A numerical description of heterogeneous propellant combustion enables us to examine the spatial and temporal fluctuations in the flow field arising from the heterogeneity. Particular focus is placed on the fluctuations in a zone intermediate between the combustion field (where reaction is important) and the chamber flow domain, for these define boundary conditions for simulations of the turbulent chamber flow. The statistics of the temperature field and the normal velocity field are described, and characteristic length scales and time scales are identified. The length scales are small compared to any relevant length scale of the chamber flow, and so the boundary conditions for this flow at any mesh point are statistically independent of those at any other mesh point. But the temporal correlations at a fixed point are significant, and affect the nature of the chamber flow in a variety of ways. We describe the fluctuations in the head-end pressure that arise because of them, and contrast these results with those calculated using a white-noise assumption.

We consider the flapping stability and response of a thin two-dimensional flag of high extensional rigidity and low bending rigidity. The three relevant non-dimensional parameters governing the problem are the structure-to-fluid mass ratio, μ = ρsh/(ρfL); the Reynolds number, Rey = VL/ν; and the non-dimensional bending rigidity, KB = EI/(ρfV2L3). The soft cloth of a flag is represented by very low bending rigidity and the subsequent dominance of flow-induced tension as the main structural restoring force. We first perform linear analysis to help understand the relevant mechanisms of the problem and guide the computational investigation. To study the nonlinear stability and response, we develop a fluid–structure direct simulation (FSDS) capability, coupling a direct numerical simulation of the Navier–Stokes equations to a solver for thin-membrane dynamics of arbitrarily large motion. With the flow grid fitted to the structural boundary, external forcing to the structure is calculated from the boundary fluid dynamics. Using a systematic series of FSDS runs, we pursue a detailed analysis of the response as a function of mass ratio for the case of very low bending rigidity (KB = 10−4) and relatively high Reynolds number (Rey = 103). We discover three distinct regimes of response as a function of mass ratio μ: (I) a small μ regime of fixed-point stability; (II) an intermediate μ regime of period-one limit-cycle flapping with amplitude increasing with increasing μ; and (III) a large μ regime of chaotic flapping. Parametric stability dependencies predicted by the linear analysis are confirmed by the nonlinear FSDS, and hysteresis in stability is explained with a nonlinear softening spring model. The chaotic flapping response shows up as a breaking of the limit cycle by inclusion of the 3/2 superharmonic. This occurs as the increased flapping amplitude yields a flapping Strouhal number (St = 2Af/V) in the neighbourhood of the natural vortex wake Strouhal number, St ≃ 0.2. The limit-cycle von Kármán vortex wake transitions in chaos to a wake with clusters of higher intensity vortices. For the largest mass ratios, strong vortex pairs are distributed away from the wake centreline during intermittent violent snapping events, characterized by rapid changes in tension and dynamic buckling.

Axisymmetric liquid-metal pipe flow passes through a quadrupole magnetic field that is generated by a pair of ‘oppositely sensed’ d.c. current coils. As a result of this arrangement, the flow experiences a degree of braking, mostly in the vicinity of the magnetic neutral point, owing to the effect of Lorentz forces acting upon the liquid-metal. Usefully, the system represents a practical and novel electromagnetic (e.m.) valve capable of regulating the flow of molten metal emanating from a tun-dish, for example. Linear theory predicts the development of a counter-intuitive unidirectional ‘slug-like’ profile throughout the liquid-metal pipe flow at large values of the Hartmann number, M, in the presence of an idealized axisymmetric neutral point that extends to infinity. We confirm that this behaviour is also apparent, but over a narrow region spanning the neutral point, in the case of a more realistic liquid-metal pipe flow acted upon by a pair of oppositely sensed d.c. current coils. The axial pressure gradient along the wall of this flow manifests a sharp peak at large M centred on the neutral point that is generated by the steep gradients in the slug profile there. In fact, the pressure drop developed across this region is approximately equal to the net braking effect of the e.m. valve.

