betaFeedback

WorksPhysics and Astronomy

Intermittency and collisions of fast sedimenting droplets in turbulence

Abstract

We study theoretically and numerically spatial distribution and collision rate of droplets that sediment in homogeneous isotropic Navier-Stokes turbulence. It is assumed that typical turbulent accelerations of fluid particles are much smaller than gravity. This was shown to imply that the particles interact weakly with individual vortices and, as a result, form a smooth flow in most of the space. In weakly intermittent turbulence with moderate Reynolds number, rare regions where the flow breaks down can be neglected in the calculation of space averaged rate of droplet collisions. However, increase of Re increases probability of rare, large quiescent vortices whose long coherent interaction with the particles destroys the flow. Thus at higher Re the space averaged collision rate forms in rare regions where the assumption of smooth flow breaks down. We describe the transition between the regimes and provide collision kernel in the case of moderate Re describable by the flow. The distribution of pairwise distances is shown to obey a separable dependence on the magnitude and the polar angle of the separation vector. Magnitude dependence obeys a power-law with a negative exponent, manifesting multifractality of the droplet's attractor. We provide the so far missing numerical confirmation of a relation between this exponent and the Lyapunov exponents and demonstrate that it holds beyond the theoretical range. The angular dependence of the RDF exhibits a maximum at small angles quantifying particle's formation of spatial columns. We derive the droplet's collision kernel using that in the considered limit the gradients of droplet's flow are Gaussian. We demonstrate that as Re increases the column's aspect ratio decreases, eventually becoming one when the isotropy is restored. We propose how the theory could be constructed at higher Re of clouds by using the example of the RDF.

Introduction

Turbulence of air in warm clouds accelerates collisions of water droplets and thus must be included in studies of precipitation [1, 2, 3, 4, 5, 6]. This inclusion is of high interest since it could help to resolve the bottleneck problem in rain formation. The bottleneck is caused by the narrowness of the size distribution of droplets created by condensation of vapor. The size proximity implies that the difference in settling velocities of the droplets is so small that their gravitational collisions would take long times, that are incompatible with the observations. Hence, the passage to the later stages of formation of large precipitating drops, where the growth occurs by gravitational collisions of different size droplets, demands something beyond gravity. Turbulence, that creates size dispersion by introducing mechanism for collisions of equal-size droplets, can be the missing factor. However, the relevance of turbulence has not been quantified so far [7] and assessment of how much turbulence influences the rain formation is an open problem. A main problem is the extremely high Reynolds number that holds in the clouds. Thus numerical studies can be performed only at much smaller Taylor microscale Reynolds numbers ReλRe_\lambda than in the clouds. This makes rigorous theoretical studies of droplets’ collisions in the Navier-Stokes turbulence (NST), which is the focus of this paper, specially significant. However theoretical studies of droplets’ behavior in high-ReλRe_\lambda NST are obstructed by having to deal with flow whose statistics is unknown. The problem is aggravated because intermittency produces significant probability of rare events that may locally accelerate the collision rates by a large factor in comparison with estimates using typical events.

The lack of knowledge of statistics of turbulence by itself does not prohibit quantitative predictions for the turbulent transport. In fact, the theory is able to provide accurate quantitative predictions for the NST [4]. This is thanks to independence of certain properties of turbulent transport of the details of the statistics (universality). As it is often the case in statistical physics, the universality holds due to the appearance of sum of a large number of independent random variables in the analysis. Here we study the case where the particle sediments so quickly through an individual vortex that the vortex perturbs its motion only weakly. Thus the droplet’s velocity is determined by effects of many independent vortices accumulated during the particle velocity relaxation time τ\tau. The issuing universality demands that intermittency is not too strong so that the probability of large quiescent vortices is not too high (intermittency both increases the probability of strong bursts and of large quiescent regions [8]). Otherwise individual vortex is so large that the particle never leaves it during the interaction time τ\tau. The motion would then be determined by interaction with a single vortex destroying the universality. This limits the study to moderate ReλRe_\lambda well below those in clouds where the statistics is very intermittent. Still we demonstrate that the results provide a useful reference point and some of them do generalize to high ReλRe_\lambda.

