Cross-streamline migration of soft particles in suspensions is essential for cell and DNA sorting, blood flow, polymer processing and so on. Pioneering work by Poiseuille on blood flow in vivo revealed an erythrocyte-free layer close to blood vessel walls. The formation of this layer is related to a viscous lift force caused by cell deformation that pushes cells towards the centre of blood capillaries. This lift force has in this case a strong impact on blood flow. In contrast, rigid spherical particles migrate from the centre towards the periphery, owing to inertia (the Segré–Silberberg effect). An important open issue is to elucidate the interplay between particle deformation and inertia. By using a capsule suspension model, Krueger, Kaoui & Harting (J. Fluid Mech., 2014, vol. 751, pp. 725–745) discovered that capsule flexibility can suppress the Segré–Silberberg effect and inertia promotes overall flow efficiency thanks to a strong inertial flow focusing effect.

A sessile droplet partially wets a planar solid support. We study the linear stability of this spherical-cap base state to disturbances whose three-phase contact line is (i) pinned, (ii) moves with fixed contact angle and (iii) moves with a contact angle that is a smooth function of the contact-line speed. The governing hydrodynamic equations for inviscid motions are reduced to a functional eigenvalue problem on linear operators, which are parameterized by the base-state volume through the static contact angle and contact-line mobility via a spreading parameter. A solution is facilitated using inverse operators for disturbances (i) and (ii) to report frequencies and modal shapes identified by a polar $k$ and azimuthal $l$ wavenumber. For the dynamic contact-line condition (iii), we show that the disturbance energy balance takes the form of a damped-harmonic oscillator with ‘Davis dissipation’ that encompasses the dynamic effects associated with (iii). The effect of the contact-line motion on the dissipation mechanism is illustrated. We report an instability of the super-hemispherical base states with mobile contact lines (ii) that correlates with horizontal motion of the centre-of-mass, called the ‘walking’ instability. Davis dissipation from the dynamic contact-line condition (iii) can suppress the instability. The remainder of the spectrum exhibits oscillatory behaviour. For the hemispherical base state with mobile contact line (ii), the spectrum is degenerate with respect to the azimuthal wavenumber. We show that varying either the base-state volume or contact-line mobility lifts this degeneracy. For most values of these symmetry-breaking parameters, a certain spectral ordering of frequencies is maintained. However, because certain modes are more strongly influenced by the support than others, there are instances of additional modal degeneracies. We explain the physical reason for these and show how to locate them.

Direct numerical simulations of statistically steady homogeneous isotropic turbulence in viscoelastic fluids described by the FENE-P model, such as those laden with polymers, are presented. It is shown that the strong depletion of the turbulence dissipation reported by previous authors does not necessarily imply a depletion of the nonlinear energy cascade. However, for large relaxation times, of the order of the eddy turnover time, the polymers remove more energy from the large scales than they can dissipate and transfer the excess energy back into the turbulent dissipative scales. This is effectively a polymer-induced kinetic energy cascade which competes with the nonlinear energy cascade of the turbulence leading to its depletion. It is also shown that the total energy flux to the small scales from both cascade mechanisms remains approximately the same fraction of the kinetic energy over the turnover time as the nonlinear energy cascade flux in Newtonian turbulence.

The influence of harmonic spanwise wall motion on bypass transition in boundary layers is investigated using direct numerical simulations. It is shown that the appropriate choice of the forcing parameters can achieve a substantial stabilization of the laminar flow regime. However, an increase of the forcing amplitude or period beyond their optimal values diminishes the stabilizing effect, and leads to breakdown upstream of the unforced case. For the optimal wall-oscillation parameters, the reduction in propulsion power substantially outweighs the power requirement of the forcing. The mechanism of transition delay is examined in detail. Analysis of the pre-transitional streaks shows that the wall oscillation substantially reduces their average amplitude, and eliminates the most energetic streaks. As a result, the secondary instabilities that precede breakdown to turbulence are substantially weakened – an effect demonstrated by linear stability analyses of flow fields from direct numerical simulations. The outcome is transition delay owing to a significant reduction in the frequency of occurrence of turbulent spots and a downstream shift in their average inception location. Finally, it is shown that the efficiency of the forcing can be further improved by replacing the sinusoidal time dependence of the wall oscillation with a square wave.