This paper presents some simple analytical and numerical models which describe the dynamics of gas flowing from a multilayered low-permeability porous rock into a fracture. The models account for the vertical flow between relatively high- and low-permeability layers. The motion of gas in a permeable rock is governed by a nonlinear diffusion equation for the gas pressure. We analyse the gas flow described by this equation in both bounded and unbounded domains. In both cases simple scalings laws are developed to determine the fluxes and the dimensions of the regions within the rock which may be depleted over a given time scale. These are compared with the results of a full numerical model.

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

The interception of two spherical particles with arbitrary size in an infinite linear ambient Stokes flow is considered. The particle surfaces allow for slip according to the Navier–Maxwell–Basset law relating the shear stress to the tangential velocity. At any instant, the flow is computed in a frame of reference with origin at the centre of one particle using a cylindrical polar coordinate system whose axis of revolution passes through the centre of the second particle. Taking advantage of the axial symmetry of the boundaries of the flow in the particle coordinates, the problem is formulated as a system of integral equations for the zeroth, first, and second Fourier coefficients of the boundary traction with respect to the meridional angle. The force and torque exerted on each particle are determined by the zeroth and first Fourier coefficients, while the stresslet is determined by the zeroth, first, and second Fourier coefficients. The derived integral equations are solved with high accuracy using a boundary element method featuring adaptive element distribution and automatic time step adjustment according to the inter-particle gap. The results strongly suggest the existence of a critical value for the slip coefficient below which the surfaces of two particle collide after a finite interception time. The critical value depends on the relative initial particle positions. The particle stress tensor and coefficients of the linear and quadratic terms in the expansion of the effective viscosity of a dilute suspension in terms of the concentration in simple shear flow are discussed and evaluated. Surface slip significantly reduces the values of both coefficients and the longitudinal particle self-diffusivity.

When a gas bubble rises in a surfactant solution, the velocity field and the distribution of surfactant affect each other. This paper gives the theory for small Reynolds and internal Péclet numbers if the surfactant is gaseous or volatile, if its mass flux across the bubble and around its surface dominates its mass flux through the bulk liquid, and if slowness of both adsorption and convective diffusion must be allowed for.

The theory is tested on the experiments of Kelsall et al. (J. Chem. Soc. Faraday Trans., vol. 92, 1996, p. 3879). Their bubbles rose as expected in a pure liquid until the apparatus was opened to the atmosphere. That significantly slowed the bubbles down. The effect is so sensitive to small concentrations of slowly adsorbing or reacting surfactants that atmospheric carbon dioxide could have caused it, even though it alters the equilibrium surface tension by less than four parts per million in pure air.

There are still unexplained discrepancies between experiment and theory. Additional experiments are suggested that would help to explain them.

The stability properties of the flow past an infinitely long circular cylinder are studied in the context of linear theory. An immersed-boundary technique is used to represent the cylinder surface on a Cartesian mesh. The characteristics of both direct and adjoint perturbation modes are studied and the regions of the flow more sensitive to momentum forcing and mass injection are identified. The analysis shows that the maximum of the perturbation envelope amplitude is reached far downstream of the separation bubble, where as the highest receptivity is attained in the near wake of the cylinder, close to the body surface. The large difference between the spatial structure of the two-dimensional direct and adjoint modes suggests that the instability mechanism cannot be identified from the study of either eigenfunctions separately. For this reason a structural stability analysis of the problem is used to analyse the process which gives rise to the self-sustained mode. In particular, the region of maximum coupling among the velocity components is localized by inspecting the spatial distribution of the product between the direct and adjoint modes. Results show that the instability mechanism is located in two lobes placed symmetrically across the separation bubble, confirming the qualitative results obtained through a locally plane-wave analysis. The relevance of this novel technique to the development of effective control strategies for vortex shedding behind bluff bodies is illustrated by comparing the theoretical predictions based on the structural perturbation analysis with the experimental data of Strykowski & Sreenivasan (J. Fluid Mech. vol. 218, 1990, p. 71).

