Search

In this work numerical simulation of the breaking induced by the modulational instability are presented. The study is focused on the energy and momentum exchanged taking place at the air-water interface during the breaking process. Simulations are performed by a two-fluids Navier-Stokes solver combined with the VOF technique, which provides a detailed description of the flow in both phases. Results obtained by two different initial conditions are presented in terms of free surface shapes, energy in air and water, vertical fluxes of horizontal momentum and vorticity.

In search for the cheapest but still reliable numerical simulation, a systematic study on the effect of the computational domain ("box") size on direct numerical simulations of Taylor-Couette flow was performed. Four boxes, with varying azimuthal and axial extents were used. The radius ratio between the inner cylinder and the outer cylinder was fixed to $\eta=r_i/r_o=0.909$, and the outer was kept stationary, while the inner rotated at a Reynolds number $Re_i=10^5$. Profiles of mean and fluctuation velocities are compared, as well as autocorrelations and velocity spectra. The smallest box is found to accurately reproduce the torque and mean azimuthal velocity profiles of larger boxes, while having smaller values of the fluctuations than the larger boxes. The axial extent of the box directly reflects on the Taylor-rolls and plays a crucial role on the correlations and spectra. The azimuthal extent is also found to play a significant role, as larger boxes allow for azimuthal wave-like patterns in the Taylor rolls to develop, which affects the statistics in the bulk region. For all boxes studied, the spectra does not reach a box independent maximum.

In this paper we show how the Immersed Boundary (IB) method can be used with the Large-Eddy-Simulation (LES) to compute moderately high Reynolds number flows in complex geometric configurations. The resulting combination gives an easy-to-use, inexpensive and accurate technique which can be an important step towards the application of computational fluid dynamics (CFD) to industrially relevant problems. This paper aims at describing the main features of the method, some of the important drawbacks and possible solutions. Several representative examples are discussed in order to show the flexibility and the range of the applicability of this technique.

Direct numerical simulations of thermal convection over grooved plates are presented and discussed, in comparison with the standard flat-plate case, in order to gain a better understanding of the altered near-wall dynamics and of the enhancement of the heat transfer. The simulations are performed in a cylindrical cell of aspect-ratio (diameter over cell height) Γ = 1/2 at fixed Prandtl number Pr = 0.7 with the Rayleigh number Ra ranging from 2 × 106 to 2 × 1011. The results show an increase of heat transfer, or in non-dimensional form the Nusselt number Nu when the mean thermal boundary-layer thickness becomes smaller than the groove height, in agreement with earlier experimental investigations available from the literature. The present increase, however, results in a steeper power law of the Nu vs. Ra law rather than a simple upward shift of the Nu law of the flat plate. This finding agrees with some studies, but it is at variance with others. Possible causes for this difference are discussed with the help of an electrical analogy.

In this paper we propose a simple model that, by comparing different time scales, allows a prediction for the mean flow structure and its dynamics in confined thermal convection in a cylindrical cell of aspect ratio (diameter over cell height) $\Gamma\,{=}\,1/2$ . It is shown that the break-up of the mean elongated recirculation into two counter-rotating unity-aspect-ratio rolls, sometimes referred to as flow bimodality, occurs only in a narrow range of Rayleigh numbers whose extrema depend on the Prandtl number. The predictions of the present model are consistent with the published literature, according to which the dual mean flow structure has been observed in numerical simulations at $\hbox{\it Pr}\,{=}\,0.7$ and experiments in gaseous helium ( $\hbox{\it Pr}\,{\approx}\,0.7$ ) but never in water at ‘ambient’ temperature ( $\hbox{\it Pr} \,{\approx}\,5$ ) and only once in water at $T\,{=}\,80\,^\circ$ C ( $\hbox{\it Pr}\,{=}\,2$ ). Another prediction of the model is that the thermal properties of the sidewall affect the mean flow unsteadiness and, sometimes, prevent transitions via a subtle anchoring mechanism that has been identified and verified by ad hoc numerical simulations.

The three-dimensional o w generated by a apping foil of nite span is numerically simulated by integrating continuity and momentum equations. The foil moves with a constant forward speed and oscillates both angularly and vertically. The numerical results are compared with o w visualisations of Ref. 13. The vortex structures shed by the foil are analysed using dieren t approaches. Even for a foil span equal to three foil chord the results are signican tly dieren t from the two-dimensional case.