As far as the case of weak intermittency is concerned, the present work continues the study [9] which provided successful quantitative predictions for correlation dimension and Lyapunov exponents of particles in the Navier-Stokes turbulence. The predictions were confirmed by direct numerical simulations at Reλ=70Re_\lambda= 70 (claimed in [10] contradiction with some data is eliminated below). Here we provide the theory of the collision kernel and perform simulations at the same Reλ=70Re_\lambda= 70, that demonstrate full agreement with the theory. We derive angle-dependent radial distribution function (RDF) that gives probability of finding a pair of particles at a given vector separation. The RDF determines the collision kernel by providing probability of a pair at collision distance. The angle dependence of the RDF allows us to obtain vertical and horizontal dimensions of particles’ columns in space whose formation is the signature of the studied fast sedimentation limit, see Figure 1. In the remainder of the Introduction we provide outlook (currently lacking in the literature) at the existing rigorous theories and describe the paper’s organization.

DNS snapshots of particles’ distributions in space at fixed $St = 1$, $Re_\lambda = 70$ and increasing $Fr$

Figure 1. DNS snapshots of particles’ distributions in space at fixed St=1St = 1, Reλ=70Re_\lambda= 70 and increasing FrFr of (a) Fr=0.0125Fr = 0.0125, (b) Fr=0.0167Fr = 0.0167, (c) Fr=0.0333Fr = 0.0333, and (d) Fr=0.05Fr = 0.05. The distributions are multifractal. The positive difference of 3 and the fractal dimension increases linearly with FrFr in the considered range. The fractal structure is anisotropic. Pronounced columns of particles predominantly point in the direction of gravity, that was slightly slanted from the yy-axis to avoid numerical artifact due to periodicity. We demonstrate in the text that characteristic dimensions of the columns are gτ2g\tau^2 in the direction of gravity and ℓc∼10η\ell_c \sim10\eta in the perpendicular direction. This is confirmed by the Figure. Indeed St=1St = 1 and Kolmogorov scale η=0.024\eta= 0.024, imply gτ2=η/Frg\tau^2 = \eta/Fr. Thus the theory predicts that characteristic longest dimension of columns is 0.024/Fr0.024/Fr and the characteristic shortest dimension is 0.240.24. This agrees with the snapshots where columns become more convoluted at increasing FrFr. There is no apparent correlation with the vorticity component in the perpendicular direction, which would hold for the centrifugal mechanism of clustering.

Theory at weak gravity and small inertia and its breakdown in clouds

Until [9], a complete rigorous theory for inertial particles in the NST existed only at negligible gravity and small but finite inertia [4, 11]. The magnitude of inertia is measured by the dimensionless Stokes number StSt which is the ratio of τ\tau and the typical turnover-time of the viscous scale vortices, the Kolmogorov time τK\tau_K, see [8]. The inertia is usually considered to be small at St≪1St \ll1 where the particle velocity would relax quickly to the local velocity of the flow and the particles would nearly trace the flow [12]. However we demonstrate in this subsection that intermittency of turbulence quite certainly makes the effect of inertia in clouds large, even at St≪1St \ll1.

The work [4] introduced a picture of motion of particles in space that holds at St≪1St \ll1 independently of the Reynolds number and, thus, would hold also in the clouds. This picture can be understood by using the concept of the local Stokes number, similar to that of the local Reynolds number [8]. This number is necessary in order to describe strong contrasts in the strength of particle-vortex interaction throughout a high Reynolds number flow. The contrasts can be described via the local energy dissipation rate ϵ\epsilon, defined as the product of the kinematic viscosity ν\nu and the square of velocity gradients. The rate undergoes strong intermittent fluctuations in clouds, see [1] and references therein. We observe that the Kolmogorov time, defining StSt, is given by τK=ν/ϵ0\tau_K = \sqrt{\nu/\epsilon_0} where ϵ0\epsilon_0 is the mean energy dissipation. Correspondingly the local relevance of inertia can be characterised by the local Stokes number τϵ/ν\tau\sqrt{\epsilon/\nu}. If the local Stokes number is not small, then strong interaction of particles with the local vortices occurs in that region. Due to the intermittency of turbulence, the magnitude of fluctuations of the local Stokes number depends on ReλRe_\lambda, implying dependence of particles’ statistics on both StSt and ReλRe_\lambda.

