|
[Sponsors] |
![]() |
![]() |
#1 | |
Senior Member
Guilherme
Join Date: Apr 2017
Posts: 246
Rep Power: 10 ![]() |
Hello,
I have a doubt related to the surface-normal gradient schemes, below is described the log.checkMesh of my mesh. Quote:
However I'm not sure which scheme to use for laplacianSchene. Should I use corrected? orthogonal? limited 1.0? I'm using 'limited 1.0', but I do not know if it's the best for my case ... What would be the difference between them in the results that I will get? |
||
![]() |
![]() |
![]() |
![]() |
#2 |
Senior Member
Guilherme
Join Date: Apr 2017
Posts: 246
Rep Power: 10 ![]() |
Could anyone help me?
|
|
![]() |
![]() |
![]() |
![]() |
#3 |
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Bad Wörishofen
Posts: 2,712
Blog Entries: 6
Rep Power: 52 ![]() ![]() ![]() |
Hi,
personally, an average non-orthogonality of 2.xy is almost zero. I would use the orthogonal scheme or the uncorrected one (limited with limiter zero -> similar to orthogonal but with some under relaxation; somehow like matrix relaxation with factor 1) or the corrected. The limited scheme with a limiter factor > 0 is generally used for higher non-orthogonal meshes because the correction term can cause problems if the angle increases too much > 70°. Thus, it is more for high non-orthogonality). All schemes can be used, but the accuracy might change from second order to first order. A better summary (more or less the same): Surface Normal Gradient Schemes The differences are related to accuracy and stability. The surface-normal gradient calculation is an important quantity.
__________________
Keep foaming, Tobias Holzmann |
|
![]() |
![]() |
![]() |
![]() |
#4 | |
Senior Member
Guilherme
Join Date: Apr 2017
Posts: 246
Rep Power: 10 ![]() |
Quote:
Thanks for the clarification. I recently completed some of the preliminary tests I was doing, and this topic was meant to ''cover'' the lack of knowledge that still existed in my setup. I used the ''limited 1.0'' scheme which is ''similar'' to ''corrected'', but I believe it is not similar due to the clarification you mentioned, in relation to the relaxation coefficients... anyway, I believe that I did not make a bad choice. Now I am thinking of using ''uncorrected'' and later ''orthogonal''. I will post the test results to your knowledge. |
||
![]() |
![]() |
![]() |
![]() |
#5 | |
Super Moderator
Tobias Holzmann
Join Date: Oct 2010
Location: Bad Wörishofen
Posts: 2,712
Blog Entries: 6
Rep Power: 52 ![]() ![]() ![]() |
Quote:
Good luck with your work.
__________________
Keep foaming, Tobias Holzmann |
||
![]() |
![]() |
![]() |
Thread Tools | Search this Thread |
Display Modes | |
|
|
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
problem with setting laplacianSchemes | Bob | OpenFOAM | 2 | April 15, 2009 16:17 |