Nature shuns homogeneity. In turbulent clouds, industrial reactors and geophysical flows, discrete particles arrange in clusters, posing difficult challenges to theory. A persistent question is whether clusters can be modelled with continuum equations. Recent evidence presented by Mitrano et al. (J. Fluid Mech., vol. 738, 2014, R2) indicates that suitable equations can predict the formation of clusters in granular flows, despite violating the simplifying assumptions upon which they are based.

We study experimentally the effect of adding an in-line oscillatory motion to the oscillatory heaving and pitching motion of flapping foils that use a power downstroke. We show that far from being a limitation imposed by the muscular structure of certain animals, in-line motion can be a powerful means to either substantially augment the mean lift, or reduce oscillatory lift and increase thrust; propulsive efficiency can also be increased. We also show that a model-based optimization scheme that is used to drive an iterative sequence of experimental runs provides exceptional ability for flapping foils to tightly vector and keep the force in a desired direction, hence improving performance in locomotion and manoeuvring. Flow visualization results, using particle image velocimetry, establish the connection of distinct wake patterns with flapping modes associated with high lift forces, or modes of high thrust and low lift forces.

New experiments are presented which explore the dynamics of a turbulent buoyant plume produced by a vertically distributed linear source of buoyancy of strength $f$ per unit height. In a uniform environment, the plume volume flux increases with height from the base of the source, $z$, as $q(z) = {2^{-1/3}} {\pi }^{2/3} \alpha ^{4/3} f^{1/3} z^2$ where the entrainment coefficient, $\alpha = 0.09\pm 0.01$. In an enclosed space, with a net upward vertical ventilation flow $Q_V$, the buoyant plume generates a steady ambient stratification. The lowest part of the space, $z<h_i$, where $q(h_i)=Q_V$, is filled with fluid supplied by the ventilation flow and there is a net upflow in the ambient. Above this, $z>h_i$, the ambient fluid is linearly stratified with a reduced gravity gradient $f/Q_V$, and has no net vertical motion. Instead, for $z>h_i$, the time-averaged volume flux in the plume equals the ventilation flow. The intermittent entrainment of ambient fluid into the plume is now matched by intermittent detrainment from the plume, and the mean buoyancy in the plume relative to the ambient remains constant. The supply of fresh ventilation fluid to the ambient in the linearly stratified zone only occurs through the local detrainment and consequent horizontal intrusion of fluid from the plume. This has key implications for design of ventilation systems, in which there may be vertically distributed sources of buoyancy.

Most classical work on the hydrodynamics of low-Reynolds-number swimming addresses deterministic locomotion in quiescent environments. Thermal fluctuations in fluids are known to lead to a Brownian loss of the swimming direction, resulting in a transition from short-time ballistic dynamics to effective long-time diffusion. As most cells or synthetic swimmers are immersed in external flows, we consider theoretically in this paper the stochastic dynamics of a model active particle (a self-propelled sphere) in a steady general linear flow. The stochasticity arises both from translational diffusion in physical space, and from a combination of rotary diffusion and so-called run-and-tumble dynamics in orientation space. The latter process characterizes the manner in which the orientation of many bacteria decorrelates during their swimming motion. In contrast to rotary diffusion, the decorrelation occurs by means of large and impulsive jumps in orientation (tumbles) governed by a Poisson process. We begin by deriving a general formulation for all components of the long-time mean square displacement tensor for a swimmer with a time-dependent swimming velocity and whose orientation decorrelates due to rotary diffusion alone. This general framework is applied to obtain the convectively enhanced mean-squared displacements of a steadily swimming particle in three canonical linear flows (extension, simple shear and solid-body rotation). We then show how to extend our results to the case where the swimmer orientation also decorrelates on account of run-and-tumble dynamics. Self-propulsion in general leads to the same long-time temporal scalings as for passive particles in linear flows but with increased coefficients. In the particular case of solid-body rotation, the effective long-time diffusion is the same as that in a quiescent fluid, and we clarify the lack of flow dependence by briefly examining the dynamics in elliptic linear flows. By comparing the new active terms with those obtained for passive particles we see that swimming can lead to an enhancement of the mean-square displacements by orders of magnitude, and could be relevant for biological organisms or synthetic swimming devices in fluctuating environmental or biological flows.

