CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > OpenFOAM Meshing & Mesh Conversion

Good skewness values in checkMesh response

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

Like Tree4Likes
  • 1 Post By vkrastev
  • 3 Post By roth

Reply
 
LinkBack Thread Tools Display Modes
Old   September 22, 2010, 06:27
Default Good skewness values in checkMesh response
  #1
Senior Member
 
Vesselin Krastev
Join Date: Jan 2010
Location: University of Tor Vergata, Rome
Posts: 361
Rep Power: 10
vkrastev is on a distinguished road
Hello everybody, my question is short and (I hope) clear: I think I've understand that the skewness values reported in the checkMesh response are something different than the "traditional" skewness definition (as, for instance, in Fluent or other commercial packages).
So, apart from their real meaning, I just want to know what is the max skewness value below which the overall mesh quality can be considered satisfactory...
Thank you in advance

V. K.
kiddmax likes this.
vkrastev is offline   Reply With Quote

Old   September 22, 2010, 07:57
Default
  #2
Senior Member
 
Vesselin Krastev
Join Date: Jan 2010
Location: University of Tor Vergata, Rome
Posts: 361
Rep Power: 10
vkrastev is on a distinguished road
Any replies?
vkrastev is offline   Reply With Quote

Old   September 23, 2010, 08:38
Default
  #3
Senior Member
 
Vesselin Krastev
Join Date: Jan 2010
Location: University of Tor Vergata, Rome
Posts: 361
Rep Power: 10
vkrastev is on a distinguished road
Any replies?
vkrastev is offline   Reply With Quote

Old   September 24, 2010, 08:05
Default
  #4
Member
 
Michael Roth
Join Date: Mar 2009
Location: Guelph, Ontario, Canada
Posts: 46
Rep Power: 8
roth is on a distinguished road
There are two sources of information. Skewness is defined in the code here:

OpenFOAM-1.7.x\src\meshTools\cellQuality\cellQuality.C

A picture and a description of how skewness impacts the solution can be found in Dr Jasak's thesis
(especially pg. 124):

http://powerlab.fsb.hr/ped/kturbo/Op...jeJasakPhD.pdf

In words, for a given face, draw a line between the centres of the two cells that share the face. Call this "d". Now find the intersection of d with the face. Call this "fi". Now draw a line between fi and the face center. Call this "m". The skewness is then just mag(m)/mag(d). The cell's skewness is just the max value found after travelling over all the faces in the cell.

Mike
lth, s.m and rajibroy like this.
roth is offline   Reply With Quote

Old   September 24, 2010, 08:21
Default
  #5
Senior Member
 
Vesselin Krastev
Join Date: Jan 2010
Location: University of Tor Vergata, Rome
Posts: 361
Rep Power: 10
vkrastev is on a distinguished road
Quote:
Originally Posted by roth View Post
There are two sources of information. Skewness is defined in the code here:

OpenFOAM-1.7.x\src\meshTools\cellQuality\cellQuality.C

A picture and a description of how skewness impacts the solution can be found in Dr Jasak's thesis
(especially pg. 124):

http://powerlab.fsb.hr/ped/kturbo/Op...jeJasakPhD.pdf

In words, for a given face, draw a line between the centres of the two cells that share the face. Call this "d". Now find the intersection of d with the face. Call this "fi". Now draw a line between fi and the face center. Call this "m". The skewness is then just mag(m)/mag(d). The cell's skewness is just the max value found after travelling over all the faces in the cell.

Mike
Thank you, your answer is very clear. But, however, following your explanation (and by your own experience), can we consider a max skewness value (from the checkMesh report) of about 1 as a good value for an unstructured mesh (mainly made by tetrahedrons with some amount of prisms)?

Thank you once again

V. K.
vkrastev is offline   Reply With Quote

Old   March 11, 2012, 06:58
Default
  #6
New Member
 
Ali Q Raeini
Join Date: Feb 2010
Posts: 21
Rep Power: 7
aliqasemi is on a distinguished road
Quote:
Originally Posted by roth View Post
There are two sources of information. Skewness is defined in the code here:

OpenFOAM-1.7.x\src\meshTools\cellQuality\cellQuality.C

A picture and a description of how skewness impacts the solution can be found in Dr Jasak's thesis
(especially pg. 124):

http://powerlab.fsb.hr/ped/kturbo/Op...jeJasakPhD.pdf

In words, for a given face, draw a line between the centres of the two cells that share the face. Call this "d". Now find the intersection of d with the face. Call this "fi". Now draw a line between fi and the face center. Call this "m". The skewness is then just mag(m)/mag(d). The cell's skewness is just the max value found after travelling over all the faces in the cell.

Mike

the skewness reported by checkMesh seems to be calculated in "Foam:rimitiveMesh::checkFaceSkewness" (see http://foam.sourceforge.net/docs/cpp/a06916_source.html) which seems to use a different algorithm, anyway it gives me max skewness of 2.4 for a uniform orthogonal hexahedral mesh !!!, any comments on why this is the case?

Last edited by aliqasemi; March 11, 2012 at 09:34.
aliqasemi is offline   Reply With Quote

Old   April 6, 2012, 20:58
Default
  #7
Senior Member
 
Gonzalo
Join Date: Mar 2011
Location: Argentina
Posts: 107
Rep Power: 7
gfoam is on a distinguished road
see at the code in http://foam.sourceforge.net/docs/cpp/a06916_source.html that the maximum skewness admissible is 4, so 2.4 may be a good value for skewness
gfoam is offline   Reply With Quote

Old   April 7, 2012, 09:35
Default
  #8
New Member
 
Ali Q Raeini
Join Date: Feb 2010
Posts: 21
Rep Power: 7
aliqasemi is on a distinguished road
Thanks Gonzalo, Yes I figured this out, I am a snappyHexMesh user and this was confusing to me because when I was using a maxInternalSkewness=3 in snappyHexMeshDict, it had no effect. Here is the results of hours of my investigation:

If I remember correctly, there was two definitions for Skewness in openfoam, one calculated for cells (src\meshTools\cellQuality\cellQuality.C) and one for faces (Foam:: PrimitiveMesh::checkFaceSkewness in primitiveMeshCheck.C). The later is reported by checkMesh utility, and is greater than one.
The one which should be used in snappyHexMeshDict is the former one, whose value is, between 0 and 1. But apparently this is not documented anywhere and as a result the snappyHexMesh users usually use a limit for maxInternalSkewness and maxBoundarySkewness greater than one, which does not cause any trouble but is useless.

Correct me if I am wrong here please; I checked these things weeks ago and not thoroughly and I certainly have made mistakes.



aliqasemi is offline   Reply With Quote

Old   April 10, 2012, 21:03
Default
  #9
Senior Member
 
Gonzalo
Join Date: Mar 2011
Location: Argentina
Posts: 107
Rep Power: 7
gfoam is on a distinguished road
Sorry, but i do not use the snappyHexMesh utility, so i can't hel you with this. I use gmsh for the generation of my meshes . But if i'm not wrong the limit for the skewness of 4 is the same for faces and cells. But in this page http://www.openfoam.org/docs/user/snappyHexMesh.php (official page of OpenFOAM, there is one table with examples of values of each keyword on the meshQualityControls sub-dictionary of snappyHexMeshDict, but in that puts a value of 4 for maxInternalSkewness (OK) and a value of 20 for maxBoundarySkewness (??), so i don't know.
gfoam 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
How to delete cells in OF based on CheckMesh? AndTen OpenFOAM Meshing & Mesh Conversion 13 June 24, 2015 04:09
Import netgen mesh to OpenFOAM hsieh Open Source Meshers: Gmsh, Netgen, CGNS, ... 32 September 13, 2011 05:50
Ensight - node values -cell values leo FLUENT 1 May 12, 2010 08:47
Reference Values Help Sham FLUENT 0 February 26, 2005 09:43
PC vs. Workstation Tim Franke Main CFD Forum 5 September 29, 1999 15:01


All times are GMT -4. The time now is 12:50.