A numerical method is developed for simulating unsteady, 3‐D, low Reynolds number flow through a bileaflet mechanical heart valve with the leaflets moved by an assigned law. The method employs a fractional‐step approach in combination with a hybrid Runge‐Kutta/Crank‐Nicolson scheme; it is second‐order accurate in space and time. An Immersed Boundary approach is also used for the treatment of the leaflets and a geometrical preprocessor is developed for updating the relative position between the computational grid and the leaflets. Calculations are carried out for the full 3‐D valve geometry under pulsatile inflow conditions on a mesh with 129×109×192 points in azimuthal, radial and streamwise directions respectively (2.7×106 nodes). The computed results show that downstream of the leaflets the flow is dominated by two counter‐rotating vortices, which originate on either side of the leaflets interacting with the boundary layer of the aortic wall. Our results show the highly 3‐D flow structure and the symmetry breaking with respect to both symmetry planes. It demonstrates the need of highly resolved, fully 3‐D simulations to accurately predict the flow in prosthetic mechanical heart valves.

The results from direct numerical simulations of turbulent Boussinesq convection are briefly presented. The flow is computed for a cylindrical cell of aspect ratio 1∕2 in order to compare with the results from recent experiments. The results span eight decades of Ra from 2×106 to 2×1014 and form the baseline data for a strictly Boussinesq fluid of constant Prandtl number (Pr=0.7). A conclusion is that the Nusselt number varies nearly as the 1∕3 power of Ra for about four decades towards the upper end of the Ra range covered.

The results of numerical experiments aimed at investigating the topology of the vortex structures shed by an oscillating foil of finite span are described. The motion of the foil and its geometry are chosen to mimic the tail of a fish using the carangiform swimming. The numerical results have been compared with the flow visualizations of Freymuth [J. Fluids Eng. 111, 217 (1989)] and those of von Ellenrieder et al. [J. Fluid Mech.490, 129 (2003)]. The results show that a vortex ring is shed by the oscillating foil every half a cycle. The dynamics of the vortex rings depends on the Strouhal number St. For relatively small values of St, the interaction between adjacent rings is weak and they are mainly convected downstream by the free stream. On the other hand, for relatively large values of St, a strong interaction among adjacent rings takes place and the present results suggest the existence of reconnection phenomena, which create pairs of longitudinal counter-rotating vortices.

Convective heat transport has important applications in engineering and meteorology and a better understanding of heat transport phenomena would lead to improvements in technological applications such as cooling of thermal machines and micro‐electronic components or cooling during a metallurgical fusion. It would also improve the prediction of geophysical motions in oceans and atmosphere. The use of rough surfaces is a way to enhance the heat flux. The interaction between the main shear flow and the rough surface creates secondary vortices that enhance the detachment of thermal plumes from the tip of the rough elements. In this work numerical simulations are conducted in a cylindrical cell heated from below and cooled from above in presence of rough surfaces. A comparison of Rayleigh versus Nusselt number scaling between rough surfaces and smooth ones shows enhanced heat fluxes. The flow is solved using a direct numerical simulation (DNS) of the three dimensional unsteady Navier Stokes equations with the Boussinesq approximation and an immersed boundary approach is used for the treatment of rough surfaces.

Numerical simulations of the Stokes boundary layer over a three‐dimensional wavy wall are performed in order to investigate the role played by infinitesimal wall imperfections in triggering transition to turbulence. Our results show flow patterns qualitatively similar to those experimentally detected. In particular the laminar, disturbed‐laminar and intermittent turbulent regimes are recovered. The characteristics of the above flow regimes are analyzed.

Numerical simulations of a truck-type configuration are carried out with the aim to investigate the effectiveness of various drag reduction devices. Unsteady Reynolds-Averaged Navier-Stokes equations are solved to capture the periodic wake motion at the base of the model and comparisons with available measurements show satisfactory agreement. Two passive devices are applied to the back of the model to investigate their effect on the pressure recovery and the base pressure. Up to 15% drag reduction is obtained using a boattail appendix. In addition, a Coanda jet system is designed and simulated. In this case the amount of drag reduction is directly related to the amount of injected mass and a zero drag condition can be obtained.

