This paper describes an experimental investigation of steady-state resonant waves. Several co-propagating short-crested wave trains are generated in a basin at the State Key Laboratory of Ocean Engineering (SKLOE) in Shanghai, and the wavefields are measured and analysed both along and normal to the direction of propagation. These steady-state resonant waves are first calculated theoretically under the exact resonance criterion with sufficiently high nonlinearity, and then are generated in the basin by means of the main wave components that contain at least 95 % of the wave energy. The steady-state wave spectra are quantitatively observed within the inherent system error of the basin and identified by means of a contrasting experiment. Both symmetrical and anti-symmetrical steady-state resonant waves are observed and the experimental and theoretical results show excellent agreement. These results offer the first experimental evidence of the existence of steady-state resonant waves with multiple solutions.

A kinetic model of the Boltzmann equation for non-vibrating polyatomic gases is proposed, based on the Rykov model for diatomic gases. We adopt two velocity distribution functions (VDFs) to describe the system state; inelastic collisions are the same as in the Rykov model, but elastic collisions are modelled by the Boltzmann collision operator (BCO) for monatomic gases, so that the overall kinetic model equation reduces to the Boltzmann equation for monatomic gases in the limit of no translational–rotational energy exchange. The free parameters in the model are determined by comparing the transport coefficients, obtained by a Chapman–Enskog expansion, to values from experiment and kinetic theory. The kinetic model equations are solved numerically using the fast spectral method for elastic collision operators and the discrete velocity method for inelastic ones. The numerical results for normal shock waves and planar Fourier/Couette flows are in good agreement with both conventional direct simulation Monte Carlo (DSMC) results and experimental data. Poiseuille and thermal creep flows of polyatomic gases between two parallel plates are also investigated. Finally, we find that the spectra of both spontaneous and coherent Rayleigh–Brillouin scattering (RBS) compare well with DSMC results, and the computational speed of our model is approximately 300 times faster. Compared to the Rykov model, our model greatly improves prediction accuracy, and reveals the significant influence of molecular models. For coherent RBS, we find that the Rykov model could overpredict the bulk viscosity by a factor of two.

We develop a process-based model for the dispersion of a passive scalar in the turbulent flow around the buildings of a city centre. The street network model is based on dividing the airspace of the streets and intersections into boxes, within which the turbulence renders the air well mixed. Mean flow advection through the network of street and intersection boxes then mediates further lateral dispersion. At the same time turbulent mixing in the vertical detrains scalar from the streets and intersections into the turbulent boundary layer above the buildings. When the geometry is regular, the street network model has an analytical solution that describes the variation in concentration in a near-field downwind of a single source, where the majority of scalar lies below roof level. The power of the analytical solution is that it demonstrates how the concentration is determined by only three parameters. The plume direction parameter describes the branching of scalar at the street intersections and hence determines the direction of the plume centreline, which may be very different from the above-roof wind direction. The transmission parameter determines the distance travelled before the majority of scalar is detrained into the atmospheric boundary layer above roof level and conventional atmospheric turbulence takes over as the dominant mixing process. Finally, a normalised source strength multiplies this pattern of concentration. This analytical solution converges to a Gaussian plume after a large number of intersections have been traversed, providing theoretical justification for previous studies that have developed empirical fits to Gaussian plume models. The analytical solution is shown to compare well with very high-resolution simulations and with wind tunnel experiments, although re-entrainment of scalar previously detrained into the boundary layer above roofs, which is not accounted for in the analytical solution, is shown to become an important process further downwind from the source.

We study the motion of a bubble driven by buoyancy and thermocapillarity in a tube with a non-uniformly heated walls, containing a so-called ‘self-rewetting fluid’; the surface tension of the latter exhibits a parabolic dependence on temperature, with a well-defined minimum. In the Stokes flow limit, we derive the conditions under which a spherical bubble can come to rest in a self-rewetting fluid whose temperature varies linearly in the vertical direction, and demonstrate that this is possible for both positive and negative temperature gradients. This is in contrast to the case of simple fluids whose surface tension decreases linearly with temperature, for which bubble motion is arrested only for negative temperature gradients. In the case of self-rewetting fluids, we propose an analytical expression for the position of bubble arrestment as a function of other dimensionless numbers. We also perform direct numerical simulation of axisymmetric bubble motion in a fluid whose temperature increases linearly with vertical distance from the bottom of the tube; this is done for a range of Bond and Galileo numbers, as well as for various parameters that govern the functional dependence of surface tension on temperature. We demonstrate that bubble motion can be reversed and then arrested only in self-rewetting fluids, and not in linear fluids, for sufficiently small Bond numbers. We also demonstrate that considerable bubble elongation is possible under significant wall confinement, and for strongly self-rewetting fluids and large Bond numbers. The mechanisms underlying the phenomena observed are elucidated by considering how the surface tension dependence on temperature affects the thermocapillary stresses in the flow.

