Using the parallel-flow approximation the prechaotic bifurcation behaviour of Blasius boundary-layer flow is studied at finite Reynolds numbers. The objective of this investigation is to search for qualitative solution changes which might be linked to the rapid breakdown at transition. As a first step this requires the computation of the two-dimensional primary equilibrium surface of nonlinear Tollmien-Schlichting waves. This two-dimensional neutral surface exhibits a period-halving bifurcation which is qualitatively different from the situation for plane Poiseuille flow. At the same time the numerically computed equilibrium solution offers the possibility of assessing the range of convergence of weakly nonlinear results.

In a second step the stability of this nonlinear equilibrium solution is investigated with respect to three-dimensional disturbances. Of particular importance is the existence of a modal degeneracy between amplified secondary instability modes, implying locally algebraic growth. On decreasing the Reynolds number, the amplification rate of this direct resonance point switches from being amplified to being damped. Interestingly, the Reynolds number corresponding to this zero-amplification point seems to be in the vicinity of the experimentally observed transition Reynolds number for Blasius flow.

Cylindrical gravity currents, produced by a full-depth lock release, in a rotating system are investigated by means of three-dimensional high-resolution simulations of the incompressible variable-density Navier–Stokes equations with the Coriolis term and using the Boussinesq approximation for a small density difference. Here, the depth of the fluid is chosen to be the same as the radius of the cylindrical lock and the ambient fluid is non-stratified. Our attention is focused on the situation when the ratio of Coriolis to inertia forces is not large, namely $0.1\leqslant {\mathcal{C}}\leqslant 0.3$, and the non-rotating case, namely ${\mathcal{C}}=0$, is also briefly considered. The simulations reproduce the major features observed in the laboratory and provide more detailed flow information. After the heavy fluid contained in a cylindrical lock is released in a rotating system, the influence of the Coriolis effects is not significant during the initial one-tenth of a revolution of the system. During the initial one-tenth of a revolution of the system, Kelvin–Helmholtz vortices form and the rotating cylindrical gravity currents maintain nearly perfect axisymmetry. Afterwards, three-dimensionality of the flow quickly develops and the outer rim of the spreading heavy fluid breaks away from the body of the current, which gives rise to the maximum dissipation rate in the system during the entire adjustment process. The detached outer rim of heavy fluid then continues to propagate outward until a maximum radius of propagation is attained. The body of the current exhibits a complex contraction–relaxation motion and new outwardly propagating pulses form regularly in a period slightly less than half-revolution of the system. Depending on the ratio of Coriolis to inertia forces, such a contraction–relaxation motion may be initiated after or before the attainment of a maximum radius of propagation. In the contraction–relaxation motion of the heavy fluid, energy is transformed between potential energy and kinetic energy, while it is mainly the kinetic energy that is consumed by the dissipation. As a new pulse initially propagates outward, the potential energy in the system increases at the expense of decreasing kinetic energy, until a local maximum of potential energy is reached. During the latter part of the new pulse propagation, the kinetic energy in the system increases at the expense of decreasing potential energy, until a local minimum of potential energy is reached and another new pulse takes form. With the use of three-dimensional high-resolution simulations, the lobe-and-cleft structure at the advancing front can be clearly observed. The number of lobes is maintained only for a limited period of time before merger between existing lobes occurs when a maximum radius of propagation is approached. The high-resolution simulations complement the existing shallow-water formulation, which accurately predicts many important features and provides insights for rotating cylindrical gravity currents with good physical assumptions and simple mathematical models.