Gudmundsson and Colonius (J. Fluid Mech., vol. 689, 2011, pp. 97–128) have recently shown that the average evolution of low-frequency, low-azimuthal modal large-scale structures in the near field of subsonic jets are remarkably well predicted as linear instability waves of the turbulent mean flow using parabolized stability equations. In this work, we extend this modelling technique to an isothermal and a moderately heated Mach 1.5 jet for which the mean flow fields are obtained from a high-fidelity large-eddy simulation database. The latter affords a rigourous and extensive validation of the model, which had only been pursued earlier with more limited experimental data. A filter based on proper orthogonal decomposition is applied to the data to extract the most energetic coherent components. These components display a distinct wavepacket character, and agree fairly well with the parabolized stability equations model predictions in terms of near-field pressure and flow velocity. We next apply a Kirchhoff surface acoustic propagation technique to the near-field pressure model and obtain an encouraging match for far-field noise levels in the peak aft direction. The results suggest that linear wavepackets in the turbulence are responsible for the loudest portion of the supersonic jet acoustic field.

Inspired by the recent experiment on erythrocytes (red blood cells, RBCs) in weak shear flows by Dupire et al. (Proc. Natl Acad. Sci. USA, vol. 109, 2012, pp. 20808–20813), we conduct a numerical investigation to study the dynamics of RBCs in low-shear-rate flows by applying a multiscale fluid–structure interaction model. By employing a spheroidal stress-free state in the cytoskeleton, we are able to numerically predict an important feature, namely that the cell maintains its biconcave shape during tank-treading motions. Furthermore, we numerically confirm the hypothesis that, as the stress-free state approaches a sphere, the threshold shear rates corresponding to the establishment of tank treading decrease. By comparing with the experimental measurements, our study suggests that the stress-free state of RBCs is a spheroid that is close to a sphere, rather than the biconcave shape applied in existing models (the implication is that the RBC skeleton is pre-stressed in its natural biconcave state). It also suggests that the response of RBCs in low-shear-rate flows may provide a measure to quantitatively determine the distribution of shear stress in the RBC cytoskeleton in the natural state.

The turbulent/non-turbulent interface in a zero-pressure-gradient turbulent boundary layer at high Reynolds number ($\mathit{Re}_\tau =14\, 500$) is examined using particle image velocimetry. An experimental set-up is utilized that employs multiple high-resolution cameras to capture a large field of view that extends $2\delta \times 1.1\delta $ in the streamwise/wall-normal plane with an unprecedented dynamic range. The interface is detected using a criteria of local turbulent kinetic energy and proves to be an effective method for boundary layers. The presence of a turbulent/non-turbulent superlayer is corroborated by the presence of a jump for the conditionally averaged streamwise velocity across the interface. The steep change in velocity is accompanied by a discontinuity in vorticity and a sharp rise in the Reynolds shear stress. The conditional statistics at the interface are in quantitative agreement with the superlayer equations outlined by Reynolds (J. Fluid Mech., vol. 54, 1972, pp. 481–488). Further analysis introduces the mass flux as a physically relevant parameter that provides a direct quantitative insight into the entrainment. Consistency of this approach is first established via the equality of mean entrainment calculations obtained using three different methods, namely, conditional, instantaneous and mean equations of motion. By means of ‘mass-flux spectra’ it is shown that the boundary-layer entrainment is characterized by two distinctive length scales which appear to be associated with a two-stage entrainment process and have a substantial scale separation.

