The paper continues our investigations of three-dimensional nonlinear resonant fluid sloshing in a square-base basin with finite depth (mean depth/tank length ratio $h\,{\ge}\, 0.3$). The sloshing is forced by a combined sway/surge resonant harmonic excitation of the two lowest natural modes. The new studies are strongly motivated by a discrepancy between previous quantitative theoretical results and experimental measurements and consist of a more precise description of the strong nonlinear amplification of higher modes. The latter is justified here by secondary resonance. Effective frequency domains of the secondary resonance are quantified. An adaptive asymptotic modal theory improves agreement with earlier and new experimental data both in the transient and steady-state conditions. Local breaking and overturning near the walls, that may lead to a ‘switch’ between distinct steady regimes, increase both global damping and generate random-like excitation of higher modes, are extensively discussed.

The evolution of streamwise vortices in a plane mixing layer and their role in the generation of small-scale three-dimensional motion are studied in a closed-return water facility. Spanwise-periodic streamwise vortices are excited by a time-harmonic wavetrain with span wise-periodic amplitude variations synthesized by a mosaic of 32 surface film heaters flush-mounted on the flow partition. For a given excitation frequency, virtually any span wise wavelength synthesizable by the heating mosaic can be excited and can lead to the formation of streamwise vortices before the rollup of the primary vortices is completed. The onset of streamwise vortices is accompanied by significant distortion in the transverse distribution of the streamwise velocity component. The presence of inflexion points, absent in corresponding velocity distributions of the unforced flow, suggests the formation of locally unstable regions of large shear in which broadband perturbations already present in the base flow undergo rapid amplification, followed by breakdown to small-scale motion. Furthermore, as a result of spanwise-non-uniform excitation the cores of the primary vortices are significantly altered. The three-dimensional features of the streamwise vortices and their interaction with the base flow are inferred from surfaces of r.m.s. velocity fluctuations and an approximation to cross-stream vorticity using three-dimensional single component velocity data. The striking enhancement of small-scale motion and the spatial modification of its distribution, both induced by the streamwise vortices, can be related to the onset of the mixing transition.

The three-dimensional time-dependent turbulent flow in the stably stratified Ekman layer over a smooth surface is computed numerically by directly solving the Navier–Stokes equations, using the Boussinesq approximation to account for buoyancy effects. All relevant scales of motion are included in the simulation so that no turbulence model is needed. The Ekman layer is an idealization of the Earth's boundary layer and provides information concerning atmospheric turbulence models. We find that, when non-dimensionalized according to Nieuwstadt's local scaling scheme, some of the simulation data agree very well with atmospheric measurements. The results also suggest that Brost & Wyngaard's ‘constant Froude number’ and Hunt's ‘shearing length’ stable layer models for the dissipation rate of turbulent kinetic energy are both valid, when Reynolds number effects are accounted for. Simple gradient closures for the temperature variance and heat flux demonstrate the same variation with Richardson number as in Mason & Derbyshire's large-eddy simulation (LES) study, implying both that the models are relatively insensitive to Reynolds number and that local scaling should work well when applied to the stable atmospheric layer. In general we find good agreement between the direct numerical simulation (DNS) results reported here and Mason & Derbyshire's LES results.

Previous work by the present authors on the onset of convection in a layered porous medium heated from below is extended to an investigation of the heat transported by convection at slightly supercritical Rayleigh numbers.

The two-dimensional convection patterns and associated values of the critical Rayleigh number, cell width and slope of the Nusselt-number graph are calculated for two- and three-layer configurations over a wide range of layer depth and permeability ratios. The results show that the commonly studied problem of a homogeneous layer bounded above and below by impermeable boundaries is a special case, in that the slope of the Nusselt-number graph at the critical point is nearly independent of cell width. For a homogeneous layer with a permeable upper boundary, and for multi-layered systems, the slope of this graph depends strongly on cell width.

A mathematical model and numerical method for studying the collective dynamics of geotactic, gyrotactic and chemotactic micro-organisms immersed in a viscous fluid is presented. The Navier–Stokes equations of fluid dynamics are solved in the presence of a discrete collection of micro-organisms. These microbes act as point sources of gravitational force in the fluid equations, and thus affect the fluid flow. Physical factors, e.g. vorticity and gravity, as well as sensory factors affect swimming speed and direction. In the case of chemotactic microbes, the swimming orientation is a function of a molecular field. In the model considered here, the molecules are a nutrient whose consumption results in an upward gradient of concentration that drives its downward diffusion. The resultant upward chemotactically induced accumulation of cells results in (Rayleigh–Taylor) instability and eventually in steady or chaotic convection that transports molecules and affects the translocation of organisms. Computational results that examine the long-time behaviour of the full nonlinear system are presented.

