We present and analyse a model for the spherical pulsations and translational motions of a pair of interacting gas bubbles in an incompressible liquid. The model is derived rigorously in the context of potential flow theory and contains all terms up to and including fourth order in the inverse separation distance between the bubbles. We use this model to study the cases of both weak and moderate applied acoustic forcing. For weak acoustic forcing, the radial pulsations of the bubbles are weakly coupled, which allows us to obtain a nonlinear time-averaged model for the relative distance between the bubbles. The two parameters of the time-averaged model classify four different dynamical regimes of relative translational motion, two of which correspond to the attraction and repulsion of classical secondary Bjerknes theory. Also predicted is a pattern in which the bubbles exhibit stable, time-periodic translational oscillations along the line connecting their centres, and another pattern in which there is an unstable separation distance such that bubble pairs can either attract or repel each other depending on whether their initial separation distance is smaller or larger than this value. Moreover, it is shown that the full governing equations possess the dynamics predicted by the time-averaged model. We also study the case of moderate-amplitude acoustic forcing, in which the bubble pulsations are more strongly coupled to each other and bubble translation also affects the radial pulsations. Here, radial harmonics and nonlinear phase shifting play a significant role, as bubble pairs near resonances are observed to translate in patterns opposite to those predicted by classical secondary Bjerknes theory. In this work, dynamical systems techniques and the method of averaging are the primary mathematical methods that are employed.

A range of two- and three-dimensional problems is explored featuring the gravity-driven flow of a continuous thin liquid film over a non-porous inclined flat surface containing well-defined topography. These are analysed principally within the framework of the lubrication approximation, where accurate numerical solution of the governing nonlinear equations is achieved using an efficient multigrid solver.

Results for flow over one-dimensional steep-sided topographies are shown to be in very good agreement with previously reported data. The accuracy of the lubrication approximation in the context of such topographies is assessed and quantified by comparison with finite element solutions of the full Navier–Stokes equations, and results support the consensus that lubrication theory provides an accurate description of these flows even when its inherent assumptions are not strictly satisfied. The Navier–Stokes solutions also illustrate the effect of inertia on the capillary ridge/trough and the two-dimensional flow structures caused by steep topography.

Solutions obtained for flow over localized topography are shown to be in excellent agreement with the recent experimental results of Decré & Baret (2003) for the motion of thin water films over finite trenches. The spread of the ‘bow wave’, as measured by the positions of spanwise local extrema in free-surface height, is shown to be well-represented both upstream and downstream of the topography by an inverse hyperbolic cosine function.

An explanation, in terms of local flow rate, is given for the presence of the ‘downstream surge’ following square trenches, and its evolution as trench aspect ratio is increased is discussed. Unlike the upstream capillary ridge, this feature cannot be completely suppressed by increasing the normal component of gravity. The linearity of free-surface response to topographies is explored by superposition of the free surfaces corresponding to two ‘equal-but-opposite’ topographies. Results confirm the findings of Decré & Baret (2003) that, under the conditions considered, the responses behave in a near-linear fashion.

Large-eddy simulation (LES) has been used to study the flow in a planar asymmetric diffuser. The wide range of spatial and temporal scales, the presence of an adverse pressure gradient, and the formation of an unsteady separation bubble in the rear part of the diffuser make this flow a challenging test case for assessing the predictive capability of LES. Simulation results for mean flow, pressure recovery and skin friction are in excellent agreement with data from two recent experiments. The inflow consists of a fully developed turbulent channel flow at a Reynolds number based on shear velocity, Reτ=500. It is found that accurate representation of the in flow velocity field is critical for accurate prediction of the flow in the diffuser. Although the simulation in the diffuser is well resolved, the subgrid-scale model plays a significant role for both mean momentum and turbulent kinetic energy balances. Subgrid-scale stresses contribute a maximum of 8% to the local value of the total shear stress with the maximum values found in the inlet duct and along the flat wall where the flow remains attached. The subgrid-scale model adapts to the enhanced turbulence levels in the rear part of the diffuser by providing more than 80% of the dissipation rate for turbulent kinetic energy. The unsteady separation excites large scales of motion which extend over the major part of the duct cross-section and penetrate deeply into the core of the flow. Instantaneous flow reversal is observed along both walls immediately behind the diffuser throat which is far upstream of the location of main separation. While the mean flow profile changes gradually as the flow enters the expansion, turbulent stresses undergo rapid changes over a short streamwise distance along the deflected wall. An explanation is offered which considers the strain field as well as the influence of geometry changes. The effect of grid resolution and spanwise domain size on the flow field prediction has been documented and this allows an assessment of the computational requirements for carrying out such simulations.