We investigate three-dimensional flagellar swimming in a fluid with a sparse network of stationary obstacles or fibres. The Brinkman equation is used to model the average fluid flow where a flow-dependent term, including a resistance parameter that is inversely proportional to the permeability, captures the effects of the fibres on the fluid. To solve for the local linear and angular velocities that are coupled to the flagellar motion, we extend the method of regularized Brinkmanlets to incorporate a Kirchhoff rod, discretized as point forces and torques along a centreline. Representing a flagellum as a Kirchhoff rod, we investigate emergent waveforms for different preferred strain and twist functions. Since the Kirchhoff rod formulation allows for out-of-plane motion, in addition to studying a preferred planar sine wave configuration, we also study a preferred helical configuration. Our numerical method is validated by comparing results to asymptotic swimming speeds derived for an infinite-length cylinder propagating planar or helical waves. Similar to the asymptotic analysis for both planar and helical bending, we observe that with small amplitude bending, swimming speed is always enhanced relative to the case with no fibres in the fluid (Stokes) as the resistance parameter is increased. For regimes not accounted for with asymptotic analysis, i.e. large amplitude planar and helical bending, our model results show a non-monotonic change in swimming speed with respect to the resistance parameter; a maximum swimming speed is observed when the resistance parameter is near one. The non-monotonic behaviour is due to the emergent waveforms; as the resistance parameter increases, the swimmer becomes incapable of achieving the amplitude of its preferred configuration. We also show how simulation results of slower swimming speeds for larger resistance parameters are actually consistent with the asymptotic swimming speeds if work in the system is fixed.

Extensive experimental studies are presented of the effects of prolonged streamline divergence on developing turbulent boundary layers. The experiment was arranged as source flow over a flat plate with a maximum divergence parameter of about 0.075. Mild, but alternating in sign, upstream-pressure-gradient effects on diverging boundary layers are also discussed.

It appears that two overlapping stages of development are involved. The initial stage covers a distance of about 20 initial boundary-layer thicknesses (δ0) from the start of divergence, where the coupled effects of pressure gradient and divergence are present. In this region there is a fairly large reduction in divergence parameter, Rθ (Reynolds number based on momentum thickness) remains constant (≈ 1400) and the boundary-layer properties change rapidly. In the second region, which lasts nearly to the end of the diverging section, the pressure-gradient effects are negligible, the rate of decrease in divergence parameter is very small and Rθ increases gradually. Up to the last measurement station (≈ 100δ0) the flow is still considered to be at a low Reynolds number (Rθ ≈ 2000). For almost the entire length of this region, the profiles of non-dimensional eddy viscosity appear to be self-similar, but have larger values than for the unperturbed flow. Also in this region, beyond 35δ0, the wake parameter, which has reduced significantly, becomes nearly constant and independent of Rθ. On the other hand the entrainment rate attains a constant value at around 50δ0. It appears that the boundary layer reaches a state of equilibrium. It is suggested that this is the result of an enhanced turbulent diffusion to the outer layer. Spectral measurements show that divergence affects mainly the low-wavenumber, large-scale motions. However, there is no change in large-eddy configurations, since the dimensionless structure parameters show only negligible deviations from the unperturbed values.

In this study, the theory of micropolar fluids is employed to study the stability problem of flow between two concentric rotating cylinders. The field equations subject to no-slip conditions (non-zero velocity and microrotation velocity components) at the wall surfaces are solved. The analytical solutions of the velocity and microrotation velocity fields as well as the shear stress difference, couple stress and strain rate for basic flow are obtained. The equations with respect to non-axisymmetric disturbances are derived and solved by a direct numerical procedure. It is found that non-zero wall-surface microrotation velocity makes the flow faster and more unstable. Moreover, it tends to reduce the limits of critical non-axisymmetric disturbances. The effect on the stability characteristics can be magnified by increasing the microstructure or couple-stress parameter or the microinertia parameter for the cases of corotating cylinders and a stationary outer cylinder or by decreasing the radius ratio or the microinertia parameter for the case of counterrotating cylinders.

