CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > ANSYS Meshing & Geometry

[ICEM] Prism layers around interior wall

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   December 7, 2009, 09:32
Default Prism layers around interior wall
  #1
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
Im having some issues with prism layers around interior walls:

1) at some places, the number of layers seems to decrease and then increase back to my specified amount. Any idea why this could be happening?

2) At the end of the interior wall, the prisms seem to taper off to nothing. I have seen other results where this is done properly, but can't seem to mimic it. I have tried lowering my minimum quality to 1e-6 which fixed some areas, but not all. I get the following message when generating my prism using the batch utility:

max_prism_angle reset to 179.943 because the min_prism_quality is 1e-006

Do I need to set the min quality even lower, or is there some other root problem?
Attached Images
File Type: png 1.PNG (23.6 KB, 177 views)
File Type: png 2.png (12.7 KB, 143 views)
rwryne is offline   Reply With Quote

Old   December 8, 2009, 03:44
Default
  #2
Member
 
zeitistgeld
Join Date: Apr 2009
Posts: 37
Rep Power: 8
zeitistgeld is on a distinguished road
keep the maximum angle 180 degs, and set advanced prism parameters: deselect stairstep
zeitistgeld is offline   Reply With Quote

Old   December 8, 2009, 09:56
Default
  #3
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
Quote:
Originally Posted by zeitistgeld View Post
keep the maximum angle 180 degs, and set advanced prism parameters: deselect stairstep
I do have the maximum angle set at 180*, but it automatically lowers it to adjust for the quality. I have it set at 1e-6, but do not know if it is "safe" to go lower.

Will try deselecting the stair step option.


Thanks,

Ryne

edit:

Your suggestion of disabling stair step did indeed fix problem 1. Still am having troubles with problem 2.

Last edited by rwryne; December 8, 2009 at 12:23.
rwryne is offline   Reply With Quote

Old   December 8, 2009, 13:07
Default Baffles...
  #4
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,662
Blog Entries: 1
Rep Power: 35
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
If it is a zero thickness baffle, then the stair step is helping you out with quality. In 12.1, you can turn the stair stepping off for baffles, but a better method would be to extrude the end of the baffle further out and call the new surface (interior). Then the baffle will have full prisms and the stair stepping can happen on the interior surface.

I will attach some pics.

Simon
Attached Images
File Type: jpg Baffle_Stair.jpg (93.8 KB, 158 views)
File Type: jpg Baffle_Stair_not.jpg (94.5 KB, 150 views)
PSYMN is offline   Reply With Quote

Old   December 8, 2009, 15:57
Default
  #5
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
Quote:
Originally Posted by PSYMN View Post
If it is a zero thickness baffle, then the stair step is helping you out with quality. In 12.1, you can turn the stair stepping off for baffles, but a better method would be to extrude the end of the baffle further out and call the new surface (interior). Then the baffle will have full prisms and the stair stepping can happen on the interior surface.

I will attach some pics.

Simon

Simon:

In response to your suggestion, the shell geometry that I am trying to generate a prism layer around is too complex to extrude further unfortunately.

I am trying to recreate someone's mesh of the same geometry, but my prism layers around the end of these shell elements is coming out quite different.

Theirs appears not to use stepping, but rather to just convergence the prisms to a point at the end. I tried to turn stepping off to recreate this, but did not get the same results.

Is there another setting that might control this end convergence?

Thanks,

Ryne
rwryne is offline   Reply With Quote

Old   December 9, 2009, 10:39
Default Version?
  #6
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,662
Blog Entries: 1
Rep Power: 35
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
Were you using 12.1? This doesn't work in 12.0.
PSYMN is offline   Reply With Quote

Old   December 9, 2009, 10:56
Default
  #7
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
Quote:
Originally Posted by PSYMN View Post
Were you using 12.1? This doesn't work in 12.0.

Ahh, great

IT already hates me! Now I'll have to pressure more for them to install 12.1 ASAP. Hopefully 12.1 will kill two birds with one stone and fix my workbench issues as well.