At St≪1St \ll1, for the more probable events turbulent vortices are slow, τϵ/ν≪1\tau\sqrt{\epsilon/\nu} \ll1. The particles’ motion in these vortices is smooth and ordered in space so that there is no crossing of trajectories of different particles. A flow of particles v(t,x)\mathbf{v}(t,\mathbf{x}) can be introduced [12], providing the changes of their coordinates x(t)\mathbf{x}(t) as x˙=v(t,x(t))\dot{\mathbf{x}} = \mathbf{v}(t,\mathbf{x}(t)). However for fast vortices with time-scale ≲τ\lesssim\tau, characterized by non-small local Stokes number, the turbulent driving creates jets of particles that separate from the turbulent flow, as in a sling, and cause the trajectories’ crossing [4]. This “sling effect” was observed experimentally [17] and at any ReλRe_\lambda, including those in clouds, it is confined to well isolated small regions of fast vortices. Increase of ReλRe_\lambda, at least in the high-ReλRe_\lambda limit, would make regions of the sling effect even more rare in space because the regions of quiescent turbulence increase in size with the Reynolds number [8]. This creates significant difficulties in experimental and numerical measurements which must have resolution that increases with ReλRe_\lambda.

The above picture implies that space averages, such as the average rate of droplets’ collisions, can be found as sums of contributions of flow and sling regions (termed below ”flow contribution” and ”sling contribution”). The contributions of these regions into the collision rate were calculated in [4]. It was demonstrated that, despite that the regions of the sling effect are rare in space, their contribution into the collision kernel can be significant because they create optimal conditions for collisions [4, 18].

In contrast with the above qualitative picture, that is valid at any Reynolds number, quantitative predictions of [4, 11] break down at increasing ReλRe_\lambda. The predictions for the RDF rely on the assumption that at St≪1St \ll1 the RDF is determined by vortices whose turnover time, similarly to the most probable turnover time τK\tau_K, is much smaller than τ\tau. Making that assumption, the theory implies that distribution of inertial particles in the turbulent flow is multifractal. Multifractality manifests itself in the RDF that obeys at small distances a power-law with an exponent −α-\alpha. Here $\alpha is positive - the RDF diverges at zero separation because the particles’ concentration is singular. An explicit formula for α\alpha via a high order moment of the turbulent velocity gradients is provided in [4, 11]. Application of the standard phenomenology of turbulence to that formula gives that at large ReλRe_\lambda the magnitude of the exponent is proportional to the product of St2St^2 and a positive power of Reynolds number. This implies that in the limit of large ReλRe_\lambda, however small StSt is, the exponent, calculated under the described assumptions, is larger than 3. This however implies that the RDF becomes non-integrable at the origin which contradicts finiteness of the total number of particles (the number’s second moment is proportional to the integral of the RDF over distance).

The reason for the described contradiction is that the assumption of the theory that the RDF is determined by vortices with turnover time smaller than τ\tau becomes inconsistent at large ReλRe_\lambda. In fact, the theory predicts its own breakdown via the formula for α\alpha. Intermittency of turbulence implies appreciable presence in the flow of vortices whose turnover time is much smaller than τK\tau_K. More precisely the probability that the turnover time is smaller than τK\tau_K by a power of the Reynolds number is appreciable [8], cf. the recent [13]. As a result, in the limit of large Reynolds numbers the gradients that determine α\alpha grow with ReλRe_\lambda according to a non-trivial power-law [16]. This implies that the time-scale of the vortices that determine α\alpha decreases with ReλRe_\lambda as τKReλ−q\tau_K Re_\lambda^{-q} with q>0q>0 (according to the standard phenomenology time scales as the inverse gradient). Thus, considering increase of ReλRe_\lambda at a fixed small StSt, at large enough ReλRe_\lambda the time-scale of the relevant vortices becomes of order τ\tau. We designate the corresponding threshold Reynolds number by Reλ∗Re_\lambda^*. This number is not sharp and is determined by order of magnitude only via the asymptotic equality τK(Reλ∗)−q∼τ\tau_K (Re_\lambda^*)^{-q}\sim\tau. Thus at Reλ≳Reλ∗Re_\lambda\gtrsim Re_\lambda^* the assumption that the relevant vortices have time-scale larger than τ\tau is inconsistent. We conclude that the RDF is determined at Reλ≳Reλ∗Re_\lambda\gtrsim Re_\lambda^* by vortices whose characteristic turnover time is of order τ\tau or less. Anomalously strong, short-lived vortices become relevant to the RDF at higher ReλRe_\lambda because intermittency makes their probability appreciable. We also conclude that an approach other than that of [4, 11] must be devised at Reλ≳Reλ∗Re_\lambda\gtrsim Re_\lambda^*.

