Divergence detected in AMG solver
 

Good evening everyone.

I am performing an internal combustion engine simulation using Fluent and I have encountered the following error, over and over again:

"Divergence detected in AMG solver: pressure correction -> Increasing relaxation sweeps!"

I followed pretty much every tutorial I found on IC engines, however I will give more information:

I am using a k-e standard model, pressure-inlet and outlet for boundary conditions, PISO is used for the pressure-velocity coupled scheme, Presto! for pressure, second-order upwind for the rest of the variables with the exception of k and epsilon, since it has some dificutlties converging in the first time-steps. Initially, the valves are closed and before they do open, the solution converges in the iterations. Afterwards, epsilon, again, has great difficulties converging, however I increase the number of iterations and it eventually converges, but very slowly. After i increase the time-step and under-relaxation factors, epsilon does not show signs of diveregence, and nontheless the error shown above appears. So I suppose it is not related to epsilon but yes with pressure or density. Should i adjust the solution-method for either or change the under-relaxation factor? I have been messing around the AMG controls: increasing pre-sweeps to 2, I changed epsilon cycle to V (which helped) and stabilization method to BCGSTAB.

Also, the maximum equi-angle skewness is about 0.79 and and minimum ortho quality is 0.25.

I have about a year of experience in Fluent, however most of what I know has been reading topics and tutorials. I would appreciate any suggestion or opinion. If you need anymore info, feel free to ask.

Thank you very much for your attention.

decrease your pressure relaxation factor to 0.6 from 0.3 and momentum to 0.4 from 0.7

if u have enabled energy, decrease that as well to 0.5 from 1

Thank you for your reply.

My URF's are currently:
Pressure - 0.3
Density - 1
Body Forces - 1
Momentum - 0.5
k - 0.4
epsilon - 0.4
energy - 1

You are suggesting for me to decrease the energy to 0.5, is that correct?
Thank you once again.
Best regards

yea..
just try like this and post the update

Pressure - 0.5
Density - 0.6
Body Forces - 0.6
Momentum - 0.4
k - 0.4
epsilon - 0.4
energy - 0.5

Will do. Thank you for your help. Let's if this time it works!

Best regards

Greatings!

I have tried your suggestion, however without success. I have no idea nor clue what is causing this divergence...
Thanjs for your help ;)

Hi
 

Hi, I am doing the similar study, modeling engine spray combustion using FLUENT! My urf is as followings:
Pressure 0.1
TKE 0.2
TDR 0.2
energy 0.2
However, I still have the same problem. Have you solved it?

Please check your mesh, It usually cause by mesh, also reduce the time step size and URF will also help

Hello Yuhan!
I previously was using PISO for pressure-velocity coupling, however attempted changing it to Coupled and it helped, but not completely. My problem occured a few degrees after the valve was activated. Now the divergence occurs due to temperature reaching its minimum limit in some cells. I have read that, as lehoanganh07 said, it can be the mesh.

Thanks for your reply! ;)