Experimental data obtained in various turbulent flows are analysed by means of orthogonal wavelet transforms. Several configurations are analysed: homogeneous grid turbulence at low and very low Reλ, and fully developed jet turbulence at moderate and high Reλ. It is shown by means of the wavelet decomposition in combination with the form of scaling named extended self-similarity that some statistical properties of fully developed turbulence may be extended to low-Reλ flows. Indeed, universal properties related to intermittency are observed down to Reλ≃10. Furthermore, the use of a new conditional averaging technique of velocity signals, based on the wavelet transform, permits the identification of the time signatures of coherent structures which may or may not be responsible for intermittency depending on the scale of the structure itself. It is shown that in grid turbulence, intermittency at the smallest scales is related to structures with small characteristic size and with a shape that may be related to the passage of vortex tubes. In jet turbulence, the longitudinal velocity component reveals that intermittency may be induced by structures with a size of the order of the integral length. This effect is interpreted as the signature of the characteristic jet mixing layer structures. The structures identified on the transverse velocity component of the jet case turn out on the other hand not to be affected by the mixing layer and the corresponding shape is again correlated with the signature of vortex tubes.

We consider bound states in the continuum (BSCs) or embedded trapped modes in two- and three-dimensional acoustic axisymmetric duct–cavity structures. We demonstrate numerically that, under variation of the length of the cavity, multiple BSCs occur due to the Friedrich–Wintgen two-mode full destructive interference mechanism. The BSCs are detected by tracing the resonant widths to the points of the collapse of Fano resonances where one of the two resonant modes acquires infinite life-time. It is shown that the approach of the acoustic coupled mode theory cast in the truncated form of a two-mode approximation allows us to analytically predict the BSC frequencies and shape functions to a good accuracy in both two and three dimensions.

The steady, incompressible, isothermal, developing flow in a square-section curved duct with smooth walls has been investigated. The 40 × 40 mm duct had a radius ratio of 2·3 with long upstream and downstream straight ducts attached. Measurements of the longitudinal and radial components of mean velocity, and corresponding components of the Reynolds-stress tensor, were obtained with a laser-Doppler anemometer at a Reynolds number of 4 × 104 and in various cross-stream planes. The secondary mean velocities, driven mainly by the pressure field, attain values up to 28% of the bulk velocity and are largely responsible for the convection of Reynolds stresses in the cross-stream plane. Production of turbulent kinetic energy predominates close to the outer-radius wall and regions with negative contributions to the production exist. Thus, at a bend angle of 90° and near the inner-radius wall, $\overline{u_{\theta}u_r}\partial U_{\theta}/\partial r$ is positive and represents a negative contribution to the generation of turbulent kinetic energy.

In spite of the complex mean flow and Reynolds stress distributions, the cross-stream flow is controlled mainly by the centrifugal force, radial pressure gradient imbalance. As a consequence, calculated mean velocity results obtained from the solution of elliptic differential equations in finite difference form and incorporating a two-equation turbulence model are not strongly dependent on the model; numerical errors are of greater importance.

