CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS (http://www.cfd-online.com/Forums/ansys/)
-   -   Cell area issue (http://www.cfd-online.com/Forums/ansys/93965-cell-area-issue.html)

General_Gee November 1, 2011 07:06

Cell area issue
 
Hey,

I have a wierd issue with FLUENT 12. In my 2D mesh I have a semi-circle shaped edge, which I have meshed with a constant spacing.

But when I look at the Face Area Magnitude (either through XY-plot, or through exporting through a udf using F_AREA(A,f,t); & area = NV_MAG(A); ) it varies along the edge. :confused:

http://img830.imageshack.us/img830/3554/spherearea.png


Surely as the mesh has a constant spacing, the area of the each cells face should be the same...if I do the same on a straight edge, the reported area is constant.

Any thoughts

Cheers

Gee

jonny_b November 2, 2011 15:23

I can't remember off the top of my head but I think the face area magnitude calculation in FLUENT is vector direction dependent. I notice that the area value in the graph is biggest for the cells that are parallel with the yellow line on the bottow of the domain and drops to zero for the cells that are perpendicular to this line. My assumption is that the vector direction it is using is normal to the yellow line. See if you can change this direction, but ultimately you will need to use a different calculation of area I think.

General_Gee November 3, 2011 10:16

OK. So I queried this with ANSYS support and this was the response. It is important to note that in my description here I forgot to say that it is axisymetric about the yellow line

Quote:

From the XY plot picture, it seems that you are using axi-symmetric model and plotting the face area magnitude. If you have selected axi-symmetric solver and axis defined, the area of each face would be different along the curve. The face which fall just above the center will have max area and this is what we expect from axi-symmetric geometry.

For 2d geometry (planar), the face area magnitude would be constant along the curve provided edge is meshed with constant size. You can check this in your model by changing axi-symmetric to Planar and axis to wall.
So it seems that the Face Area Magnitude in axi-symmetric simulations is more of an area magnitude projected in the radial direction. Not sure how much this helps me, but at least I know the cause now.

EDIT:
Thinking about this, I understand it a bit more, so will clarify for future reference.

In the 2D case the area of the cell is calculated based on the length of the edge of the cell times by the distance in the third (non-existent) direction. As for the values of C_D and C_L in 2D, FLUENT uses a unit vector in the third direction and so the area is calculated based on #cell_edge_length# x #unit_in_z_direction#. In the axi-symmetric case the 3D version of the cell is actually a wedge shape, which is narrower the closer you get to the axis, hence making the cell area dependent on distance away from the axis.


All times are GMT -4. The time now is 13:34.