Direct numerical simulations (DNS) of turbulent thermal convection in a box-shaped domain with regular surface roughness at the heated bottom and cooled top surfaces are conducted for Prandtl number $\mathit{Pr}=0.786$ and Rayleigh numbers $\mathit{Ra}$ between $10^{6}$ and $10^{8}$. The surface roughness is introduced by four parallelepiped equidistantly distributed obstacles attached to the bottom plate, and four obstacles located symmetrically at the top plate. By varying $\mathit{Ra}$ and the height and width of the obstacles, we investigate the influence of the regular wall roughness on the turbulent heat transport, measured by the Nusselt number $\mathit{Nu}$. For fixed $\mathit{Ra}$, the change in the value of $\mathit{Nu}$ is determined not only by the covering area of the surface, i.e. the obstacle height, but also by the distance between the obstacles. The heat flux enhancement is found to be largest for wide cavities between the obstacles which can be ‘washed out’ by the flow. This is also manifested in an empirical relation, which is based on the DNS data. We further discuss theoretical limiting cases for very wide and very narrow obstacles and combine them into a simple model for the heat flux enhancement due to the wall roughness, without introducing any free parameters. This model predicts well the general trends and the order of magnitude of the heat flux enhancement obtained in the DNS. In the $\mathit{Nu}$ versus $\mathit{Ra}$ scaling, the obstacles work in two ways: for smaller $\mathit{Ra}$ an increase of the scaling exponent compared to the smooth case is found, which is connected to the heat flux entering the cavities from below. For larger $\mathit{Ra}$ the scaling exponent saturates to the one for smooth plates, which can be understood as a full washing-out of the cavities. The latter is also investigated by considering the strength of the mean secondary flow in the cavities and its relation to the wind (i.e. the large-scale circulation), that develops in the core part of the domain. Generally, an increase in the roughness height leads to stronger flows both in the cavities and in the bulk region, while an increase in the width of the obstacles strengthens only the large-scale circulation of the fluid and weakens the secondary flows. An increase of the Rayleigh number always leads to stronger flows, both in the cavities and in the bulk.

The excitation of instability modes in the wake generated behind a discrete roughness element in a boundary layer at Mach 6 is analysed through numerical simulations of the compressible Navier–Stokes equations. Recent experimental observations show that transition to turbulence in high-speed boundary layers during re-entry flight is dominated by wall roughness effects. Therefore, understanding the roughness-induced transition to turbulence in this flow regime is of primary importance. Our results show that a discrete roughness element with a height of about half the local boundary-layer thickness generates an unstable wake able to sustain the growth of a number of modes. The most unstable of these modes are a sinuous mode (mode SL) and two varicose modes (modes VL and VC). The varicose modes grow approximately 17 % faster than the most unstable Mack mode and their growth persists over a longer streamwise distance, thereby leading to a notable acceleration of the laminar–turbulent transition process. Two main mechanisms are identified for the excitation of wake modes: the first is based on the interaction between the external disturbances and the reverse flow regions induced by the roughness element and the second is due to the interaction between the boundary-layer modes (first modes and Mack modes) and the non-parallel roughness wake. An important finding of the present study is that, while being less unstable, mode SL is the preferred instability for the first of the above excitation mechanisms, which drives the wake modes excitation in the absence of boundary-layer modes. Modes VL and VC are excited through the second mechanism and, hence, become important when first modes and Mack modes come into interaction with the roughness wake. The new mode VC presents similarities with the Mack mode instability, including the tuning between its most unstable wavelength and the local boundary-layer thickness, and it is believed to play a fundamental role in the roughness-induced transition of high-speed boundary layers. In contrast to the smooth-wall case, wall cooling is stabilising for all the roughness-wake modes.

