taylor microscale Reynolds number $R_{\unicode[STIX]{x1D706}}$. It is found that the contribution of the large-scale terms to the scale by scale energy budget differs from flow to flow. For a fixed $R_{\unicode[STIX]{x1D706}}$, this contribution is largest on the centreline of a fully developed channel flow but smallest for stationary forced periodic box turbulence. For decaying-type flows, the contribution lies between the previous two cases. Because of the difference in the large-scale term between flows, the third-order longitudinal velocity structure function at $r=\unicode[STIX]{x1D706}$ differs from flow to flow at small to moderate $R_{\unicode[STIX]{x1D706}}$. The effect on the second-order velocity structure functions appears to be negligible. More importantly, the effect of $R_{\unicode[STIX]{x1D706}}$ on the scaling range exponent of the longitudinal velocity structure function is assessed using measurements of the streamwise velocity fluctuation $u$, with $R_{\unicode[STIX]{x1D706}}$ in the range 500–1100, on the axis of a plane jet. It is found that the magnitude of the exponent increases as $R_{\unicode[STIX]{x1D706}}$ increases and the rate of increase depends on the order $n$. The trend of published structure function data on the axes of an axisymmetric jet and a two-dimensional wake confirms this dependence. For a fixed $R_{\unicode[STIX]{x1D706}}$, the exponent can vary from flow to flow and for a given flow, the larger $R_{\unicode[STIX]{x1D706}}$ is, the closer the exponent is to the value predicted by Kolmogorov (Dokl. Akad. Nauk SSSR, vol. 30, 1941a, pp. 299–303) (hereafter K41). The major conclusion is that the finite Reynolds number effect, which depends on the flow, needs to be properly accounted for before determining whether corrections to K41, arising from the intermittency of the energy dissipation rate, are needed. We further point out that it is imprudent, if not incorrect, to associate the finite Reynolds number effect with a consequence of the modified similarity hypothesis introduced by Kolmogorov (J. Fluid Mech., vol. 13, 1962, pp. 82–85) (K62); we contend that this association has misled the vast majority of post K62 investigations of the consequences of K62.

The time-asymptotic linear stability of pulsatile flow in a channel with compliant walls is studied together with the evaluation of modal transient growth within the pulsation period of the basic flow as well as non-modal transient growth. Both one (vertical-displacement) and two (vertical and axial) degrees-of-freedom compliant-wall models are implemented. Two approaches are developed to study the dynamics of the coupled fluid–structure system, the first being a Floquet analysis in which disturbances are decomposed into a product of exponential growth and a sum of harmonics, while the second is a time-stepping technique for the evolution of the fundamental solution (monodromy) matrix. A parametric study of stability in the non-dimensional parameter space, principally defined by Reynolds number ($Re$), Womersley number ($Wo$) and amplitude of the applied pressure modulation ($\unicode[STIX]{x1D6EC}$), is then conducted for compliant walls of fixed geometric and material properties. The flow through a rigid channel is shown to be destabilized by pulsation for low $Wo$, stabilized due to Stokes-layer effects at intermediate $Wo$, while the critical $Re$ approaches the steady Poiseuille-flow result at high $Wo$, and that these effects are made more pronounced by increasing $\unicode[STIX]{x1D6EC}$. Wall flexibility is shown to be stabilizing throughout the $Wo$ range but, for the relatively stiff wall used, is more effective at high $Wo$. Axial displacements are shown to have negligible effect on the results based upon only vertical deformation of the compliant wall. The effect of structural damping in the compliant-wall dynamics is destabilizing, thereby suggesting that the dominant inflectional (Rayleigh) instability is of the Class A (negative-energy) type. It is shown that very high levels of modal transient growth can occur at low $Wo$, and this mechanism could therefore be more important than asymptotic amplification in causing transition to turbulent flow for two-dimensional disturbances. Wall flexibility is shown to ameliorate mildly this phenomenon. As $Wo$ is increased, modal transient growth becomes progressively less important and the non-modal mechanism can cause similar levels of transient growth. We also show that oblique waves having non-zero transverse wavenumbers are stable to higher values of critical $Re$ than their two-dimensional counterparts. Finally, we identify an additional instability branch at high $Re$ that corresponds to wall-based travelling-wave flutter. We show that this is stabilized by the inclusion of structural damping, thereby confirming that it is of the Class B (positive-energy) instability type.

Transport of dense fluid by an inclined gravity current can control the vertical density structure of the receiving basin in many natural and industrial settings. A case familiar to many is a lake fed by river water that is dense relative to the lake water. In laboratory experiments, we pulsed dye into the basin inflow to visualise the transport pathway of the inflow fluid through the basin. We also measured the evolving density profile as the basin filled. The experiments confirmed previous observations that when the turbulent gravity current travelled through ambient fluid of uniform density, only entrainment into the dense current occurred. When the gravity current travelled through the stratified part of the ambient fluid, however, the outer layers of the gravity current outflowed from the current by a peeling detrainment mechanism and moved directly into the ambient fluid over a large range of depths. The prevailing model of a filling box flow assumes that a persistently entraining gravity current entrains fluid from the basin as the current descends to the deepest point in the basin. This model, however, is inconsistent with the transport pathway observed in visualisations and poorly matches the stratifications measured in basin experiments. The main contribution of the present work is to extend the prevailing filling box model by incorporating the observed peeling detrainment. The analytical expressions given by the peeling detrainment model match the experimental observations of the density profiles more closely than the persistently entraining model. Incorporating peeling detrainment into multiprocess models of geophysical systems, such as lakes, will lead to models that better describe inflow behaviour.

