betaFeedback

WorksPhysics and Astronomy

Spatio-temporal fluctuations of interscale and interspace energy transfer dynamics in homogeneous turbulence

Abstract

We study fluctuations of all co-existing energy exchange/transfer/transport processes in stationary periodic turbulence including those which average to zero and are not present in average cascade theories. We use a Helmholtz decomposition of accelerations which leads to a decomposition of all terms in the Kármán-Howarth-Monin-Hill (KHMH) equation (scale-by-scale two-point energy balance) causing it to break into two energy balances, one resulting from the integrated two-point vorticity equation and the other from the integrated two-point pressure equation. The various two-point acceleration terms in the Navier-Stokes difference (NSD) equation for the dynamics of two-point velocity differences have similar alignment tendencies with the two-point velocity difference, implying similar characteristics for the NSD and KHMH equations. We introduce the two-point sweeping concept and show how it articulates with the fluctuating interscale energy transfer as the solenoidal part of the interscale transfer rate does not fluctuate with turbulence dissipation at any scale above the Taylor length but with the sum of the time-derivative and the solenoidal interspace transport rate terms. The pressure fluctuations play an important role in the interscale and interspace turbulence transfer/transport dynamics as the irrotational part of the interscale transfer rate is equal to the irrotational part of the interspace transfer rate and is balanced by two-point fluctuating pressure work. We also study the homogeneous/inhomogeneous decomposition of interscale transfer. The statistics of the latter are skewed towards forward cascade events whereas the statistics of the former are not. We also report statistics conditioned on intense forward/backward interscale transfer events.

Introduction

Modeling of turbulence dissipation is a cornerstone of one-point turbulent flow prediction methods based on the Reynolds Averaged Navier Stokes (RANS) equations such as the widely used k−εk-\varepsilon and the k−ωk-\omega models [see 29, 19] and also of two-point turbulence flow prediction methods based on filtered Navier Stokes equations, namely Large Eddy Simulations (LES) [see 29, 30]. The mechanism of turbulence dissipation away from walls is the turbulence cascade [29 Vassilicos] 2015). Indeed, without cascade, the larger turbulent eddies which constitute the most energetic part of the turbulence, would take much longer to lose their energy than just one or a few eddy turnover times. The physical understanding of this cascade which, to this day, has underpinned these prediction methods is based on Kolmogorov’s average cascade in statistically homogeneous and stationary turbulence. Notwithstanding recent advances which have shown that the turbulence dissipation and cascade are different from Kolmogorov’s both in non-stationary (see e.g. [38, 11, 33]) and in non-homogeneous turbulence ([4, 5]), Kolmogorov’s cascade is in fact valid only as an average cascade even in homogeneous stationary turbulence. Turbulence has been known to be intermittent since the late 1940s (see [9] and references therein), and this intermittency has mainly been taken into account as structure function exponent corrections to Kolmogorov’s average picture. However, Yasuda & Vassilicos [42] studied intermittent fluctuations without reference to structure function exponents which require high Reynolds numbers to be well defined and to be predicted from Kolmogorov’s theory or various intermittency-accounting variants of this theory (see [9] and references therein). They concentrated their attention on the actual fundamental basis of Kolmogorov’s theory which is scale-by-scale equilibrium for statistically homogeneous and stationary turbulence, and not on the theory’s structure function and energy spectrum scaling consequences. The scale-by-scale equilibrium implied by statistical homogeneity and stationarity is that the average interscale turbulence energy transfer rate is balanced by nothing more than average scale-by-scale viscous diffusion rate, average turbulence dissipation rate and average energy input rate by a stirring force, irrespective of Reynolds number (except that the Reynolds number needs to be large enough for the presence of random fluctuations). It is most natural for a study of intermittency to start with the fluctuations around this balance, which means that along with the fluctuations of interscale transfer, dissipation, diffusion and energy input, all other fluctuating turbulent energy change rates need to be taken into account as well even if their spatio-temporal average is zero in statistically stationary homogeneous turbulence. The intermittency corrections to Kolmogorov’s average cascade theory which have been developed since the 1960s (e.g. see [9]; [32]) are often based on the intermittent fluctuations of the local (in space and time) turbulence dissipation rate, yet Yasuda & Vassilicos [42] demonstrated that these dissipation fluctuations are much less intense than the fluctuations of other turbulent energy change rates such as the non-linear interspace energy transfer rate (which is a scale-by-scale rate of turbulent transport in physical space), the fluctuating work resulting from the correlation of the fluctuating pressure gradient with the fluctuating velocity and the time-derivative of the scale-by-scale turbulent kinetic energy. Yasuda & Vassilicos [42] made these observations using Direct Numerical Simulations (DNS) of statistically stationary periodic turbulence at low to moderate Taylor length-based Reynolds numbers from about 80 to 170. Even though their Reynolds numbers were not high enough to test the high Reynolds number scaling consequences of Kolmogorov’s theory, they observed an energy spectrum with a near-decade power law range where the power law exponent was not too far from Kolmogorov’s −5/3-5/3. However, they did not observe a significant range of scales where the scale-by-scale equilibrium reduces to a scale-by-scale balance between average interscale turbulence energy transfer rate and average turbulence dissipation as predicted by the Kolmogorov theory for statistically stationary homogeneous turbulence at asymptotically high Reynolds numbers. This high Reynolds number scale-by-scale equilibrium is the trademark of the Kolmogorov average cascade and is typically not put in question by existing intermittency corrections to Kolmogorov’s theory (e.g. see [9]).