We consider two-dimensional one-sided convection of a solute in a fluid-saturated porous medium, where the solute decays via a first-order reaction. Fully nonlinear convection is investigated using high-resolution numerical simulations and a low-order model that couples the dynamic boundary layer immediately beneath the distributed solute source to the slender vertical plumes that form beneath. A transient-growth analysis of the boundary layer is used to characterise its excitability. Three asymptotic regimes are investigated in the limit of high Rayleigh number $\mathit{Ra}$, in which the domain is considered deep, shallow or of intermediate depth, and for which the Damköhler number $\mathit{Da}$ is respectively large, small or of order unity. Scaling properties of the flow are identified numerically and rationalised via the analytic model. For fully established high-$\mathit{Ra}$ convection, analysis and simulation suggest that the time-averaged solute transfer rate scales with $\mathit{Ra}$ and the plume horizontal wavenumber with $\mathit{Ra}^{1/2}$, with coefficients modulated by $\mathit{Da}$ in each case. For large $\mathit{Da}$, the rapid reaction rate limits the plume depth and the boundary layer restricts the rate of solute transfer to the bulk, whereas for small $\mathit{Da}$ the average solute transfer rate is ultimately limited by the domain depth and the convection is correspondingly weaker.

This research implements a recently proposed framework for meander migration, in order to explore the coevolution of planform and channel width in a freely meandering river. In the model described here, width evolution is coupled to channel migration through two submodels, one describing bank erosion and the other describing bank deposition. Bank erosion is modelled as erosion of purely non-cohesive bank material damped by natural armouring due to basal slump blocks, and bank deposition is modelled in terms of a flow-dependent rate of vegetal encroachment. While these two submodels are specified independently, the two banks interact through the medium of the intervening channel; the morphodynamics of which is described by a fully nonlinear depth-averaged morphodynamics model. Since both banks are allowed to migrate independently, channel width is free to vary locally as a result of differential bank migration. Through a series of numerical runs, we demonstrate coevolution of local curvature, width and streamwise slope as the channel migrates over time. The correlation between the local curvature, width and bed elevation is characterized, and the nature of this relationship is explored by varying the governing parameters. The results show that, by varying a parameter representing the ratio between a reference bank erosion rate and a reference bank deposition rate, the model is able to reproduce the broad range of river width–curvature correlations observed in nature. This research represents a step towards providing general metrics for predicting width variation patterns in river systems.

The linear global instability and resulting transition to turbulence induced by an isolated cylindrical roughness element of height $h$ and diameter $d$ immersed within an incompressible boundary layer flow along a flat plate is investigated using the joint application of direct numerical simulations and fully three-dimensional global stability analyses. For the range of parameters investigated, base flow computations show that the roughness element induces a wake composed of a central low-speed region surrounded by a three-dimensional shear layer and a pair of low- and high-speed streaks on each of its sides. Results from the global stability analyses highlight the unstable nature of the central low-speed region and its crucial importance in the laminar–turbulent transition process. It is able to sustain two different global instabilities: a sinuous and a varicose one. Each of these globally unstable modes is related to a different physical mechanism. While the varicose mode has its root in the instability of the whole three-dimensional shear layer surrounding the central low-speed region, the sinuous instability turns out to be similar to the von Kármán instability in the two-dimensional cylinder wake and has its root in the lateral shear layers of the separated zone. The aspect ratio of the roughness element plays a key role on the selection of the dominant instability: whereas the flow over thin cylindrical roughness elements transitions due to a sinuous instability of the near-wake region, for larger roughness elements the varicose instability of the central low-speed region turns out to be the dominant one. Direct numerical simulations of the flow past an aspect ratio ${\it\eta}=1$ (with ${\it\eta}=d/h$) roughness element sustaining only the sinuous instability have revealed that the bifurcation occurring in this particular case is supercritical. Finally, comparison of the transition thresholds predicted by global linear stability analyses with the von Doenhoff–Braslow transition diagram provides qualitatively good agreement.

A rotating fluid flow between differentially translating parallel plates, which induce uniform suction and injection, is studied as a canonical model of swirling flow where suction, shear and Coriolis effects compete. This relatively simple modelling yields several reduced equations that are valid for asymptotically large suction, shear and/or rotation rates. The linear stability problems derived from the full Navier–Stokes and reduced problems are numerically solved and compared. In addition to Taylor-vortex modes, transverse-roll-type instabilities are found in Rayleigh-stable and -unstable parameter regions when weak suction is applied. These instabilities, separated by the so-called Rayleigh line, are characterised by vortices attached to the suction wall. Another type of instability, which exists beyond the Rayleigh line and shows inviscid motion in the fluid core, is found when suction is sufficiently strong. The relation of this instability to the stability results by Gallet, Doering & Spiegel (Phys. Fluids, vol. 22, 2010, 034105) is discussed. Our nonlinear analyses indicate subcritical and supercritical bifurcations of finite-amplitude solutions for the near-wall and fluid-core instabilities, respectively.