Off-surface integral solutions to an inhomogeneous wave equation based on acoustic analogy could suffer from spurious wave contamination when volume integrals are ignored for computation efficiency and vortical/turbulent gusts are convected across the integration surfaces, leading to erroneous far-field directivity predictions. Vortical gusts often exist in aerodynamic flows and it is inevitable their effects are present on the integration surface. In this work, we propose a new sound extrapolation method for acoustic far-field directivity prediction in the presence of vortical gusts, which overcomes the deficiencies in the existing methods. The Euler equations are rearranged to an alternative form in terms of fluctuation variables that contains the possible acoustical and vortical waves. Then the equations are manipulated to an inhomogeneous wave equation with source terms corresponding to surface and volume integrals. With the new formulation, spurious monopole and dipole noise produced by vortical gusts can be suppressed on account of the solenoidal property of the vortical waves and a simple convection process. It is therefore valid to ignore the volume integrals and preserve the sound properties. The resulting new acoustic inhomogeneous convected wave equations could be solved by means of the Green’s function method. Validation and verification cases are investigated, and the proposed method shows a capacity of accurate sound prediction for these cases. The new method is also applied to the challenging airfoil leading edge noise problems by injecting vortical waves into the computational domain and performing aeroacoustic studies at both subsonic and transonic speeds. In the case of a transonic airfoil leading edge noise problem, shocks are present on the airfoil surface. Good agreements of the directivity patterns are obtained compared with direct computation results.

In this paper we examine the characteristics of the interfaces that demarcate regions of relatively uniform streamwise momentum in turbulent boundary layers. The analysis utilises particle image velocimetry databases that span more than an order of magnitude of friction Reynolds number ($Re_{\unicode[STIX]{x1D70F}}=10^{3}$–$10^{4}$), enabling us to provide a detailed description of the interfacial layers as a function of Reynolds number. As reported by Adrian et al. (J. Fluid Mech., vol. 422, 2000, pp. 1–54), these interfaces appear as persistent regions of strong shear with distinct patches of vorticity consistent with a packet-like structure. Here, however, we treat these interfaces as continuous lines, thus averaging the properties of the vortical patches, and find that their geometry is highly contorted and exhibits self-similarity across a wide range of scales. Specifically, the lengths of the edges of uniform momentum zones exhibit a power-law behaviour with a fractal scaling that has a constant exponent across the boundary layer, while the topmost edge or the turbulent/non-turbulent interface shows a sudden increase in the exponent. The accompanying sharp changes in velocity that occur at these edges are found to change in magnitude as a function of wall-normal height, being larger closer to the wall. Further, a Reynolds number invariance is exhibited when the magnitude of the step-like changes in velocity is scaled by the skin-friction velocity, meanwhile, the width across which it occurs is shown to be of the order of the Taylor microscale. Based on these quantitative measures, the Reynolds number scaling observed and the persistent presence of sharp changes in momentum in turbulent boundary layers, a simple model is used to reconstruct the mean velocity profile. Insight gained from the model enhances our understanding of how instantaneous phenomena (such as a zonal-like structural arrangement) manifests in the averaged flow statistics and confirms that the instantaneous momentum in a turbulent boundary layer appears to mainly consist of a step-like profile as a function of wall-normal distance.

The accurate description of the growth or dissolution dynamics of a soluble gas bubble in a super- or undersaturated solution requires taking into account a number of physical effects that contribute to the instantaneous mass transfer rate. One of these effects is the so-called history effect. It refers to the contribution of the local concentration boundary layer around the bubble that has developed from past mass transfer events between the bubble and liquid surroundings. In Part 1 of this work (Peñas-López et al., J. Fluid Mech., vol. 800, 2016b, pp. 180–212), a theoretical treatment of this effect was given for a spherical, isolated bubble. Here, Part 2 provides an experimental and numerical study of the history effect regarding a spherical bubble attached to a horizontal flat plate and in the presence of gravity. The simulation technique developed in this paper is based on a streamfunction–vorticity formulation that may be applied to other flows where bubbles or drops exchange mass in the presence of a gravity field. Using this numerical tool, simulations are performed for the same conditions used in the experiments, in which the bubble is exposed to subsequent growth and dissolution stages, using stepwise variations in the ambient pressure. Besides proving the relevance of the history effect, the simulations highlight the importance that boundary-induced advection has to accurately describe bubble growth and shrinkage, i.e. the bubble radius evolution. In addition, natural convection has a significant influence that shows up in the velocity field even at short times, although given the supersaturation conditions studied here, the bubble evolution is expected to be mainly diffusive.

We present a Lagrangian analysis of nonlinear surface waves propagating zonally on a zonal current in the presence of the Earth’s rotation that shows the existence of two modes of wave motion. The first, ‘fast’ mode, one with wavelengths commonly found for wind waves and swell in the ocean, represents the wave–current interaction counterpart of the rotationally modified Gerstner waves found first by Pollard (J. Geophys. Res., vol. 75, 1970, pp. 5895–5898) that quite closely resemble Stokes waves. The second, slower, mode has a period nearly equal to the inertial period and has a small vertical scale such that very long, e.g. $O(10^{4}~\text{km})$ wavelength, waves have velocities etc. that decay exponentially from the free surface over a scale of $O(10~\text{m})$ that is proportional to the strength of the mean current. In both cases, the particle trajectories are closed in a frame of reference moving with the mean current, with particle motions in the second mode describing inertial circles. Given that the linear analysis of the governing Eulerian equations only captures the fast mode, the slow mode is a fundamentally nonlinear phenomenon in which very small free surface deflections are manifestations of an energetic current.