CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ICEM] Meshing by bottom up approach (

saisanthoshm88 January 17, 2011 13:51

Meshing by bottom up approach
1 Attachment(s)
I have a heater coil geometry to mesh in ICEM , the attached image shows the geometry. i'd like to consider a bottom up approach for blocking the geometry to generate a Hex mesh. Can some one please suggest me if it is fine to proceed this way?

Another question:

I've tried to coarsen a mesh in ICEM by the Adjust mesh density option but that didn't work out what can be the reason for this ?

PSYMN January 17, 2011 14:36

Yes, bottom up would probably be best/easiest. You could try the automatic surface blocking with a fill... (aka MultiZone, which you could also try in ANSYS Meshing (workbench)) or you could start with just a 2D block of one end and extrude it along the curve...

As for adjust mesh density not working, you have not given enough info to guess anything but user error at this point. What did you try and what was the result?

saisanthoshm88 January 18, 2011 14:09

oh sorry regarding the coarsening of mesh i didn't know about the option coarsen surface mesh elements , later when i turned it on everything went well.

Any way I found this option of coarsening the mesh to be very useful in my case
but it's for the first time that I used it so I wonder if it is a commonly used and a recommended option in ICEM

and again as I try to smoothen the resultant coarse mesh i see that the number of elements in the mesh goes on decreasing further as the smoothing progresses, what can be the reason for this ? I have the option allow node merging turned on while smoothing.

PSYMN January 18, 2011 17:28

Yes, the mesh coarsening works pretty well because ICEM CFD is aware of your edges and other sharp features... So when it coarsens, you don't actually lose these important details. Also, the limiting factor is the quality, so you know your final mesh will be good.

Some users like using a very fine mesh to make sure all CFD details are captured and then they coarsen to get the cell count down.

Other FEA users may start with a durability model (say 150,000 elements) and then coarsen it to get the 15K NVH model in one operation and without losing the features...

All times are GMT -4. The time now is 05:18.