The flapping coupling between two filaments is studied theoretically and experimentally in this paper. A temporal linear instability analysis is carried out based on a simplified hydrodynamic model. The dispersion relationship between the eigen-frequency ω and wavenumber k is expressed by a quartic equation. Two special cases of flapping coupling, i.e. two identical filaments having the same length and two filaments having different lengths, are studied in detail. In the case of two identical filaments, the theoretical analysis predicts four coupling modes, i.e. the stretched-straight mode, the antisymmetrical in-phase mode, the symmetrical out-of-phase mode and the indefinite mode. The theory also predicts the existence of an eigenfrequency jump during transition between the in-phase and out-of-phase modes, which has been observed in previous experiments and numerical simulations. In the case of two filaments having different lengths, four modes similar to those in the former case are identified theoretically. The distribution of coupling modes for both the cases is shown in two planes. One is a dimensionless plane of S vs. U, where S is the density ratio of solid filament to fluid and U2 is the ratio of fluid kinetic energy to solid elastic potential energy. The other is a dimensional plane of the half-distance (h) between two filaments vs. the filament length (L). Relevant experiments are carried out in a soap-film tunnel and the stable and unstable modes are observed. Theory and experiment are compared in detail. It should be noted that the model used in our analysis is a very simplified one that can provide intuitional analytical results of the coupling modes as well as their qualitative distributions. The factors neglected in our model, such as vortex shedding, viscous and nonlinear effects, do not allow the model to predict results precisely consistent with the experiments. Moreover, the Strouhal numbers of the flapping filaments are found to be generally around a fixed value in the experiments for both cases, implying that the filaments try to maintain a lower potential energy state.

Turbulent Taylor–Couette flow with arbitrary rotation frequencies ω1, ω2 of the two coaxial cylinders with radii r1 < r2 is analysed theoretically. The current Jω of the angular velocity ω(x,t) = uϕ(r,ϕ,z,t)/r across the cylinder gap and and the excess energy dissipation rate ϵw due to the turbulent, convective fluctuations (the ‘wind’) are derived and their dependence on the control parameters analysed. The very close correspondence of Taylor–Couette flow with thermal Rayleigh–Bénard convection is elaborated, using these basic quantities and the exact relations among them to calculate the torque as a function of the rotation frequencies and the radius ratio η = r1/r2 or the gap width d = r2 − r1 between the cylinders. A quantity σ corresponding to the Prandtl number in Rayleigh–Bénard flow can be introduced, . In Taylor–Couette flow it characterizes the geometry, instead of material properties of the liquid as in Rayleigh–Bénard flow. The analogue of the Rayleigh number is the Taylor number, defined as Ta ∝ (ω1 − ω2)2 times a specific geometrical factor. The experimental data show no pure power law, but the exponent α of the torque versus the rotation frequency ω1 depends on the driving frequency ω1. An explanation for the physical origin of the ω1-dependence of the measured local power-law exponents α(ω1) is put forward. Also, the dependence of the torque on the gap width η is discussed and, in particular its strong increase for η → 1.

We present a simple flow model and solution to describe ‘horizontal convection’ driven by a gradient of temperature or heat flux along one horizontal boundary of a rectangular box. Following laboratory observations of the steady-state convection, the model is based on a localized vertical turbulent plume from a line or point source that is located anywhere within the area of the box and that maintains a stably stratified interior. In contrast to the ‘filling box’ process, the convective circulation involves vertical diffusion in the interior and a stabilizing buoyancy flux distributed over the horizontal boundary. The stabilizing flux forces the density distribution to reach a steady state. The model predictions compare well with previous laboratory data and numerical solutions. In the case of a point source for the plume (the case which best mimics the localized sinking in the large-scale ocean overturning) the thermal boundary layer is much thicker than that given by the two-dimensional boundary layer scaling of H. T. Rossby (Tellus, vol. 50, 1965, p. 242).

An energy criterion for conditional stability is proposed, based on the definition of generalized energies, obtained through a perturbation of the classical L2 (kinetic) energy. This perturbation is such that the contribution of the linear term in the perturbation equation to the generalized energy time derivative is negative definite. A critical amplitude threshold is then obtained by imposing the monotonic decay of the generalized energy. The capabilities of the procedure are appraised through the application to three different low-dimensional models. The effects of different choices in the construction of the generalized energy on the prediction of the critical amplitude threshold in the subcritical regime are also discussed.

