Convergence problems- what can Id o?
 

Hello Foamer,

I tried to simulate a centrifugal fan. After importing the mesh from fluent, I used the AMIMixer-Tutorial for my case-file. Now he solution starts. But after 120iterations the residuen turn up!!
What could be a reason for this phenomena? And waht can I do?

Thanks in advance!
Lancester

Hi,

I'm afraid you will have to supply more information about your case in order for someone to be able to help you. Please have a look at this post to get an idea of what information is usually needed:

http://www.cfd-online.com/Forums/ope...-get-help.html

Cheers,

A

Ok, I try to give you some more information, because at he weekend I also have problems during my simulation.

First the geometry. I try to simulate a radial fan. I used two interfaces. So in sumary I have 3 zones. I attach a picture of the geometry.

First I create the mesh with ansys and export in .msh. So I used the fluent3DMeshToFoam for importing in OpenFOam. When I check the mesh, there is only a fault of skewness, but I read in other threads that his is no problem. I also run the simulation with Ansys CFX and it worked. So I think it couldn't be a problem of mesh.

I tried a lot of options in the case files but without succes. Attached my last residual plot.

What else could I do for a good convergence? After 200 iterations my time step continuity error crows up!

I also give you some more information about my used solvers an options. So I attached the case-file but I delete the mesh-information because of file-size.

It would be really great if anyone can help me!

Thank you,
Lancester

Maybe I have a solution:

I modified my relaxationsFactors and after 800 iterations my solution converge and there are no diverge effects.

For p and U I used relaxationsFactors 0.5!
And for k and omega I used 0.01!

I think for a completely converging solutio I must wait a little bit but I think now it works!

Glad to see you found the problem. I am a bit concerned with you using relaxation factors as low as 0.01 I think it might cause your turbulent properties to remain virtually unchanged throughout the solution process and thus leading to an incorrect solution or very very long convergence times. Perhaps you can try it with something like 0.5 as well and see if it will still be stable?