The actual dynamical system consisting of fluid and suspended swimming organisms is obviously three-dimensional, as are the basic modelling equations. While the computations presented in this paper are two-dimensional, they provide results that match remarkably well the spatial patterns and long-time temporal dynamics of actual experiments; various physically applicable assumptions yield steady states, chaotic states, and bottom-standing plumes. The simplified representation of microbes as point particles allows the variation of input parameters and modelling details, while performing calculations with very large numbers of particles (≈104–105), enough so that realistic cell concentrations and macroscopic fluid effects can be modelled with one particle representing one microbe, rather than some collection of microbes. It is demonstrated that this modelling framework can be used to test hypotheses concerning the coupled effects of microbial behaviour, fluid dynamics and molecular mixing. Thus, not only are insights provided into the differing dynamics concerning purely geotactic and gyrotactic microbes, the dynamics of competing strategies for chemotaxis, but it is demonstrated that relatively economical explorations in two dimensions can deliver striking insights and distinguish among hypotheses.

The transition from laminar to turbulent of the natural-convection flow inside a square, differentially heated cavity with adiabatic horizontal walls is calculated, using the finite-volume method. The purpose of this study is firstly to determine the dependence of the laminar-turbulent transition on the Prandtl number and secondly to investigate the physical mechanisms responsible for the bifurcations observed. It is found that in the square cavity, for Prandtl numbers between 0.25 and 2.0, the transition occurs through periodic and quasi-periodic flow regimes. One of the bifurcations is related to an instability occurring in a jet-like fluid layer exiting from those corners of the cavity where the vertical boundary layers are turned horizontal. This instability is mainly shear-driven and the visualization of the perturbations shows the occurrence of vorticity concentrations which are very similar to Kelvin–Helmholtz vortices in a plane jet, suggesting that the instability is a Kelvin–Helmholtz-type instability. The other bifurcation for Prandtl numbers between 0.25 and 2.0 occurs in the boundary layers along the vertical walls. It differs however from the related instability in the natural-convection boundary layer along an isolated vertical plate: the instability in the cavity is shear-driven whereas the instability along the vertical plate is mainly buoyancy-driven. For Prandtl numbers between 2.5 and 7.0, it is found that there occurs an immediate transition from the steady to the chaotic flow regime without intermediate regimes. This transition is also caused by instabilities originating and concentrated in the vertical boundary layers.

Rayleigh instability on an axisymmetric viscous film around a vertical fibre produces localized wave structures (pulses) on a flat substrate film that can grow by an order of magnitude to form large capillary drops. We show that this drop formation process is driven by a unique mechanism in the form of an ever-growing pulse which leaves behind a trailing film thinner than the one it advances into. In addition to accumulating liquid from the film, the growing pulse also captures smaller and slower pulses in coalescence cascades. We construct this supercritical growing pulse by matched asymptotics and show that it will eventually evolve into a capillary drop if the local film thickness h is larger than hc=1.68R3H−2 where R is the fibre radius and H the capillary length. We also show that subcritical pulses with h<hc will equilibrate into stationary pulses that do not grow or coalesce readily to form drops.

State estimation of turbulent near-wall flows based on wall measurements is one of the key pacing items in model-based flow control, with low-Re channel flow providing the canonical testbed. Model-based control formulations in such settings are often separated into two subproblems: estimation of the near-wall flow state via skin friction and pressure measurements at the wall, and (based on this estimate) control of the near-wall flow field fluctuations via actuation of the fluid velocity at the wall. In our experience, the turbulent state estimation sub-problem has consistently proven to be the more difficult of the two. Though many estimation strategies have been tested on this problem (by our group and others), none have accurately captured the turbulent flow state at the outer boundary of the buffer layer (5 ≤ y+ ≤ 30), which is deemed to be an important milestone, as this is the approximate range of the characteristic near-wall turbulent structures, the accurate estimation of which is important for the control problem. Leveraging the ensemble Kalman filter (an effective variant of the Kalman filter which scales well to high-dimensional systems), the present paper achieves at least an order of magnitude improvement (in the near-wall region) over the best results available in the published literature on the estimation of low-Reynolds number turbulent channel flow based on wall information alone.

