CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Siemens (https://www.cfd-online.com/Forums/siemens/)
-   -   Divergence problem (https://www.cfd-online.com/Forums/siemens/55946-divergence-problem.html)

George March 23, 2007 02:51

Divergence problem
 
Hello to everybody,

I have a problem with my simulation. My solution do not diverge and there are a lot of oscillations after an iteration. I use standard k-e turbulence model. Do someone can advice me what to do? I tried to change relaxation factors of k and e but my solution again diverge. Many thanks in advance.

George

usker March 23, 2007 04:04

Re: Divergence problem
 
Could you please explain in more detail?

tom March 23, 2007 11:47

Re: Divergence problem
 
George, many things can cause divergence. One is the model itself. For example, the inlet may not be able to see the outlet. Check you model using the check tool for cracks and for connectivity.

The other thing is solution method. CG is more stable than AMG and UD is more stable than MARS or others. When changing relaxation factors, also reduce the momentum and pressure relaxation.

Like usker, I would need more info to give any further advice.

Tom

George March 24, 2007 05:19

Re: Divergence problem
 
Yes sorry. I had a problem with the solution. When my run started the monitoring cell show that the solution was going very well. After the 300 iterations the variables started to oscillate and for example velocity u was going from 4 to 6 and then to -1 and then to 6 again. I let the run to finish and after I checked the residuals and I saw that for velocity u it was falling down until 10^-4 and then started to make this oscillations. I read your email tom and tried to change the relaxation factors for u,v,w from 0.7 to 0.65 and my run was ok then. Do you believe this was the problem?

George

Tom March 24, 2007 12:46

Re: Divergence problem
 
George, Yes, proabably was. I am at the point that before I even start, I set momentum at 0.5, pressure at 0.1 and turbulance at 0.5. Tom

George March 24, 2007 13:33

Re: Divergence problem
 
But, Tom is this allowed to reduce all the relaxation factors so much? Because the manual say that no more than 0.05-0.1 must be the reduction.

Tom March 24, 2007 13:34

Re: Divergence problem
 
When I had my first divergence problem, this is what adapco tech support recommended. It worked, so I have continued. Tom


All times are GMT -4. The time now is 06:51.