This paper describes a new computational model developed based on the filtered mass density function (FMDF) for large-eddy simulation (LES) of two-phase turbulent reacting flows. The model is implemented with a unique Lagrangian–Eulerian–Lagrangian computational methodology. In this methodology, the resolved carrier gas velocity field is obtained by solving the filtered form of the compressible Navier–Stokes equations with high-order finite difference (FD) schemes. The gas scalar (temperature and species mass fractions) field and the liquid (droplet) phase are both obtained by Lagrangian methods. The two-way interactions between the phases and all the Eulerian and Lagrangian fields are included in the new two-phase LES/FMDF methodology. The results generated by LES/FMDF are compared with direct numerical simulation (DNS) data for a spatially developing non-reacting and reacting evaporating mixing layer. Results for two more complex and practical flows (a dump combustor and a double-swirl burner) are also considered. For all flows, it is shown that the two-phase LES/FMDF results are consistent and accurate.

The evolution of an initially localized disturbance in polymeric channel flow is investigated, with the FENE-P model used to characterize the viscoelastic behaviour of the flow. In the linear growth regime, the flow response is stabilized by viscoelasticity, and the maximum attainable disturbance-energy amplification is reduced with increasing polymer concentration. The reduction in the energy growth rate is attributed to the polymer work, which plays a dual role. First, a spanwise polymer-work term develops, and is explained by the tilting action of the wall-normal vorticity on the mean streamwise conformation tensor. This resistive term weakens the spanwise velocity perturbation thus reducing the energy of the localized disturbance. The second action of the polymer is analogous, with a wall-normal polymer work term that weakens the vertical velocity perturbation. Its indirect effect on energy growth is substantial since it reduces the production of Reynolds shear stress and in turn of the streamwise velocity perturbation, or streaks. During the early stages of nonlinear growth, the dominant effect of the polymer is to suppress the large-scale streaky structures which are strongly amplified in Newtonian flows. As a result, the process of transition to turbulence is prolonged and, after transition, a drag-reduced turbulent state is attained.

In this study, we revisit the consequence of assuming equilibrium between the rates of production ($P$) and dissipation $({\it\epsilon})$ of the turbulent kinetic energy $(k)$ in the highly anisotropic and inhomogeneous near-wall region. Analytical and dimensional arguments are made to determine the relevant scales inherent in the turbulent viscosity (${\it\nu}_{t}$) formulation of the standard $k{-}{\it\epsilon}$ model, which is one of the most widely used turbulence closure schemes. This turbulent viscosity formulation is developed by assuming equilibrium and use of the turbulent kinetic energy $(k)$ to infer the relevant velocity scale. We show that such turbulent viscosity formulations are not suitable for modelling near-wall turbulence. Furthermore, we use the turbulent viscosity $({\it\nu}_{t})$ formulation suggested by Durbin (Theor. Comput. Fluid Dyn., vol. 3, 1991, pp. 1–13) to highlight the appropriate scales that correctly capture the characteristic scales and behaviour of $P/{\it\epsilon}$ in the near-wall region. We also show that the anisotropic Reynolds stress ($\overline{u^{\prime }v^{\prime }}$) is correlated with the wall-normal, isotropic Reynolds stress ($\overline{v^{\prime 2}}$) as $-\overline{u^{\prime }v^{\prime }}=c_{{\it\mu}}^{\prime }(ST_{L})(\overline{v^{\prime 2}})$, where $S$ is the mean shear rate, $T_{L}=k/{\it\epsilon}$ is the turbulence (decay) time scale and $c_{{\it\mu}}^{\prime }$ is a universal constant. ‘A priori’ tests are performed to assess the validity of the propositions using the direct numerical simulation (DNS) data of unstratified channel flow of Hoyas & Jiménez (Phys. Fluids, vol. 18, 2006, 011702). The comparisons with the data are excellent and confirm our findings.

We report on the structure and dynamics of gaseous detonation stabilized in a supersonic flow emanating radially from a central source. The steady-state solutions are computed and their range of existence is investigated. Two-dimensional simulations are carried out in order to explore the stability of the steady-state solutions. It is found that both collapsing and expanding two-dimensional cellular detonations exist. The latter can be stabilized by putting several rigid obstacles in the flow downstream of the steady-state sonic locus. The problem of initiation of standing detonation stabilized in the radial flow is also investigated numerically.