Experimental and computational results are presented on an aerofoil undergoing pitch oscillations in ground effect, that is, close to a solid boundary. The time-averaged thrust is found to increase monotonically as the mean position of the aerofoil approaches the boundary while the propulsive efficiency stays relatively constant, showing that ground effect can enhance thrust at little extra cost for a pitching aerofoil. Vortices shed into the wake form pairs rather than vortex streets, so that in the mean a momentum jet is formed that angles away from the boundary. The time-averaged lift production is found to have two distinct regimes. When the pitching aerofoil is between 0.4 and 1 chord lengths from the ground, the lift force pulls the aerofoil towards the ground. In contrast, for wall proximities between 0.25 and 0.4 chord lengths, the lift force pushes the aerofoil away from the ground. Between these two regimes there is a stable equilibrium point where the time-averaged lift is zero and thrust is enhanced by approximately 40 %.

The high-Reynolds-number behaviour of the canonical incompressible turbulent channel flow is investigated through large-scale direct numerical simulation (DNS). A Reynolds number is achieved ($Re_{\tau } = h/\delta _v \approx 4000$, where $h$ is the channel half-height, and $\delta _v$ is the viscous length scale) at which theory predicts the onset of phenomena typical of the asymptotic Reynolds number regime, namely a sensible layer with logarithmic variation of the mean velocity profile, and Kolmogorov scaling of the velocity spectra. Although higher Reynolds numbers can be achieved in experiments, the main advantage of the present DNS study is access to the full three-dimensional flow field. Consistent with refined overlap arguments (Afzal & Yajnik, J. Fluid Mech. vol. 61, 1973, pp. 23–31; Jiménez & Moser, Phil. Trans. R. Soc. Lond. A, vol. 365, 2007, pp. 715–732), our results suggest that the mean velocity profile never achieves a truly logarithmic profile, and the logarithmic diagnostic function instead exhibits a linear variation in the outer layer whose slope decreases with the Reynolds number. The extrapolated value of the von Kármán constant is $k \approx 0.41$. A near logarithmic layer is observed in the spanwise velocity variance, as predicted by Townsend’s attached eddy hypothesis, whereas the streamwise variance seems to exhibit a shoulder, perhaps being still affected by low-Reynolds-number effects. Comparison with previous DNS data at lower Reynolds number suggests enhancement of the imprinting effect of outer-layer eddies onto the near-wall region. This mechanisms is associated with excess turbulence kinetic energy production in the outer layer, and it reflects in flow visualizations and in the streamwise velocity spectra, which exhibit sharp peaks in the outer layer. Associated with the outer energy production site, we find evidence of a Kolmogorov-like inertial range, limited to the spanwise spectral density of $u$, whereas power laws with different exponents are found for the other spectra. Finally, arguments are given to explain the ‘odd’ scaling of the streamwise velocity variances, based on the analysis of the kinetic energy production term.

This paper reports results from a direct numerical simulation of the flow past a circular cylinder with axial curvature. The main objective is to explore the effects of spanwise curvature on the stability of the shear layers and the turbulent wake at the subcritical Reynolds number of 3900. The bluff-body geometry is adapted from a previous study conducted at lower Reynolds numbers, in which a quarter segment of a ring represented the deformed cylinder. A convex configuration in which the free-stream direction is towards the outer face of the ring is adopted here. The present results show a striking distinction between the upper and lower wake regions. Despite the turbulent character of the wake, the upper wake region is more coherent due to the periodic vortex shedding of primary vortical structures, which are in close alignment with the axial curvature. A mild axial flow develops upwards along the lee face of the curved cylinder, displacing the vortex formation region further downstream from the location expected for a straight cylinder at the same Reynolds number. In the lower wake region the vortex shedding strength is drastically reduced due to larger local inclination, resulting in higher three-dimensionality and loss of coherence. A strong downdraft with a swirling pattern is the dominating feature in the lower base region. This is associated with a substantial decrease of the base suction, and the suppression of the characteristic recirculating backflow.