Viscoelastic properties of complex fluids are usually extracted by applying an oscillatory shear rate ( where is a constant which is small for a linear response to make sense) to the fluid. This leads to a complex effective viscosity where its real part carries information on viscous effects while its imaginary part informs us on elastic properties. We show here theoretically, by taking a dilute vesicle suspension as an example, that application of a pure shearing oscillation misses several interesting microscopic features of the suspension. It is shown that if, in addition to the oscillatory part, a basic constant shear rate is applied to the suspension (so that the total shear rate is , with a constant), then the complex viscosity reveals much more insightful properties of the suspension. First, it is found that the complex viscosity exhibits a resonance for tank-treading vesicles as a function of the frequency of oscillation. This resonance is linked to the fact that vesicles, while being in the stable tank-treading regime (with their main axis having a steady orientation with respect to the flow direction), possess damped oscillatory modes. Second, in the region of parameter space where the vesicle exhibits either vacillating-breathing (permanent oscillations of the main axis about the flow direction and breathing of the shape) or tumbling modes, the complex viscosity shows an infinite number of resonances as a function of the frequency. It is shown that these behaviours markedly differ from that obtained when only the classical oscillation is applied. The results are obtained numerically by solution of the analytical constitutive equation of a dilute vesicle suspension and confirmed analytically by a linear-response phenomenological theory. It is argued that the same type of behaviour is expected for any suspension of soft entities (capsules, red blood cells, etc.) that exhibit periodic motion under constant shear flow. We shall also discuss the reason why this type of behaviour could not have been captured by existing constitutive laws of complex fluids.

The vertical heat transfer in Bénard–Marangoni convection of a fluid layer with infinite Prandtl number is studied by means of upper bounds on the Nusselt number $Nu$ as a function of the Marangoni number $Ma$. Using the background method for the temperature field, it has recently been proved by Hagstrom & Doering (Phys. Rev. E, vol. 81, 2010, art. 047301) that $Nu\leqslant 0.838Ma^{2/7}$. In this work we extend previous background method analysis to include balance parameters and derive a variational principle for the bound on $Nu$, expressed in terms of a scaled background field, that yields a better bound than Hagstrom & Doering’s formulation at a given $Ma$. Using a piecewise-linear, monotonically decreasing profile we then show that $Nu\leqslant 0.803Ma^{2/7}$, lowering the previous prefactor by 4.2 %. However, we also demonstrate that optimisation of the balance parameters does not affect the asymptotic scaling of the optimal bound achievable with Hagstrom & Doering’s original formulation. We subsequently utilise convex optimisation to optimise the bound on $Nu$ over all admissible background fields, as well as over two smaller families of profiles constrained by monotonicity and convexity. The results show that $Nu\leqslant O(Ma^{2/7}(\ln Ma)^{-1/2})$ when the background field has a non-monotonic boundary layer near the surface, while a power-law bound with exponent $2/7$ is optimal within the class of monotonic background fields. Further analysis of our upper-bounding principle reveals the role of non-monotonicity, and how it may be exploited in a rigorous mathematical argument.

This work reports an experimental investigation of the flow through a random array of fixed solid spheres. The volume fraction of the spheres is 2 %, and the Reynolds number $Re$ based on the sphere diameter and the average flow velocity is varied from 120 to 1040. Using time and spatial averaging, the fluctuations have been decomposed into two contributions of different natures: a spatial fluctuation that accounts for the strong inhomogeneity of the flow around each sphere, and a time fluctuation that comes from the instability of the flow at large enough Reynolds numbers. The evolutions of these two contributions with the Reynolds number are different, so that their relative importance varies. However, when each is normalized by using its own variance and the integral length scales of the fluctuations, their spectra and probability density functions (PDFs) are almost independent of $Re$ . The spatial fluctuation mostly comes from the velocity deficit in the wakes of the spheres, and is thus dominated by scales larger than one or two sphere diameters. It is found to be responsible for the asymmetry of the PDFs of the vertical fluctuations and of the major part of the anisotropy level between the vertical and the horizontal components of the fluctuations. The time fluctuation dominates at scales smaller than the integral length scale. It is isotropic and its PDFs, well described by an exponential distribution, are non-Gaussian. The spectra of the spatial and the time fluctuations both show an evolution as the power $-3$ of the wavenumber, but not exactly in the same subrange. All these properties are found in remarkable agreement with the results of both experimental investigations and large eddy simulations (LES) of a homogeneous bubble swarm. This confirms that the main mechanism responsible for the production of bubble-induced fluctuations is the interaction of the velocity disturbances caused by obstacles immersed in a flow and that the structure of this agitation is weakly dependent on the precise nature of the obstacles. The understanding and the modelling of the agitation generated by the motion of a dispersed phase, such as the bubble-induced agitation, therefore require one to distinguish between the roles of these two contributions.