Sound transmission through ducts of constant cross-section with a uniform inviscid mean flow and a constant acoustic lining (impedance wall) is classically described by a modal expansion, where the modes are eigenfunctions of the corresponding Laplace eigenvalue problem along a duct cross-section. A natural extension for ducts with cross-section and wall impedance that are varying slowly (compared to a typical acoustic wavelength and a typical duct radius) in the axial direction is a multiple-scales solution. This has been done for the simpler problem of circular ducts with homentropic irrotational flow. In the present paper, this solution is generalized to the problem of ducts of arbitrary cross-section. It is shown that the multiple-scales problem allows an exact solution, given the cross-sectional Laplace eigensolutions. The formulation includes both hollow and annular geometries. In addition, the turning point analysis is given for a single hard-wall cut-on, cut-off transition. This appears to yield the same reflection and transmission coefficients as in the circular duct problem.

This paper deals with the development of the flow in a curved tube near the inlet. The solution is obtained by the method of matched asymptotic expansions. Two inlet conditions are considered: (i) the condition of constant dynamic pressure at the entrance, which may be of practical interest in applications to blood flow in the aorta; and (ii) a uniform entry condition. It is shown that the geometry and the nature of the entry condition appreciably influence the initial development of the flow. The effect of the secondary flow due to the curvature on the wall shear is discussed and it is shown that the cross-over between shear maxima on the inside and the outside of the tube occurs at a downstream distance which is 1·9 times the radius of the tube for entry condition (i) while in the case of entry condition (ii) it is 0·95 times the radius, which is half the distance required in case (i). It is found that the pressure distribution is not significantly influenced by the secondary flow during the initial development of the motion. The analysis, which is developed for steady motion, can be extended to pulsatile flows, which are of greater physiological interest.

A novel reduced-order model for time-varying nonlinear flows arising from a resolvent decomposition based on the time-mean flow is proposed. The inputs required for the model are the mean-flow field and a small set of velocity time-series data obtained at isolated measurement points, which are used to fix relevant frequencies, amplitudes and phases of a limited number of resolvent modes that, together with the mean flow, constitute the reduced-order model. The technique is applied to derive a model for the unsteady three-dimensional flow in a lid-driven cavity at a Reynolds number of 1200 that is based on the two-dimensional mean flow, three resolvent modes selected at the most active spanwise wavenumber, and either one or two velocity probe signals. The least-squares full-field error of the reconstructed velocity obtained using the model and two point velocity probes is of the order of 5 % of the lid velocity, and the dynamical behaviour of the reconstructed flow is qualitatively similar to that of the complete flow.

A theoretical description is given for infinitesimal non-axisymmetric disturbances of a shear flow caused by differential rotation. It is assumed that the deviations from the state of rigid rotation are small corresponding to the case of small Rossby number. The shear flow becomes unstable at a finite critical value of the Rossby number, at which the inertial forces overcome the friction in the Ekman boundary layers and the constraint imposed by the variation in depth of the container. The governing equation is closely related to the Rayleigh stability equation in the theory of hydrodynamic stability of plane parallel flow. Experimental observations by R. Hide show reasonable agreement with the theoretical predictions.

We argue that the hypothesis of preservation of shape of dimensionless second- and third-order correlations during decay of incompressible homogeneous magnetohydrodynamic (MHD) turbulence requires, in general, at least two independent similarity length scales. These are associated with the two Elsässer energies. The existence of similarity solutions for the decay of turbulence with varying cross-helicity implies that these length scales cannot remain in proportion, opening the possibility for a wide variety of decay behaviour, in contrast to the simpler classic hydrodynamics case. Although the evolution equations for the second-order correlations lack explicit dependence on either the mean magnetic field or the magnetic helicity, there is inherent implicit dependence on these (and other) quantities through the third-order correlations. The self-similar inertial range, a subclass of the general similarity case, inherits this complexity so that a single universal energy spectral law cannot be anticipated, even though the same pair of third-order laws holds for arbitrary cross-helicity and magnetic helicity. The straightforward notion of universality associated with Kolmogorov theory in hydrodynamics therefore requires careful generalization and reformulation in MHD.

The impact of micron-size drops on a smooth, flat, chemically homogeneous solid surface is studied using a diffuse-interface model (DIM). The model is based on the Cahn–Hilliard theory that couples thermodynamics with hydrodynamics, and is extended to include non-90° contact angles. The (axisymmetric) equations are numerically solved using a combination of finite- and spectral-element methods. The influence of various process and material parameters such as impact velocity, droplet diameter, viscosity, surface tension and wettability on the impact behaviour of drops is investigated. Relevant dimensionless parameters are defined and, depending on the values of the Reynolds number, the Weber number and the contact angle, which for the cases considered here range from 1.3 to 130, 0.43 to 150 and 45° to 135°, respectively, the model predicts the spreading of a droplet with or without recoil or even rebound of the droplet, totally or partially, from the solid surface. The wettability significantly affects the impact behaviour and this is particularly demonstrated with an impact at Re = 130 and We = 1.5, where for θ < 60° the droplet oscillates a few times before attaining equilibrium while for θ ≥ 60° partial rebound of the droplet occurs, i.e. the droplet breaks into two unequal sized drops. The size of the part that remains in contact with the solid surface progressively decreases with increasing θ until at a value θ ≈ 120° a transition to total rebound happens. When the droplet rebounds totally, it has a top-heavy shape.

One of the principal mechanisms by which surfaces and interfaces affect microbial life is by perturbing the hydrodynamic flows generated by swimming. By summing a recursive series of image systems, we derive a numerically tractable approximation to the three-dimensional flow fields of a stokeslet (point force) within a viscous film between a parallel no-slip surface and a no-shear interface and, from this Green’s function, we compute the flows produced by a force- and torque-free micro-swimmer. We also extend the exact solution of Liron & Mochon (J. Engng Maths, vol. 10 (4), 1976, pp. 287–303) to the film geometry, which demonstrates that the image series gives a satisfactory approximation to the swimmer flow fields if the film is sufficiently thick compared to the swimmer size, and we derive the swimmer flows in the thin-film limit. Concentrating on the thick-film case, we find that the dipole moment induces a bias towards swimmer accumulation at the no-slip wall rather than the water–air interface, but that higher-order multipole moments can oppose this. Based on the analytic predictions, we propose an experimental method to find the multipole coefficient that induces circular swimming trajectories, allowing one to analytically determine the swimmer’s three-dimensional position under a microscope.

It is shown that, according to the linearized theory of long waves in a rotating, unbounded sea, if there is a discontinuity in depth along a straight line separating two regions each of uniform depth, then wave motions may exist which are propagated along the discontinuity and whose amplitude falls off exponentially to either side. Thus the discontinuity acts as a kind of wave-guide.

The period of the waves is always greater than the inertial period. The wave period also exceeds the period of Kelvin waves in the deeper medium. As the ratio of the depth tends to infinity, the wave period tends to the inertial period or to the Kelvin wave period, whichever is the greater. On the other hand as the wavelength decreases (within the limits of shallow-water theory) so the waves tend to the non-divergent planetary waves found recently by Rhines.

In an infinite ocean of uniform depth free waves with period greater than a pendulum-day cannot normally be propagated without attenuation (if the Coriolis parameter is constant). But non-uniformities of depth provide a means whereby such energy may be channelled over great distances with little attenuation.

It is suggested that a gradually diminishing discontinuity will act as a chromatograph, each position along the discontinuity being marked by waves of a particular period.

Mass transfer through the flat shear-free surface of a turbulent open-channel flow is investigated over a wide range of Schmidt number (1 $ \le $Sc$ \le $ 200) by means of large-eddy simulations using a dynamic subgrid-scale model. In contrast with situations previously analysed using direct numerical simulation, the turbulent Reynolds number Re is high enough for the near-surface turbulence to be fairly close to isotropy and almost independent of the structure of the flow in the bottom region (the statistics of the velocity field are identical to those described by I. Calmet & J. Magnaudet J. Fluid Mech. vol. 474, 2003, p. 355). The main statistical features of the concentration field are analysed in connection with the structure of the turbulent motion below the free surface, characterized by a velocity macroscale $u$ and an integral length scale $L$. All near-surface statistical profiles are found to be Sc-independent when plotted vs. the dimensionless coordinate Sc$^{1 / 2}yu$/$\nu $ ($y$ is the distance to the surface and $\nu $ is the kinematic viscosity). Mean concentration profiles are observed to be linear throughout an inner diffusive sublayer whose thickness is about one Batchelor microscale, i.e. LSc$^{ - 1 / 2 }$Re$^{ - 3 / 4}$. In contrast, the concentration fluctuations are found to reach their maximum near the edge of the outer diffusive layer which scales as LSc$^{ - 1 / 2}$Re$^{ - 1 / 2}$. Instantaneous views of the near-surface isovalues of the concentration and vertical velocity are used to reveal the influence of the Schmidt number. In particular, it is observed that at high Schmidt number, the tiny concentration fluctuations that subsist in the diffusive sublayer just mirror the divergence of the two-component surface velocity field. Co-spectra of concentration and vertical velocity fluctuations indicate that the main contribution to the turbulent mass flux is provided by eddies whose horizontal size is close to $L$, which strongly supports the view that the mass transfer is governed by large-scale structures. The dimensionless mass transfer rate is observed to be proportional to Sc$^{ - 1 / 2}$ over the whole range of Schmidt number. Based on a frequency analysis of the concentration equation and on the Sc$^{ - 1 / 2}$Re$^{ - 3 / 4 }$scaling of the diffusive sublayer, it is shown that the mass transfer rate at a given Sc is proportional to $\langle {\beta ^2}\rangle ^{1 / 4}$, $\langle {\beta ^2}\rangle $ being the variance of the divergence of the surface velocity field. This yields dimensionless mass transfer rates of the form $\alpha$Sc$^{ - 1 / 2}$Re$^{ - 1 / 4}$, where the value of $\alpha$ is shown to result from both the kinematic blocking of the vertical velocity and the viscous damping of the horizontal vorticity components induced by the free surface.

