CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS (http://www.cfd-online.com/Forums/ansys/)
-   -   yplus (wall shear) saw-like shape (http://www.cfd-online.com/Forums/ansys/86680-yplus-wall-shear-saw-like-shape.html)

potac March 30, 2011 03:32

yplus (wall shear) saw-like shape
 
Hi all,
I have been simulating a air flow around a 2D curved roof building in CFX. When plotting values of wall shear or yplus along the roof I obtain saw-like steps. I have tried different mesh resolutions (yplus<50, <1000), turbulence models (KE, SST), schemes (upwind, high resolution), but those steps or waves remain. I ran it steady and transient using different time steps but without any improvement.
Can someone give me some hint I should focus on, please?
Thank you. J.

potac April 3, 2011 04:12

2 Attachment(s)
I have been doing some test to obtain better results. It seems to be independent on numerics. I have attached picture of near wall mesh and resulting plot of yplus. In that case, the it converges pretty well but the yplus values are very weird. When I change my mesh significantly (keep the vertical spacing and increased the prism length from 0.2 m to 0.5 or 1 m) the saw-like steps decreases or vanishes. But the previous convergence is gone.
Any suggestion, please?
Thank you. J.

potac April 14, 2011 09:19

I think, I have found the cause of the yplus behavior. I tried to smooth my mesh and that showed me the cause. In ICEM, the arch of the roof is split into several sections to capture the curvature. And the mesh is built on this sections. I found all the peaks in the yplus correspond to nodes where these sections meet. The smoothing in ICEM is able to remove these edges but it is quite time consuming and the result is not always acceptable.
It seems the solution could be in some setup of DesignModeler or ICEM to cover the curvatures more precisely. I have been searching in setup menus of both DM and ICEM but cannot find anything related. BTW, I created kind of similar mesh in Ansys Meshing and the resulting yplus was smooth. Thank you. J.


All times are GMT -4. The time now is 02:25.