Compressible hydrodynamic turbulence is studied under the assumption of a polytropic closure. Following Kolmogorov, we derive an exact relation for some two-point correlation functions in the asymptotic limit of a high Reynolds number. The inertial range is characterized by: (i) a flux term implying in particular the enthalpy; and (ii) a purely compressible term $\mathcal{S}$ which may act as a source or a sink for the mean energy transfer rate. At subsonic scales, we predict dimensionally that the isotropic $k^{-5/3}$ energy spectrum for the density-weighted velocity field ($\rho ^{1/3} \boldsymbol {v}$), previously obtained for isothermal turbulence, is modified by a polytropic contribution, whereas at supersonic scales $\mathcal{S}$ may impose another scaling depending on the polytropic index. In both cases, it is shown that the fluctuating sound speed is a key ingredient for understanding polytropic compressible turbulence.

When a coin falls in water, its trajectory is one of four types, determined by its dimensionless moment of inertia $I^\ast $ and Reynolds number $\text {Re}$: (A) steady; (B) fluttering; (C) chaotic; or (D) tumbling. The dynamics induced by the interaction of the water with the surface of the coin, however, makes the exact landing site difficult to predict a priori. Here, we describe a carefully designed experiment in which a coin is dropped repeatedly in water to determine the probability density functions (p.d.f.s) associated with the landing positions for each of the four trajectory types, all of which are radially symmetric about the centre drop-line. In the case of the steady mode, the p.d.f. is approximately Gaussian distributed with small variances, indicating that the coin is most likely to land at the centre, right below the point from which it is dropped. For the other falling modes, the centre is one of the least likely landing sites. Indeed, the p.d.f.s of the fluttering, chaotic and tumbling modes are characterized by a ‘dip’ around the centre. In the tumbling mode, the p.d.f. is a ring configuration about the centreline whereas in the chaotic mode, the p.d.f. is generally a broadband distribution spread out radially symmetrically about the centreline. For the steady and fluttering modes, the coin never flips, so the coin lands with the same side up as when it was dropped. The probability of heads or tails is close to 0.5 for the chaotic mode and, in the case of the tumbling mode, the probability of heads or tails is based on the height of the drop which determines whether the coin flips an even or odd number of times during descent.

Non-normal transient growth of disturbances is considered as an essential prerequisite for subcritical transition in shear flows, i.e. transition to turbulence despite linear stability of the laminar flow. In this work we present numerical and analytical computations of linear transient growth covering all linearly stable regimes of Taylor–Couette flow. Our numerical experiments reveal comparable energy amplifications in the different regimes. For high shear Reynolds numbers $\mathit{Re}$, the optimal transient energy growth always follows a $\mathit{Re}^{2/3}$ scaling, which allows for large amplifications even in regimes where the presence of turbulence remains debated. In co-rotating Rayleigh-stable flows, the optimal perturbations become increasingly columnar in their structure, as the optimal axial wavenumber goes to zero. In this limit of axially invariant perturbations, we show that linear stability and transient growth are independent of the cylinder rotation ratio and we derive a universal $\mathit{Re}^{2/3}$ scaling of optimal energy growth using Wentzel–Kramers–Brillouin theory. Based on this, a semi-empirical formula for the estimation of linear transient growth valid in all regimes is obtained.