The above considerations demonstrate that the actual validity condition of the small inertia theory of [4, 11] is St≪Reλ−qSt \ll Re_\lambda^{-q} and not St≪1St \ll1, cf. [14, 15]. The value of qq can hardly be predicted and must be fixed numerically. Thus the theory of [4, 11] is a so-called asymptotic theory at small StSt i.e. it holds at St→0St \to0 when ReλRe_\lambda is fixed.

The above demonstration of limitations of the theory of [4] due to intermittency is performed by using the laws holding in the limit of large ReλRe_\lambda. This does not signify that intermittency becomes relevant only at Reλ→∞Re_\lambda\to\infty. Simulations of [14] of the particles’ motion in the NST demonstrated that for St∼0.1St \sim0.1 the small inertia theory of the RDF of [4] applies accurately at the moderate Reλ=21Re_\lambda=21. However, considering higher ReλRe_\lambda, already at Reλ=47Re_\lambda=47, that theory fails. We see that smallness of StSt does not guarantee the smallness of inertial effects already at rather moderate ReλRe_\lambda. Despite that the criterion of validity of the theory, St≪Reλ−qSt \ll Re_\lambda^{-q}, derived at Reλ→∞Re_\lambda\to\infty, cannot be used at ReλRe_\lambda of [14], we can use it to get very rough idea of magnitude of qq. We find that q∼0.5q\sim0.5 i.e. it is a number of order one. Using q=0.5q=0.5 we find that for Reλ∼104Re_\lambda\sim10^4, that can hold in the clouds, the effects of inertia on the RDF are appreciable at StSt as small as 0.01 (which covers the whole range of relevant droplet sizes, see Sec. II). More precise considerations demand extensive future numerical work.

The above conclusion transfers to the collision kernel which, under the assumption of small inertia, is given by a sum of two terms [4]. One term describes the contribution of vortices whose turnover time is much smaller than τ\tau. That term is proportional to the RDF and, as we saw, it becomes determined by vortices with turnover time of order τ\tau at higher ReλRe_\lambda. The other term in the collision kernel describes the contribution of the sling effect which by definition is due to vortices with time-scale τ\tau or smaller. Thus in the limit of high ReλRe_\lambda the collision kernel is due to vortices with turnover time τ\tau or smaller.

We summarize how increase of ReλRe_\lambda results in the breakdown of the theory of [4]. The theory demonstrates that the clustering rate, holding in the regions of the smooth motion of the droplets, is determined by vortices whose characteristic time-scale decreases with ReλRe_\lambda as a power-law. The theory holds at moderate ReλRe_\lambda where this time-scale is much larger than τ\tau as necessary for the self-consistency of the assumption of smooth motion. When ReλRe_\lambda increases, faster vortices with decreasing turnover time become relevant, until their time-scale becomes comparable with τ\tau at some Reλ∗Re_\lambda^*, cf. [14]. When this happens the separation of the total rate of collisions into the contributions of the regions of the smooth motion and of the sling effect, that was used in [4] for the calculations, breaks down. Both contributions are determined by ”resonant” vortices with lifetime of order τ\tau. The corresponding changes in the theory will be published elsewhere.