Temporal statistics of temperature and velocity fluctuations are studied in a highly turbulent convective flow developed within a cylindrical cell of aspect ratio Γ=1/2. Numerical data are analyzed for Ra up to 2×1011 at fixed Prandtl number (Pr=0.7). Temperature and velocity time series are collected from numerical probes placed within the fluid volume in the bulk and close to the boundaries. It is shown that the effects of the boundaries and of the large scale recirculation are reflected in the temporal statistics of the analyzed quantities. Spectra and structure functions show that the temperature statistics follow the Bolgiano scaling while the velocity fluctuations exhibit scaling laws which are surprisingly very close to those of the temperature. These results are in contrast with the Bolgiano or Kolmogorov behaviors expected for the velocity statistics and a possible theoretical explanation is presented.

The moderate Reynolds number flow developing in a cylindrical unbaffled stirred tank is studied using direct numerical simulation (DNS) and the results are compared with available experimental data and Reynolds Averaged Navier–Stokes (RANS) solutions. The geometry of the impeller is handled using an immersed-boundary procedure implemented in a solver written for cylindrical coordinates. This allows efficient simulation of the flow at a reasonable computational cost and accurate prediction of the mean and rms velocity fields. For this configuration, RANS performs poorly because of the low Reynolds and strongly unsteady and inhomogeneous nature of the flow: many different flow structures are produced, ranging from small-scale vortices generated at the blade tips to large-scale meridional recirculation. It is shown, in addition, that inaccurate results are produced by the wrong (computational) assumption that an impeller with n blades produces instantaneous fields with an n-fold symmetry. Because simple stirred-tank configurations (like the present one) have been recently used to assess the performance of several RANS closures, the main message of this study is that simple geometrical configurations and low Reynolds numbers are not benign parameters for such a task. © 2004 American Institute of Chemical Engineers AIChE J, 50: 1109–1118, 2004

The mixing process in the formation of an axisymmetric vortex ring is analyzed by direct numerical simulation. After the initial phase, when the impulse is imparted to the fluid, the mixing increases linearly in time with a slope that is directly proportional to the vortex circulation (Γ) and to reactants diffusivity to the power −1/3 (Pe−0.33). In addition, for reacting flows with equivalence ratio (s) different from 1, it has been found that as s decreases the mixing first decreases and then increases. An explanation for this behavior is given.

Numerical experiments are conducted to study high-Rayleigh-number convective turbulence (Ra ranging from 2 × 10 6 up to 2 × 10 11 ) in a Γ = 1/2 aspect-ratio cylindrical cell heated from below and cooled from above and filled with gaseous helium (Pr = 0.7). The numerical approach allows three-dimensional velocity, vorticity and temperature fields to be analysed. Furthermore, several numerical probes are placed within the fluid volume, permitting point-wise velocity and temperature time series to be extracted. Taking advantage of the data accessibility provided by the direct numerical simulation the flow dynamics has been explored and separated into its mean large-scale and fluctuating components, both in the bulk and in the boundary layer regions

Axisymmetric numerical experiments of round jets are reported with the aim of showing the nozzle geometry effects on the jet evolution. It is shown in particular that, although the simulation of the nozzle flow has a non-negligible computational cost, its effect on the jet evolution can be very important. It is also demonstrated that the flow configuration resulting at the nozzle exhaust can not be assigned at the boundary of a computational domain without considering also the flow inside the nozzle since it is a combination of radial and axial velocity components supported by an appropriate pressure distribution.

The interaction between two shielded monopolar vortices has been investigated experimentally in a nonrotating linearly stratified fluid and by full three-dimensional (3D) numerical simulations. The characteristic Reynolds and Froude numbers in the experiments are approximately Re [[approximate]] 500?1000 and F [[approximate]] 0.2?0.4. In a first set of numerical simulations the flow is initialized with the experimentally obtained two-dimensional horizontal velocity field, which is measured in the horizontal symmetry plane of the vortices, and the lacking initial data such as the density perturbation and the full 3D vorticity field are provided by the so-called diffusion model [Beckers et al., J. Fluid Mech. 433, 1 (2001)]. A comparison of the experimental and numerical data shows that using the diffusion model to supply the full 3D initial data is reliable for the Reynolds and Froude numbers considered in present experiments. Conjecturing a wider range of applicability, a second set of numerical simulations has been performed using initial conditions for the flow field and the density perturbation entirely based on the diffusion model. These numerical simulations enable an investigation of the role of Reynolds numbers (up to Re = 10 000) and Froude numbers (F = 0.30, 0.65, and 1.0) which are outside the experimentally accessible range. These simulations provide a better understanding of the dipole formation process, the influence of the vortex shields and the density perturbation in this process, and of the 3D structure of the dipoles. Compactness of the dipole, entrainment of irrotational fluid from the rear, and tail formation behind the dipole have been discussed. Additionally, experimental and numerical results are reported on the interaction between two shielded monopoles of the same sign in a linearly stratified fluid. It is found that for the full range of Reynolds and Froude numbers mentioned above the presence of the vortex shields prevents the merging of these monopoles.

