Wind and boundary layers in Rayleigh-Benard convection. Part 2: boundary layer character and scaling

TL;DR

This study uses direct numerical simulations to analyze the complex dual laminar-turbulent nature of boundary layers in Rayleigh-Benard convection, revealing how plume dynamics influence boundary layer scaling and turbulence characteristics.

Contribution

It introduces a detailed analysis of boundary layer behavior in Rayleigh-Benard convection, highlighting the dual laminar-turbulent features and proposing scaling laws for boundary layer thickness and friction factor.

Findings

Boundary layers exhibit both laminar and turbulent features.

Reynolds stresses have minimal impact on the friction factor at studied Ra.

Scaling laws for boundary layer thickness and friction factor are validated.

Abstract

The effect of the wind of Rayleigh-Benard convection on the boundary layers is studied by direct numerical simulation of an L/H=4 aspect-ratio domain with periodic side boundary conditions for Ra={10^5, 10^6, 10^7} and Pr=1. It is shown that the kinetic boundary layers on the top- and bottom plate have some features of both laminar and turbulent boundary layers. A continuous spectrum, as well as significant forcing due to Reynolds stresses indicates undoubtedly a turbulent character, whereas the classical integral boundary layer parameters -- the shape factor and friction factor (the latter is shown to be dominated by the pressure gradient) -- scale with Reynolds number more akin to laminar boundary layers. This apparent dual behavior is caused by the large influence of plumes impinging onto and detaching from the boundary layer. The plume-generated Reynolds stresses have a negligible effect on the friction factor at the Rayleigh numbers we consider, which indicates that they are passive with respect to momentum transfer in the wall-parallel direction. However, the effect of Reynolds stresses cannot be neglected for the thickness of the kinetic boundary layer. Using a conceptual wind model, we find that the friction factor C_f should scale proportional to the thermal boundary layer thickness as C_f ~ lambda_Theta, while the kinetic boundary layer thickness lambda_u scales inversely proportional to the thermal boundary layer thickness and wind Reynolds number lambda_u ~ lambda_Theta^{-1} Re^{-1}. The predicted trends for C_f and \lambda_u are in agreement with DNS results.