Fast sedimentation theory

Applicability of the above theory to clouds is limited by three factors. The first factor is the Stokes number which is not necessarily small for the droplets taking part in the rain formation. The second factor is gravity, which in clouds is strong and not weak. The magnitude of gravity can be measured by the Froude number FrFr given by the ratio of the typical (Lagrangian) acceleration of the fluid particles ϵ03/4/ν1/4\epsilon_0^{3/4}/\nu^{1/4} and gravitational acceleration gg, i.e. Fr=ε03/4/(gν1/4)Fr = \varepsilon_0^{3/4}/(g\nu^{1/4}). Thus, [4] applies at Fr→∞Fr \to\infty when gravity is negligible. In contrast, in the clouds even rather strong cloud turbulence with ε0=2000cm2/s3\varepsilon_0 = 2000\mathrm{cm}^2/\mathrm{s}^3 gives Fr=0.5Fr = 0.5. For typical turbulence with smaller ε0\varepsilon_0, the Froude number is yet smaller. The last limiting factor is the high ReλRe_\lambda of the clouds, as described above. The two former limitations were overcome in [9] who constructed the theory that holds at strong gravity, Fr→0Fr \to0, and any StSt including St≫1St \gg1. However this theory, similarly to that considered in the previous subsection, is also asymptotic and has the validity condition Fr≪Reλ−q′Fr \ll Re_\lambda^{-q'} with q′>0q' > 0, as will be discussed in this work. Thus how small FrFr must be, for the theory to be valid at a given ReλRe_\lambda, is not known. It was found in numerical simulations of [9] that at Reλ=70Re_\lambda= 70 the validity of the theory demands Fr≤0.033Fr \le0.033 which implies quite weak typical accelerations ε03/4/ν1/4\varepsilon_0^{3/4}/\nu^{1/4} as compared with the gravity. The theory was also shown to apply reasonably well at Fr=0.05Fr = 0.05.

The main limitation of [9] is again the Reynolds number - the intermittency increases the fluctuations of the gradients and destroys the approximations made. Nevertheless this theory, in comparison with that of [4], incorporates the effects of the gravitational sedimentation of the droplets, that are strong in precipitating clouds, giving us qualitative insight and the possibility of consistent interpolation. It seems reasonable, given that clouds are indeed characterized by small FrFr, to make the small FrFr theory a starting point for approaching collisions in clouds. It is to this theory that this paper is devoted. Here we provide more detailed predictions than in [9], confirm them numerically and describe how the theory breaks down at higher Reynolds number.

Before continuing the development of the approach of [9], a controversy must be faced. The theoretical formula of [9] for the correlation codimension α\alpha, studied in detail below, was tested numerically in [10] at Reλ=398Re_\lambda= 398. The comparison was done for Fr=0.052Fr = 0.052 and a range of StSt. They observed that while [9] describe correctly the independence of α\alpha of StSt for St≥1St \ge1, quantitatively the predictions are wrong by about 50 per cent. It is unfortunate that the performed comparison contained two mistakes. The equation (4.21) of [10] which the authors considered as the prediction of [9] for α\alpha, is factor of 2 smaller than the actual prediction made in [9]. If the correct formula is used then the discrepancy is about 30 and not 50 per cent. This seems to be as much as one can hope for, because [10] use the predictions outside the domain of validity of the theory, which is α≪1\alpha\ll1, cf. [11]. In fact, the result of [10] completely agrees with the simulations of [9] that demonstrated that at Fr=0.05Fr = 0.05 there are significant deviations from the theory. The inequalities α≪1\alpha\ll1 and Fr≪1Fr \ll1 differ much because of the large numerical factor: we have α≈13Fr\alpha\approx13Fr, see below and Sec. IV.

