Divergence probleme

dahmen · January 28, 2015, 03:09

Hi everybody;
I have a problem that i can't resolve. PLEASE any help is appreciated.
The calculation was stopped and showed me a message "Divergence detected in AMG solver: x-momentum", and in the console the error message was screened and suggested to increase the relaxation sweeps "Divergence detected in AMG solver: x-momentum -> Increasing relaxation sweeps!". What that means, and how to do it?

CeesH · January 30, 2015, 09:33

It means that, for whatever reason, your solution doesn't converge (and in this case the x-momentum equation is the problem).

There are a few things you can try; like decreasing your under-relaxation factors, decreasing your timestep size (in transient), decreasing your mesh size (increases accuracy, but also computation time), or possibly change your model equations.

For example, using 1st order upwind instead of a higher order discretization scheme may stabilize the computation (but at the cost of accuracy). Sometimes first solving the problem in 1st order, and switching to 2nd order once the first order solution converged helps (although it may still diverge then - it's not a magic solution).

Hope that helps!
Good luck,
Cees

dahmen · February 1, 2015, 04:31

thank u very much, Cees. I have tried all these possibilities but the divergence is still happening.

dahmen · February 1, 2015, 08:44

Hallo Cees; apparently with change of mesh density it works (with much longer time), but it works.
So thank u a lot for ur advices.

January 28, 2015, 03:09	Divergence probleme	#1
dahmen New Member adam Join Date: Apr 2013 Posts: 9 Rep Power: 12	Hi everybody; I have a problem that i can't resolve. PLEASE any help is appreciated. The calculation was stopped and showed me a message "Divergence detected in AMG solver: x-momentum", and in the console the error message was screened and suggested to increase the relaxation sweeps "Divergence detected in AMG solver: x-momentum -> Increasing relaxation sweeps!". What that means, and how to do it?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Floating Point Exception Error	nyox	FLUENT	11	November 30, 2018 12:31
[ANSYS Meshing] Help with element size	sandri_92	ANSYS Meshing & Geometry	14	November 14, 2018 07:54
Divergence problem	Smaras	FLUENT	13	February 21, 2013 05:03
3d vof	Smaras	FLUENT	2	February 19, 2013 06:58
Quarter Burner mesh with periosic condition	SamCanuck	FLUENT	2	August 31, 2011 11:34

January 30, 2015, 09:33		#2
CeesH Senior Member Cees Haringa Join Date: May 2013 Location: Delft Posts: 607 Rep Power: 0	It means that, for whatever reason, your solution doesn't converge (and in this case the x-momentum equation is the problem). There are a few things you can try; like decreasing your under-relaxation factors, decreasing your timestep size (in transient), decreasing your mesh size (increases accuracy, but also computation time), or possibly change your model equations. For example, using 1st order upwind instead of a higher order discretization scheme may stabilize the computation (but at the cost of accuracy). Sometimes first solving the problem in 1st order, and switching to 2nd order once the first order solution converged helps (although it may still diverge then - it's not a magic solution). Hope that helps! Good luck, Cees

February 1, 2015, 04:31		#3
dahmen New Member adam Join Date: Apr 2013 Posts: 9 Rep Power: 12	thank u very much, Cees. I have tried all these possibilities but the divergence is still happening.

February 1, 2015, 08:44		#4
dahmen New Member adam Join Date: Apr 2013 Posts: 9 Rep Power: 12	Hallo Cees; apparently with change of mesh density it works (with much longer time), but it works. So thank u a lot for ur advices.