Given the low to moderate Reynolds numbers of the DNS used by Yasuda & Vassilicos (2018), their observations concern interscale turbulence energy transfers more than the turbulence cascade per se if the concept of turbulence cascade is taken to have meaning only at very large Reynolds numbers. They demonstrated that an interscale transfer picture appears that is radically different from Kolmogorov’s if the average is lifted and all spatio-temporal intermittent fluctuations are taken into account. This different picture involves highly fluctuating processes which vanish on average in statistically stationary and homogeneous turbulence and are not taken into account by the Kolmogorov theory for that very reason. We stress once more that Yasuda & Vassilicos [42] made this demonstration in statistically homogeneous and stationary turbulence, the very type of turbulence where Kolmogorov’s theory has been designed for.

It is hard to imagine that the complex turbulence energy transfer picture educed by the DNS of Yasuda & Vassilicos [42] does not survive at asymptotically high Reynolds numbers because it is known that the small-scale turbulence becomes increasingly intermittent with increasing Reynolds number (e.g. see Frisch (1995); Sreenivasan & Antonia (1997)). A DNS study at higher Reynolds numbers is nevertheless needed to ascertain this point. However, this is not the study proposed in this paper. In this paper our aim is to gain deeper insight into the fluctuating energy transfer picture revealed by the DNS of Yasuda & Vassilicos [42] and we do this in terms of Helmholtz decomposed solenoidal and irrotational acceleration fields. Given that the computational cost involved in this Helmholtz decomposition is high (see following two sections) it is not possible for us to carry out our study for a variety of increasing Reynolds numbers and thereby combine it with a Reynolds number dependence study. We therefore limit ourselves to Reynolds numbers comparable to those of Yasuda & Vassilicos [42].

The radically different turbulence energy transfer picture which appears when all intermittent turbulence fluctuations are taken into account exhibits correlations between fluctuations of different processes: in particular, the fluctuating pressure-velocity term mentioned above is correlated with the interscale energy transfer rate, and the time derivative of the turbulent kinetic energy below a certain two-point length rr is correlated with the inter-space energy transport rate at the same length rr. Yasuda & Vassilicos [42] explained the former correlation as resulting from the link between non-linearity and non-locality (via the pressure field) and the latter correlation as reflecting the passive sweeping of small turbulent eddies by large ones (Tennekes [36]). However, this sweeping (also termed “random Taylor hypothesis”) has been studied by reference to the one-point incompressible Navier-Stokes equation [e.g. 36, 37] rather than the two-point Kármán-Howarth-Monin-Hill (KHMH) equation, used by Yasuda & Vassilicos [42] in their study of the fluctuating turbulence cascade. The KHMH equation is a scale-by-scale energy budget local in space and time, directly derived from the incompressible Navier-Stokes equations for the instantaneous velocity field (see Hill [14]) without decomposition (e.g. Reynolds decomposition), without averages (e.g. Reynolds averages), and without any assumption made about the turbulent flow (e.g. homogeneity, isotropy, etc.). The initial trigger of the present paper is to substantiate the claim of Yasuda & Vassilicos [42] concerning correlations being caused by random sweeping by translating the sweeping analysis of Tsinober et al. [37] to the KHMH equation. It is in doing so that we espouse the Helmholtz decomposition which Tsinober et al. [37] introduced for the analysis of the acceleration field. We apply it to the two-point Navier-Stokes difference (NSD) equation (which is the equation governing the dynamics of two-point velocity differences) and the KHMH equation which derives from it. This decomposition into solenoidal and irrotational terms breaks the Navier-Stokes equation into two equations, one being the irrotational balance between non-linearity and non-locality (pressure) and the other being the solenoidal balance between local unsteadiness and advection which encapsulates the sweeping. With this decomposition we substantiate all the correlations observed by Yasuda & Vassilicos [42] between different KHMH terms representing different energy change processes, not only the ones caused by sweeping. In fact, we educe the relation between interspace turbulence energy transfer/transport and two-point sweeping (i.e. the random Taylor hypothesis that we generalise to two-point statistics), and we extend the correlation study to solenoidal and irrotational sub-terms of the KHMH equation which leads to even stronger correlations than those found by Yasuda & Vassilicos [42]. This approach sheds some light on the way that two-point sweeping and interscale energy transfer relate to each other. We then ask whether the scale-by-scale equilibrium which is at the basis of Kolmogorov’s theory and which disappears when the average is lifted does nevertheless exist locally at relatively high energy transfer events, a question which leads us to consider whether two-point sweeping also holds at such events. Finally, we study the recently introduced decomposition (Alves Portela et al. [1]) of the interscale transfer rate into a homogeneous and an inhomogeneous interscale transfer component. We analyse their fluctuations and the correlations of these fluctuations, both unconditionally and conditionally on relatively rare intense interscale transfer events.

In the following section we introduce our direct numerical simulations (DNSs) of forced periodic turbulence. Subsection 3.1 is a reminder of the application of this decomposition to the one-point Navier-Stokes equation by Tsinober et al. [37]. In this sub-section we also validate our DNS by retrieving the conclusions of Tsinober et al. [37] on sweeping and by comparing our DNS results on one-point acceleration dynamics to theirs. In subsections 3.2–3.3 we apply the Helmholtz decomposition to the two-point NSD equation for the case of homogeneous/periodic turbulence and in subsection 3.4 we derive from the Helmholtz decomposed Navier-Stokes difference equations corresponding KHMH equations. Subsection 3.4 formalises the connection between the NS and KHMH dynamics, clarifies under which conditions a link exists between NS and KHMH dynamics and provides results on scale and Reynolds number dependencies of the KHMH dynamics. By considering the NSD dynamics in terms of solenoidal and irrotational dynamics, we derive two new KHMH equations. In section 4 we use these two new KHMH equations to obtain new results on the fluctuating cascade dynamics across scales both unconditionally and conditionally on rare extreme interscale energy transfer events. In section 5 we analyse the inhomogeneous and homogeneous contributions to the interscale energy transfer rate. Finally, section 6 summarises our results.

DNS of body-forced period turbulence

Our study requires turbulence data from a turbulent flow where the Kolmogorov equilibrium theory for statistically homogeneous and stationary turbulence is applicable. We therefore follow Yasuda & Vassilicos [42] and perform Direct Numerical Simulations of body-forced periodic Navier-Stokes turbulence with the same pseudo-spectral code that they used. This code solves numerically the vorticity equation

∂ω∂t=∇x×(u×ω)+ν∇x2ω+∇x×f,(1)\frac{\partial\boldsymbol{\omega}}{\partial t} = \nabla_x \times\left(\mathbf{u} \times\boldsymbol{\omega}\right) + \nu\nabla_x^2 \boldsymbol{\omega} + \nabla_x \times\mathbf{f}, \tag*{(1)}

subjected to the continuity equation

∇x⋅u=0,(2)\nabla_x \cdot\mathbf{u} = 0, \tag*{(2)}

where u(x,t)\mathbf{u}(\mathbf{x},t), f(x,t)\mathbf{f}(\mathbf{x},t) and ω(x,t)\boldsymbol{\omega}(\mathbf{x},t) are the velocity, force and vorticity fields respectively and ν\nu is the kinematic viscosity. All fields are 2π2\pi-periodic in each one of the three orthogonal spatial coordinates x1x_1, x2x_2 and x3x_3, and x=(x1,x2,x3)\mathbf{x}=(x_1,x_2,x_3). The pseudo-spectral method is fully dealiased with a combination of phase-shifting and spherical truncation (Patterson & Orszag

NN⟨Reλ⟩t\langle Re_{\lambda}\rangle_tν/103\nu/10^3kmax⁡⟨η⟩tk_{\max}\langle\eta\rangle_t2π/⟨L⟩t2\pi/\langle L\rangle_t⟨λ⟩/⟨L⟩t\langle\lambda\rangle/\langle L\rangle_tTs/TT_s/TΔT/T\Delta T/T
2561121.801.885.63.5210.01
5121740.721.895.45.2270.12

caption: tbl:1: Specifications of the numerial simulations. NN denotes the number of grid points in each Cartesian coordinate, ReλRe_{\lambda} the Taylor-scale Reynolds number, ν\nu the kinematic viscosity, kmax⁡=2/3Nk_{\max}=\sqrt{2/3}N is the highest resolved wavenumber, η\eta and λ\lambda are, respectively, the Kolmogorov and Taylor lengths and ⟨…⟩t\langle\ldots\rangle_t denotes a time-average. LL is the integral lengths calculated from the three-dimensional energy spectrum E(k,t)E(k,t): L(t)=(3π/4)−1∫0∞k−1E(k,t) dk/K(t)L(t)=(3\pi/4)^{-1}\int_0^\infty k^{-1}E(k,t)\,dk/K(t) where K(t)K(t) is the kinetic energy per unit mass. TsT_s denotes the total sampling time over which converged statistics are calculated by sampling randomly in space-time, ΔT\Delta T denotes the time between samples and T≡⟨L⟩t/2/3⟨K⟩tT\equiv\langle L\rangle_t/\sqrt{2/3\langle K\rangle_t} is the turnover time.

:::

1971). The forcing method is a negative damping forcing (Linkmann & Morozov 2015; McComb et al. 2015b)

f^(k,t)=(ϵW2Kf)u^(k,t)for 0<∣k∣<kf,(3)\hat{\mathbf{f}}(\mathbf{k},t)=\left(\frac{\epsilon_W}{2K_f}\right)\hat{\mathbf{u}}(\mathbf{k},t) \qquad\text{for }0<|\mathbf{k}|<k_f, \tag*{(3)}
=0otherwise,(4)=0 \qquad\text{otherwise}, \tag*{(4)}

where f^(k,t)\hat{\mathbf{f}}(\mathbf{k},t) and u^(k,t)\hat{\mathbf{u}}(\mathbf{k},t) are the Fourier transforms of f(x,t)\mathbf{f}(\mathbf{x},t) and u(x,t)\mathbf{u}(\mathbf{x},t) respectively, kfk_f is the cutoff wavenumber, ϵW\epsilon_W is the energy input rate per unit mass and KfK_f is the kinetic energy per unit mass in the wavenumber band 0<∣k∣<kf0<|\mathbf{k}|<k_f. Note that this forcing is incompressible and has therefore no irrotational part. The addition of a potential, i.e. irrotational, term to the forcing would effectively just be subsumed into the pressure required to keep the flow incompressible.

We perform two DNS of forced periodic/homogeneous turbulence with forcing parameters ϵW=0.1\epsilon_W=0.1 and kf=2.5k_f=2.5 at both simulation sizes 2563256^3 and 5123512^3. Average statistics are given in table 1. For these two simulation sizes respectively, deviations around these averages are as follows: the standard deviations of LL are 0.007Lb0.007L_b and 0.006Lb0.006L_b (where Lb=2πL_b=2\pi) and the maximum LL values are 0.188Lb0.188L_b and 0.202Lb0.202L_b; the standard deviations of λ\lambda are 2.5%2.5\% and 3.7%3.7\% of ⟨λ⟩t\langle\lambda\rangle_t; and the standard deviation of kmax⁡ηk_{\max}\eta are 0.025 and 0.035.

McComb et al. (2015a) performed DNS with identical combinations of NN, ν\nu and forcing as in our simulations. The time-averaged Taylor-scale Reynolds numbers ⟨Reλ⟩t\langle Re_{\lambda}\rangle_t, the ratios of the box size to the time-averaged integral length 2π/⟨L⟩t2\pi/\langle L\rangle_t and the time-averaged Kolmogorov microscales ⟨η⟩t\langle\eta\rangle_t are all very similar (and ⟨…⟩t\langle\ldots\rangle_t denotes a time-average). This study reports slightly poorer small-scale resolution kmax⁡⟨η⟩tk_{\max}\langle\eta\rangle_t than ours due to their more severe spherical truncation.

We have also verified that the results do not significantly change when the flow is forced at small wavenumbers with an ABC forcing with A=B=CA=B=C [28]. In contrast to the negative damping forcing, this forcing is independent of time and of the velocity field and is also maximally helical as ∇x×f\nabla_{\mathbf{x}}\times\mathbf{f} is parallel to f\mathbf{f} [10]. The helicity input of the ABC forcing has been studied in the context of the energy cascade in terms of its effect on the dissipation coefficient in Linkmann [20].

Our Reynolds numbers are relatively limited due to the high computational expense of our NSD and KMHH post-processing (which is typically at least one order of magnitude more expensive than the DNS). We detail the computational expense of the post-processing once the relevant terms have been introduced in § 3.4.

In the following section we show how we apply the Helmholtz decomposition to the KMHH equation. We start in subsection 3.1 by applying this decomposition to the one-point Navier-Stokes equation following Tsinober et al. [37]. In this sub-section we also validate our DNS by retrieving the conclusions of Tsinober et al. [37], in particular on sweeping, and by comparing our DNS results on one-point acceleration dynamics to theirs. In subsections 3.2 and 3.3 we apply the Helmholtz decomposition to the two-point Navier-Stokes difference equation for the case of homogeneous/periodic turbulence and in subsection 3.4 we derive from the Helmholtz-decomposed Navier-Stokes difference equations corresponding KHMH equations.

Helmholtz decomposition of two-point Navier-Stokes dynamics and corresponding turbulent energy exchanges

Solenoidal and irrotational acceleration fluctuations

The Helmholtz decomposition states that a twice continuously differentiable 3D vector field q(x,t)\boldsymbol{q}(\boldsymbol{x},t) defined on a domain V⊆R3V \subseteq\mathbb{R}^{3} can be expressed as the sum of an irrotational vector field qI(x,t)\boldsymbol{q}_{I}(\boldsymbol{x},t) and a solenoidal vector field qS(x,t)\boldsymbol{q}_{S}(\boldsymbol{x},t) [13, 34, 3]

qI(x,t)=−∇xϕ(x,t),qS(x,t)=∇x×B(x,t),(5)\boldsymbol{q}_{I}(\boldsymbol{x},t)=-\nabla_{\boldsymbol{x}}\phi(\boldsymbol{x},t),\qquad\boldsymbol{q}_{S}(\boldsymbol{x},t)=\nabla_{\boldsymbol{x}}\times\boldsymbol{B}(\boldsymbol{x},t), \tag*{(5)}

where ϕ(x,t)\phi(\boldsymbol{x},t) is a scalar potential and B(x,t)\boldsymbol{B}(\boldsymbol{x},t) is a vector potential. The Helmholtz decomposition and its interpretation can be applied to any vector field q(x,t)\boldsymbol{q}(\boldsymbol{x},t) satisfying the above conditions, and Tsinober et al. [37] applied it to fluid accelerations and the incompressible Navier-Stokes equation.

The solenoidal and irrotational Navier-Stokes equations in homogeneous/periodic turbulence can be derived from the incompressible Navier-Stokes equation in Fourier space (see appendix § 7). After transformation back to physical space, one obtains

∂u∂t+(u⋅∇xu)T=ν∇x2u+fT,(6)\frac{\partial\boldsymbol{u}}{\partial t}+\left(\boldsymbol{u}\cdot\nabla_{\boldsymbol{x}}\boldsymbol{u}\right)^{T}=\nu\nabla_{\boldsymbol{x}}^{2}\boldsymbol{u}+\boldsymbol{f}^{T}, \tag*{(6)}
(u⋅∇xu)L=−1ρ∇xp+fL,(7)\left(\boldsymbol{u}\cdot\nabla_{\boldsymbol{x}}\boldsymbol{u}\right)^{L}=-\frac{1}{\rho}\nabla_{\boldsymbol{x}}p+\boldsymbol{f}^{L}, \tag*{(7)}

where superscripts LL and TT denote fields obtained from longitudinal and transverse parts of respective Fourier vector fields (see appendix § 7 for precise definitions and [29, 34]), p=p(x,t)p=p(\boldsymbol{x},t) is the pressure field and ρ\rho is the density. For any periodic vector field q\boldsymbol{q}, qL\boldsymbol{q}^{L} equals the irrotational field qI\boldsymbol{q}_{I} and qT\boldsymbol{q}^{T} equals the solenoidal field qS\boldsymbol{q}_{S} (see appendix § 7 and Stewart [34]). From equations (6)–(7), one arrives at Tsinober et al. [37]

∂u∂t+(u⋅∇xu)S=ν∇x2u+fS,(8)\frac{\partial\boldsymbol{u}}{\partial t}+\left(\boldsymbol{u}\cdot\nabla_{\boldsymbol{x}}\boldsymbol{u}\right)_{S}=\nu\nabla_{\boldsymbol{x}}^{2}\boldsymbol{u}+\boldsymbol{f}_{S}, \tag*{(8)}
(u⋅∇xu)I=−1ρ∇xp+fI,(9)\left(\boldsymbol{u}\cdot\nabla_{\boldsymbol{x}}\boldsymbol{u}\right)_{I}=-\frac{1}{\rho}\nabla_{\boldsymbol{x}}p+\boldsymbol{f}_{I}, \tag*{(9)}

which we refer to as Tsinober equations. (8) contains only solenoidal vector fields and (9) contains only irrotational vector fields. Note that in the case of an incompressible periodic velocity field, the velocity field is solenoidal, i.e. u=uS\boldsymbol{u}=\boldsymbol{u}_{S}. This follows immediately from the scalar potential ϕ\phi being the solution to ∇x2ϕ=0\nabla_{\boldsymbol{x}}^{2}\phi=0 with periodic boundary conditions for ∇xϕ\nabla_{\boldsymbol{x}}\phi, yielding ϕ=const\phi=\mathrm{const}.

In appendix C we show that (8) is the integrated vorticity equation and that (9) is the integrated Poisson equation for pressure. The procedure presented in appendix C for obtaining the Tsinober equations is also used in this same appendix to obtain generalised

References

  1. [1]Alves Portela, F., Papadakis, G. & Vassilicos, J. C. 2020 The role of coherent structures and inhomogeneity in near-field interscale turbulent energy transfers. J. Fluid Mech. 896, A16.
  2. [2]Bardina, J., Ferziger, J. & Reynolds, W. 1980 Improved subgrid-scale models for large-eddy simulation. In 13th Fluid and Plasma Dynamics Conference. AIAA.
  3. [3]Bhatia, H., Norgard, G., Pascucci, V. & Bremer, P. 2013 The Helmholtz-Hodge decomposition—a survey. IEEE Trans. Vis. Comput. Graph. 19 (8), 1386–1404.DOI
  4. [4]Chen, J. G., Cuvier, C., Foucaut, J.-M., Ostovan, Y. & Vassilicos, J. C. 2021 A turbulence dissipation inhomogeneity scaling in the wake of two side-by-side square prisms. J. Fluid Mech. 924, A4.
  5. [5]Chen, J. G. & Vassilicos, J. C. 2022 Scalings of scale-by-scale turbulence energy in non-homogeneous turbulence. J. Fluid Mech. 938, A7.
  6. [6]Chevillard, L., Roux, S. G., Lévêque, E., Mordant, N., Pinton, J.-F. & Arnèodo, A. 2005 Intermittency of velocity time increments in turbulence. Phys. Rev. Lett. 95 (6), 64501.
  7. [7]Cimarelli, A., Abbà, A. & Germano, M. 2019 General formalism for a reduced description and modelling of momentum and energy transfer in turbulence. J. Fluid Mech. 866, 865–896.
  8. [8]Dairay, Thibault, Lamballais, Eric, Laizet, Sylvain & Vassilicos, John Christos 2017 Numerical dissipation vs. subgrid-scale modelling for large eddy simulation. J. Comput. Phys. 337, 252–274.DOI
  9. [9]Frisch, U. 1995 Turbulence: the legacy of A. N. Kolmogorov. Cambridge University Press.
  10. [10]Galanti, B. & Tsinober, A. 2000 Self-amplification of the field of velocity derivatives in quasi-isotropic turbulence. Phys. Fluids 12 (12), 3097–3099.
  11. [11]Goto, S. & Vassilicos, J. C. 2016 Unsteady turbulence cascades. Phys. Rev. E 94 (5), 53108.pubmed.ncbi.nlm.nih.gov/27967192
  12. [12]Gulitski, G., Kholmyansky, M., Kinzelbach, W., Lüthi, B., Tsinober, A. & Yorish, S. 2007 Velocity and temperature derivatives in high-Reynolds-number turbulent flows in the atmospheric surface layer. Part 2. Accelerations and related matters. J. Fluid Mech. 589, 83–102.
  13. [13]Helmholtz, H. 1867 On Integrals of the hydrodynamical equations, which express vortex-motion. Lond. Edinb. Dubl. Phil. Mag. 33 (226), 485–512.
  14. [14]Hill, R. J. 2002 Exact second-order structure-function relationships. J. Fluid Mech. 468, 317–326.DOI
  15. [15]Hill, R. J. & Thordorsson, S. T. 1997 Experimental evaluation of acceleration correlations for locally isotropic turbulence. Phys. Rev. E 55 (2), 1600–1606.
  16. [19]Leschziner, M. 2016 Statistical turbulence modelling for fluid dynamics, demystified : an introductory text for graduate engineering students. Imperial College Press.