References

  1. [1]B. J. Devenish, P. Bartello, J.-L. Brenguier, L. R. Collins, W. W. Grabowski, R. H. A. IJzermans, S. P. Malinowski, M. W. Reeks, J. C. Vassilicos, L.-P. Wang, and Z. Warhaft, Droplet growth in warm turbulent clouds, Q. J. R. Meteorol. Soc. 138, 1401 (2012).DOI
  2. [2]A. P. Khain and M. Pinsky, Physical Processes in Clouds and Cloud Modeling, (Cambridge University Press, 2018).DOI
  3. [3]H. R. Ruppacher and J. D. Klet, Microphysics of Clouds and Precipitation, (Kluwer Academic Publishers, New York, 2004).
  4. [4]G. Falkovich, A. Fouxon and M. Stepanov, Acceleration of rain initiation by cloud turbulence, Nature 419, 151-154 (2002).pubmed.ncbi.nlm.nih.gov/12226661
  5. [5]M. Pinsky, A. Khain, and H. Krugliak, Collisions of cloud droplets in a turbulent flow. Part V: Application of detailed tables of turbulent collision rate enhancement to simulation of droplet spectra evolution, J. Atmos. Sci. 65, 357 (2008).
  6. [6]N. Benmoshe, A. Khain, M. Pinsky, and A. Pokrovsky, Turbulent effects on cloud microstructure and precipitation of deep convective clouds as seen from simulations with a 2-D spectral microphysics cloud model, J. Geophys. Res. 117, D06220 (2012).
  7. [7]M. L. Larsen, R. A. Shaw, A. B. Kostinski, and S. Glienke, Fine-scale droplet clustering in atmospheric clouds: 3D radial distribution function from airborne digital holography, Phys. Rev. Lett. 121, 204501 (2018).DOI
  8. [8]U. Frisch, Turbulence: The Legacy of A. N. Kolmogorov, (Cambridge University Press, New York, 1995).
  9. [9]I. Fouxon, Y. Park, R. Harduf, and C. Lee, Inhomogeneous distribution of water droplets in cloud turbulence, Phys. Rev. E 92, 033001 (2015).DOI
  10. [10]P. J. Ireland, A. D. Bragg, and L. R. Collins, The effect of Reynolds number on inertial particle dynamics in isotropic turbulence. Part 2. Simulations with gravitational effects, J. Fluid Mech. 796, 659 (2016).
  11. [11]I. Fouxon, Distribution of particles and bubbles in turbulence at a small Stokes number, Phys. Rev. Lett., 108, 134502 (2012).pubmed.ncbi.nlm.nih.gov/22540704
  12. [12]M. R. Maxey, The gravitational settling of aerosol particles in homogeneous turbulence and random flow fields, J. Fluid Mech. 174, 441 (1987).DOI
  13. [13]D. Briaria, A. Pumir, E. Bodenschatz, and P. K. Yeung, Extreme velocity gradients in turbulent flows, New J. Phys. 21, 043004 (2019).
  14. [14]G. Falkovich and A. Pumir, Intermittent distribution of heavy particles in a turbulent flow, Phys. Fluids 16, L47 (2004).DOI
  15. [15]G. Falkovich and A. Pumir, Sling effect in collisions of water droplets in turbulent clouds, J. Atmos. Sci. 64, 4497 (2007).DOI
  16. [16]J. Schumacher, J. D. Scheel, D. Krasnov, D. A. Donzis, V. Yakhot and K. R. Sreenivasan, Small-scale universality in fluid turbulence, Proc. Nat. Acad. Soc. 111, 10961 (2014).DOI
  17. [17]G. P. Bewley, E. W. Saw, and E. Bodenschatz, Observation of the sling effect, New J. Phys. 15, 083051 (2013).
  18. [18]M. Vosskuhle, A. Pumir, E. Leveque, and M. Wilkinson, Prevalence of the sling effect for enhancing collision rates in turbulent suspensions, J. Fluid Mech. 749, 841 (2014).
  19. [19]Y. Park and C. Lee, Gravity-driven clustering of inertial particles in turbulence, Phys. Rev. E 89, 061004(R) (2014).DOI
  20. [20]R. Harduf, Master thesis, undef supervision of I. Fouxon, submitted and approved in 2011, Tel-Aviv University.
  21. [21]K. Gustavsson, S. Vajedi, and B. Mehlig, Clustering of particles falling in a turbulent flow, Phys. Rev. Lett. 112, 214501 (2014).DOI
  22. [22]J. Bec, H. Homann, and S. S. Ray, Gravity-driven enhancement of heavy particle clustering in turbulent flow, Phys. Rev. Lett. 112, 184501 (2014).DOI
  23. [23]U. Frisch and J. Bec, Burgulence, in New trends in turbulence (Les Houches, 2000, Eds. M. Lesieur, A. Yaglom, F. David).
  24. [24]I. Fouxon and P. Horvai, Separation of heavy particles in turbulence, Phys. Rev. Lett. 100, 040601 (2008).pubmed.ncbi.nlm.nih.gov/18352248
