Mixed convection study on the influence of low Prandtl numbers and buoyancy in turbulent heat transfer using DNS

Abstract

Direct numerical simulation (DNS) is performed to study turbulent heat transfer in Poiseuille-Rayleigh-Bénard (PRB) flows with low Prandtl numbers in this article. The mesh is Cartesian and a highly accurate finite difference sixth-order compact scheme is chosen to discretize the incompressible Navier–Stokes equations to perform DNS. Liquids with a fixed Richardson number of 0.25 and four different Prandtl number (Pr = 0.025, 0.05, 0.1, 0.71) are simulated and compared with Poiseuille flow to investigate the influence of Prandtl number and buoyancy on –PRB flows. Constant fluid properties and Boussinesq approximation are assumed. The obtained results are discussed and analysed in an extensive way in this study. Specifically, buoyancy initiate large scale circulation and the scale shrinks with the increasing of Prandtl number. Velocity fluctuations become stronger with PRB flow which indicate that buoyancy can strongly enhance the turbulent intensity. Reτ is increased in the cases of low Pr. Moreover, when Pr decreases, temperature distributions are found to be more homogeneous and mixing of the fluids is more sufficient in the middle of the channel. Additionally, the scale of the large-scale structures is enlarged in mixed convection compared with forced convection. This can be observed in the temperature field of low-Prandtl-number fluids. It is also observed that Reynolds analogy cannot be used to predict the thermal field under mixed convection or forced convection with low Prandtl number. The research results can be used for the R&D of Gen IV nuclear fast reactors.