We consider a thin liquid film flowing down an inclined plate in the presence of a counter-current turbulent gas. By making appropriate assumptions, Tseluiko & Kalliadasis (J. Fluid Mech., vol. 673, 2011, pp. 19–59) developed low-dimensional non-local models for the liquid problem, namely a long-wave (LW) model and a weighted integral-boundary-layer (WIBL) model, which incorporate the effect of the turbulent gas. By utilising these models, along with the Orr–Sommerfeld problem formulated using the full governing equations for the liquid phase and associated boundary conditions, we explore the linear stability of the gas–liquid system. In addition, we devise a generalised methodology to investigate absolute and convective instabilities in the non-local equations describing the gas–liquid flow. We observe that at low gas flow rates, the system is convectively unstable with the localised disturbances being convected downwards. As the gas flow rate is increased, the instability becomes absolute and localised disturbances spread across the whole domain. As the gas flow rate is further increased, the system again becomes convectively unstable with the localised disturbances propagating upwards. We find that the upper limit of the absolute instability region is close to the ‘flooding’ point associated with the appearance of large-amplitude standing waves, as obtained in Tseluiko & Kalliadasis (J. Fluid Mech., vol. 673, 2011, pp. 19–59), and our analysis can therefore be used to predict the onset of flooding. We also find that an increase in the angle of inclination of the channel requires an increased gas flow rate for the onset of absolute instability. We generally find good agreement between the results obtained using the full equations and the reduced models. Moreover, we find that the WIBL model generally provides better agreement with the results for the full equations than the LW model. Such an analysis is important for an understanding of the ranges of validity of the reduced model equations. In addition, a comparison of our theoretical predictions with the experiments of Zapke & Kröger (Intl J. Multiphase Flow, vol. 26, 2000, pp. 1439–1455) shows a fairly good agreement. We supplement our stability analysis with time-dependent computations of the linearised WIBL model. To provide some insight into the mechanisms of instability, we perform an energy budget analysis.

Interaction between surface and interfacial waves with continuous energy spectra in a two-layer density stratified fluid system is investigated numerically. For an initial wave field which consists only of the surface waves all propagating in the same direction, it is confirmed that the spectra $S_{s}(k)$ of the surface waves and $S_{i}(k)$ of the interfacial waves change significantly due to the recently found class 3 triad resonance. When the bulk of the surface wave spectrum $S_{s}(k)$ is initially located well above the critical wavenumber $k_{crit}$, below which the class 3 triad resonance is prohibited, $S_{s}(k)$ downshifts gradually toward the lower wavenumber during the initial stage of evolution. However, this downshift halts when the peak of $S_{s}(k)$ reaches around $k_{crit}$, and after that a steep peak forms in $S_{s}(k)$ around $k_{crit}$. It is confirmed that the timescale of the spectral evolution is of $O(1/{\it\epsilon}^{2})$ (${\it\epsilon}$ is a characteristic non-dimensional wave amplitude) in most of the $k$ space, consistent with the prediction of the wave turbulence theory for a system with a decay-type dispersion relation. However, it is also found that the timescale of the formation and growth of the sharp peak in $S_{s}(k)$ around $k_{crit}$ is of $O(1/{\it\epsilon})$, i.e. the timescale of the deterministic three-wave resonance.

We propose an analytical model for the determination of the microstructure and stresses in a sheared suspension that consists of a dense monolayer of identical spheres in a viscous fluid. We calculate the anisotropy in the orientational distribution of spheres, associated with a short-range repulsive force assumed to act between the spheres, and a particle pressure and normal stress difference that result from this anisotropy. The microstructure and stresses are similar to those measured in Stokesian dynamics simulations.

We investigate the forces and unsteady flow structures associated with harmonic oscillations of an airfoil in the streamwise (surging) and transverse (plunging) directions in two-dimensional simulations at low Reynolds number. For the surging case, we show that there are specific frequencies where the wake instability synchronizes with the unsteady motion of the airfoil, leading to significant changes in the mean forces. Resonant behaviour of the time-averaged forces is observed near the vortex shedding frequency and its subharmonic; the behaviour is reminiscent of the dynamics of the generic nonlinear oscillator known as the Arnol’d tongue or the resonance horn. Below the wake instability frequency, there are two regimes where the fluctuating forces are amplified and attenuated, respectively. A detailed study of the flow structures associated with leading-edge vortex (LEV) growth and detachment are used to relate this behaviour with the LEV acting either in phase with the quasi-steady component of the forces for the amplification case, or out of phase for the attenuation case. Comparisons with wind tunnel measurements show that phenomenologically similar dynamics occur at higher Reynolds number. Finally, we show that qualitatively similar phenomena occur during both surging and plunging.

