Comparison of low Reynolds number turbulence and conjugate heat transfer modelling for pin-fin roughness elements

This study contrasts Reynolds-Averaged Navier-Stokes (RANS) and (Numerical) Large-Eddy Simulation (NLES) for their use in predicting conjugate heat transfer a low Reynolds number flow over surface roughness element. The (N)LES predictions are good agreement with experimental data, the estimate is within 7.7% error. linear RANS model shows larger errors up to 40%, especially modelling turbulent stresses. Using one-equation LES turbulence slightly increases prediction compared numerical LES. indicates that more advanced models might not give accurate predictions. Additionally, this investigates time dependant development of temperature fields how long data needs be collected statistically stationary results. develop approximately 1800 through-flow times, TL, 150 TL required collection statistics. A significant range length scale between was found. scales appeared physically reasonable should inform mesh resolution future studies. Two additional cases were run different cube heights showing changes field greater sensitivity modelling.