betaFeedback

WorksPhysics and Astronomy

Realizing the ultimate scaling of the convection turbulence by spatially decoupling the thermal and viscous boundary layers

Abstract

Turbulent convection plays a crucial role in many natural environments, ranging from Earth ocean, mantle and outer core, to various astrophysical systems. For such flows with extremely strong thermal driving, an ultimate scaling was proposed for the heat flux and velocity. Despite numerous experimental and numerical studies, a conclusive observation of the ultimate regime has not been reached yet. Here we show that the ultimate scaling can be perfectly realized once the thermal boundary layer is fully decoupled from the viscous boundary layer and locates inside the turbulent bulk. The heat flux can be greatly enhanced by one order of magnitude. Our results provide concrete evidences for the appearance of the ultimate state when the entire thermal boundary layer is embedded in the turbulent region, which is probably the case in many natural convection systems.

Introduction

The paradigmatic model for turbulent convection is the Rayleigh-Bénard (RB) flow, namely the buoyancy-driven flow within a fluid layer heated from below and cooled from above. Such flow system can be found in many natural environments [1, 2, 3, 4]. For a given fluid with certain Prandtl number Pr=ν/κPr = \nu/\kappa, the driving force is measured by the Rayleigh number Ra=gαΔH3/νκRa = g\alpha\Delta H^3/\nu\kappa. Here, ν\nu is kinematic viscosity, κ\kappa is thermal diffusivity, gg is the gravitational acceleration, α\alpha is thermal expansion coefficient, Δ\Delta is the temperature difference across the layer, and HH is the layer height, respectively. The most fundamental question is how the heat flux and the turbulent level of the flow depend on RaRa and PrPr [5, 6, 7, 8]. Here, the heat flux is usually measured by the Nusselt number NuNu, i.e. the ratio between the convective flux to the conductive flux. The turbulent level of the flow by the Reynolds number Re=UH/νRe = UH/\nu, with UU being some characteristic value of the flow velocity.

In RB flow, thermal and viscous boundary layers develop next to the top and bottom plates, with the convective bulk in between and full of thermal plumes. In most early experiments and simulations the viscous boundary layer is laminar, and the heat flux follows the “classic” scaling law Nu∼RaγNu \sim Ra^\gamma with γ≈1/3\gamma\approx1/3 [9]. It was hypothesized that when RaRa is large enough, saying the thermal driving is extremely strong, the viscous boundary layer becomes fully turbulent due to the shearing exposed by the vigorous bulk flow, and the system enters the “ultimate” state where the heat flux no longer depends on the viscosity of the fluid [10, 11, 12]. The heat flux then follows the ultimate scaling with γ=1/2\gamma= 1/2, which predicts NuNu many orders of magnitude larger than the classic scaling at very large RaRa as in various natural environments.

Tremendous efforts have been made to achieve the ultimate scaling in RB flow by both experiments and simulations. Fluid with very small PrPr, such as mercury, is used so that the thermal boundary layer is much thicker than the viscous one and extends into the turbulent bulk [13, 14, 15]. Recent state-of-art simulations and experiments push RaRa to very high values [15, 16, 17, 18]. Some of these studies reported evidences of the transition to the ultimate regime [13, 16, 17]. Wall roughness was introduced to trigger the transition of the momentum boundary layer to fully turbulent state, but this only works for a limit range of RaRa for a given wall roughness [19, 20]. Homogeneous convection in fully periodic domain exhibits the ultimate scaling, since in such flow all boundary layers are removed and only the turbulent bulk plays a role [21, 22]. The radiative convection with internal heating and the convection with background oscillation can both greatly enhance the exponent γ\gamma from the classic-scaling value, but such treatments change the dynamics of the convection system by introducing extra source terms for the thermal field or the momentum field [23, 24, 25, 26].