Thanks,

Ryne
rwryne is offline   Reply With Quote

Old   December 9, 2009, 12:11
Wink
  #8
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,662
Blog Entries: 1
Rep Power: 35
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
Send the guys (or dolls) in IT some donuts.
PSYMN is offline   Reply With Quote

Old   December 9, 2009, 12:12
Default
  #9
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
Simon:

Just to clarify, I made a few pictures of what I would ideally like, and what I have seen (refereed to below as "Acceptable") that I cannot recreate.

Is there a way to mimic these pictures? All my attempts end with a very weird ending for the prism, sometimes with the prism layers converging to a point OFF of the shell element.

Thanks again,

Ryne


Edit: I think I have achieved the "acceptable" one by doing a single prism layer and splitting it into my number of layers.
Attached Images
File Type: png desired.png (3.5 KB, 97 views)
File Type: png accept.png (8.4 KB, 97 views)

Last edited by rwryne; December 9, 2009 at 14:33.
rwryne is offline   Reply With Quote

Old   December 10, 2009, 18:10
Default Sure you can, with 12.1.
  #10
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,662
Blog Entries: 1
Rep Power: 35
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
Sure, see my galleon pics higher up in this thread. With 12.1 and stair stepping off, your prisms will go to the end of the baffle.

By default, it will then create a bunch of pyramids all terminating along a line. If you don't like that, just delete all the tetras and pyramids in the model (keep the prisms, Hexas and shells) and then generate bottom up tetra (delaunay TGlib with the AF option). This will recreate the rest of the volume and build individual pyramids on the side of each prism.

(Alternatively, you could select all the pyramids with a subset and then add attached tetras to them (using the subset modify controls), then delete just the elements in the subset and remesh using the fill holes option. This will be quicker and won't require remeshing the whole model. )

A word of warning here, a very thin prism has a high side aspect ratio and will produce a low quality pyramid. That is just reality. On the other hand, your solver might not mind.

Simon
PSYMN is offline   Reply With Quote

Old   December 11, 2009, 00:55
Default
  #11
jsm
Senior Member
 
JSM
Join Date: Mar 2009
Location: India
Posts: 155
Rep Power: 11
jsm is on a distinguished road
Hi Simon,

I have one question in your last reply. Kindly clarify this.
From your reply, I understood that we can delete the pyramids and tetra elements without deleting shells and prisms. I checked the deleting elements option by element type in edit mesh tab. I could not find it. If all volume elements (prism, pyramids and tetra) are in single part name, then how to do this?


Thanks in advance.
__________________
With regards,
JSM
jsm is offline   Reply With Quote

Old   December 14, 2009, 16:02
Default
  #12
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
My original issue (number 1) is reoccurring in a stronger fashion. The prism just drops randomly along the interior wall and then picks back up. I have tried turning stair stepping off, but the same thing happens.

Any ideas?

(sorry for bad pic quality, can't get it under file size as a PNG)
Attached Images
File Type: jpg crap.jpg (74.1 KB, 82 views)
rwryne is offline   Reply With Quote

Old   December 14, 2009, 16:44
Default How did you do that?
  #13
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,662
Blog Entries: 1
Rep Power: 35
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
I don't think I could make ICEM CFD do that on purpose...

Why is prism growing in the one direction but no the other?

Perhaps this is something I could solve if I had the model in front of me, but I am not understanding how you arrived at this point.

Simon
PSYMN is offline   Reply With Quote

Old   December 15, 2009, 09:12
Default
  #14
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
Quote:
Originally Posted by PSYMN View Post
I don't think I could make ICEM CFD do that on purpose...

Why is prism growing in the one direction but no the other?

Perhaps this is something I could solve if I had the model in front of me, but I am not understanding how you arrived at this point.

