Numerical Study of Laminar and Turbulent Natural Convection from a Stack of Solid Horizontal Cylinders

TL;DR

This study numerically investigates natural convection around stacks of horizontal cylinders in laminar and turbulent regimes, revealing how geometry and arrangement influence heat transfer, and introduces a new correlation for Nusselt number prediction.

Contribution

The paper provides a comprehensive numerical analysis of natural convection in cylinder stacks across laminar and turbulent regimes, introducing a new Nusselt number correlation based on Ra and L/D.

Findings

Nusselt number increases with Rayleigh number for all L/D ratios.

Short cylinders exhibit higher heat transfer than longer ones at the same Ra.

The new correlation accurately predicts Nusselt number for various configurations.

Abstract

Natural convection from a stack of isothermal solid horizontal cylinders has been investigated numerically in a three dimensional computational domain. Simulations were conducted in both laminar and turbulent flow regimes of Rayleigh number (Ra) spanning in the range (10^4 to 10^8) and (10^10 to 10^13), respectively. In the present study, the length to diameter ratio of the cylinders has been varied in the range 0.5 to 20. Three different stack arrangements were considered for the numerical simulations by arranging three, six and ten number of cylinders in a triangular manner. The present computational study is able to appraise very interesting thermo-buoyant plume structures around the stack of cylinders. The average Nusselt number (Nu) shows a positive dependence on Ra for all L/D. The average Nu for a stack of three-cylinders is marginally higher than that of six-cylinders followed by ten-cylinders. Furthermore, at a particular Ra, Nu is significantly higher for short cylinders (low L/D) and decreases with increase in L/D up to 10 or 15 and remain constant for long cylinders. In addition, the present numerical results are also compared with the stack of hollow cylinders. A new Nusselt number correlation has been developed for different stacks as a function of Ra and L/D, which would be useful to industrial practitioners and academic researchers.