CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ICEM] Problem with Pyramid mesh elements (

mezah August 4, 2013 14:17

Problem with Pyramid mesh elements
I am working on simulating the flow around wind turbine with three blades.
The blade is divided in to two upper and lower parts to study a brake system.
The upper part is rotated by one degree

After meshing using ICEM, and try to make prism layer around the blade i notice that there are pyramid mesh elements around the place where the blade is divided to upper and lower.

I have tried several changes in the prism option -according to a pdf i found here on the forum- like:
Min prsim quality
Ortho weight , etc but still there are pyramid elements.

One more thing, i already perform the meshing to the blade with no change in upper part angel, using the same meshing option and it works great.

I wonder if treated the blade as two different parts will solve the problem, since the prism layers is try to capture the topology of the blade.

Any help is appreciated.

Thanks in advance.

PSYMN August 26, 2013 10:37

Yea, if you have that sort of intersection, prism can not calculate what to do... There are a number of good work around, including a script tech support in Germany (Bernd Brasas) wrote to sort it out. Bernds script creates a small (several elements across) sphere centered on the "null point" intersections that are giving prism trouble... It then puts a different material point inside the sphere. When he grows prism, he grows it from the original surfaces, plus the sphere surfaces, but only into the outer fluid, but not the inner fluid...

Other solutions I have seen include a script (or manually) to edit the mesh at those corners to add some surface elements around the outside of the tetras that touch the corner... The tetras that touch the corner are put into the "other" volume part and prism is grown outward similar to the sphere case.

mezah August 26, 2013 16:48

Thanks alot for replying to my question.

I can hardly understand your explanation :confused: , i don't know if i make any sense when saying that i can understand the method yet i can't imagine it.

If you wouldn't mind providing me with a link or tell me under which category should i search.

Thanks in advance

PSYMN August 27, 2013 08:18

Maybe these solutions work better when the the two adjacent bars meet cleanly (like they do in many situations) and form that nice diamond shaped contact region with 4 null points...

Perhaps a better solution for your case would just be to "tweak" the geometry.
Simplify the geometry a bit to make it easier to mesh...

Sorry, it is a busy morning and I don't have time to get into this any deeper.

mezah August 27, 2013 17:59

Thanks alot for taking time and respond to my questions.

I will try to modify the geometry as you said, and whenever you have the time and mood
try to explain more :D

Best regards

mezah September 3, 2013 13:09

I have tried several modifications on the geometry but it all faild to provide the required results.

I am now following a new approach to simulate and mesh the problem by using non conformal mesh between upper part and lower part.

i will post it as soon as i finish it and obtain CFD results.

All times are GMT -4. The time now is 07:35.