Published online by Cambridge University Press: 16 March 2021

We report an experimental study of the Prandtl-number effects in quasi-two-dimensional (quasi-2-D) Rayleigh–Bénard convection. The experiments were conducted in four rectangular convection cells over the Prandtl-number range of  $11.7 \leqslant Pr \leqslant 650.7$ and over the Rayleigh-number range of
$11.7 \leqslant Pr \leqslant 650.7$ and over the Rayleigh-number range of  $6.0\times 10^8 \leqslant Ra \leqslant 3.0\times 10^{10}$. Flow visualization reveals that, as
$6.0\times 10^8 \leqslant Ra \leqslant 3.0\times 10^{10}$. Flow visualization reveals that, as  $Pr$ increases from 11.7 to 145.7, thermal plumes pass through the central region much less frequently and their self-organized large-scale motion is more confined along the periphery of the convection cell. The large-scale flow is found to break down for higher
$Pr$ increases from 11.7 to 145.7, thermal plumes pass through the central region much less frequently and their self-organized large-scale motion is more confined along the periphery of the convection cell. The large-scale flow is found to break down for higher  $Pr$, resulting in a regime transition in the Reynolds number
$Pr$, resulting in a regime transition in the Reynolds number  $Re$. For the
$Re$. For the  $Pr$ range with a large-scale flow of system size, the
$Pr$ range with a large-scale flow of system size, the  $Re$ number, Nusselt number
$Re$ number, Nusselt number  $Nu$ and local temperature fluctuations were investigated systematically. It is found that
$Nu$ and local temperature fluctuations were investigated systematically. It is found that  $Re$ scales as
$Re$ scales as  $Re \sim Ra^{0.58}Pr^{-0.82}$ in the present geometry, which suggests that it is in line with the behaviour in the 2-D configuration. On the other hand, the measured
$Re \sim Ra^{0.58}Pr^{-0.82}$ in the present geometry, which suggests that it is in line with the behaviour in the 2-D configuration. On the other hand, the measured  $Nu(Ra, Pr)$ relation
$Nu(Ra, Pr)$ relation  $Nu \sim Ra^{0.289}Pr^{-0.02}$ tends to be compatible with the finding in a three-dimensional (3-D) system. For the temperature fluctuations in the cell centre and near the sidewall, they exhibit distinct
$Nu \sim Ra^{0.289}Pr^{-0.02}$ tends to be compatible with the finding in a three-dimensional (3-D) system. For the temperature fluctuations in the cell centre and near the sidewall, they exhibit distinct  $Ra$-dependent scalings that could not be accounted for with existing theories, but their
$Ra$-dependent scalings that could not be accounted for with existing theories, but their  $Pr$ dependences for
$Pr$ dependences for  $Pr \lesssim 50$ are in agreement with the predictions by Grossmann & Lohse (Phys. Fluids, vol. 16, 2004, pp. 4462–4472). These results enrich our understanding of quasi-2-D thermal convection, and its similarities and differences compared to 2-D and 3-D systems.
$Pr \lesssim 50$ are in agreement with the predictions by Grossmann & Lohse (Phys. Fluids, vol. 16, 2004, pp. 4462–4472). These results enrich our understanding of quasi-2-D thermal convection, and its similarities and differences compared to 2-D and 3-D systems.
 $10^{-1}$ and
$10^{-1}$ and  $10^4$ and Rayleigh numbers between
$10^4$ and Rayleigh numbers between  $10^5$ and
$10^5$ and  $10^9$. J. Fluid Mech. 662, 409–446.CrossRefGoogle Scholar
$10^9$. J. Fluid Mech. 662, 409–446.CrossRefGoogle Scholar $Pr >1$. Phys. Fluids 32, 015101.CrossRefGoogle Scholar
$Pr >1$. Phys. Fluids 32, 015101.CrossRefGoogle Scholar $Nu \sim Ra^{1/2}$ scaling enabled by multiscale wall roughness in Rayleigh–Bénard turbulence. J. Fluid Mech. 869, R4.CrossRefGoogle Scholar
$Nu \sim Ra^{1/2}$ scaling enabled by multiscale wall roughness in Rayleigh–Bénard turbulence. J. Fluid Mech. 869, R4.CrossRefGoogle Scholar