  25. [26]H. G. E. Hentschel and I. Procaccia, The infinite number of generalized dimensions of fractals and strange attractors, Phys. D 8, 435 (1983).DOI
  26. [27]P. Saffman and J. S. Turner, On the collision of drops in turbulent clouds, J. Fluid Mech., 1, 16 (1956).DOI
  27. [29]J. Bec, Fractal clustering of inertial particles in random flows, Phys. Fluids 15, L81 (2003).DOI
  28. [30]J. Bec, M. Cencini, and R. Hillerbrand, Heavy particles in incompressible flows: the large Stokes number asymptotics, Phys. D 226, 11 (2007).DOI
  29. [32]N. Mordant, A. M. Crawford, and E. Bodenschatz, Experimental Lagrangian acceleration probability density function measurement, Phys. D 193, 245 (2004).DOI
  30. [33]C. Meneveau and K. R. Sreenivasan, The multifractal nature of turbulent energy dissipation, J. Fluid Mech. 224, 429 (1991).DOI
  31. [35]J. L. Kaplan and J. A. Yorke, in Functional Differential Equations and Approximation of Fixed Points, ed. by H.-O. Peitgen and H.-O. Walther, 730, 204 (Springer, Berlin Heidelberg, 1979).
  32. [36]G. Falkovich, K. Gawȩdzki, and M. Vergassola, Particles and fields in fluid turbulence, Rev. Mod. Phys. 73, 913 (2001).DOI
  33. [38]E. Balkovsky, G. Falkovich, and A. Fouxon, Intermittent distribution of inertial particles in turbulent flows, arxiv:9912027, see also E. Balkovsky, G. Falkovich and A. Fouxon, Phys. Rev. Lett. 86, 2790 (2001).pubmed.ncbi.nlm.nih.gov/11290040
  34. [39]O. Ayala, B. Rosa, and L.-P. Wang, Effects of turbulence on the geometric collision rate of sedimenting droplets. Part 2. Theory and parameterization, New J. Phys. 10, 075016 (2008).
  35. [40]P. L. Johnson and C. Meneveau, Large-deviation joint statistics of the finite-time Lyapunov spectrum in isotropic turbulence, Phys. Fluids 27, 085110 (2015).DOI
  36. [41]G. K. Batchelor and J. T. Green, The hydrodynamic interaction of two small freely-moving spheres in a linear flow field, J. Fluid Mech. 56, 375 (1972).DOI
  37. [42]S. Chen, M. K. Yau, and P. Bartello, Turbulence effects of collision efficiency and broadening of droplet size distribution in cumulus clouds, J. Atmos. Sci. 75, 203 (2018).DOI
  38. [43]M. A. Vyavuz, R. P. J. Kun, G. J. F. Van Heijst, and H. J. H. Clercx, Extreme small-scale clustering of droplets in turbulence driven by hydrodynamic interactions, Phys. Rev. Lett. 20, 244504 (2018).
  39. [44]A. Crisanti, M. H. Jensen, A. Vulpiani, and G. Paladin, Intermittency and predictability in turbulence, Phys. Rev. Lett. 70, 166 (1993).pubmed.ncbi.nlm.nih.gov/10053719
  40. [45]J. Bec, L. Biferale, G. Boffetta, M. Cencini, S. Musacchio, and F. Toschi, Lyapunov exponents of heavy particles in turbulence, Phys. Fluids 18, 091702 (2006).DOI
  41. [47]K. R. Sreenivasan and R. A. Antonia, The phenomenology of small-scale turbulence, Ann. Rev. Fluid Mech. 29, 435 (1997).DOI
  42. [48]H. Siebert, K. Lehmann, and M. Wendisch, Observations of small-scale turbulence and energy dissipation rates in the cloudy boundary layer, J. Atmos. Sci.63, 1451 (2006).DOI
  43. [53]T. Ishihara, T. Gotoh, and Y. Kaneda, Study of high Reynolds number isotropic turbulence by direct numerical simulation, Ann. Rev. Fluid Mech. 41, 165 (2009).DOI
  44. [54]J. Bec, A. Celani, M. Cencini, and S. Musacchio, Clustering and collisions of heavy particles in random smooth flows, Phys. Fluids 17, 073301 (2005).DOI
  45. [55]O. Ayala, B. Rosa, L.-P. Wang, and W. W. Grabowski, Effects of turbulence on the geometric collision rate of sedimenting droplets. Part 1. Results from direct numerical simulation, New J. Phys. 10, 075015 (2008).DOI
  46. [56]S. Sundaram and L. R. Collins, Collision statistics in an isotropic particle-laden turbulent suspension. Part 1. Direct numerical simulations, J. Fluid Mech. 335, 75 (1997).DOI

Paper details

vlnc:W_dffj27jq
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 3

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents