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

[GAMBIT] Few newbie questions

Register Blogs Community New Posts Updated Threads Search

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   February 25, 2011, 20:18
Default Few newbie questions
  #1
New Member
 
Join Date: Dec 2010
Posts: 13
Rep Power: 15
ajkratos is on a distinguished road
Hi all
I just have few basic mesh related qeustion and hoping someone will answer them. If i have 3 different meshes to simulate a jet flow and 1 of the meshes gives me the best result for a single phase jet, will the same mesh give me best result also if simulating two-phase jet flow? I am getting some strange results if i make the mesh finer. For the most coarse mesh (40,000 mesh faces) i get a nicely shaped jet, spreading downstream but as i make the mesh finer (640,000 mesh faces at most), the shape changes significantly and starts looking very uneven. I do not have any data to compare my results with at the moment but for a rough guide what is the relationship between the mesh and results? is finer mesh always better?
ajkratos is offline   Reply With Quote

Old   February 26, 2011, 06:46
Default
  #2
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,663
Blog Entries: 1
Rep Power: 47
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
No finer mesh is not always better. Sometimes the quality drops, depending on the mesh topology. For instance, imagine a case where you have mapped quads to a circle. At the corners, you will have elements where there are 3 out of the 4 nodes of a quad on the perimeter of the circle. If your mesh only had 4 quads, the elements would be decent quality (the angle between the quads would be minimal, but the finer your mesh gets, the more that corner angle opens up... Ultimately as the number of nodes approaches infinity, the corner element angle approaches 180 degrees.

However, if your quad elements against the wall only had 2 nodes each on the circle and were orthogonal (topologically), they would not get worse with refinement...

Refinement usually allows for more smoother flexibility because a smaller percentage of the nodes are constrained by the geometry, so the quality is usually better.
PSYMN is offline   Reply With Quote

Old   February 27, 2011, 11:09
Default
  #3
New Member
 
Join Date: Dec 2010
Posts: 13
Rep Power: 15
ajkratos is on a distinguished road
Hi
thanks for your reply. I get your point although geometery of my problem is very simple and the mesh is a structured mesh but the result changes significantly with increasing the mesh faces. I have attached the pics of the results produced by 3 diffrerent meshes (40,000, 160,000 and 640,000 mesh faces). Also attached the pic of 1 of meshes with 40,000 mesh faces, other 2 meshes look the same except that they are denser. If i dont have anythng to compare the results with how do i determine which mesh provides the best reults. As you can see in the pictures the geometry is very simple and therefore resulting mesh is simple also, so in this case will it be right to say finer mesh=better results ?
Attached Images
File Type: jpg 40000 mesh faces.jpg (42.8 KB, 23 views)
File Type: jpg 160000 mesh faces.jpg (43.0 KB, 20 views)
File Type: jpg 640000 mesh faces.jpg (41.2 KB, 22 views)
File Type: jpg 40000 (mesh).jpg (35.5 KB, 15 views)
ajkratos is offline   Reply With Quote

Old   February 27, 2011, 14:33
Default
  #4
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,663
Blog Entries: 1
Rep Power: 47
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
If the mesh is just mapped quads, then finer should be better... But looking at your post images, the first looks best...

How is the mesh quality in the finer examples?
PSYMN is offline   Reply With Quote

Old   February 28, 2011, 14:00
Default
  #5
New Member
 
Join Date: Dec 2010
Posts: 13
Rep Power: 15
ajkratos is on a distinguished road
Hi
All three meshes are quad mapped meshes. Everytime i made the mesh finer I just increased the number of grid points propotionately in both directions. I have attached 2 pics of zoomed view of the most fine mesh which has 640,000 mesh faces. The result from the simulation using this mesh is in my last post. Thanks
Attached Images
File Type: jpg 640000.jpg (101.2 KB, 7 views)
File Type: jpg 640000(2).jpg (100.7 KB, 5 views)
ajkratos is offline   Reply With Quote

Old   February 28, 2011, 22:26
Default Perfect...
  #6
Senior Member
 
PSYMN's Avatar
 
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,663
Blog Entries: 1
Rep Power: 47
PSYMN has a spectacular aura aboutPSYMN has a spectacular aura about
So the mesh quality in both cases is pretty much perfect... Perfect angles, perfect (lack of ) skewness, perfect uniform (lack of) transition, aspect ratio of 1, etc. Perfect, maybe pretty inefficient (more elements than you need in some directions), but perfect quality.

The mesh doesn't get better when it is refined because it was already perfect, but taking smaller chunks at a time should let you resolve the flow features better...

You have nothing wrong with your mesh.

Perhaps it is just capturing solver problems better?

Now you need to post this interesting question on the solver boards ( if you have not already)... Please let us know how it turns out.
PSYMN is offline   Reply With Quote

Reply


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 Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Newbie to OpenFoam - meshing questions monti OpenFOAM 2 January 3, 2011 10:07
total newbie - Basic Questions caslor Main CFD Forum 9 December 10, 2009 11:49
Newbie Navier Stokes Questions + Grid Jigar Main CFD Forum 4 February 6, 2006 12:14
Some Newbie Questions Please Ogbeni FLUENT 1 February 3, 2005 04:08
Newbie questions Anders Olofsson Main CFD Forum 7 September 1, 2000 16:19


All times are GMT -4. The time now is 15:53.