  17. [20]Linkmann, Moritz 2018 Effects of helicity on dissipation in homogeneous box turbulence. J. Fluid Mech. 856, 79–102.
  18. [21]Linkmann, Moritz, Buzzicotti, Michele & Biferale, Luca 2018 Multi-scale properties of large eddy simulations: correlations between resolved-scale velocity-field increments and subgrid-scale quantities. J. Turb. 19 (6), 493–527.
  19. [25]Monin, A. S, Yaglom, A. M. & Lumley, J. L. 1975 Statistical fluid mechanics, vol. 2. M.I.T. Press.
  20. [26]Moser, Robert D, Haering, Sigfried W & Yalla, Gopal R 2021 Statistical properties of subgrid-scale turbulence models. Annu. Rev. Fluid Mech. 53 (1), 255–286.DOI
  21. [27]Patterson, G. S. & Orszag, S. A. 1971 Spectral calculations of isotropic turbulence: efficient removal of aliasing interactions. Phys. Fluids 14 (11), 2538–2541.
  22. [28]Podvigina, O. & Pouquet, A. 1994 On the non-linear stability of the 1:1:1 ABC flow. Physica D 75 (4), 471–508.
  23. [29]Pope, S. B. 2000 Turbulent flows. Cambridge University Press.DOI
  24. [30]Sagaut, P. 2000 Large eddy simulation for incompressible flows: an introduction, 1st edn. Springer.
  25. [31]Sprössig, W. 2010 On Helmholtz decompositions and their generalizations—an overview. Math. Methods Appl. Sci. 33 (4), 374–383.DOI
  26. [32]Sreenivasan, K. R. & Antonia, R. A. 1997 The phenomenology of small-scale turbulence. Annu. Rev. Fluid Mech. 29 (1), 435–472.DOI
  27. [33]Steiros, K. 2022 Balanced nonstationary turbulence. Phys. Rev. E 105 (3), 35109.DOI
  28. [34]Stewart, A. M. 2012 Longitudinal and transverse components of a vector field. Sri Lankan J. Phys. 12 (0), 33.DOI
  29. [35]Tang, S L, Antonia, R A & Djendi, L 2022 Transport equations for the normalized nth-order moments of velocity derivatives in grid turbulence. J. Fluid Mech. 930, A31.
  30. [36]Tennekes, H. 1975 Eulerian and lagrangian time microscales in isotropic turbulence. J. Fluid Mech. 67 (3), 561–567.DOI
  31. [37]Tsinober, A., Vedula, P. & Yeung, P. K. 2001 Random Taylor hypothesis and the behavior of local and convective accelerations in asotrophic turbulence. Phys. Fluids 13 (7), 1974–1984.DOI
  32. [38]Vassilicos, J. C. 2015 Dissipation in turbulent flows. Annu. Rev. Fluid Mech. 47 (1), 95–114.DOI
  33. [39]Vedula, P. & Yeung, P. K. 1999 Similarity scaling of acceleration and pressure statistics in numerical simulations of isotropic turbulence. Phys. Fluids 11 (5), 1208–1220.
  34. [40]Vela-Martín, Alberto 2022 Subgrid-scale models of isotropic turbulence need not produce energy backscatter. J. Fluid Mech. 937, A14.
  35. [42]Yasuda, T. & Vassilicos, J. C. 2018 Spatio-temporal intermittency of the turbulent energy cascade. J. Fluid Mech. 853, 235–252.DOI
  36. [43]Yeung, P. K., Pope, S. B., Lamorgese, A. G. & Donzis, D. A. 2006 Acceleration and dissipation statistics of numerically simulated isotropic turbulence. Phys. Fluids 18 (6), 65103.

Paper details

vlnc:W_sxvp8djx
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

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