The scaling exponents of the velocity structure functions are predicted by a simple model based on the probability density function (PDF) of the local turbulent kinetic energy fluctuations with no hypotheses on the shape of the most intermittent coherent structures. The energy fluctuations are presumed to be induced by statistically independent coherent structures and, therefore, are characterized by exponential-like PDFs. Accounting also for the scaling of the energy with the scale, a recursive formula is obtained which permits the calculation of the scaling exponents of the velocity structure functions. The predicted exponents are checked to be in good agreement with previous theoretical and experimental results. The main hypotheses behind the present approach are also validated by analyzing experimental velocity time series acquired in a jet flow at moderately high Reλ. It is also shown that present approach allows the calculation of the scaling exponents also in cases of nonhomogeneous turbulence.

The effect of Prandtl number on the dynamics of a convective turbulent flow is studied by numerical experiments. In particular, three series of experiments have been performed; in two of them the Rayleigh number spanned about two decades while the Prandtl number was set equal to 0.022 (mercury) and 0.7 (air). In the third series, in contrast, we fixed the Rayleigh number at 6×105 and the Prandtl number was varied from 0.0022 up to 15. The results have shown that, depending on the Prandtl number, there are two distinct flow regimes; in the first (Pr[lsim ]0.35) the flow is dominated by the large-scale recirculation cell that is the most important ‘engine’ for heat transfer. In the second regime, on the other hand, the large-scale flow plays a negligible role in the heat transfer which is mainly transported by the thermal plumes.For the low-Pr regime a model for the heat transfer is derived and the predictions are in qualitative and quantitative agreement with the results of the numerical simulations and of the experiments. All the hypotheses and the consequences of the model are directly checked and all the findings are consistent with the predictions and with experimental observations performed under similar conditions. Finally, in order to stress the effects of the large-scale flow some counter examples are shown in which the large-scale motion is artificially suppressed.

The turbulent regime of a Rayleigh–Benard convective flow is studied through the analysis of temporal series and three-dimensional data fields obtained by direct numerical simulations. Numerical probes are positioned in the homogeneous (in the center) and non–homogeneous (close to the boundaries) turbulence regions of the cell and the corresponding data are obtained at low Prandtl (Pr=0.022) for a range of Rayleigh numbers 5×104⩽Ra⩽106. We discuss the dependence upon the Rayleigh number of some quantities characterizing the turbulence in the low and high frequency range such as length scales, dissipation and flow isotropy. The analysis of the velocity and vorticity spectra indicates that, in the center of the cell, the turbulence remains locally isotropic for the whole range of Ra considered. Furthermore, the dynamics of the large convective cell and the turbulence parameters, depend upon Ra in a manner consistent with experimental results at higher Ra. In conclusion, none of the indicators considered sup...

Low Prandtl number thermal convection is being studied, owing to the interest in geophysical applications and in many technical problems. Earth’s liquid-core convection (at Pr ~ 0.1) and sodium heat exchangers are only two examples among many. As pointed out by Thual (1992), however, large attention is devoted to this problem also because its study contributes to a large amount of theoretical progress in nonlinear dynamics and turbulence. In addition, many numerical and experimental works are available in literature for Prandtl numbers in the range 0.7–7, while less is known when this parameter assumes lower values; this is the main motivation for the present study.