The effect of a blunt leading edge on bypass transition is studied by numerical simulation. A mixed direct and large-eddy simulation of a flat plate with a super-ellipse leading edge is carried out at various conditions. Onset and completion of transition is seen to move upstream with increasing bluntness. For sharper leading edges, at lower levels of turbulence, transition usually occurs through instabilities on low-speed streaks as observed by Jacobs & Durbin (2001) and Brandt et al. (2004) whereas increasing either the turbulence intensity or the leading-edge bluntness brings into play another mechanism. Free-stream vortices are amplified at the leading edge because of stretching. In the case of particularly strong vortices, this interaction induces a localized streamwise vortical disturbance in the boundary layer which then grows as it convects downstream and eventually breaks down to form a turbulent spot. These disturbances, which are localized and hence wavepacket-like, move at speeds in the range 0.55 U∞–0.65 U∞ and occur in the lower portion of the boundary layer. Simulations conducted with isolated vortices confirm such a response of the boundary layer.

Experiments were performed on concentrated suspensions of large (0·97–1·78 mm mean diameters) neutrally buoyant spherical particles sheared in a concentric-cylinder Couette-flow apparatus in which the inner cylinder rotated while the outer one was fixed. The variations of shear stress with apparent shear rate, concentration, particle diameter and wall roughness were studied, and the results are compared with related experiments of Bagnold. Generally the shear stresses measured in the present experiments were larger than those of Bagnold. The difference can be attributed to differences in the experimental arrangements; Bagnold's flexible-walled inner cylinder was fixed while the outer cylinder rotated. A strong effect of wall roughness was observed. The higher stresses generated with rough walls imply that particle ‘slip’ may have occurred in the smooth wall tests. The larger stresses might also be due to an increase in strength of the interparticle collisions caused by the roughness. No dependence of stress upon particle diameter d was observed for concentrations of about 0·3, but a strong dependence (> d2) was found at the highest concentrations with the rough walls.

A fictitious domain method is used to perform fully resolved numerical simulations of particle-laden turbulent flow in a horizontal channel. The effects of large particles of diameter 0.05 and 0.1 times the channel height on the turbulence statistics and structures are investigated for different settling coefficients and volume fractions (0.79 %–7.08 %) for the channel Reynolds number being 5000. The results indicate the following. (a) When the particle sedimentation effect is negligible (i.e. neutrally buoyant), the presence of particles decreases the maximum r.m.s. of streamwise velocity fluctuation near the wall by weakening the intensity of the large-scale streamwise vortices, while increasing the r.m.s. of the streamwise fluctuating velocity in the region very close to the wall and in the centre region. On the other hand, the particles increase the r.m.s. of transverse and spanwise fluctuating velocities in the near-wall region by inducing the small-scale vortices. (b) When the particle settling effect is so substantial that most particles settle onto the bottom wall and form a particle sediment layer (SL), the SL plays the role of a rough wall and parts of the vortex structures shedding from the SL ascend into the core region and substantially increase the turbulence intensity there. (c) When the particle settling effect is moderate, the effects of particles on the turbulence are a combination of the former two situations, and the Shields number is a good parameter for measuring the particle settling effects (i.e. the particle concentration distribution in the transverse direction). The average velocities of the particle are smaller in the lower half-channel and larger in the upper half-channel compared to the local fluid velocities in the presence of gravity effects. The effects of the smaller particles on the turbulence are found to be stronger at the same particle volume fractions.

Two-dimensional flow of a surface-active monolayer on a thin viscous film is considered. Simplifications of negligible gravity and pressure forces are made. Interfacial properties are described by simple model equations of state. Solutions are obtained for when the monolayer is scraped along the interface by a barrier and a steady state exists where surface advection is balanced by surface diffusion. Surface velocity, film thickness and spreading rate dependence on surface diffusivity are examined.

