The rheology of the steady uniform shear flow of smooth inelastic spheres is analysed by choosing the anisotropic/triaxial Gaussian as the single-particle distribution function. An exact solution of the balance equation for the second-moment tensor of velocity fluctuations, truncated at the ‘Burnett order’ (second order in the shear rate), is derived, leading to analytical expressions for the first and second ($\unicode[STIX]{x1D615}_{1}$ and $\unicode[STIX]{x1D615}_{2}$) normal stress differences and other transport coefficients as functions of density (i.e. the volume fraction of particles), restitution coefficient and other control parameters. Moreover, the perturbation solution at fourth order in the shear rate is obtained which helped to assess the range of validity of Burnett-order constitutive relations. Theoretical expressions for both $\unicode[STIX]{x1D615}_{1}$ and $\unicode[STIX]{x1D615}_{2}$ and those for pressure and shear viscosity agree well with particle simulation data for the uniform shear flow of inelastic hard spheres for a large range of volume fractions spanning from the dilute regime to close to the freezing-point density (${\it\nu}\sim 0.5$). While the first normal stress difference $\unicode[STIX]{x1D615}_{1}$ is found to be positive in the dilute limit and decreases monotonically to zero in the dense limit, the second normal stress difference $\unicode[STIX]{x1D615}_{2}$ is negative and positive in the dilute and dense limits, respectively, and undergoes a sign change at a finite density due to the sign change of its kinetic component. It is shown that the origin of $\unicode[STIX]{x1D615}_{1}$ is tied to the non-coaxiality (${\it\phi}\neq 0$) between the eigendirections of the second-moment tensor $\unicode[STIX]{x1D648}$ and those of the shear tensor $\unicode[STIX]{x1D63F}$. In contrast, the origin of $\unicode[STIX]{x1D615}_{2}$ in the dilute limit is tied to the ‘excess’ temperature ($T_{z}^{ex}=T-T_{z}$, where $T_{z}$ and $T$ are the $z$-component and the average of the granular temperature, respectively) along the mean vorticity ($z$) direction, whereas its origin in the dense limit is tied to the imposed shear field.

A linear theory, based on wide-spacing and high-frequency approximations, is developed to describe resonant behaviour in two-dimensional water-wave problems involving a freely floating half-immersed cylinder in the presence of a vertical rigid wall. The theory is not able to describe the lowest-frequency resonance, but otherwise yields explicit approximations for the locations of resonances in the complex plane and for their corresponding residues. Two problems are investigated in detail: the time-domain motion following a vertical displacement of the cylinder from equilibrium, and the time-harmonic motion of a cylinder that is excited by an incident plane wave.

Time-resolved tomographic particle image velocimetry experiments show that new hairpin vortices are generated within a fully developed and unperturbed turbulent boundary layer. The measurements are taken at a Reynolds number based on the momentum thickness of 2038, and cover the near-wall region below $y^{+}=140$, where $y^{+}$ is the wall-normal distance in wall units. Instantaneous visualizations of the flow reveal near-wall low-speed streaks with associated quasi-streamwise vortices, retrograde inverted arch vortices, hairpin vortices and hairpin packets. The hairpin heads are observed as close to the wall as $y^{+}=30$. Examples of hairpin packet evolution reveal the development of new hairpin vortices, which are created upstream and close to the wall in a manner consistent with the auto-generation model (Zhou et al., J. Fluid Mech., vol. 387, 1999, pp. 353–396). The development of the new hairpin appears to be initiated by an approaching sweep event, which perturbs the shear layer associated with the initial packet. The shear layer rolls up, thereby forming the new hairpin head. The head subsequently connects to existing streamwise vortices and develops into a hairpin. The time scale associated with the hairpin auto-generation is 20–30 wall units of time. This demonstrates that hairpins can be created over short distances within a developed turbulent boundary layer, implying that they are not simply remnants of the laminar-to-turbulent transition process far upstream.

The effects of wing–body interaction (WBI) on aerodynamic performance and vortex dynamics have been numerically investigated in the forward flight of cicadas. Flapping wing kinematics was reconstructed based on the output of a high-speed camera system. Following the reconstruction of cicada flight, three models, wing–body (WB), body-only (BD) and wings-only (WN), were then developed and evaluated using an immersed-boundary-method-based incompressible Navier–Stokes equations solver. Results have shown that due to WBIs, the WB model had a 18.7 % increase in total lift production compared with the lift generated in both the BD and WN models, and about 65 % of this enhancement was attributed to the body. This resulted from a dramatic improvement of body lift production from 2 % to 11.6 % of the total lift produced by the wing–body system. Further analysis of the associated near-field and far-field vortex structures has shown that this lift enhancement was attributed to the formation of two distinct vortices shed from the thorax and the posterior of the insect, respectively, and their interactions with the flapping wings. Simulations are also used to examine the new lift enhancement mechanism over a range of minimum wing–body distances, reduced frequencies and body inclination angles. This work provides a new physical insight into the understanding of the body-involved lift-enhancement mechanism in insect forward flight.

The potential for anticipated destruction of a counter-rotating vortex pair using the linear optimal perturbation of the Crow instability is assessed. Direct numerical simulation is used to study the development of the Crow instability and the subsequent evolution of the flow up to 30 characteristic times at a circulation-based Reynolds number of 1000. The conventional development of the instability leads to multiple contortions of the vortices including the linear growth of sinusoidal deformation, vortex linking and the formation of vortex rings. A new evolution stage is identified, succeeding this well-established sequence: the vortex rings undergo periodic oscillation. Two complete periods are simulated during which the vortical system is hardly altered, thereby demonstrating the extraordinary resilience of the vortices. The possibility of preventing these dynamics using the linear optimal perturbation of the Crow instability, the adjoint mode, is analysed. By appropriately setting the forcing amplitude, the lifetime of the vortices until their loss of coherence is reduced to approximately 13 characteristic times, which is less than half that of the natural Crow behaviour observed with infinitesimal forcing. The dynamics of the flow induced by the linear optimal perturbation that enable this result are connected to processes already known to efficiently alter vortical flows, in particular transient growth and four-vortex dynamics.

The head-on collision, merging and internal mixing dynamics of two unequal-sized droplets were experimentally studied and interpreted, using water, $n$-decane and $n$-tetradecane to identify the distinguishing effects of surface tension and liquid viscosity on the merging and mixing patterns. It is shown that, upon merging of water and $n$-decane droplets, mushroom-like jets of dissimilar characteristics develop within the merged mass for small and large values of the impact Weber number (We), and that such jets are not developed for intermediate values of We. Furthermore, such jet-like mixing patterns were not observed for droplets of $n$-tetradecane, which has smaller surface tension and larger viscosity as compared to water. A regime nomogram relating the Ohnesorge and symmetric Weber numbers is constructed, providing a unified interpretation of the internal mixing patterns. Numerical simulations based on an improved volume-of-fluid method and an adaptive mesh refinement algorithm provide auxiliary diagnoses of the flow fields and the observed phenomena.

Measurements of the thermal creep flow through a single rectangular microchannel connected to two tanks maintained initially at the same pressure, but at different temperatures, are carried out for five noble gas species, over a large range of pressure and for two temperature differences between the tanks. The time-dependent pressure variations in both cold and hot tanks are investigated, and the temperature-driven (thermal creep) mass flow rate between two tanks is calculated from these data for the rarefaction parameter ranging from the transitional to slip flow regime. The measured mass flow rate is compared with the numerical solution of the S-model kinetic equation, and they show good agreement. A novel approximate expression to calculate the temperature-driven mass flow rate in the transitional and slip flow regimes is proposed. This expression provides results in good agreement with the measured values of the mass flow rate. In the slip flow regime, the thermal slip coefficient is calculated by employing the previously reported methodology, and the influence of the nature of the gas on this coefficient is investigated. The measured values of the thermal slip coefficient agree well with the values available in the literature, indicating that this coefficient is independent of the shape of a channel.

It has been recently shown that the energy-containing motions (i.e. coherent structures) in turbulent channel flow exist in the form of Townsend’s attached eddies by a numerical experiment which simulates the energy-containing motions only at a prescribed spanwise length scale using their self-sustaining nature (Hwang, J. Fluid Mech., vol. 767, 2015, pp. 254–289). In the present study, a detailed investigation of the self-sustaining process of the energy-containing motions at each spanwise length scale (i.e. the attached eddies) in the logarithmic and outer regions is carried out with an emphasis on its relevance to ‘bursting’, which refers to an energetic temporal oscillation of the motions (Flores & Jiménez, Phys. Fluids, vol. 22, 2010, 071704). It is shown that the attached eddies in the logarithmic and outer regions, composed of streaks and quasi-streamwise vortical structures, bear the self-sustaining process remarkably similar to that in the near-wall region: i.e. the streaks are significantly amplified by the quasi-streamwise vortices via the lift-up effect; the amplified streaks subsequently undergo a ‘rapid streamwise meandering motion’, reminiscent of streak instability or transient growth, which eventually results in breakdown of the streaks and regeneration of new quasi-streamwise vortices. For the attached eddies at a given spanwise length scale ${\it\lambda}_{z}$ between ${\it\lambda}_{z}^{+}\simeq 100$ and ${\it\lambda}_{z}\simeq 1.5h$, the single turnover time period of the self-sustaining process is found to be $Tu_{{\it\tau}}/{\it\lambda}_{z}\simeq 2$ ($u_{{\it\tau}}$ is the friction velocity), which corresponds well to the time scale of the bursting. Two additional numerical experiments, designed to artificially suppress the lift-up effect and the streak meandering motions, respectively, reveal that these processes are essential ingredients of the self-sustaining process of the attached eddies in the logarithmic and outer regions, consistent with several previous theoretical studies. It is also shown that the artificial suppression of the lift-up effect of the attached eddies in the logarithmic and outer regions leads to substantial amounts of turbulent skin-friction reduction.

We study the influence of hydrodynamic, thermodynamic and interparticle forces on the diffusive motion of a Brownian probe driven by a constant external force through a dilute colloidal dispersion. The influence of these microscopic forces on equilibrium self-diffusivity (passive microrheology) is well known: all three act to hinder the short- and long-time self-diffusion. Here, via pair-Smoluchowski theory, we explore their influence on self-diffusion in a flowing suspension, where particles and fluid have been set into motion by an externally forced probe (active microrheology), giving rise to non-equilibrium flow-induced diffusion. The probe’s motion entrains background particles as it travels through the bath, deforming the equilibrium suspension microstructure. The shape and extent of microstructural distortion is set by the relative strength of the external force $F^{\mathit{ext}}$ to the entropic restoring force $kT/a_{\mathit{th}}$ of the bath particles, defining a Péclet number $\mathit{Pe}\equiv F^{\mathit{ext}}/(2kT/a_{\mathit{th}})$; and also by the strength of hydrodynamic interactions, set by the range of interparticle repulsion ${\it\kappa}=(a_{\mathit{th}}-a)/a$, where $kT$ is the thermal energy and $a_{\mathit{th}}$ and $a$ are the thermodynamic and hydrodynamic sizes of the particles, respectively. We find that in the presence of flow, the same forces that hinder equilibrium diffusion now enhance it, with diffusive anisotropy set by the range of interparticle repulsion ${\it\kappa}$. A transition from hindered to enhanced diffusion occurs when diffusive and advective forces balance, $\mathit{Pe}\sim 1$, where the exact value is a sensitive function of the strength of hydrodynamics, ${\it\kappa}$. We find that the hindered to enhanced transition straddles two transport regimes: in hindered diffusion, stochastic forces in the presence of other bath particles produce deterministic displacements (Brownian drift) at the expense of a maximal random walk. In enhanced diffusion, driving the probe with a deterministic force through an initially random suspension leads to fluctuations in the duration of probe–bath particle entrainment, giving rise to enhanced, flow-induced diffusion. The force-induced diffusion is anisotropic for all $\mathit{Pe}$, scaling as $D\sim \mathit{Pe}^{2}$ in all directions for weak forcing, regardless of the strength of hydrodynamic interactions. When probe forcing is strong, $D\sim \mathit{Pe}$ in all directions in the absence of hydrodynamic interactions, but the picture changes qualitatively as hydrodynamic interactions grow strong. In this nonlinear regime, microstructural asymmetry weakens while the anisotropy of the force-induced diffusion tensor increases dramatically. This behaviour owes its origins to the approach toward Stokes flow reversibility, where diffusion along the direction of probe force scales advectively while transverse diffusion must vanish.

We study the response of a steady free-falling liquid curtain perturbed by focused air jets blowing perpendicularly against it. Asymmetric and symmetric perturbations are applied by using either a single pulsed jet or two identical steady jets facing each other. The response strongly depends on the curtain flow rate, and the location and strength of the perturbation. For pulsed asymmetric perturbations of increasing amplitude, sinuous wave, drop ejection, bubble ejection and hole opening are successively observed. For steady symmetric perturbations, a steady hole forms downstream in the wake. For this latter case, we present a model for the curtain thickness and the location of the hole in the wake in terms of the curtain flow rate and the size, flow rate and location of the jets. The adjustable-parameter-free expression we obtain compares favourably to the experiments provided that the perturbation is sufficiently small (jet stagnation pressure smaller than curtain stagnation pressure) and the liquid viscosity is negligible.

This study is motivated by controlling transient growth and subsequent bypass transition of the laminar boundary layer to turbulence. In experiments employing a model problem, an active roughness element is used to introduce steady/unsteady streak disturbances in a Blasius boundary layer. This tractable arrangement enables a systematic investigation of the evolution of the disturbances and of potential methods to control them in real time. The control strategy utilizes wall-shear-stress sensors, upstream and downstream of a plasma actuator, as inputs to a model-based controller. The controller is designed using empirical input/output data to determine the parameters of simple models, approximating the boundary layer dynamics. The models are used to tune feedforward and feedback controllers. The control effect is examined over a range of roughness-element heights, free stream velocities, feedback sensor positions, unsteady disturbance frequencies and control strategies; and is found to nearly completely cancel the steady-state disturbance at the downstream sensor location. The control of unsteady disturbances exhibits a limited bandwidth of less than 1.3 Hz. However, concurrent modelling demonstrates that substantially higher bandwidth is achievable by improving the feedforward controller and/or optimizing the feedback sensor location. Moreover, the model analysis shows that the difference in the convective time delay of the roughness- and actuator-induced disturbances over the control domain must be known with high accuracy for effective feedforward control. This poses a limitation for control effectiveness in a stochastic environment, such as in bypass transition beneath a turbulent free stream; nonetheless, feedback can remedy some of this limitation.

The stability of base gravitational convection in a layer of ferrofluid confined between two vertical wide and tall non-magnetic plates, heated from one side, cooled from the other and placed in a uniform oblique external magnetic field is studied. Two distinct mechanisms, thermo-gravitational and thermo-magnetic, are found to be responsible for the appearance of various stationary and wave-like instability modes. The characteristics of all instability modes are investigated as functions of the orientation angles of the applied magnetic field and its magnitude for various values of magnetic parameters when both the thermo-magnetic and gravitational buoyancy mechanisms are active. The original three-dimensional problem is cast in an equivalent two-dimensional form using generalised Squire’s transformations, which significantly reduces a computational cost. Subsequently, full three-dimensional instability patterns are recovered using the inverse Squire’s transformation, and the optimal field and pattern orientations are determined.

We consider displacement flows in slightly diverging or converging plane channels. The two fluids are miscible and buoyancy is significant. We assume that the channel is oriented close to horizontal. Employing a classical lubrication approximation, we simplify the governing equations to furnish a semi-analytical solution for the flux functions. Then, we demonstrate how the non-uniformity of the displacement flow geometry can affect the propagation of the interface between the heavy and light fluids in time, for various parameters studied, e.g. the viscosity ratio, a buoyancy number and rheological features. By setting the molecular diffusion effects to zero, certain solution behaviours at longer times can be practically predicted through the associated hyperbolic problem, using which it becomes possible to directly compute the interfacial features of interest, e.g. leading and trailing front heights and speeds. For a Newtonian displacement flow in a converging or uniform channel, as the buoyancy number increases from zero, we are able to classify three flow regimes based on the behaviour of the trailing front near the top of the channel: a no-back-flow regime, a stationary interface flow regime, and a sustained back-flow regime. For the case of a diverging channel flow, the sustained back-flow regime is replaced by an eventually stationary interface flow regime. In addition, as the displacement flow progresses, the leading front speed typically increases (decreases) in a converging (diverging) channel, while the opposite is usually true for the front height. For the no-back-flow regime (i.e. with small buoyancy), the solution of the displacement flow at long times in all the geometries considered converges to a similarity form, while no similarity form is found for the other flow regimes. As the displacement flow develops, frontal diffusive effects are reduced (enhanced) in a converging (diverging) channel and multiple fronts are progressively less (more) present in a converging (diverging) channel. Regarding non-Newtonian effects, a shear-thinning fluid displacing a Newtonian fluid exhibits an increasingly fast front that has a short height in a converging channel. When a yield stress is present in the displaced fluid, it is possible to find residual wall layers of displaced fluid that are completely static. These layers disappear at a certain critical downstream distance in a converging channel while they appear at a critical distance in a diverging channel. Finally, the combination of strong buoyant and yield-stress effects can modify the destiny of a second front that follows the leading front.

