April 26, 2007, 06:04
|
BSL Automatic Wall Function - Shifted koordinates
|
#1
|
Guest
|
Dear forum members,
I currently play around with calculations of the turbulent flow in a straight pipe (cfx10). The grid consists of a 2d crossection (ogrid) orthogonal to the main flow direction. I tested different high-reynolds and low-reynolds grids. The k-epsilon model (on the high-re grid) sets the velocity for the the first inner vertex as expected (matches with the logarithmic law). With BSL-komega+automatic wall model, there is always an apparent shift in the coordinates. In the low-re Case , the asymptotic behaviour u+=y+ is matched if y+ is shifted by 1 if the dimension free wall distance of the first vertex is 1, by 2 if it is 2 and so on. I can only guess that u+=1 is always assumed for the first vertex (on the wall) in the low-re case? I dont understand how the manual explains the automatic wall function: "the location of the first mesh point is now virtually moved down through the viscous sublayer as the mesh is refined in the low-Re mode." Maybe somebody could explain this differently? In the high-re grid, the profile from the bsl calculation seems also shifted by some value but i dont quite see how.
Since these are simple calculations i guess that there is a simple answer, but i just dont see it at the moment.
Thanks a lot, Sali
|
|
|