The present work focuses on the collective effect on both bubble dynamics and mass transfer in a dense homogeneous bubble swarm for gas volume fractions ${\it\alpha}$ up to 30 %. The experimental investigation is carried out with air bubbles rising in a square column filled with water. Bubble size and shape are determined by means of a high-speed camera equipped with a telecentric lens. Gas volume fraction and bubble velocity are measured by using a dual-tip optical probe. The combination of these two techniques allows us to determine the interfacial area between the gas and the liquid. The transfer of oxygen from the bubbles to the water is measured from the time evolution of the concentration of oxygen dissolved in water, which is obtained by means of the gassing-out method. Concerning the bubble dynamics, the average vertical velocity is observed to decrease with ${\it\alpha}$ in agreement with previous experimental and numerical investigations, while the bubble agitation turns out to be weakly dependent on ${\it\alpha}$. Concerning mass transfer, the Sherwood number is found to be very close to that of a single bubble rising at the same Reynolds number, provided the latter is based on the average vertical bubble velocity, which accounts for the effect of the gas volume fraction on the bubble rise velocity. This conclusion is valid for situations where the diffusion coefficient of the gas in the liquid is very low (high Péclet number) and the dissolved gas is well mixed at the scale of the bubble. It is understood by considering that the transfer occurs at the front part of the bubbles through a diffusion layer which is very thin compared with all flow length scales and where the flow remains similar to that of a single rising bubble.

T1 topological rearrangement, i.e. switching of neighbouring bubbles in a liquid foam, is the elementary process of foam dynamics, and it involves film disappearance and generation. It has been studied extensively as it is crucial in foam rheology or foam collapse. T1 dynamics depends mainly on the surfactants used to generate the foam, and several models taking into account surface viscosity and/or elasticity have been proposed. By performing experiments in a cubic assembly of films, we go a step forward in this global analysis and investigate experimentally the mechanism of formation of the new film. In particular, the flow velocity field is probed by particle tracking and the film thickness is measured by light absorption and interferometric measurements. Two limit behaviours for the film are reported: it may (i) undergo an homogeneous extension, or (ii) resist elongation and remain at rest, new film being created from liquid exchange with connecting meniscus. Both T1 dynamics and film thickness are shown to depend on the competition between these two behaviours. Interestingly, their balance is set by the surfactant solution used, but it is also shown to vary during a single T1 relaxation process.

This paper presents the results of numerical stability analysis of the wake of an elliptical cylinder. Aspect ratios where the ellipse is longer in the streamwise direction than in the transverse direction are considered. The focus is on the dependence on the aspect ratio of the ellipse of the various bifurcations to three-dimensional flow from the two-dimensional Kármán vortex street. It is shown that the three modes present in the wake of a circular cylinder (modes A, B and QP) are present in the ellipse wake, and that in general they are all stabilized by increasing the aspect ratio of the ellipse. Two new pertinent modes are found: one long-wavelength mode with similarities to mode A, and a second that is only unstable for aspect ratios greater than approximately 1.75, which has similar spatiotemporal symmetries to mode B but has a distinct spatial structure. Results from fully three-dimensional simulations are also presented confirming the existence and growth of these two new modes in the saturated wakes.

The onset of circulating waves, i.e. waves with a circulating eddy in the main wave hump, and the onset of flow reversal, i.e. a vortex in the first capillary minimum, in inclined falling films is investigated as a function of the Reynolds number and inclination number using the weighted integral boundary layer (WIBL) model and direct numerical simulations (DNS). Analytical criteria for the onset of circulating waves and flow reversal based on the wave celerity, the average film thickness and the maximum and minimum film thickness have been approximated using self-similar parabolic velocity profiles. This approximation has been validated by second-order WIBL and DNS simulations. It is shown that the onset of circulating waves in the phase diagram for homoclinic solutions (waves of infinite wavelength) is strongly dependent on the inclination, but independent of the streamwise viscous dissipation effect. On the contrary, the onset of flow reversal shows a clear dependence on the viscous dissipation. Furthermore, simulation results for limit cycles (finite wavelength) reveal a strong increase of the corresponding critical Reynolds number with the excitation frequency. Additionally, a critical ratio between the maximum and substrate film thickness (value of approximately 2.5) was found for the onset of circulating waves, which is independent of wavelength, inclination, viscous dissipation and Reynolds number.

