Convergence problems

saisanthoshm88 · June 1, 2012, 06:42

Could some one please explain if the lack of prisms and the use of a pure tet mesh can be a reason for convergence problems in CFD

Far · June 1, 2012, 09:42

No. May be accuracy is lost, but shouldn't affect the convergence.

cdegroot · June 1, 2012, 15:48

I respectfully disagree. A poor quality mesh can hurt convergence in several ways. Assuming a FV discretization, integration points will be further from the vector connecting neighboring cells, meaning the method becomes more "explicit" and will take more coefficient iterations to converge. Also limiters can be more active, which hurts convergence.

Now I am not saying this in general for tets. It is more mesh quality that can be a problem.

Far · June 2, 2012, 02:39

Quote:

It is more mesh quality that can be a problem.

I agree

Far · June 2, 2012, 11:46

Here is update from CFX forum

Quote:

Originally Posted by ghorrocks

If you are saying you have a turbulent simulation and are simply using a tet mesh with no inflation layer and it is not converging - then no, inflation usually makes convergence harder, not easier. It is unlikely the lack of inflation layers is causing the problem. Something else is causing your convergence issue.

cdegroot · June 2, 2012, 12:51

Another good point. Tets are certainly less accurate in boundary layers in part because of greater false diffusion and as we know a more diffusive scheme is easier to converge. Like I said before if the mesh has very bad quality that could cause convergence issues, but most likely it is something else. You should absolutely be able to get convergence without prisms.

June 1, 2012, 06:42	Convergence problems	#1
saisanthoshm88 Senior Member --------- Join Date: Oct 2010 Posts: 303 Rep Power: 17	Could some one please explain if the lack of prisms and the use of a pure tet mesh can be a reason for convergence problems in CFD __________________ Best regards, Santhosh.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convergence Problems using Spalart Allmaras	recnice	OpenFOAM Running, Solving & CFD	3	October 9, 2013 12:19
Gas-liquid vertical separator, problems with convergence	juliom	Main CFD Forum	0	October 5, 2011 20:20
NACA0012 Convergence Problems	StudentAndrew	CFX	6	November 21, 2005 06:49
Convergence problems	Simone	Siemens	5	June 29, 2005 10:48
Convergence problems	Chetan	FLUENT	3	April 15, 2004 19:13

June 1, 2012, 09:42		#2
Far Super Moderator Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,553 Blog Entries: 6 Rep Power: 54	No. May be accuracy is lost, but shouldn't affect the convergence.

June 1, 2012, 15:48		#3
cdegroot Senior Member Chris DeGroot Join Date: Nov 2011 Location: Canada Posts: 414 Rep Power: 17	I respectfully disagree. A poor quality mesh can hurt convergence in several ways. Assuming a FV discretization, integration points will be further from the vector connecting neighboring cells, meaning the method becomes more "explicit" and will take more coefficient iterations to converge. Also limiters can be more active, which hurts convergence. Now I am not saying this in general for tets. It is more mesh quality that can be a problem.

June 2, 2012, 12:51		#6
cdegroot Senior Member Chris DeGroot Join Date: Nov 2011 Location: Canada Posts: 414 Rep Power: 17	Another good point. Tets are certainly less accurate in boundary layers in part because of greater false diffusion and as we know a more diffusive scheme is easier to converge. Like I said before if the mesh has very bad quality that could cause convergence issues, but most likely it is something else. You should absolutely be able to get convergence without prisms.