Heat Transfer in Turbulent Rayleigh-Benard Convection below the Ultimate Regime

TL;DR

This study precisely measures heat transfer in turbulent Rayleigh-Benard convection over a range of Prandtl and Rayleigh numbers, revealing flow bimodality, side-wall effects, and negligible Prandtl dependence.

Contribution

It provides the first high-accuracy measurements of Nu(Ra,Pr) in cryogenic helium-4, clarifying the Prandtl dependence and flow behavior in turbulent convection.

Findings

Nu(Ra) shows bimodality with 4-7% difference.

Side-wall effects influence heat transfer.

Nu(Pr) dependence is negligible, slope <0.03.

Abstract

A Rayleigh-B\'enard cell has been designed to explore the Prandtl (Pr) dependence of turbulent convection in the cross-over range $0.7<Pr<21$ and for the full range of soft and hard turbulences, up to Rayleigh number $Ra\simeq 10^{11}$. The set-up benefits from the favourable characteristics of cryogenic helium-4 in fluid mechanics, in-situ fluid property measurements, and special care on thermometry and calorimetric instrumentation. The cell is cylindrical with $diameter/height=0.5$. The effective heat transfer $Nu(Ra,Pr)$ has been measured with unprecedented accuracy for cryogenic turbulent convection experiments in this range of Rayleigh numbers. Spin-off of this study include improved fits of helium thermodynamics and viscosity properties. Three main results were found. First the $Nu(Ra)$ dependence exhibits a bimodality of the flow with $4-7 %$ difference in $Nu$ for given $Ra$ and $Pr$. Second, a systematic study of the side-wall influence reveals a measurable effect on the heat transfer. Third, the $Nu(Pr)$ dependence is very small or null : the absolute value of the average logarithmic slope $(dlnNu/dlnPr)_{Ra}$ is smaller than 0.03 in our range of $Pr$, which allows to disciminate between contradictory experiments [Ashkenazi \textit{et al.}, Phys. Rev.Lett. 83:3641 (1999)][Ahlers \textit{et al.}, Phys.Rev.Lett. 86:3320 (2001)].