A new statistical coherent structure (CS), the velocity–vorticity correlation structure (VVCS), using the two-point cross-correlation coefficient $R_{ij}$ of velocity and vorticity components, $u_i$ and $\omega _j~ (i, j = 1, 2, 3)$, is proposed as a useful descriptor of CS. For turbulent channel flow with the wall-normal direction $y$, a VVCS study consists of using $u_i$ at a fixed reference location $y_r$, and using $|R_{ij} (y_r; x, y, z)|\geqslant R_0$ to define a topologically invariant high-correlation region, called $\mathit{VVCS}_{ij}$. The method is applied to direct numerical simulation (DNS) data, and it is shown that the $\mathit{VVCS}_{ij}$ qualitatively and quantitatively captures all known geometrical features of near-wall CS, including spanwise spacing, streamwise length and inclination angle of the quasi-streamwise vortices and streaks. A distinct feature of the VVCS is that its geometry continuously varies with $y_r$. A topological change of $\mathit{VVCS}_{11}$ from quadrupole (for smaller $y_r$) to dipole (for larger $y_r$) occurs at $y^{+}_r=110$, giving a geometrical interpretation of the multilayer nature of wall-bounded turbulent shear flows. In conclusion, the VVCS provides a new robust method to quantify CS in wall-bounded flows, and is particularly suitable for extracting statistical geometrical measures using two-point simultaneous data from hotwire, particle image velocimetry/laser Doppler anemometry measurements or DNS/large eddy simulation data.

The temporal evolution of nonlinear large-scale internal gravity waves, in a two-layer flow affected by background rotation, is studied via laboratory experiments conducted in a cylindrical tank, mounted on a rotating turntable. The internal wave field is excited by the relaxation of an initial forced tilt of the density interface ($\eta _{i}$), which generates internal waves, such as Kelvin and Poincaré waves, in response to rotation effects. The behaviour of $\eta _{i}$, in the shore region, is analysed in terms of the background rotation and the nonlinear steepening of the basin-scale waves. The results show that the degeneration of the fundamental Kelvin wave into a solitary-type wave packet is caused by nonlinear steepening and it is influenced by the background rotation. In addition, the physical scales of the leading solitary-type wave are closer to Korteweg–de Vries theory as the rotation increases. Moreover, the nonlinear interaction between the Kelvin wave and the Poincaré wave can transfer energy to higher or lower frequencies than the frequency of the fundamental Kelvin wave, as a function of the background rotation. In particular, a specific normal mode in the off-shore region could be energized by this interaction. Finally, the bulk decay rate of the fundamental Kelvin wave, $\tau _{dk}$, was investigated. The results exhibit that $\tau _{dk}$ is concordant with the Ekman damping time scale when there is no evidence of steepening in the basin-scale waves. However, as nonlinear processes increase, $\tau _{dk}$ shows a strong decrease. In this context, the nonlinear processes play an important role in the decay of the fundamental Kelvin wave, via the energy radiation to other modes. The results reported demonstrate that the background rotation and nonlinear processes are essential aspects in understanding the degeneration and the decay of large-scale internal gravity waves on enclosed basins.

In this study we investigate vortex structures which lead to the maximum possible growth of palinstrophy in two-dimensional incompressible flows on a periodic domain. The issue of palinstrophy growth is related to a broader research program focusing on extreme amplification of vorticity-related quantities which may signal singularity formation in different flow models. Such extreme vortex flows are found systematically via numerical solution of suitable variational optimization problems. We identify several families of maximizing solutions parameterized by their palinstrophy, palinstrophy and energy and palinstrophy and enstrophy. Evidence is shown that some of these families saturate estimates for the instantaneous rate of growth of palinstrophy obtained using rigourous methods of mathematical analysis, thereby demonstrating that this analysis is in fact sharp. In the limit of small palinstrophies the optimal vortex structures are found analytically, whereas for large palinstrophies they exhibit a self-similar multipolar structure. It is also shown that the time evolution obtained using the instantaneously optimal states with fixed energy and palinstrophy as the initial data saturates the upper bound for the maximum growth of palinstrophy in finite time. Possible implications of this finding for the questions concerning extreme behaviour of flows are discussed.

