CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ICEM] Why Prism Incomplete? (http://www.cfd-online.com/Forums/ansys-meshing/118319-why-prism-incomplete.html)

m5edr May 24, 2013 23:03

Why Prism Incomplete?
 
3 Attachment(s)
Hi Experts

I built a simple prism around wind turbine blade (see picture 1) BUT the Prism is Incomplete along the blade (see picture 2) , for the prism parameters (see picture 3). Why?

Thanks in advance

siw May 25, 2013 03:32

Just as a test - is the prism layer incomplete if you let ICEM float them when they are generated? To do this set zero values for the initial height and the total height. By doing this ICEM will nicely size the prisms so that the transition from the last prism to the first tetrahedral is smooth. After that you can split and re-distribute the prism to suit (to a certain extent). It may not give you the exact prism mesh you what but it may go some way to getting a complete prism mesh.

m5edr May 25, 2013 07:34

1 Attachment(s)
Quote:

Originally Posted by siw (Post 429922)
Just as a test - is the prism layer incomplete if you let ICEM float them when they are generated? To do this set zero values for the initial height and the total height. By doing this ICEM will nicely size the prisms so that the transition from the last prism to the first tetrahedral is smooth.

I did this step But the same error in prism also produced (see figure)

siw May 25, 2013 15:22

Are you generating the prisms into a volume full of tetrahedral elements or are you generating the prisms first and then filling the remainder of the volume with elements?

Whenever I make a hybrid mesh in ICEM I always follow these steps:

1) Use the Octree method to fill the volume. Delete the volume elements. Check and smooth the surface elements.
2) Use the Delaunay method (TGlib and Use AF = ON) to remake the volume elements. Check and smooth.
3) Generate the prism elements. Check and smooth.

Upload your files as it may be easier to get a fix.

BrolY May 27, 2013 05:40

prism angle = 180 might help.

otherwise, you can find a link to download Simon's tips and tricks for ICEM.
this document would help y ou.

m5edr June 1, 2013 04:19

1 Attachment(s)
Quote:

Originally Posted by siw (Post 430018)

1) Use the Octree method to fill the volume. Delete the volume elements. Check and smooth the surface elements.
2) Use the Delaunay method (TGlib and Use AF = ON) to remake the volume elements. Check and smooth.
3) Generate the prism elements. Check and smooth.

I tried these steps BUT nothing new , the same results of incomplete prism

Quote:

Originally Posted by siw (Post 430018)
Upload your files as it may be easier to get a fix.

I attached my files (the prism setup is found there)

m5edr June 1, 2013 04:25

Quote:

Originally Posted by BrolY (Post 430215)
prism angle = 180 might help.

Otherwise, you can find a link to download Simon's tips and tricks for ICEM.
this document would help y ou.

Yes I read this tips before But 180 prism angle where this angle prevent pyramids but also make worst prism quality.. However thanks for your recommendations :)

BrolY June 3, 2013 02:51

You don't really mind about the prism quality; for example, CFD++ accepts prism with quality < 0.001 ... And prism are almost always bad quality ...So, the min angle would help to fill the prism around the blade, which looks like to be your main issue here !

m5edr June 3, 2013 10:47

Quote:

Originally Posted by BrolY (Post 431564)
You don't really mind about the prism quality; for example, CFD++ accepts prism with quality < 0.001 ... And prism are almost always bad quality

I knew the prisms usually have low quality but Fluent not always comfortable with this low quality and always notify for this issue , and during my little expiercnce , Fluent may be not accept the mesh in some cases , Could you explain more about this?

Quote:

Originally Posted by BrolY (Post 431564)
...So, the min angle would help to fill the prism around the blade, which looks like to be your main issue here !

O.K Thanks, but let me ensure that I'm understand right "The only solution to my case (not all cases) is kept my min. angle = 180"

Thanks in advance

BrolY June 3, 2013 10:51

Not the only one ! But it might help.
Also, there are old thread about that issue ...


All times are GMT -4. The time now is 19:26.