Governing equations and numerical methods

Here by specially designed numerical experiments, we confirm the physical conjecture of the ultimate paradigm. Once the whole thermal boundary layer is spatially decoupled from the viscous one and locate entirely in the turbulent flow region, the ultimate scaling of the heat flux can be perfectly realized even at low to moderate Rayleigh numbers. Specifically, we consider the RB flow between two parallel plates which are perpendicular to the direction of the gravity and separated by a distance HH. A homogeneous layer of height hh is introduced at top and bottom of the domain. Within this layer the temperature is uniform and equal to the temperature of the adjacent plate, which can be readily realized by using an immersed-boundary technique. The buoyancy-force term takes effect only between the two homogeneous layers. By doing so the thermal boundary layer is lifted by a height hh from the plate where the viscous boundary layer occurs. If hh is large enough, the two boundary layers can be fully decoupled.

Specifically, we consider a fluid layer bounded by two parallel plates which are perpendicular to the gravity and separated by a distance HH. The Oberbeck-Boussinesq approximation is utilized to take care of the buoyancy effects. First, the fluid density depends linearly on the scalar field as ρ=ρ0(1−αθ)\rho=\rho_0(1-\alpha\theta). Here ρ0\rho_0 is the reference density at the reference temperature T0T_0, and θ\theta is the temperature deviation from T0T_0. Second, the variation of density is relatively small so that all the fluid properties can be treated as constants, and only the buoyancy force needs to be included. The governing equations then read

∂tui+u⋅∇u=−1ρ∇p+ν∇2u+αθg,(1)\partial_t u_i+\mathbf{u}\cdot\nabla\mathbf{u}=-\frac{1}{\rho}\nabla p+\nu\nabla^2\mathbf{u}+\alpha\theta\mathbf{g}, \tag*{(1)}
∂tθ+u⋅∇θ=κ∇2θ,(2)\partial_t\theta+\mathbf{u}\cdot\nabla\theta=\kappa\nabla^2\theta, \tag*{(2)}
∇⋅u=0.(3)\nabla\cdot\mathbf{u}=0. \tag*{(3)}

Here u\mathbf{u} is three dimensional velocity, pp is pressure, and ∂t\partial_t is the partial derivative with respect to time, respectively. The above equations are non-dimensionalized by the layer height HH, the temperature difference between two plates Δ\Delta, and the free-fall velocity U=gαΔHU=\sqrt{g\alpha\Delta H}. The two plates are no-slip and the periodic boundary conditions are applied to the two horizontal directions.

Direct numerical simulations were conducted with the in-house code, which employs the finite difference scheme and a fraction-time-step method. The code has been tested for various wall-turbulence and convection turbulence [27]. To introduce the homogeneous thermal layer, an extra source term is included in the temperature equation (2) based on the idea of the immersed boundary method [28]. Also the buoyancy force in the momentum equation (1) is turned off inside the homogeneous thermal layer. Grid-independence test has been conducted in order to ensure the resolution is adequate. The three-dimensional volume rendering in figure 1 is generated by the open source software VisIt [29].

Main results

In figure 1 we compare one thermal field of the normal RB flow and that with homogeneous layers at Ra=106Ra=10^6 and Pr=1Pr=1. Here hh is chosen as 0.2H0.2H. We only show the fluid with temperature larger than 90%Δ90\%\Delta and smaller than 10%Δ10\%\Delta. For normal RB flow at this relatively low RaRa, the large thermal anomaly cannot be transported very far from the plates, which limits the global heat flux. On the contrary, with two homogeneous thermal layers the thermal plumes with large anomaly are significantly stronger, and spread over the entire convection bulk. Accordingly, the heat flux is greatly enhanced in the flow with homogeneous thermal layers.

Image not extracted

Figure 1. Volume rendering of the thermal field at Ra=106Ra = 10^{6} and Pr=1Pr = 1. Yellow and blue indicate the temperature larger than 90%Δ90\%\Delta and smaller than 10%Δ10\%\Delta, respectively. A: the normal Rayleigh-Bénard flow. B: the modified flow with two homogeneous thermal layers of the height h=0.2Hh = 0.2H. The flow in B has much stronger thermal plumes in the bulk, while for the flow in A the large temperature anomaly can hardly be transported into the bulk.

We then present the main results of the current study, i.e. the ultimate scalings of heat flux and flow velocity, which in their full form read

Nu∼Pr1/2Rab1/2,Re∼Pr−1/2Rab1/2.(4)Nu \sim Pr^{1/2}Ra_b^{1/2}, \qquad Re \sim Pr^{-1/2}Ra_b^{1/2}. \tag*{(4)}

To account for the height of two homogeneous thermal layers, the bulk Rayleigh number is calculated as Rab=Ra(1−2h/H)3Ra_b = Ra(1-2h/H)^3. The Nusselt number is defined as Nu=q(H−2h)/(κΔ)Nu = q(H-2h)/(\kappa\Delta) with qq being the total dimensional heat flux. When h=0h = 0, both definitions recover the usual form of RB flow. The Reynolds number is Re=urmsH/νRe = u_{\mathrm{rms}}H/\nu and urmsu_{\mathrm{rms}} is the root-mean-square velocity of the whole domain, since the flow motion is confined by two plates, including the homogeneous thermal layers. We first look at the dependences of NuNu and ReRe on RabRa_b for fixed Pr=1Pr = 1. Three decades of RaRa are covered from 10510^{5} up to 10810^{8}. The height of the homogeneous thermal layer is set at h=0.2Hh = 0.2H so that for all simulations the momentum boundary layer is fully embedded inside the homogeneous thermal layer. Figure 2A and Figure 2B show that both NuNu and ReRe follow the ultimate scaling law (4) perfectly, even though the Rayleigh number is not very large. Moreover, both the heat flux and flow velocity are greatly enhanced by introducing the homogeneous thermal layers. For Rab≥106Ra_b \ge10^{6}, the modified flow generates a heat flux more than one order of magnitude higher than that for the normal RB flow. For instance, the Nusselt number is comparable for the modified flow at Ra≈108Ra \approx10^{8} and the normal RB flow at Ra≈1011Ra \approx10^{11}.

Image not extracted

Figure 2. The ultimate scaling of the heat flux and flow velocity. Solid symbols and lines depict the ultimate scalings of the modified flow with homogeneous thermal layers. Open symbols and dashed lines show the normal RB flow and the classic scaling. The corresponding exponents are determined by linear fitting. A and B: The Nusselt number NuNu and Reynolds number ReRe versus the increasing bulk Rayleigh number RabRa_b for a fixed Prandtl number Pr=1Pr = 1. C and D: NuNu and ReRe versus the increasing PrPr for fixed Ra=3×106Ra = 3 \times10^{6}.

For the second group of simulations, we fix Ra=3×106Ra = 3 \times10^{6} and increase PrPr from 0.01 to 10, again covering three orders of magnitude. The height of the homogeneous layers is h=0.2Hh = 0.2H. As shown in Figure 2C and Figure 2D, the behaviors of NuNu and ReRe are also very close to the ultimate scaling (4). The scaling exponent for NuNu given by the linear fitting is indistinguishable from the theoretically predicted value 1/2. The exponent for ReRe obtained from the numerical data is slightly different from −1/2-1/2. For the moderate Ra=3×106Ra = 3 \times10^{6} considered here, the flow is not fully turbulent at high Prandtl number regime. Therefore ReRe decreases faster than that predicted by the ultimate scaling. Another interesting observation is that at low PrPr region, the difference between the modified flow and the normal RB flow is very small. This is expectable since for small PrPr, the viscous boundary layer is thinner than the thermal one in the normal RB flow, and part of the thermal boundary layer is already in the turbulent bulk.

To further demonstrate that the ultimate scaling is satisfied when the thermal and viscous boundary layers are decoupled from each other, we run simulations with increasing hh for three different Rayleigh numbers Ra=105Ra = 10^{5}, 10610^{6}, and 10710^{7}. The Prandtl number is fixed at Pr=1Pr = 1. We focus on the behaviors of Nusselt number, which are shown in Figure 3. As hh increases, i.e. the thermal boundary layers are gradually decoupled from the viscous ones, the heat flux is enhanced and the scaling exponent transits from the classic value of 1/3 to the ultimate value 1/2. In Figure 3 we plot the scaling exponent γ\gamma versus the height of the homogeneous thermal layer hh. For small hh, γ\gamma increases rapidly to a value larger than 1/21/2. At this overshoot the thermal boundary layer already enters the turbulent bulk for large RaRa but is still coupled with the viscous boundary layer for low RaRa. When hh is large enough, for all three RaRa’s the thermal boundary layer is fully within the turbulent region, and the exponent γ\gamma gradually approaches the ultimate value 1/21/2.

Image not extracted

Figure 3. The effect of the thermal layer height on the heat-flux scaling. A: The heat flux versus RabRa_b for different hh, which increases from 0 to 0.2 as color changes from blue to red. B: The scaling exponent γ\gamma versus the height hh. In both panels the solid and dashed lines mark the ultimate scaling with the exponent 1/2 and the classic scaling with the exponent 1/3, respectively.

Conclusion and discussion

In summary, we demonstrate that once the thermal boundary layers decouple with the viscous ones and locate within the turbulent convection bulk, the ultimate scaling of the heat flux and flow velocity can be perfectly realized even at relatively low Rayleigh numbers. Our results support the physical picture of the ultimate state of convection turbulence, namely when the momentum boundary layers become fully turbulent, the heat flux is independent of the fluid viscosity. In line with the unifying model for RB convection [11, 12], the current flow corresponds to the IVl\mathrm{IV}_{l} regime. Compared to the homogeneous convection [21, 22], the current flow configuration is more close to the ultimate regime since the thermal boundary layers are included and the ultimate scaling is achieved.

A definitive proof for the existence of the ultimate regime relies on the direct observation of the ultimate scaling in the experiments and simulations of the normal RB convection. Nevertheless, one can anticipate that, based on the present study, convection flows will eventually enter the ultimate state when the boundary layer region becomes fully turbulent at extremely high thermal driving. Such conditions are very likely satisfied in natural systems such as the Earth outer core and stars. Transition towards the ultimate regime has been reported in recent studies [16, 17], and we may expect that a fully developed ultimate state will be observed in experiments in near future.

Acknowledgements: This work is supported by the Major Research Plan of National Natural Science Foundation of China for Turbulent Structures under the Grants 91852107 and 91752202. Y.Yang also acknowledges the partial support from the Strategic Priority Research Program of Chinese Academy of Sciences under the Grant No. XDB42000000.