For the first time, we used computer simulations to study lift forces on two static disks placed side-by-side within a two-dimensional granular flow and found them to be either repulsive or attractive depending on the flow velocity and separation between the disks. Our simulations results reveal that differences in the flow velocity between the disks and outside of that region are closely correlated with the lift force. We propose an empirical function for the lift force based on this correlation and our dimensional analysis. The specific region where the measured velocity exhibits this correlation suggests that attractive lift is not a Bernoulli-like effect. Instead, we speculate that it might be explained by a force balance based on Coulomb’s theory of passive failure in a Mohr–Coulomb material. Our results confirm that repulsive lift is due to the jamming of particles flowing between the disks.

We examine the dynamics of slender, rigid rods in direct numerical simulation of isotropic turbulence. The focus is on the statistics of three quantities and how they vary as rod length increases from the dissipation range to the inertial range. These quantities are (i) the steady-state rod alignment with respect to the perceived velocity gradients in the surrounding flow, (ii) the rate of rod reorientation (tumbling) and (iii) the rate at which the rod end points move apart (stretching). Under the approximations of slender-body theory, the rod inertia is neglected and rods are modelled as passive particles in the flow that do not affect the fluid velocity field. We find that the average rod alignment changes qualitatively as rod length increases from the dissipation range to the inertial range. While rods in the dissipation range align most strongly with fluid vorticity, rods in the inertial range align most strongly with the most extensional eigenvector of the perceived strain-rate tensor. For rods in the inertial range, we find that the variance of rod stretching and the variance of rod tumbling both scale as $l^{-4/3}$, where $l$ is the rod length. However, when rod dynamics are compared to two-point fluid velocity statistics (structure functions), we see non-monotonic behaviour in the variance of rod tumbling due to the influence of small-scale fluid motions. Additionally, we find that the skewness of rod stretching does not show scale invariance in the inertial range, in contrast to the skewness of longitudinal fluid velocity increments as predicted by Kolmogorov’s $4/5$ law. Finally, we examine the power-law scaling exponents of higher-order moments of rod tumbling and rod stretching for rods with lengths in the inertial range and find that they show anomalous scaling. We compare these scaling exponents to predictions from Kolmogorov’s refined similarity hypotheses.

High resolution direct numerical simulation (DNS) (512×1024×512) and large-eddy simulation (LES) of a shear-free mixing layer are presented. The geometry of the flow consists of two layers with different turbulence intensities that are in contact and interact through a fairly thin mixing layer. This geometry is used to explore the influence of inhomogeneities in the characteristic length scales, times scales and energy scales on the turbulence properties. Comparison of DNS results is made with the Veeravalli & Warhaft (J. Fluid Mech.207, 191–229, 1989) experiment. The LES is performed on a 32×64×32 grid using an eddy-viscosity model. The use of such a model appears to be justified by the very weak departures from isotropy that are observed in the shear-free mixing layer. The LES predictions are compared with the filtered DNS data and show that the eddy viscosity model performs very well in predicting the energy profile as well as the deviation from Gaussianity in the turbulent velocity field statistics.

The orientation of spheroidal particles dispersed in a fluid flow is known to influence the particles’ rotation mode. Rod-like and disk-like particles orient themselves differently and accordingly also rotate differently. In order to explore the role of the deterministic factors, i.e. mean shear and vorticity anisotropy, on the orientational behaviour of inertialess tracer spheroids, we adopted a purpose-made Couette–Poiseuille flow simulated numerically by Yang et al. (Intl J. Heat Fluid Flow, vol. 63, 2017, pp. 14–27). Typical wall turbulence with streamwise-oriented streaky structures caused by the locally high mean shear rate was observed only at one of the walls. The absence of mean shear at the other wall gave rise to an atypical turbulence field. Over a relatively wide and quasi-homogeneous core region, a modest mean shear rate made the vorticity field anisotropic. In spite of the mean shear, rod-like tracers were spinning and disk-like spheroids were tumbling, just as in homogeneous isotropic turbulence. We explained this phenomenon by the isotropic particle orientation and concluded that zero mean shear is not necessary for rod spinning and disk tumbling. The orientation and rotation of the Lagrangian tracer spheroids near the high shear wall were almost indistinguishable from the well-known behaviour in turbulent channel flows. Near the opposite wall, where the mean shear was negligibly small, disk-like particles aligned preferentially in the wall-normal direction and rotated similarly as in the presence of high shear. Rod-like particles, on the contrary, aligned more randomly and accordingly rotated similarly as in the core region. These observations revealed that the degree of particle orientation anisotropy has a major impact on the particle rotation mode, whereas mean shear, irrespective of the actual shear rate, only plays a secondary role in particle rotation. Deduction of the eigenvectors of the left Cauchy–Green tensor showed that the preferential orientation of the tracer spheroids was caused by the alignment of rods and disks with the strongest Lagrangian stretching and compression directions, respectively. Lagrangian stretching/compression determines the particle orientations and the particle orientation affects the particle rotation mode.