A thin rigid plate is submerged beneath the free surface of deep water. The plate performs small-amplitude oscillations. The problem of calculating the radiated waves can be reduced to solving a hypersingular boundary integral equation. In the special case of a horizontal circular plate, this equation can be reduced further to one-dimensional Fredholm integral equations of the second kind. If the plate is heaving, the problem becomes axisymmetric, and the resulting integral equation has a very simple structure; it is a generalization of Love's integral equation for the electrostatic field of a parallel-plate capacitor. Numerical solutions of the new integral equation are presented. It is found that the added-mass coefficient becomes negative for a range of frequencies when the disc is sufficiently close to the free surface.

We present experimental observations of the velocity and spatial distribution of inertial particles dispersed in turbulent downward flow through a vertical channel at friction Reynolds numbers $\mathit{Re}_{\unicode[STIX]{x1D70F}}=235$ and 335. The working fluid is air laden with size-selected glass microspheres, having Stokes numbers $St=\mathit{O}(10)$ and $\mathit{O}(100)$ when based on the Kolmogorov and viscous time scales, respectively. Cases at solid volume fractions $\unicode[STIX]{x1D719}_{v}=3\times 10^{-6}$ and $5\times 10^{-5}$ are considered. In the more dilute regime, the particle concentration profile shows near-wall and centreline maxima compatible with a turbophoretic drift down the gradient of turbulence intensity; the particles travel at speed similar to that of the unladen flow except in the near-wall region; and their velocity fluctuations generally follow the unladen flow level over the channel core, exceeding it in the near-wall region. The denser regime presents substantial differences in all measured statistics: the near-wall concentration peak is much more pronounced, while the centreline maximum is absent; the mean particle velocity decreases over the logarithmic and buffer layers; and particle velocity fluctuations and deposition velocities are enhanced. An analysis of the spatial distributions of particle positions and velocities reveals different behaviours in the core and near-wall regions. In the channel core, dense clusters form which are somewhat elongated, tend to be preferentially aligned with the vertical/streamwise direction and travel faster than the less concentrated particles. In the near-wall region, the particles arrange in highly elongated streaks associated with negative streamwise velocity fluctuations, several channel heights in length and spaced by $\mathit{O}(100)$ wall units, supporting the view that these are coupled to fluid low-speed streaks typical of wall turbulence. The particle velocity fields contain a significant component of random uncorrelated motion, more prominent for higher $St$ and in the near-wall region.

A direct numerical simulation (DNS) of turbulent flow in a three-dimensional diffuser at Re = 10000 (based on bulk velocity and inflow-duct height) was performed with a massively parallel high-order spectral element method running on up to 32768 processors. Accurate inflow condition is ensured through unsteady trip forcing and a long development section. Mean flow results are in good agreement with experimental data by Cherry et al. (Intl J. Heat Fluid Flow, vol. 29, 2008, pp. 803–811), in particular the separated region starting from one corner and gradually spreading to the top expanding diffuser wall. It is found that the corner vortices induced by the secondary flow in the duct persist into the diffuser, where they give rise to a dominant low-speed streak, due to a similar mechanism as the ‘lift-up effect’ in transitional shear flows, thus governing the separation behaviour. Well-resolved simulations of complex turbulent flows are thus possible even at realistic Reynolds numbers, providing accurate and detailed information about the flow physics. The available Reynolds stress budgets provide valuable references for future development of turbulence models.