Simon

Glad it is causing confusing for more people than just me for once

It was generated as 1 layer and then cut into 17 or so, but this occurs if I generate 17 initially too. It always happens in the same spot, on both sides (somewhat symmetric through the centerline). The really weird part is that if I lower the overall height of the prism, it generates in this area but messes up at the end of the interior wall (see attached)

I'll look into emailing it to you, but I am not sure if I will be able to.
Attached Images
File Type: jpg weird.jpg (66.7 KB, 88 views)
rwryne is offline   Reply With Quote

Old   December 15, 2009, 10:36
Default Max Height over Base.
  #15
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,662
Blog Entries: 1
Rep Power: 35
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
Now seeing this latest image and knowing that the other was meshed as a single layer, I would check for max height over base. (under global prism params) Perhaps you have this set to something like 0.5 and it is stopping the prism columns to prevent exceeding this value.
PSYMN is offline   Reply With Quote

Old   December 15, 2009, 11:13
Default
  #16
Super Moderator
 
Ryne Whitehill
Join Date: Aug 2009
Posts: 313
Rep Power: 9
rwryne is on a distinguished road
As always, thanks for the help Simon.

I think I got it coming out how I wanted it now.

-Ryne
rwryne is offline   Reply With Quote

Old   February 1, 2013, 09:44
Default
  #17
New Member
 
Faisal
Join Date: Sep 2012
Posts: 2
Rep Power: 0
caeiro_faisal is on a distinguished road
How to delete pyramids without deleting tetras and prisms, meaning they are all a part of subset volume and are not separately defined. So how to delete them individually
caeiro_faisal is offline   Reply With Quote

Old   February 1, 2013, 10:25
Default
  #18
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,662
Blog Entries: 1
Rep Power: 35
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
Quote:
Originally Posted by caeiro_faisal View Post
How to delete pyramids without deleting tetras and prisms, meaning they are all a part of subset volume and are not separately defined. So how to delete them individually
In the model tree on the left, you can expand the volume mesh branch and turn off elements by type... If you display only the pyramids, you can select them with a box or use the "Select all visible" option from the selection tool bar.

Actually, if you are using a newer version of ICEM CFD, you can use the selection tool bar directly... On the right side of the Select Mesh Elements tool bar, you should see the options to Select all volume elements. If you have a more recent version (such as 14.5) you will see a pull down on that option that lets you check or uncheck what sort of volume elements you want it to select. You could unselect the tetrahedral and prism elements and then click the button to select just the pyramids...

EditMesh_SelectPyramids.jpg

That said, I am not sure you want to delete all the pyramids... This will leave voids and uncovered faces. What is your actual goal?
__________________
-----------------------------------------
Please help guide development at ANSYS by filling in these surveys

Public ANSYS ICEM CFD Users Survey

This second one is more general (Gambit, TGrid and ANSYS Meshing users welcome)...

CFD Online Users Survey
PSYMN is offline   Reply With Quote

Old   February 1, 2013, 11:57
Default
  #19
New Member
 
Faisal
Join Date: Sep 2012
Posts: 2
Rep Power: 0
caeiro_faisal is on a distinguished road
I am having 12.1 on one of the computers and 13 on the other .. Possibly I would be able to update to 14.5 as both are perpetual licences but that will take some time. I am having highly skewed tetras in the volume cos im growing nearly 40 layers by marching them. Also Im talking of meshes of nearly 30 million cells including prism ... so Ill take your advice and put on tetras in the tree and delete them :P i know my system will hang for a while ill switch to wireframe How do I grow tetras over the prism after deleting tetras? Is delauney possible in icem13?
caeiro_faisal is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
[ICEM] Surface/Volume orientation errors growing prism layers jlichtwa ANSYS Meshing & Geometry 9 April 1, 2012 11:03
Getting prism to inflate into mixed tet-hex meshes Joe CFX 16 October 10, 2011 07:06
ICEM Prism Layer transition between surface with prism layers and one without TWaung ANSYS Meshing & Geometry 2 October 12, 2009 14:56
How many prism layers should be added in a pump simulation? Jasmine CFX 3 June 17, 2009 16:05
Icemcfd: Preventing prism inflation collisions? Joe CFX 1 July 31, 2007 09:13


All times are GMT -4. The time now is 03:01.