Abstract
Direct numerical simulations (DNS) and large-eddy simulations (LES) of a turbulent channel flow with transverse square bars on one wall have been carried out at Re = 6.9 x 103. Both the Smagorinsky, with Van Driest damping, and dynamic subgrid models have been used. There is satisfactory agreement between the two types of simulations for the pressure and skin friction on the wall and the rms velocity near the wall. Far from the wall, LES gives only an approximation of the rms distributions obtained by DNS. However, the subgrid models are a significant improvement relative to when no model is used (which corresponds to an unresolved DNS). Another DNS at Re = 1.2 x 104 has been performed with the aim of comparing the results with the experiment by (Hanjalic, K., and Launder, B. E., Fully Developed Asymmetric Flow in Plane Channel, Journal of Fluid Mechanics, Vol. 51, 1972, pp. 301-335). The agreement between mean and rms distribution is good. The Reynolds-number dependence (Re ranging from 2.8 x 103 to 1.2 x 104) has been discussed. The pressure on the wall and hence the form drag do not depend on Reynolds number, and the velocity profile changes slightly for Re > 6.9 x 103.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call