[jsm]

Hi,

In ICEM CFD, how to avoid the pyramid elements generation while making prism layers. There is any best practices are available?

Kindly give your suggestions.

[rikio]

Generate one layer of prism first, and then split this layer into as many layers as u want.

[jsm]

Hi,

Thanks for your suggestion.

[PSYMN]

I did intend to get to this, but I am traveling now and my internet connection is not good enough to post the ppt...

I will try later this week.

Simon

[jsm]

Hi Simon,

Thanks for your efforts. Please post your presentation at your convenient time.

[apantune]

JSM,
this might help you....

Mesh -> Global Mesh Setup -> Prism Meshing Parameter -> Advanced Prism Meshing Parameters -> Turn Off : Stop Columns, Refine Prism Boundary Turn On : Auto Reduction.

This setup will stop your prism layer before it intends to generate a pyramid element. You might not get the desired prism height.

Alex

[apantune]

Originally Posted by rikio

Generate one layer of prism first, and then split this layer into as many layers as u want.

Rikio,
if you have a very complex geometry and try to raise one single prism layer (with the total height to afterwards split it) the tetrahedrals sometimes can not be adequately acomodated. With this approach you smash the tetrahedrals by a certain amount in just one single steps (on the step by step approach the tetra are smoothed during the proccess). For complex geometries this can bring you some problems.

Regards,
Alex

[jsm]

Hi Alex,

Thanks for your suggestion. I am looking for similar kind of tips and tricks for prism mesh generation.

FYI : Simon will also post one presentation for prism mesh generation.

[PSYMN]

Sorry for the long delay… The ppt is here ftp://ftp.ansys.com/outgoing/simon/Prism_NoPyramids.pdf

The ppt was originally made for a customer using a combination of Fluent and CFD++, neither of which like pyramids very much, on models that were typically hundreds of millions of elements (they didn't want to just "try things").

Other solvers, such as CFX, would prefer pyramids to poor quality prisms, so this may not be the best settings for everyone... Which is why they are not the hard wired defaults... ICEM CFD is all about flexibility.

There are also some general prism tips in there.

[Far]

Hi PSYMN

I want to write mesh for CFD++. Would you please guide me to set the boundary conditions for CFD++ and then output the mesh.

[dnsantosh]

hi
am trying to generate tri/prism on my geometry which ha V shape , wen i try to generate prism by icem cfd it generates penetrating elements and pyramids, can anybody tell me how to avoid penetrating elements between V ( V angle 73 deg sides 80mm, 65mm and need 30mm prism height)

[rikio]

Set the smoothing step to 0 under Global Prism Setting tab. Manually smooth after the prism generation.

[PSYMN]

Another option (depending on your version) would be to go into Prism advanced settings and click the option for PrismV10... I think that was sorted out in later versions, but I forget when exactly.

[shinde.gopal]

Originally Posted by PSYMN

Sorry for the long delay… The ppt is here ftp://ftp.ansys.com/outgoing/simon/Prism_NoPyramids.pdf

The ppt was originally made for a customer using a combination of Fluent and CFD++, neither of which like pyramids very much, on models that were typically hundreds of millions of elements (they didn't want to just "try things").

Other solvers, such as CFX, would prefer pyramids to poor quality prisms, so this may not be the best settings for everyone... Which is why they are not the hard wired defaults... ICEM CFD is all about flexibility.

There are also some general prism tips in there.

Thanks for preparing the stuff. But, I am not able to access the ftp. Can you please let me know if the file is still available at the same path and ftp is up. Thanks for the help.

[Far]

http://www.cfd-online.com/Forums/ans...efinement.html

Enjoy !!!

[PSYMN]

@ Shinde Gopal,

That pdf was from 2009... But I guess it is still relevant.

That whole FTP site was taken down. We have switched to Google.

Here is the google link.

https://docs.google.com/file/d/0ByIL...it?usp=sharing

[shinde.gopal]

Thanks a lot

Originally Posted by PSYMN

@ Shinde Gopal,

That pdf was from 2009... But I guess it is still relevant.

That whole FTP site was taken down. We have switched to Google.

Here is the google link.

https://docs.google.com/file/d/0ByIL...it?usp=sharing

[shinde.gopal]

Originally Posted by PSYMN

@ Shinde Gopal,

That pdf was from 2009... But I guess it is still relevant.

That whole FTP site was taken down. We have switched to Google.

Here is the google link.

https://docs.google.com/file/d/0ByIL...it?usp=sharing

Thanks alot

[shinde.gopal]

@ Admin : Please delete this post

[haristrawberry]

Hi,

Can anyone post the PPT or send it to haristrawberry@gmail.com


Thanks
Hari