It has recently been observed that intense elongated vortices (‘worms’) are present among the small scales of many turbulent flows (Siggia, 1981), (Vincent & Meneguzzi, 1991), (Jimenez et al., 1993). Jimenez et al. (1993) have shown that worms are natural features of both decaying and forced turbulence and, even when they are artificially removed, they reform from the background vorticity. Following the explanation given in Jimenez & Wray (1994) for the formation of elongated vortices in isotropic turbulence, three numerical experiments are presented to show that axial waves exist and form vorticity filaments in turbulent flows. In the first experiment angled short vortices (‘sticks’) are considered showing that waves homogenize the core also for not aligned vortices, which is most likely the case of turbulence. In the second, to balance the viscous diffusion, a uniform strain has been superimposed and a Burgers vortex has been obtained. However the model with a uniform strain is in contrast with the statistical turbulence properties. A non uniform strain with zero spatial mean is compatible with the strain field along worm axes observed in numerically simulated isotropic turbulence. This strain has been used in the third experiment producing axial waves and forming an elongated vortex. In this case the equilibrium solution was independent of the initial conditions, in agreement with real worms whose properties depend only on Re Г and Re λ.

This paper describes a study of the centred collision between a dipolar vortex and a solid circular cylinder. The flow was analysed experimentally by using dye visualizations and streak photography. Flow characteristics such as vorticity fields and the transport of passive tracers were compared with numerical simulations. Observations revealed that thin layers of vorticity, created at the cylinder wall are advected by the primary dipole halves, which, while rolling up into compact patches, give rise to the formation of two new asymmetric dipoles that move away along curved trajectories. The structure of the vorticity distribution inside the dipole, before and after the collision, has been investigated. Both the numerical and the experimental results indicate that the vorticity patches originating from the original primary dipole approximately preserve their original functional relationship ω=f(ψ), while the secondary vorticity patches show a tendency to organize into structures attaining a similar relationship.

Direct numerical simulations of viscous columnar vortices subject to a non uniform strain have shown that more or less uniform vortices can form even when the strain has zero spatial average, thus having compressing regions. According to the arguments of Jimenez & Wray (1994) this is due to the effect of axial waves that homogenize the core of the vortices regardless the fact that they are initially straight or not. In the present study it has been shown that the same mechanism works for unsteady strains, provided axial waves have enough time to develop along the vortex core.

Direct and large eddy simulations of a temporal evolving round jet at low Reynolds number have been performed using a second-order finite differences scheme in cylindrical coordinates. Both a Smagorinsky model and a dynamical model has been used. The introduction of the models yields a substantial improvement to the results when coarse grids are used.

We report on the results of a combined experimental and numerical study on the fluid motion generated by the controlled approach and arrest of a solid sphere moving towards a solid wall at moderate Reynolds number. The experiments are performed in a small tank filled with water for a range of Reynolds numbers for which the flow remains axisymmetric. The fluid agitation of the fluid related to the kinetic energy is obtained as function of time in the experiment in a volume located around the impact point. The same quantities are obtained from the numerical simulations for the same volume of integration as in the experiments and also for the entire volume of the container. As shown in previous studies, this flow is characterized by a vortex ring, initially in the wake of the sphere, that spreads radially along the wall, generating secondary vorticity of opposite sign at the sphere surface and wall. It is also observed that before the impact, the kinetic energy increases sharply for a small period of time and then decreases gradually as the fluid motion dies out. The measure of the relative agitation of the collision is found to increase weakly with the Reynolds number Re. The close agreement between the numerics and experiments is indicative of the robustness of the results. These results may be useful in light of a potential modelling of particle-laden flows. Movies illustrating the spatio-temporal dynamics are provided with the online version of this paper

Accurate numerical simulations of vortex rings impinging on flat boundaries have revealed the same features observed in the experiment of Walker et al. (1987). They observed at Re v > 1250 the formation of azimuthal instabilities in the secondary ring during its compression within the primary ring. In the present numerical simulation the number of waves agrees very well with those observed in the experiment. The distributions of the vorticities together with the distribution of each term in the voriicity equations give insights on the formation of azimuthal instabilities.

A direct numerical simulation of a three-dimensional free evolving vortex ring has been performed to understand some features of the dynamics of vortex rings. In previous experiments, a persistent wake has been observed and the direct simulation has clarified the causes of this wake formation. A kinematic alignment analysis of the vorticity vector with the eigenvectors of the strain-rate tensor, together with vorticity visualisations, have been employed to explain the growth of the ‘Widnall’ instabilities. The analysis has shown the presence of two different zones, the first characterised by intense vorticity and low strain rate and the second with opposite features. In the former, there is a ring core which maintains its coherence even in the long term, in the latter there are elongated structures contributing to the deformation of the ring.