References

  1. [1]E. A. Spiegel, Convection in stars I. basic Boussinesq convection, Annu. Rev. Astro. Astrophys. 9, 323 (1971).DOI
  2. [2]P. J. Tackley, D. J. Stevenson, G. A. Glatzmaier, and G. Schubert, Effects of an endothermic phase transition at 670 km depth in a spherical model of convection in the Earth’s mantle, Nature 361, 699 (1993).DOI
  3. [3]J. Marshall and F. Schott, Open-ocean convection: Observations, theory, and models,, Rev. Geophys 37, 1 (1999).DOI
  4. [4]Heimpel, M., Aurnou, J., & Wicht, J. (2012). Simulation of equatorial and high-latitude jets on Jupiter in a deep convection model. Phys. Rev. Lett., 108, 024502.
  5. [5]G. Ahlers, S. Grossmann, and D. Lohse, Heat transfer and large scale dynamics in turbulent Rayleigh-Bénard convection, Rev. Mod. Phys. 81, 503 (2009).DOI
  6. [6]F. Chillà and J. Schumacher, New perspectives in turbulent Rayleigh?Bénard convection., Eur. Phys. J. E 35, 58 (2012).DOI
  7. [7]Xia, K.-Q. (2013). Current trends and future directions in turbulent thermal convection. Theor. Appl. Mech. Lett., 3, 052001.
  8. [8]M. Plumley and K. Julien, Scaling laws in Rayleigh-Bénard convection, Earth and Space Science 34, 1580 (2019).DOI
  9. [9]M. V. Malkus, The heat transport and spectrum of thermal turbulence, Proc. R. Soc. Lond. A 225, 196 (1954).DOI
  10. [10]R. H. Kraichnan, Turbulent thermal convection at arbitrary prandtl number, Phys. Fluids 5, 1374 (1962).DOI
  11. [11]Grossmann, S., & Lohse, D. (2000). Scaling in thermal convection: a unifying theory. J. Fluid Mech., 407, 27.
  12. [12]S. Grossmann and D. Lohse, Thermal convection for large Prandtl numbers, Phys. Rev. Lett. 86, 3316 (2001).DOI
  13. [13]Chavanne, X., Chilla, F., Castaing, B., Hebral, B., Chabraud, B., & Chaussy, J. (1997). Observation of the ultimate regime in Rayleigh-Bénard convection. Phys. Rev. Lett., 79, 3648.
  14. [14]Glazier, J. A., Segawa, T., Naert, A., & Sano, M. (n.d.). Evidence against ‘ultrahard’ thermal tur-.
  15. [15]J. J. Niemela, L. Skrbek, K. R. Sreenivasan, and R. J. Donnelly, Turbulent convection at very high Rayleigh numbers, Nature 404, 837 (2000).pubmed.ncbi.nlm.nih.gov/10786783
  16. [16]X. He, D. Funfschilling, H. Nobach, E. Bodenschatz, and G. Ahlers, Transition to the ultimate state of turbulent Rayleigh-Bénard convection, Phys. Rev. Lett. 108, 024502 (2012).pubmed.ncbi.nlm.nih.gov/22324688
  17. [17]X. Zhu, V. Mathai, R. J. A. M. Stevens, R. Verzicco, and D. Lohse, Transition to the ultimate regime in two-dimensional Rayleigh-Bénard convection, Phys. Rev. Lett. 120, 144502 (2018).DOI
  18. [18]K. P. Iyer, J. D. Scheel, J. Schumacher, and K. R. Sreenivasan, Classical 1/3 scaling of convection holds up to Ra=10¹⁵, Proc. Natl. Acad. Sci. U.S.A. 117, 7594 (2020).DOI
  19. [19]X. Zhu, R. Stevens, R. Verzicco, and D. Lohse, Roughness-facilitated local 1/2 scaling does not imply the onset of the ultimate regime of thermal convection, Phys. Rev. Lett. 119, 154501 (2017).DOI
  20. [20]Zhu, X., Mathai, V., Stevens, R. J. A. M., Verzicco, R., & Lohse, D. (2019). $Nu \sim Ra^{1/2}$ scaling enabled by multiscale wall roughness in Rayleigh-Bénard convection. J. Fluid Mech., 869, R4.
  21. [21]D. Lohse and F. Toschi, The ultimate state of thermal convection, Phys. Rev. Lett. 90, 034503 (2003).DOI
  22. [22]E. Calzavarini, D. Lohse, F. Toschi, and R. Tripiccione, Rayleigh and Prandtl number scaling in the bulk of Rayleigh-Bénard turbulence, Phys. Fluids 17, 055107 (2005).DOI
  23. [23]S. Lepot, S. Aumaitre, and B. Gallet, Radiative heating achieves the ultimate regime of thermal convection, Proc. Natl Acad. Sci. USA 115, 8973 (2018).DOI
  24. [24]V. Bouillaut, S. Lepot, S. Aumaitre, and B. Gallet, Transition to the ultimate regime in a radiatively driven convection experiment, J. Fluid Mech. 861, R5 (2019).DOI
  25. [25]M. Creyssels, Model for classical and ultimate regimes of radiatively driven turbulent convection, J. Fluid Mech. 900, A39 (2020).DOI
  26. [26]B.-F. Wang, Q. Zhou, and C. Sun, Vibration-induced boundary-layer destabilization achieves massive heat-transport enhancement, Sci. Adv. 6, ezza8239 (2020).DOI
  27. [27]R. Ostilla-Mónico, Y. Yang, E. P. van der Poel, D. Lohse, and R. Verzicco, A multiple resolutions strategy for direct numerical simulation of scalar turbulence., J. Comput. Phys. 301, 308 (2015).DOI
  28. [28]Fadlun, E. A., Verzicco, R., Orlandi, P., & Mohd-Yusof, J. (n.d.). Combined immersed. Journal of Computational Physics, 161, 35.
  29. [29]H. Childs, E. Brugger, B. Whitlock, J. Meredith, S. Ahern, D. Pugmire, K. Biagas, M. Miller, C. Harrison, G. H. Weber, H. Krishnan, T. Fogal, A. Sanderson, C. Garth, E. W. Bethel, D. Camp, O. Rübel, M. Durant, J. M. Favre, and P. Navrátil, VisIt: An End-User Tool For Visualizing and Analyzing Very Large Data, in High Performance Visualization–Enabling Extreme-Scale Scientific Insight (2012) pp. 357–372.

Paper details

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

views · 5

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