This paper shows that a long vertical columnar vortex pair created by a double flap apparatus in a strongly stratified fluid is subjected to an instability distinct from the Crow and short-wavelength instabilities known to occur in homogeneous fluid. This new instability, which we name zigzag instability, is antisymmetric with respect to the plane separating the vortices. It is characterized by a vertically modulated twisting and bending of the whole vortex pair with almost no change of the dipole's cross- sectional structure. No saturation is observed and, ultimately, the vortex pair is sliced into thin horizontal layers of independent pancake dipoles. For the largest Brunt–Väisälä frequency N = 1.75 rad s−1 that may be achieved in the experiments, the zigzag instability is observed only in the range of Froude numbers: 0.13 < Fh0 < 0.21 (Fh0 = U0/NR, where U0 and R are the initial dipole travelling velocity and radius). When Fh0 > 0.21, the elliptic instability develops resulting in three-dimensional motions which eventually collapse into a relaminarized vortex pair. Irregular zigzags are then also observed to grow. The threshold for the inhibition of the elliptic instability Fh0 = 0.2±0.01 is independent of N and in good agreement with the theoretical study of Miyazaki & Fukumoto (1992). Complete stabilization for Fh0 < 0.13 is probably due to viscous effects since the associated Reynolds number is low, Re0 < 260. In geophysical flows characterized by low Froude numbers and large Reynolds numbers, we conjecture that this viscous stabilization will occur at much lower Froude number.

It is tentatively argued that this new type of instability may explain the layering widely observed in stratified turbulent flows.

The steady flow in rectangular cavities is investigated both numerically and experimentally. The flow is driven by moving two facing walls tangentially in opposite directions. It is found that the basic two-dimensional flow is not always unique. For low Reynolds numbers it consists of two separate co-rotating vortices adjacent to the moving walls. If the difference in the sidewall Reynolds numbers is large this flow becomes unstable to a stationary three-dimensional mode with a long wavelength. When the aspect ratio is larger than two and both Reynolds numbers are large, but comparable in magnitude, a second two-dimensional flow exists. It takes the form of a single vortex occupying the whole cavity. This flow is the preferred state in the present experiment. It becomes unstable to a three-dimensional mode that subdivides the basic streched vortex flow into rectangular convective cells. The instability is supercritical when both sidewall Reynolds numbers are the same. When they differ the instability is subcritical. From an energy analysis and from the salient features of the three-dimensional flow it is concluded that the mechanism of destabilization is identical to the destabilization mechanism operative in the elliptical instability of highly strained vortices.

The three-dimensional waving plate theory is developed to investigate the swimming performance of fish undulatory motion. In particular, the propulsive effectiveness is discussed. The unsteady potential flow over model rectangular and triangular flexible plates performing a motion which consists of a progressing wave with variable amplitudes is calculated by the vortex ring panel method. It is found that the undulatory motion can reduce three-dimensional effects. It is this important hydrodynamic phenomenon that may be one of the main reasons why such undulation is widely used as the swimming method by a large number of aquatic animals. When the span of the undulating plate is nearly unchanged and the wave amplitude is constant or increases slightly along the chord, and the wavelength is close to the body length, theoretical results show that the swimming performance is best and the flow around the plate has a quasi-two-dimensional property. This swimming method may be observed in many fishes, especially those with an anguilliform mode of propulsion. The modification of the anguilliform mode of propulsion to the carangiform mode is also discussed. It is confirmed that the pronounced necking of the body anterior to the tail, which acts to improve the propulsive performance, is a major morphological adaptation of fishes using the carangiform mode, in which the characteristic nature of flexural movement confined to the rear part of the body is that the amplitude of undulation increases posteriorly and no complete wavelength is at any time apparent.