This study considers the natural convection flow in a water body subjected to heating by solar radiation. The investigation into this type of natural convection flow has been motivated by the fact that it is known to play a crucial role in the daytime heat and mass transfer in shallow regions of natural water reservoirs and lakes, with a resultant impact on biological activity. An analytical solution for temperature in such an internally heated system shows that the temperature stratification consists of an upper stable stratification and a lower unstable stratification. One of the important consequences of such a nonlinear temperature stratification is the limitation of the mixing driven by rising thermal plumes with the penetration length scale of the plumes determining the lower mixed layer thickness. A theoretical analysis conducted in the present study suggests that in relatively deep waters, the lower mixed layer thickness is equal to the attenuation length of the radiation, which has important implications for water quality, including the transport of pollutants and nutrients in the water body. Scalings are also obtained for the quasi-steady boundary layer. The theoretical analysis is validated against numerical simulations.

In the presence of electric fields, pairs of liquid drops can be rapidly drawn together such that, at contact, the deformed interface resembles a double-cone. Following contact, these drop pairs are observed to either coalesce or recoil. Experimental and theoretical results suggest that the transition between coalescence and recoil is due to the conical drop topology rather than charge effects. However, even with this assumption, existing models disagree on how the transition develops, leading to different predictions of the critical cone angle and bridge morphology. Here we use high-resolution numerical simulations to highlight the impact of the initial double-cone angle on drop coalescence and reconcile the differences in the previous models. The results demonstrate a self-similar behaviour at intermediate scales for both coalescence and recoil that is independent of the other length scales in the problem. We calculate a critical polar angle of ${\it\theta}_{c}=1.14$ rad ($65.3^{\circ }$), or a complementary angle of ${\it\beta}=90^{\circ }-{\it\theta}_{c}=25^{\circ }$. This calculated critical angle for morphological transition is in agreement with previous experimental observations of ${\it\beta}\approx 27\pm 2^{\circ }$.

We present a model for the rheological behaviour of non-dilute suspensions of initially spherical viscoelastic particles in viscous fluids under uniform Stokes flow conditions. The particles are assumed to be neutrally buoyant Kelvin–Voigt solids undergoing time-dependent finite deformations and exhibiting generalized neo-Hookean behaviour in their purely elastic limit. We investigate the effects of the shape dynamics and constitutive properties of the viscoelastic particles on the macroscopic rheological behaviour of the suspensions. The proposed model makes use of known homogenization estimates for composite material systems consisting of random distributions of aligned ellipsoidal particles with prescribed two-point correlation functions to generate corresponding estimates for the instantaneous (incremental) response of the suspensions, together with appropriate evolution laws for the relevant microstructural variables. To illustrate the essential features of the model, we consider two special cases: (i) extensional flow and (ii) simple shear flow. For each case, we provide the time-dependent response and, when available, the steady-state solution for the average particle shape and orientation, as well as for the effective viscosity and normal stress differences in the suspensions. The results exhibit shear thickening for extensional flows and shear thinning for simple shear flows, and it is found that the volume fraction and constitutive properties of the particles significantly influence the rheology of the suspensions under both types of flows. In particular, for extensional flows, suspensions of particles with finite extensibility constraints are always found to reach a steady state, while this is only the case at sufficiently low strain rates for suspensions of (less realistic) neo-Hookean particles, as originally reported by Roscoe (J. Fluid Mech., vol. 28, 1967, pp. 273–293) and Gao et al. (J. Fluid Mech., vol. 687, 2011, pp. 209–237). For shear flows, viscoelastic particles with high viscosities can experience a damped oscillatory motion of decreasing amplitude before reaching the steady state.