The flow associated with a synthetic jet transitioning to turbulence in an otherwise quiescent external flow is examined using time-accurate three-dimensional numerical simulations. The incompressible Navier–Stokes solver uses a second-order accurate scheme for spatial discretization and a second-order semi-implicit fractional step method for time integration. The simulations are designed to model the experiments of C. S. Yao et al. (Proc. NASA LaRC Workshop, 2004) which have examined, in detail, the external evolution of a transitional synthetic jet in quiescent flow. Although the jet Reynolds and Stokes numbers in the simulations match with the experiment, a number of simplifications have been made in the synthetic jet actuator model adopted in the current simulations. These include a simpler representation of the cavity and slot geometry and diaphragm placement. Despite this, a reasonably good match with the experiments is obtained in the core of the jet and this indicates that for these jets, matching of these key non-dimensional parameters is sufficient to capture the critical features of the external jet flow. The computed results are analysed further to gain insight into the dynamics of the external as well as internal flow. The results indicate that near the jet exit plane, the flow field is dominated by the formation of counter-rotating spanwise vortex pairs that break down owing to the rapid growth of spanwise instabilities and transition to turbulence a short distance from the slot. Detailed analyses of the unsteady characteristics of the flow inside the jet cavity and slot provide insights that to date have not been available from experiments.

We develop a complete set of equations governing the evolution of a sharp interface separating two fluid phases undergoing transformation. In addition to the conventional balances for mass, linear momentum and energy these equations include also a counterpart of the Gibbs–Thomson equation familiar from theories for crystal growth. This additional equation arises from a consideration of configurational forces within a thermodynamical framework. Although the notion of configurational forces is well-developed and understood for the description of materials, such as crystalline solids, that possess natural reference configurations, little has been done regarding their role in materials, such as viscous fluids, that do not possess preferred reference states. We therefore provide a comprehensive discussion of configurational forces, the balance of configurational momentum, and configurational thermodynamics that does not require a choice of reference configuration. The general evolution equations arising from our theory account for the thermodynamic structure of the bulk phases and the interface and for viscous and thermal dissipation in the bulk phases and for viscous dissipation on the interface. Because of the complexity of these equations, we provide a reduced system of equations based on simplified constitutive assumptions and approximations common in the literature on phase transformations. Using these reduced equations, we apply the theory to the radially symmetric problem for the condensation of a liquid drop into the vapour phase.

The water-shipping problem is modelled in a two-dimensional framework and studied experimentally and numerically for the case of a fixed barge-shaped structure. The analysis represents the second step of the research discussed in Greco et al. (J. Fluid Mech., vol. 525, 2005, p. 309). The numerical investigation is performed by using both a boundary element method and a domain-decomposition strategy. The model tests highlight the occurrence of dam-breaking-type water on deck, (a) with and (b) without an initial plunging phase, and (c) an unusual type of water shipping connected with blunt water–deck impacts here called a hammer-fist type event never documented before. Cases (a) and (c) are connected with the most severe events and the related features and green-water loads are discussed in detail. A parametric analysis of water-on-deck phenomena has also been carried out in terms of the local incoming waves and bow flow features. We classify such phenomena in a systematic way to provide a basis for further investigations of water-on-deck events. The severity of (a)-type water-on-deck events is analysed in terms of initial cavity area and water-front velocity along the deck. The former increases as the square power of the modified incoming-wave (front-crest) steepness while the latter scales with its square-root. The two-dimensional investigation gives useful quantitative information in terms of water-front velocity for comparison with three-dimensional water-on-deck experiments on fixed bow models interacting with wave packets.

A local linear stability analysis is performed for a round inviscid jet with constant density that is injected into a uniform crossflow of the same density. The baseflow is obtained from a modified version of the inviscid transverse jet near-field solution of Coelho & Hunt (J. Fluid Mech. vol. 200, 1989, p. 95) which is valid for small values of the crossflow-to-jet velocity ratio λ. A Fourier expansion in the azimuthal direction is used to couple the disturbances with the three-dimensional crossflow. The spatial growth rates of the modes corresponding to the axisymmetric and first helical modes of the free jet as λ → 0 increase as λ increases. The diagonal dominance of the dispersion relation matrix is used as a quantitative criterion to estimate the range of velocity ratios (0 < λ < λc) within which the transverse jet instability can be considered to have a structure similar to that of the free jet. Further, we show that for λ>0 positive and negative helical modes have different growth rates, suggesting an inherent weak asymmetry in the transverse jet.