The structure and dynamics of vorticity ω and rate of strain S are studied using direct numerical simulations (DNS) of incompressible homogeneous isotropic turbulence. In particular, characteristics of the pressure Hessian Π, which describe non-local interaction of ω and S, are presented. Conditional Lagrangian statistics which distinguish high-amplitude events in both space and time are used to investigate the physical processes associated with their evolution. The dynamics are examined on the principal strain basis which distinguishes vortex stretching and induced rotation of the principal axes of S. The latter mechanism is associated with misaligned ω with respect to S, a condition which predominates in isotropic turbulence and is dynamically significant, particularly in rotation-dominated regions of the flow. Locally-induced rotation of the principal axes acts to orient ω towards the direction of either the intermediate or most compressive principal strain. The tendency towards compressive straining of ω is manifested at the termini of the high-amplitude tube-like structures in the flow. Non-locally-induced rotation, associated with Π, tends to counteract the locally-induced rotation. This is due to the strong alignment between ω and the eigenvector of Π corresponding to its smallest eigenvalue and is indicative of the controlling influence of the proximate structure on the dynamics. High-amplitude rotation-dominated regions deviate from Burgers vortices due to the misalignment of ω. Although high-amplitude strain-dominated regions are promoted primarily by local dynamics, the associated spatial structure is less organized and more discontinuous than that of rotation-dominated regions.

Numerical solutions have been obtained for steady streaming flow past an axisymmetric drop over a wide range of Reynolds numbers (0.005 [les ] Re [les ] 250), Weber numbers (0.005 [les ] We [les ] 14), viscosity ratios (0.001 [les ] λ [les ] 1000), and density ratios (0.001 [les ] ζ [les ] 1000). Our results indicate that at lower Reynolds numbers the shape of the drop tends toward a spherical cap with increasing We, but at higher Re the body becomes more disk shaped with increasing We. Unlike the recirculating wake behind an inviscid bubble or solid particle, the eddy behind a drop is detached from the interface. The size of the eddy and the separation distance from the drop depend on the four dimensionless parameters of the problem. The motion of the fluid inside the drop appears to control the behaviour of the external flow near the body, and even for cases when λ and ζ [Lt ] 1 (a ‘real’ bubble), a recirculating wake remains unattached.

The response of a turbulent boundary layer to suddenly applied concave surface curvature with δ/R = 0.01–0.02 is investigated. The main conclusion of this and the companion paper by Muck, Hoffmann & Bradshaw (1985) is that the effects of concave (destabilizing) and convex (stabilizing) curvature on boundary layers – and presumably on other shear layers – are totally different, even qualitatively. As shown in Muck, Hoffmann & Bradshaw (1985), convex curvature tends to attenuate the pre-existing turbulence and, at least in the case of mild curvature, there are no large changes in statistical average eddy shape. Concave curvature, on the other hand, can lead to the quasi-inviscid generation of longitudinal (‘Taylor–Görtler’) vortices, and we show that significant changes in the turbulence structure are induced both directly by the curvature and indirectly by the vortices.

Fine-scale scalar mixing in gas-phase planar turbulent jets is studied using measurements of three-component scalar gradient and scalar energy dissipation rate fields. Simultaneous planar Rayleigh scattering and planar laser-induced fluorescence, applied in parallel planes, yield the three-dimensional scalar field measurements. The spatial resolution is sufficient to permit differentiation in all three spatial directions. The data span a range of outer-scale Reynolds numbers from 3290 to 8330. Direct measurement of the thicknesses of scalar dissipation structures (layers) shows that the thicknesses scale with outer-scale Reynolds number as $\hbox{\it Re}_\delta^{-3/4}$, consistent with Kolmogorov/Batchelor scaling. Average layer thicknesses are described by the relation $\lambda_D=14.5\,\delta\, \hbox{\it Re}_\delta^{-3/4}\hbox{\it Sc}^{-1/2}$. There is no evidence here that Taylor scaling ($\lambda_D\propto\delta\, \hbox{\it Re}_\delta^{-1/2}$) plays a significant role in the scalar dissipation process. The present data resolve a range of length scales from the dissipation scales up to nearly the jet full width, and thus can be used in a priori testing of subgrid models for scalar mixing in large-eddy simulations (LES). Comparison of two models for subgrid scalar variance, a scale-similarity model and a gradient-based model, indicates that the scale-similarity model is more accurate at larger LES filter sizes.