Starting from a kinetic theory (KT) model for monodisperse granular flow, the exact Reynolds-averaged (RA) equations are derived for the particle phase in a collisional fluid–particle flow. The corresponding equations for a constant-density fluid phase are derived from a model that includes drag and buoyancy coupling with the particle phase. The fully coupled macroscale/hydrodynamic model, rigorously derived from a kinetic equation for the particles, is written in terms of the particle-phase volume fraction, the particle-phase velocity and the granular temperature (or total granular energy). As derived from the hydrodynamic model, the RA turbulence model solves for the RA particle-phase volume fraction, the phase-averaged (PA) particle-phase velocity, the PA granular temperature and the PA turbulent kinetic energy of the particle phase. Thus, unlike in most previous derivations of macroscale turbulence models for moderately dense granular flows, a clear distinction is made between the PA granular temperature $\Theta _\textit {p}$, which appears in the KT constitutive relations, and the particle-phase turbulent kinetic energy $k_\textit {p}$, which appears in the turbulent transport coefficients. The exact RA equations contain unclosed terms due to nonlinearities in the hydrodynamic model and we briefly discuss the available closures for these terms. Finally, we demonstrate by comparing model predictions with direct numerical simulation results that even for non-collisional fluid–particle flows it is necessary to provide separate models for $\Theta _\textit {p}$ and $k_\textit {p}$ in order to correctly account for the effect of the particle Stokes number and mass loading.

A theory is developed that describes acoustic microstreaming around a gas bubble undergoing small radial and translational oscillations in the presence of a distant rigid wall. It is shown that the presence of the wall can change the amplitude and the phase of the bubble oscillations in such a way that the intensity of acoustic microstreaming is increased considerably as compared with that generated by the same bubble in an infinite liquid. This occurs if the driving frequency is close to the resonance frequency that the bubble has in the presence of the wall. Equations for acoustic microstreaming in the boundary layer at the wall are also provided.

Results from a free-surface magnetohydrodynamic (MHD) flow experiment are presented detailing the modification of vortices in the wake of a circular cylinder with its axis parallel to the applied magnetic field. Experiments were performed at Reynolds numbers of the order of ${\mathit{Re}}\sim 10^4$ as the interaction parameter ${\mathit{N}}$, representing the ratio of electromagnetic forces to inertial forces, was increased through unity. The von Kármán vortex street in the wake of the cylinder was observed by simultaneously sampling the gradient of the induced electric potential, $ \boldsymbol {\nabla }{\phi }$, at 16 cross-stream locations as a proxy for the streamwise fluid velocity. An ensemble of vortex velocity profiles was measured as a function of the applied magnetic field strength. Results indicate a significant change in the circulation of vortices and the deviations from the average profile as ${\mathit{N}}$ was increased. By sampling the fluctuations in $\boldsymbol {\nabla }{\phi }$ at three locations in the wake, the decay of the vortices was examined and the effective viscosity was found to decrease as ${\mathit{N}}^{-0.49 \pm 0.04}$. Using temperature as a passive tracer, qualitative observations were made with an infrared (IR) camera that showed significant changes in the wake, including the absence of small-scale structures at high magnetic field strengths. Collectively, the results suggest that the reduction in effective viscosity was due to the suppression of the small-scale eddies by the magnetic field. The slope of the power spectrum was observed to change from a $k^{-1.8}$ power law at low ${\mathit{N}}$ to a $k^{-3.5}$ power law for ${\mathit{N}}> 1$. Together, these results suggest the flow smoothly transitioned from a hydrodynamic state to a magnetohydrodynamic regime over the range of $0 < {\mathit{N}}< 1$.

Direct numerical simulation (DNS), based on high-order numerical schemes, is used to study the effects of distributed pressure gradients on the redistribution of fluctuating kinetic energy in supersonic nozzle and diffuser flow with incoming fully developed turbulent pipe flow. Axisymmetric geometries and flow parameters have been selected such that shock waves are avoided and streamline curvature remains unimportant. Although mean extra rates of strain are quite small, strong changes in Reynolds stresses and their production/redistribution mechanisms are observed, in agreement with findings of Bradshaw (J. Fluid Mech., vol. 63, 1974, pp. 449–464). The central role of pressure–strain correlations in changing the Reynolds stress anisotropy is highlighted. A Green’s function-based analysis of pressure–strain correlations is presented, showing remarkable agreement with DNS data.