CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Internal Error (

Stephen December 4, 2001 14:16

Internal Error
I am running a combustion case. My case runs well when I am using the first order for my residuals. But when I switch them to second order the case will run for about 100 iterations and then I get this error.

Error: Internal Error at line 743 in file 'amgif.c' divergence detected in AMG solver Error object: ()

Can anyone tell me what this means and what I can do to correct this problem? Like I said the case runs perfect when I am using first order, but it only does this in second order. I have also set my under-relaxations to a very low value of 0.1.

Armin Gips December 5, 2001 03:31

Re: Internal Error
I think it means that the solver finds no solution so it stops the iteration. I had the same probleme, I forgot to scale my geometry. Did you scale it?

Yours Armin

Stephen December 5, 2001 16:20

Re: Internal Error
Yes I scaled it. All the measurments are correct. I'm running it using a under relaxation of 0.1 and using first order for all my residuals and everything is running fine and all my residuals are decreasing. However I am still afraid to try second order since the same thing will happen.

Josť Carlos Espinosa December 10, 2001 07:13

Re: Internal Error
Hello Stephen. The error what appears is the typical error of divergence solution. When the boundary conditions are very critical the convergence of the solution is difficult. For example when I begin a calculate, I usully make it step by step. First I turn off equations of energy and turbulence, first order upwind and low parameters of under relaxation. With this previous calculate convergence, I turn energy and turbulnce equation but I continuous with low parameters. When the solution is convergence again, go up the parameters "in default". And finally after this solution I turn second order for the pressure and momentum variable. Sometimes I conect second order for turbulence and for energy, if the calculate is necessary. Thus, step by step simulation can not convergence with second order solution,and first order solution can be Ok... Regards JC PD: Excuse me my English, please... :)

Stephen December 12, 2001 00:05

Re: Internal Error
Thanks I appreciate the help. The case is working well in first order still. I'm going to run it for a while until I switch it to 2nd order.

All times are GMT -4. The time now is 10:24.