Particle image velocimetry (PIV) is used to investigate the three-dimensional wakes of rigid pitching panels with a trapezoidal geometry, chosen to model idealized fish caudal fins. Experiments are performed for Strouhal numbers from 0.17 to 0.56 for two different trailing edge pitching amplitudes. A Lagrangian coherent structure (LCS) analysis is employed to investigate the formation and evolution of the panel wake. A classic reverse von Kármán vortex street pattern is observed along the mid-span of the near wake, but the vortices realign and exhibit strong interactions near the spanwise edges of the wake. At higher Strouhal numbers, the complexity of the wake increases downstream of the trailing edge as the spanwise vortices spread transversely and lose coherence as the wake splits. This wake transition is shown to correspond to a qualitative change in the LCS pattern surrounding each vortex core, and can be identified as a quantitative event that is not dependent on arbitrary threshold levels. The location of this transition is observed to depend on both the pitching amplitude and free stream velocity, but is not constant for a fixed Strouhal number. On the panel surface, the trapezoidal planform geometry is observed to create additional vortices along the swept edges that retain coherence for low Strouhal numbers or high sweep angles. These additional swept-edge structures are conjectured to add to the complex three-dimensional flow near the tips of the panel.

Convective dynamos in a rotating spherical shell feature steady zonal flows. This process is studied numerically for Prandtl numbers of 0.1 and 1, Ekman numbers in the range $E=\te{-4}$–$\te{-5}$, magnetic Prandtl number from 0.5 to 10 and Rayleigh numbers up to 100 times supercritical. The zonal flow is mainly of thermal wind origin, and minimizes the shear of the axisymmetric poloidal magnetic field lines, according to Ferraro's law of corotation. The dissipation in the interior of the fluid is mainly ohmic, while the introduction of rigid velocity boundary conditions confines viscous dissipation in the Ekman boundary layers. The root-mean-square amplitude $U_\varphi$ of the zonal flow in the spherical shell scales as $U_\varphi=(F/\Omega)^{0.5}$, F being the buoyancy flux through the shell and $\Omega$ the rotation rate. As a consequence of the corotation law, this scaling relationship is remarkably independent of the magnetic field amplitude. It does not depend on thermal, kinematic and magnetic diffusivities, owing to the large-scale and steady nature of forcing and dissipative processes. The scaling law is in agreement with the zonal-flow amplitude at the external boundary of the Earth's liquid core.

In this study we consider the separated flow of an inviscid fluid around a moving flat plate. The motion of the plate, which is initially started from rest, is prescribed and unconstrained and we set ourselves the task of fully characterizing the resulting motion in the surrounding fluid. To do this we use a boundary integral representation for the complex-conjugate velocity field $\Phi(z,t)$ and require that the force and torque on the plate be determined as part of the solution. The flow solution is assumed to consist of a bound vortex sheet coincident with the plate and two free vortex sheets that emanate from each of the plate's two sharp edges. The time evolution of these vortex sheets is then considered in general. For physical reasons, the flow solution is required to satisfy the unsteady Kutta condition, which states that $\Phi(z,t)$ must be bounded everywhere, and the rigorous imposition of this condition then yields two types of additional constraint. The first governs the rate at which circulation is shed from the plate's edges and the second ensures that the free vortex sheets are shed tangentially. In fact, all the familiar flow characteristics associated with the imposition of the steady Kutta condition are rigorously shown to have exact parallels in the unsteady case. In addition, explicit expressions for the normal force and torque on the plate are derived. An asymptotic solution to the full system of evolution equations is developed for small times $t > 0$ and a simplified version of this solution is used as an initial condition for a desingularized numerical treatment of the full problem. A fast numerical algorithm is proposed and implemented and the results of several example calculations are presented. The featured examples are limited to high effective angles of attack due to the occurrence of a specific type of event that prevents further time-integration of the evolution equations using the current numerical method. The event corresponds physically to a situation in which a Lagrangian point placed at one of the plate's edges moves onto instead of away from the edge.