[Smaras]

I am doing air injection on water using VOF model with axisymmetric condition. But when i start the calculation it gives me this error


Error: Divergence detected in AMG solver: x-momentum

Please can anyone help

Thanks in advance

Regards,
Smaras

[cfd seeker]

Reduce the under relaxation factors

[Smaras]

Originally Posted by cfd seeker

Reduce the under relaxation factors

Thanks for the prompt answer

Already done that....but still the same problem.
Is it something related to axial or radial direction as i am doing axisymmetric VOF?

Thanks in advance.

[Marion]

Hi Smaras,
Does it run for a few iterations before getting this error?
If you are running 2D-axi, are the mesh checks ok? have you set your domain in the X-Y plane, only positive Y and defined the axis (if you have nodes on Y=0)?
Marion.

[ghost82]

Originally Posted by Smaras

I am doing air injection on water using VOF model with axisymmetric condition. But when i start the calculation it gives me this error


Error: Divergence detected in AMG solver: x-momentum

Please can anyone help

Thanks in advance

Regards,
Smaras

Make x your axisymmetric axis if it isn't.

Daniele

[Smaras]

Originally Posted by ghost82

Make x your axisymmetric axis if it isn't.

Daniele

Thanks Daniele and Marion i was doing a mistake by not taking x-axis as axisymmetric which was giving me problem. Now everything is fine...
Thanks everyone

[niloogh]

Originally Posted by ghost82

Make x your axisymmetric axis if it isn't.

Daniele

hi Daniele
is taking x axis as axisymmetric needed in 3D too???

[jziha]

Try to change the size mesh and if you are employing a tetrahedral mesh change to hexahedral

[niloogh]

Originally Posted by jziha

Try to change the size mesh and if you are employing a tetrahedral mesh change to hexahedral

tnx for your help
but it,s not possible for my geometry to use hexahedral

[zcarral]

Change the mesh, time step or relaxation factors

[niloogh]

Originally Posted by zcarral

Change the mesh, time step or relaxation factors

tnx dear
but i tried a lot but i coudnt make my mesh better(using gambit) and i,m using steady solver.
i reduced the relaxation factors too but it didn,t work.
i,m sure that problem is related to my periodic boundaries.but i don't know what is the problem and what to change about them

[oj.bulmer]

Originally Posted by niloogh

hi Daniele
is taking x axis as axisymmetric needed in 3D too???

No you don't, but you won't run the axisymmetric case in 3D though, because running it as axisymmetric is cheaper.

Can you post your domain/BCs and describe your problem more?

OJ

[niloogh]

Originally Posted by oj.bulmer

No you don't, but you won't run the axisymmetric case in 3D though, because running it as axisymmetric is cheaper.

Can you post your domain/BCs and describe your problem more?

OJ

tnx dear oj
i found that the problem was in one of my walls.i set its coroct boundary
and it is working now

[Honey]

Hello,

I have the same problem though!

For my problem, I am defining most of the boundary conditions (like inlet velocity, turbulence boundary conditions) with UDF file which is successfully interpreted into the Fluent. However, the solution is not converging and the error of "Divergence detected in AMG solver: x-momentum" is showing up.

I have reduced the under relaxation factors to almost -2 order of magnitude for pressure & momentum+ changed the numerical scheme as well as the descretization order+ lowered the initial guess value for the velocity+ converted the tetrahedral cells to hexahedral one BUT still the divergence error appears without any iterations!! nevertheless, after lowering the initial velocity value, the case run for few iterations (less than 15) then again the error came on the screen

So, I have gone through many ways to solve this problem and now have no more idea.

I will really appreciate to hear some suggestion/help from you forum people as I am stuck on my project.

Thank you in advance

[flotus1]

Does your simulation run if you use (reasonable) constant values at the boundary conditions instead of the UDFs?

[Honey]

Thank you dear Alex.

After what you have asked I have tried to (one by one) replace the UDF with the a constant value wherever necessary in order to find out the problem. Finally, I've noticed that there is something wrong with the UDF that defines a resistance profile in the porous zones. However, I was not able to figure it out what causes the problem!

Any idea now??

[flotus1]

You could try to figure it out by yourself, or you could post the UDF here along with all relevant information and see if someone can help.

[niloogh]

Originally Posted by Honey

Hello,

I have the same problem though!

For my problem, I am defining most of the boundary conditions (like inlet velocity, turbulence boundary conditions) with UDF file which is successfully interpreted into the Fluent. However, the solution is not converging and the error of "Divergence detected in AMG solver: x-momentum" is showing up.

I have reduced the under relaxation factors to almost -2 order of magnitude for pressure & momentum+ changed the numerical scheme as well as the descretization order+ lowered the initial guess value for the velocity+ converted the tetrahedral cells to hexahedral one BUT still the divergence error appears without any iterations!! nevertheless, after lowering the initial velocity value, the case run for few iterations (less than 15) then again the error came on the screen

So, I have gone through many ways to solve this problem and now have no more idea.

I will really appreciate to hear some suggestion/help from you forum people as I am stuck on my project.

Thank you in advance

hi dear
my problem was related to the periodic boundaries.after changing some things in gambit ;i had solved my problem.if your problem has periodic boundaries i,ll explan it in details for you.

[dinesh]

Originally Posted by Marion

Hi Smaras,
Does it run for a few iterations before getting this error?
If you are running 2D-axi, are the mesh checks ok? have you set your domain in the X-Y plane, only positive Y and defined the axis (if you have nodes on Y=0)?
Marion.

Hi
My geometry is sym about Y axis, i am giving it sym. But i am getting AMG solver divergence X momentum. I tried changing URF. Also it dont run for even 1 iteration. Can you suggest me the way.
https://www.dropbox.com/sh/udmvmhbiaxsjrl1/Dd9RRnu7OI

thanks

[ghost82]

Originally Posted by dinesh

Hi
My geometry is sym about Y axis, i am giving it sym.

thanks

You must assign X as the axisymmetric axis.

Daniele