Nonlinear effects on the receptivity of cross-flow in the swept Hiemenz boundary layer are investigated. Numerical simulations are generated using a vorticity form of the Navier–Stokes equations. Steady perturbations are established using surface suction and blowing distributed along the spanwise direction as either a periodic strip or a band of small holes. The method of excitation, the size and the location of the prescribed forcing are shown to have a significant influence on the receptivity of the boundary layer. Blowing holes are found to excite perturbations with considerably larger magnitudes than those generated using a periodic suction and blowing strip. A semi-logarithmic relationship is derived that relates the initial amplitude of the linear-only disturbances with the location at which the absolute magnitude of the chordwise primary Fourier harmonic attains a stationary point or a size of approximately one-tenth of the free-stream spanwise velocity. Furthermore, the size of the physical chordwise velocity perturbation about this position can be estimated directly from the linear-only solutions. This would suggest that, for sufficiently small initial amplitudes, the onset of some nonlinear flow development properties can be predicted directly from a linear receptivity analysis.

This work is a computational study of a rarefied non-reacting hypersonic flow past a forward-facing step at zero-degree angle of attack in thermal non-equilibrium. Effects on the flow field structure and on the aerodynamic surface quantities due to changes in step frontal-face height are investigated by employing the direct simulation Monte Carlo method. The work focuses the attention of designers of hypersonic configurations on the fundamental parameter of surface discontinuity, which can have an important impact on even initial design. The results presented highlight the sensitivity of the primary flow field properties, velocity, density, pressure and temperature, to changes in the step frontal-face height. In addition, the behaviour of heat transfer, pressure and skin friction coefficients with variation of the step frontal-face height is detailed. The analysis shows that hypersonic flow past a forward-facing step in the transition flow regime is characterized by a strong compression ahead of the frontal face, which influences the aerodynamic surface properties upstream and adjacent to the frontal face. The analysis also shows that the extension of the upstream disturbance depends on the step frontal-face height. It was found that the recirculation region ahead of the step is also a function of the frontal-face height. A sequence of Moffatt eddies of decreasing size and intensity is observed in the concave step corner. Locally high heating and pressure loads were observed at three locations along the surface, i.e. on the lower surface, on the frontal face and on the upper surface. The results showed that both loads rely on the frontal-face height. The peak values for the heat transfer coefficient on the frontal-face surface were at least one order of magnitude larger than the maximum value observed for a smooth surface, i.e. a flat plate without a step. A comparison of the present simulation results with numerical and experimental data showed close agreement concerning the wall pressure acting on the step surface.

We study the physics of unsteady turbulent jets using direct numerical simulation (DNS) by introducing an instantaneous step change (both up and down) in the source momentum flux. Our focus is on the propagation speed and rate of spread of the resulting front. We show that accurate prediction of the propagation speed requires information about the energy flux in addition to the momentum flux in the jet. Our observations suggest that the evolution of a front in a jet is a self-similar process that accords with the classical dispersive scaling $z\sim \sqrt{t}$. In the analysis of the problem we demonstrate that the use of a momentum–energy framework of the kind used by Priestley & Ball (Q. J. R. Meteorol. Soc., vol. 81, 1955, pp. 144–157) has several advantages over the classical mass–momentum formulation. In this regard we generalise the approach of Kaminski et al. (J. Fluid Mech., vol. 526, 2005, pp. 361–376) to unsteady problems, neglecting only viscous effects and relatively small boundary terms in the governing equations. Our results show that dispersion originating from the radial dependence of longitudinal velocity plays a fundamental role in longitudinal transport. Indeed, one is able to find dispersion in the steady state, although it has received little attention because its effects can then be absorbed into the entrainment coefficient. Specifically, we identify two types of dispersion. Type I dispersion exists in a steady state and determines the rate at which energy is transported relative to the rate at which momentum is transported. In unsteady jets type I dispersion is responsible for the separation of characteristic curves and thus the hyperbolic, rather than parabolic, nature of the governing equations, in the absence of longitudinal mixing. Type II dispersion is equivalent to Taylor dispersion and results in the longitudinal mixing of the front. This mixing is achieved by a deformation of the self-similar profiles that one finds in steady jets. Using a comparison with the local eddy viscosity, and by examining dimensionless fluxes in the vicinity of the front, we show that type II dispersion provides a dominant source of longitudinal mixing.