We consider the full three-dimensional dynamics of a thin falling liquid film on a flat plate inclined at some non-zero angle to the horizontal. In addition to gravitational effects, the flow is driven by an electric field which is normal to the substrate far from the flow. This extends the work of Tseluiko & Papageorgiou (J. Fluid Mech., vol. 556, 2006b, pp. 361–386) by including transverse dynamics. We study both the cases of overlying and hanging films, where the liquid lies above or below the substrate, respectively. Starting with the Navier–Stokes equations coupled with electrostatics, a fully nonlinear two-dimensional Benney equation for the interfacial dynamics is derived, valid for waves that are long compared to the film thickness. The weakly nonlinear evolution is governed by a Kuramoto–Sivashinsky equation with a non-local term due to the electric field effect. The electric field term is linearly destabilising and produces growth rates proportional to $|\unicode[STIX]{x1D743}|^{3}$, where $\unicode[STIX]{x1D743}$ is the wavenumber vector of the perturbations. It is found that transverse gravitational instabilities are always present for hanging films, and this leads to unboundedness of nonlinear solutions even in the absence of electric fields – this is due to the anisotropy of the nonlinearity. For overlying films and a restriction on the strength of the electric field, the equation is well-posed in the sense that it possesses bounded solutions. This two-dimensional equation is studied numerically for the case of periodic boundary conditions in order to assess the effects of inertia, electric field strength and the size of the periodic domain. Rich dynamical behaviours are observed and reported. For subcritical Reynolds number flows, a sufficiently strong electric field can promote non-trivial dynamics for some choices of domain size, leading to fully two-dimensional evolutions of the interface. We also observe two-dimensional spatiotemporal chaos on sufficiently large domains. For supercritical flows, such two-dimensional chaotic dynamics emerges in the absence of a field, and its presence enhances the amplitude of the fluctuations and broadens their spectrum.

Surface waves superimposed upon a larger-scale flow are blocked and reflected at the points where the group velocities balance the convection by the larger-scale flow. In this study, we first extended the theory of Shyu & Phillips (1990) to the situation when short deep-water gravity waves propagate obliquely upon a steady unidirectional irrotational current and are reflected by it. In this case, the uniformly valid solution and the WKBJ solution of the short waves were derived from the Laplace equation and the kinematical and dynamical boundary conditions. These solutions in terms of some parameters (the expressions for which have also been deduced in this case) take the same forms as those derived by Shyu & Phillips, which by referring to Smith's (1975) theory can even be proved to be valid for gravity waves in an intermediate-depth region and near a curved moving caustic induced by an unsteady multidirectional irrotational current. In this general case, the expressions for certain parameters in these solutions cannot be obtained so that their values must be estimated in a numerical calculation. The algorithm for estimates of some of these parameters that are responsible for the amplitude of the reflected wave not being equal to that of the incident wave in the vicinity of the caustic and therefore are crucial for the computer calculation of the ray solution to be continued after reflection, was illustrated through numerical tests. This algorithm can avoid the error magnification phenomenon that occurred in the previous estimates of the reflected wave in the vicinity of the caustic using the action conservation principle directly. The forms of the solutions have also been utilized to clarify the wave profiles near caustics in a general situation, which indicate that in storm conditions freak waves characterized by a steeper forward face preceded by a deep trough will probably occur in the caustic regions.