If classical quasigeostrophic (QG) flow breaks down at smaller scales, it gives rise to questions of whether higher-order nonlinear balance can be maintained, to what scale and for how long. These are naturally followed by asking how this is affected by stratification and rotation. To address these questions, we perform non-hydrostatic Boussinesq simulations where the initial data is balanced using the Baer–Tribbia nonlinear normal mode initialization scheme (NNMI), which is accurate to second order in the Rossby number, as the next-order improvement to first-order QG theory. The NNMI procedure yields an ageostrophic contribution to the energy spectrum that has a very steep slope. However, as time passes, a shallow range emerges in the ageostrophic spectrum when the Rossby number is large enough for a given Reynolds number. It is argued that this shallow range is the unbalanced part of the motion that develops spontaneously in time and eventually dominates the energy at small scales. If the initial flow is not nonlinearly balanced, the shallow range emerges at even lower Rossby number and it appears at larger scales. Through numerous simulations at different rotation and stratification, this study gives a clear picture of how energy is cascaded in different initially balanced regimes of rotating stratified flow. We find that at low Rossby number the flow mainly consists of a geostrophic part and a balanced ageostrophic part with a steep spectrum. As the Rossby number increases, the unbalanced part of the ageostrophic energy increases at a rate faster than the balanced part. Hence, the total energy spectrum displays a shallow range above a transition wavenumber. This wavenumber evolves to smaller values as rotation weakens.

Some celestial bodies such as planets, moons and comets (here referred to as moons for simplicity) emit jets of material at speeds that in some instances are large enough to escape gravity. Previous investigations have shown this problem to be highly complex, e.g. involving multi-phase flows, phase changes, radiation and gas rarefaction effects. In order to learn from exploring a manageable parameter space, and to provide a limiting case, the present study considers a much simpler model situation in which the material of the jet is an inviscid, non-heat-conducting, perfect gas that issues radially at the surface of the moon with sonic velocity. Theoretical considerations show that the escape velocity of a gas is much smaller than that of a solid body. An analytical solution is obtained for the maximum height reached by a jet in steady flow. A computational parameter study of unsteady, inviscid, axisymmetric flow, including the effect of an atmosphere, provides a rich picture of the features and behaviour of the model jet. The deficit of the computed maximum steady-state penetration height below the isentropic theoretical value may be explained by the effect of the atmosphere and of dissipation in shock waves that occur in the computed flows. Many of the features of the gas jet are qualitatively mirrored in an experiment using a water flow analogy in which the gravitational field is simulated by a surface of suitable shape.

Parallel shear flows cannot be kinematic dynamos on their own (Zel’dovich, Sov. Phys. JETP, vol. 4, 1957, pp. 460–462), but the addition of small flow perturbations can trigger dynamo action. Using an optimization algorithm inspired by Willis (Phys. Rev. Lett., vol. 109 (25), 2012, 251101) and Chen et al. (J. Fluid Mech., vol. 783, 2015, pp. 23–45), we identify the smallest perturbation that when added to Kolmogorov flow can trigger dynamo action at some fixed value of the magnetic Reynolds number. In this way we numerically measure the fragility of the Zel’dovich anti-dynamo theorem. The minimal perturbations have surprisingly simple spatial structures. Their magnitudes vary inversely proportional to the magnetic Reynolds number and are always much larger than theoretical lower bounds calculated here using the methods of Proctor (Geophys. Astrophys. Fluid Dyn., vol. 98 (3), 2004, pp. 235–240; J. Fluid Mech., vol. 697, 2012, pp. 504–510).

The present paper reports on our effort to characterize vortical interactions in complex fluid flows through the use of network analysis. In particular, we examine the vortex interactions in two-dimensional decaying isotropic turbulence and find that the vortical-interaction network can be characterized by a weighted scale-free network. It is found that the turbulent flow network retains its scale-free behaviour until the characteristic value of circulation reaches a critical value. Furthermore, we show that the two-dimensional turbulence network is resilient against random perturbations, but can be greatly influenced when forcing is focused towards the vortical structures, which are categorized as network hubs. These findings can serve as a network-analytic foundation to examine complex geophysical and thin-film flows and take advantage of the rapidly growing field of network theory, which complements ongoing turbulence research based on vortex dynamics, hydrodynamic stability, and statistics. While additional work is essential to extend the mathematical tools from network analysis to extract deeper physical insights of turbulence, an understanding of turbulence based on the interaction-based network-theoretic framework presents a promising alternative in turbulence modelling and control efforts.

We present two spanwise-localized travelling-wave solutions in the asymptotic suction boundary layer, obtained by continuation of solutions of plane Couette flow. One of the solutions has the vortical structures located close to the wall, similar to spanwise-localized edge states previously found for this system. The vortical structures of the second solution are located in the free stream far above the laminar boundary layer and are supported by a secondary shear gradient that is created by a large-scale low-speed streak. The dynamically relevant eigenmodes of this solution are concentrated in the free stream, and the departure into turbulence from this solution evolves in the free stream towards the walls. For invariant solutions in free-stream turbulence, this solution thus shows that the source of energy of the vortical structures can be a dynamical structure of the solution itself, instead of the laminar boundary layer.