This experimental study examines the relationship between transverse jet structural characteristics and the shear layer instabilities forming on the upstream side of the jet column. Jets composed of mixtures of helium and nitrogen were introduced perpendicularly into a low-speed wind tunnel using several alternative injectors: convergent circular nozzles mounted either flush with or elevated above the tunnel floor, and a flush-mounted circular pipe. Both non-intrusive optical diagnostics (planar laser-induced fluorescence (PLIF) and particle image velocimetry (PIV)) and intrusive probe-based (hot-wire anemometry) measurements were used to explore a range of jet-to-crossflow momentum flux ratios and density ratios for which previous studies have identified upstream shear layer transition from convective to absolute instability. Remarkable correspondences were identified between formation of the well-known counter-rotating vortex pair (CVP) associated with the jet cross-section and conditions producing strong upstream shear layer vorticity rollup, arising typically from absolute instability in the shear layer. In contrast, asymmetries in the jet mean cross-sectional shape and/or lack of a clear CVP were observed to correspond to weaker, convectively unstable jet shear layers.

Local conservation laws are systematically constructed for three-dimensional time-dependent viscous and inviscid incompressible fluid flows, in primitive variables and vorticity formulation, using the direct construction method. Complete sets of local conservation laws in primitive variables are derived for the case of conservation law multipliers depending on derivatives up to the second order. In the vorticity formulation, there exists an infinite family of vorticity-dependent conservation laws involving an arbitrary differentiable function of space and time, holding for both viscous and inviscid cases. The infinite conservation law family is used to generate further independent hierarchies of conservation laws that essentially involve vorticity and arbitrary flow parameters, which are determined by known evolution equations such as those for momentum, energy or helicity, though not necessarily in the form of a conservation law. The new conservation laws are not restricted to any reduced flow geometry such as planar or axisymmetric limits. Examples are considered.

This study is concerned with the stability of a flow of viscous conducting liquid driven by a pressure gradient in the channel between two parallel walls subject to a transverse magnetic field. Although the magnetic field has a strong stabilizing effect, this flow, similarly to its hydrodynamic counterpart – plane Poiseuille flow – is known to become turbulent significantly below the threshold predicted by linear stability theory. We investigate the effect of the magnetic field on two-dimensional nonlinear travelling-wave states which are found at substantially subcritical Reynolds numbers starting from $\mathit{Re}_{n}=2939$ without the magnetic field and from $\mathit{Re}_{n}\sim 6.50\times 10^{3}\mathit{Ha}$ in a sufficiently strong magnetic field defined by the Hartmann number $\mathit{Ha}$. Although the latter value is a factor of seven lower than the linear stability threshold $\mathit{Re}_{l}\sim 4.83\times 10^{4}\mathit{Ha}$, it is still more than an order of magnitude higher than the experimentally observed value for the onset of turbulence in magnetohydrodynamic (MHD) channel flow.

A pinned or free-floating rigid plate lying on the free surface of a thin film of viscous fluid, which itself lies on top of a horizontal substrate that is moving to the right at a constant speed is considered. The focus of the present work is to describe how the competing effects of the speed of the substrate, surface tension, viscosity, and, in the case of a pinned plate, the prescribed pressure in the reservoir of fluid at its upstream end, determine the possible equilibrium positions of the plate, the free surface, and the flow within the film. The present problems are of interest both in their own right as paradigms for a range of fluid–structure interaction problems in which viscosity and surface tension both play an important role, and as a first step towards the study of elastic effects.

Many physical problems containing rotating symmetry exhibit azimuthal waves, from electromagnetic waves in nanophotonic crystals to seismic waves in giant stars. When this symmetry is broken, clockwise (CW) and counter-clockwise (CCW) waves are split into two distinct modes which can become unstable. This paper focuses on a theoretical study of symmetry breaking in annular cavities containing a number $N$ of flames prone to azimuthal thermo-acoustic instabilities. A general dispersion relation for non-perfectly-axisymmetric cavities is obtained and analytically solved to provide an explicit expression for the frequencies and growth rates of all azimuthal modes of the configuration. This analytical study unveils two parameters affecting the stability of the mode: (i) a coupling strength corresponding to the cumulative effects of the $N$ flames and (ii) a splitting strength due to the symmetry breaking when the flames are different. This theory has been validated using a 3D Helmholtz solver and good agreement is found. When only two types of flames are introduced into the annular cavity, the splitting strength is found to depend on two parameters: the difference between the two burner types and the pattern used to distribute the flames along the azimuthal direction. To first order, this theory suggests that the most stable configuration is obtained for a perfectly axisymmetric configuration. Therefore, breaking the symmetry by mixing different flames cannot improve the stability of an annular combustor independently of the flame distribution pattern.

Results of direct numerical simulations of the transitional processes that characterise the evolution of a breaking internal gravity wave to a fully developed and essentially steady turbulent patch are presented. The stationary lee wave was forced by the imposition of an appropriate bottom boundary shape within a density-stratified domain having a uniform upstream velocity and density gradient, and with the ratio of momentum to thermal (or other) diffusivity defined by $\mathit{Pr}=1$. An earlier paper considered the eventual, fully developed turbulent patch arising after the breaking process is complete (Yakovenko et al., J. Fluid Mech., vol. 677, 2011, pp. 103–133); the focus in this paper is on the instabilities in the breaking process itself. The flow is analysed using streamlines, density contours and temporal and spatial spectra, as well as second moments of the velocity and density fluctuations, for a Reynolds number of 4000 based on the height of the bottom topography and the upstream velocity. The computations (on a grid using in excess of $10^{9}$ mesh points) yielded sufficient resolution to capture the fine-scale transition processes as well as the subsequent fully developed turbulence discussed earlier. It is shown that the major instability is of Rayleigh–Taylor type (RTI) with a resulting mixing region depth growing in a manner consistent with more classical RTI studies, despite the much more complicated environment. The resolution was sufficient to capture secondary Kelvin–Helmholtz-type instabilities on the developing RTI structures. Overall evolution towards the fully turbulent state characterised by a significant region of $-\frac{5}{3}$ subrange in both velocity and density spectra is very rapid. It is much faster than the long time scale characterising the subsequent evolution of the turbulent patch; this latter time scale is sufficiently large that the turbulent patch can itself be viewed as essentially steady.

Direct numerical simulations of stably stratified Ekman layers are conducted to study the effect of increasing static stability on turbulence dynamics and modelling in wall-bounded flows at three moderate Reynolds numbers. The flow field is analysed by examining the mean profiles of wind speed, potential temperature and momentum flux, as well as streamwise velocity and temperature spectra. The maximum stabilizing buoyancy flux that a flow can sustain while remaining fully turbulent is found to depend on the Reynolds number. The flows with the highest Reynolds number display a relatively well-developed inertial range and logarithmic layer, and are found to bear similarities to much higher-Reynolds-number flows like the ones encountered in the atmospheric boundary layer. In particular, the near-wall mean profiles follow the Monin–Obukhov similarity theory. However, several flow features, such as the critical Richardson number and the stress–strain alignment, are found to maintain significant dependence on the Reynolds number. The budgets of turbulence kinetic energy (TKE), vertical velocity variance, momentum and buoyancy fluxes, and temperature variance are analysed. The results indicate that the effect of stability on turbulence is first directly manifested in the vertical velocity variance budget, and results in damping of vertical motions. This then leads to a reduction in the downward transport of horizontal momentum components towards the surface, and consequently to a decrease in the shear production term in the TKE budget: changes in the vertical profile of TKE shear production with increasing Richardson number are significant and have a larger impact on TKE than direct buoyancy destruction. The reduction in vertical velocity variance also results in significant drops in the production terms in the momentum flux, buoyancy flux and temperature variance budgets. Various assumptions and parameters related to low-order turbulence closures are investigated. The results suggest that the vertical velocity variance is a more appropriate parameter than the full TKE on which to base eddy-diffusivity and viscosity models.

Compressibility exerts a stabilizing influence on a variety of high-speed shear flows such as turbulent mixing layers, transitioning boundary layers and homogeneously sheared turbulence. An important stabilizing feature that is common amongst all shear flows is the velocity–pressure interaction dynamics. In this study, velocity–pressure interactions of individual perturbation or fluctuation modes are investigated using direct numerical simulations and linear analysis in high-Mach-number homogeneous shear flow. For a given perturbation wave mode, the action of pressure is shown to depend on two important factors: the orientation of the perturbation wavevector with respect to the shear plane and the Mach number. It is shown that the streamwise perturbation wave mode rapidly develops a high level of kinetic energy but is self-limiting owing to the action of pressure. On the other hand, the energy of spanwise perturbation wave modes grows unaffected by pressure or Mach number. Oblique modes combine spanwise and streamwise characteristics and are shown to be chiefly responsible for stabilizing effects seen in shear flows. Three regimes of obliqueness of different linear stability characteristics are identified. The critical role of perturbation obliqueness on stabilization is established.