Two-dimensional simulations of the flow around a circular cylinder, performed at several values of the Reynolds number, revealed that when a constant velocity is assigned at the outer boundary, accurate solutions are obtained only for large distances of the boundary from the body. On the other hand, at a reasonable distance from the cylinder, radiative boundary conditions produce results which compare well with the experimental observations, also if the external boundary is quite close to the cylinder.

In this paper the simulation of a free ring together with a passive scalar has been performed. The axisymmetric simulation has shown that the pseudo-steady solution is obtained when a single functional relationship between ω = ω θ /r and the Stokes streamfunction is achieved. The relationship is linear around the centre of the ring and the slope depends on the initial vorticity distribution. The passive scalar does not correlate with ω θ and correlates well with ω. In the 3D case the simulation has shown completely different distributions between passive scalar and vorticity field. The passive scalar tends to be concentrated in regions with low vorticity and low rate of strain. This is a further proof that flow visualisations can be misleading when conclusions on the vorticity field are withdrawn from markers.

We study the effects of severe non-Boussinesq conditions on thermal convection at the moderate Rayleigh numbers of Ra = 2◊10 8 and 2◊10 9 by resorting to direct numerical computations of the full governing equations. We illustrate the effects by considering low temperature gaseous helium. The properties of helium are allowed to depend on the temperature around the mean of 5.4K. The Nusselt number is shown to decrease as the system departs from the Boussinesq approximation. For the Rayleigh numbers chosen here, the role of viscosity in thermal convection is limited to smudging the plume generation at the bottom surface, whereas the thermal expansion coefficient is demonstrated to have a larger impact on heat transport.

A reply to the comment on the anisotropy in the optical spectrum of the GaAs(110) surface is given. The difference between polarized-refractivity spectrum, was attributed to different sizes of supercell in the calculations. It is pointed out that there are other important differences between the two calculations. (AIP)

In Rayleigh–Bénard convection experiments, the thermal coupling between the sidewall and fluid is unavoidable. As a result, the thermal properties of the sidewall can influence the flow structure that develops. To get a better understanding of the influence of the sidewall, we performed a one-to-one comparison between experiments and direct numerical simulations (DNS) in aspect ratio (diameter over height) $\unicode[STIX]{x1D6E4}=1.00$ samples. We focus on the global heat transport, i.e. the Nusselt number $Nu$, and the local vertical temperature gradients near the horizontal mid-plane on the cylinder axis and close to the sidewall. The data cover the range $10^{5}\lesssim Ra\lesssim 10^{10}$ where $Ra$ is the Rayleigh number. The $Nu$ number obtained from experimental measurements and DNS, in which we use an adiabatic sidewall, agree well. The experiments are performed with several gases, which have widely varying thermal conductivities, but all have a Prandtl number $Pr\approx 0.7$. For $Ra\gtrsim 10^{7}$, both experiments and DNS reveal a stabilizing (positive) temperature gradient at the cylinder axis. This phenomenon was known for high $Pr$, but had not been observed for small $Pr\approx 0.7$ before. The experiments reveal that the temperature gradient decreases with decreasing $Ra$ and eventually becomes destabilizing (negative). The decrease appears at a higher $Ra$ when the sidewall admittance, which measures how easily the heat transfers from the fluid to the wall, is smaller. However, the simulations with an adiabatic sidewall do not reproduce the destabilizing temperature gradient at the cylinder axis in the low $Ra$ number regime. Instead, these simulations show that the temperature gradient increases with decreasing $Ra$. We find that the simulations can reproduce the experimental findings on the temperature gradient at the cylinder axis qualitatively when we consider the physical properties of the sidewall and the thermal shields. However, the temperature gradients obtained from experiments and simulations do not agree quantitatively. The reason is that it is incredibly complicated to reproduce all experimental details accurately due to which it is impossible to reproduce all experimental measurement details. The simulations show, in agreement with the models of Ahlers (Phys. Rev. E, vol. 63 (1), 2000, 015303) and Roche et al. (Eur. Phys. J. B, vol. 24 (3), 2001, pp. 405–408), that the sidewall can act as an extra heat conductor, which absorbs heat from the fluid near the bottom plate and releases it into the fluid near the top plate. The importance of this effect decreases with increasing $Ra$. A crucial finding of the simulations is that the thermal coupling between the sidewall and fluid can strongly influence the flow structure, which can result in significant changes in heat transport. Since this effect goes beyond a simple short circuit of the heat transfer through the sidewall, it is impossible to correct experimental measurements for this effect. Therefore, careful design of experimental set-ups is required to minimize the thermal interaction between the fluid and sidewall.