The paper addresses a set of new equations concerning the scale-by-scale balance of turbulent fluctuations in dilute polymer solutions. The main difficulty is the energy associated with the polymers, which is not of a quadratic form in terms of the traditional descriptor of the micro-structure. A different choice is however possible, which, at least for mild stretching of the polymeric chains, directly leads to an L2 structure for the total free-energy density of the system thus allowing the extension of the classical method to polymeric fluids. On this basis, the energy budget in spectral space is discussed, providing the spectral decomposition of the energy of the system. New equations are also derived in physical space, to provide balance equations for the fluctuations in both the kinetic field and the micro-structure, thus extending, in a sense, the celebrated Kármán–Howarth and Kolmogorov equations of classical turbulence theory. The paper is limited to the context of homogeneous turbulence. However the necessary steps required to expand the treatment to wall-bounded flows of polymeric liquids are indicated in detail.

An experimental investigation into the influence of Brownian motion on shear-induced particle migration of monodisperse suspensions of micrometre-sized colloidal particles is presented. The suspension is pumped through a 50 μm × 500 μm rectangular cross-section glass channel. The experiments are characterized chiefly by the sample volume fraction (φ = 0.1 − 0.4), and the flow rate expressed as the Péclet number (Pe = 10 − 400). For each experiment we measure the entrance length, which is the distance from the inlet of the channel required for the concentration profile to develop to its non-uniform steady state. The entrance length increases strongly with increasing Pe for Pe ≪ 100, in marked contrast to non-Brownian flows for which the entrance length is flow-rate independent. For larger Pe, the entrance length reaches a constant value which depends on the other experimental parameters. Additionally, the entrance length decreases with increasing φ; this effect is strongest for low φ. Modelling of the migration based on spatial variation of the normal stresses due to the particles captures the primary features observed in the axial evolution over a range of Pe and φ.

We investigate experimentally the force generated by the unsteady vortex formation of low-aspect-ratio normal flat plates with one end free. The objective of this study is to determine the role of the free end, or tip, vortex. Understanding this simple case provides insight into flapping-wing propulsion, which involves the unsteady motion of low-aspect-ratio appendages. As a simple model of a propulsive half-stroke, we consider a rectangular normal flat plate undergoing a translating start-up motion in a towing tank. Digital particle image velocimetry is used to measure multiple perpendicular sections of the flow velocity and vorticity, in order to correlate vortex circulation with the measured plate force. The three-dimensional wake structure is captured using flow visualization. We show that the tip vortex produces a significant maximum in the plate force. Suppressing its formation results in a force minimum. Comparing plates of aspect ratio six and two, the flow is similar in terms of absolute distance from the tip, but evolves faster for aspect ratio two. The plate drag coefficient increases with decreasing aspect ratio.

We consider the slow deformation of a relatively inviscid conducting drop surrounded by a viscous insulating fluid subject to a uniform electric field. The general behaviour is to deform and elongate in the direction of the field. Detailed numerical computations, based on a boundary integral formulation, are presented. For fields below a critical value, we obtain the evolution of the drop to an equilibrium shape; above the critical value, we calculate the drop evolution up to breakup. At breakup it appears that smaller droplets are emitted from the ends of the drop with a charge greater than the Rayleigh limit. As the electric field strength is increased the ejected droplet size decreases. A further increase in field strength results in the mode of breakup changing to a thin jet-like structure being ejected from the end. The shape of all drops is very close to spheroidal up to aspect ratios of about 5. Also, for fields just above the critical value there is a period of slow deformation which increases in duration as the critical field strength is approached from above. Slender-body theory is also used to model the drop behaviour. A similarity solution for the slender drop is obtained and a finite-time singularity is observed. In addition, the general solution for the slender-body equations is presented and the solution behaviour is examined. The slender-body results agree only qualitatively with the full numerical computations. Finally, a spheroidal model is briefly presented and compared with the other models.