Statistics of thermal plumes and dissipation rates in turbulent Rayleigh-Benard convection in a cubic cell

We investigate the statistics of dissipation rates in turbulent Rayleigh-Benard convection inside a cubic cell for air ( Pr = 0 . 7 ) in the Rayleigh number range 2 x 10(6) <= Ra <= 10(9) using direct numerical simulations. Based upon the product of the vertical velocity and the temperature fluctuation (v'theta'), the entire cell volume is decomposed into plume and background dominated regions. Different cutoff values (C-t = 0 - 10% of the global maximum of v'theta') are used to demarcate these regions, and for all Ct, the volume fraction of the plume dominated region (v'theta' > C-t) decreases with the increase in Ra, while that of the background (v'theta' <= C-t) increases. The contribution of both the regions to the thermal dissipation rate decreases with Rayleigh number with a power-law behaviour. At C-t = 0 , the thermal dissipation rates from the plume and background approach the global scaling for higher Rayleigh numbers (Ra >= 10(8)). For lower cutoffs, the plume contributions scale with Reynolds number as predicted by the GrossmannLohse theory, while significant deviations are observed in the background counterpart. We observe that the probability density functions (PDF) of the thermal dissipation rates depart considerably from a lognormal distribution, while for viscous dissipation the PDFs approach log-normality at higher Rayleigh numbers. (c) 2021 Elsevier Ltd. All rights reserved.