Numerical simulations are presented of the long time behaviour of viscous columnar vortices subject to non-uniform axial stretching. The relevant result is that the vortices reach a steady state even when the axial average of the strain is zero, such that they are being compressed during half of their extent. The structure of the flow is analysed and shown to range from local Burgers equilibrium to massive separation. For an intermediate range of Reynolds numbers the vortices are more or less uniform and compact, and it is suggested that this condition is related to the strong vortices observed in turbulent flows. The reason for the survival of the vortices under compression is traced to induced axial pressure gradients and to the viscous cancellation of outgoing vorticity. Theoretical analyses of the linear Burgers’ regime and of the onset of separation are presented and compared to the numerical experiments. The results are related to the observation of intermittency in turbulence, and shown to be consistent both with the observed scaling of vortex diameter, and with the lack of intermittency of the velocity signal.

In this paper, Direct Numerical Simulation (DNS) and Lagrangian Particle Tracking are used to investigate dispersion and deposition of particles swarms in convective flow confined in a cylindrical domain with aspect ratio (diameter over height) 0.5. The numerical simulations are carried out with Prandtl and Rayleigh numbers respectively equal to Pr = 0.7 and Ra = 2108. For these values of aspect ratio, Rayleigh and Prandtl numbers the flow is turbulent and time-dependent. In such flow, three sets of particles with Stokes numbers, based on the large eddy time scale, equal to Stf = 0.01, Stf = 0.005 and Stf = 0.001 are randomly dispersed. Particles distribution in turbulent convective flow is highly inhomogeneous and shows the clustering correlated with the vortical structures. The level of clustering is computed with the deviation of particles probability density function (PDF) from Poisson distribution. With this technique is available the size of the cluster but their geometry is unknown. The organisation along lines, planes and surfaces was investigated using the fractal dimension of the cluster. Finally, the flow topology is studied to relate the particles dispersion to coherent flow structures.

This paper discusses numerical experiments in which an initially uniform columnar vortex is subject to several types of axisymmetric forcing that mimic the strain field of a turbulent flow. The mean value of the strain along the vortex axis is in all cases zero, and the vortex is alternately stretched and compressed. The emphasis is on identifying the parameter range in which the vortex survives indefinitely. This extends previous work in which the effect of steady single-scale non-uniform strains was studied. In a first series of experiments the effect of the unsteadiness of the forcing is analysed, and it is found that the vortex survives as a compact object if the ratio between the oscillation frequency and the strain itself is low enough. A theoretical explanation is given which agrees with the numerical results. The strain is then generalized to include several spatial scales and oscillation frequencies, with characteristics similar to those in turbulent flows. The largest velocities are carried by the large scales, while the highest gradients and faster time scales are associated with the shorter wavelengths. Also in these cases ‘infinitely long’ vortices are obtained which are more or less uniform and compact. Vorticity profiles averaged along their axes are approximately Gaussian. The radii obtained from these profiles are proportional to the Burgers' radius of the r.m.s. (small-scale) axial strain, while the azimuthal velocities are proportional to the maximum (large-scale) axial velocity differences. The study is motivated by previous observations of intense vortex filaments in turbulent flows, and the scalings found in the present experiments are consistent with those found in the turbulent simulations.

Rayleigh-Bénard (RB) convection, the flow between a cold top plate and a hot bottom plate, is commonly used to model natural convection. Experimental realizations of RB usually consist of low aspect-ratio geometries and are limited to no-slip velocity boundary conditions. This is not representative of most geo- and astrophysical systems we wish to model. In this contribution, the effect of aspect-ratio and velocity boundary conditions on a 2D RB system have been studied by varying the sidewall boundary conditions between no-slip, stress-free and periodic, and the differences in the heat transport are measured. Heat transport is found to be larger for stress-free and periodic cells than for the no-slip case. Aspect ratio plays a small role in periodic cells, as long as the aspect ratio is larger than a certain threshold. For stress-free and no-slip cells, aspect ratio plays opposing roles. Thinner cells have larger heat transport than square cells when the sidewalls are stress-free. In contrast, for cells with no-slip walls, square cells have a larger heat transport than thin cells.