CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Phoenics (https://www.cfd-online.com/Forums/phoenics/)
-   -   Convergence (https://www.cfd-online.com/Forums/phoenics/52225-convergence.html)

CD January 22, 2008 09:09

Convergence
 
Hi, Wonder if anyone can give me some tips on how to get your simulation to converge? I assume the mesh quality is essential for convergence? Also how do you know if its converged? If it flatlines can you assume it has converged? Any help would be great thanks!

PHOENICS January 23, 2008 07:14

Re: Convergence
 
Some advice is given here:

http://www.cham.co.uk/phoenics/d_pol...s/converge.htm


Patti January 25, 2008 13:13

Re: Convergence
 
I've found that convergence varies widely with the types of models you're using and the types of simulation you're trying to run. For instance, I've found that combustion models benefit from the use of the (old) SARAH setting.

EF March 18, 2008 18:15

Re: Convergence
 
Hy,

Linear relaxation should be used for Pressure and ke / ep. I found 0.4 for P1, 0.3 for KE and 0.4 for ep work nice for much cases.

Velocity relaxation depends much on the cell diameter and the velocity in the smalles cell. Choose the velocity relaxation factor as a false time step equal or lower than the shortest time the flow needs to cross the cell. Example: For a cell of 100mm and a flow speed of 10m/s the relaxation factor has to be lower than 0.01

Temperature relaxation can be a factor of 10 - 50 higher than velocity relaxation.

Most problems with relaxation come from small cells or a bad grid. Try without parsol and use higher tolerance to find problems.

best regards, ef

PattiMichelle Sheaffer May 6, 2008 16:45

Re: Convergence
 
Yes, this is consistent with what I've found (though I've never used parsol with chemkin) - thanks for the summary!

Patti


All times are GMT -4. The time now is 13:43.