Divergence in amg-solver: uds-0

Dranzer · June 29, 2015, 06:48

Hello,
I am working on mixing tank simulation. I have developed the flow field for the tank but having problems when using User defined Scalars. i am providing uds source through udf.
Divergence detected in amg solver: uds-0
Divergence detected in amg solver: uds-1
Divergence detected in amg solver: uds-2

I am using urf 0.9 for all. Need help
Thanks in advance.

mvee · June 30, 2015, 00:52

Dear Dranzer

I think the problem is not due to Source1, Source2 and Source3. You can check it by putting them in off mode. UDF code for source 1, 2 and 3 are correct.

The problem is due to return 0. It should be return m_lg.

Best wishes
Vaze

Dranzer · June 30, 2015, 03:23

Hi Vaze,
Thanks for the reply. The problem for me is that even though I am assigning the Scalars in liquid phase, Fluent is still solving it for the regions where there is no liquid. Thats why I returned 0 for that regions, I know my approach is wrong. Can you suggest any other alternatives.
I tried return m_lg but it is not working

pakk · June 30, 2015, 09:02

If you initialized UDS0, UDS1 and UDS2 all to zero, the calculated sources are indetermined (0/0).

So if that is the case, initialize them at a non-zero value, and your problem might go away.

Dranzer · June 30, 2015, 09:09

I am first developing the flow field and then introducing the scalars and then I am patching the zones to a non-zero value for the scalars. Should I initialize them?

pakk · June 30, 2015, 09:12

Quote:

Originally Posted by Dranzer

I am first developing the flow field and then introducing the scalars and then I am patching the zones to a non-zero value for the scalars. Should I initialize them?

If you are patching the zones to a non-zero value, you are initializing them to a non-zero value. Patching is a form of initializing.
So my advice was not helpful this time!

mvee · June 30, 2015, 23:26

Dear Drazner

m_lg will transfer the correct value to mass equation. It may not solve your problem.

Can you please tell me How do you assign particular scalar to liquid phase?

Best wishes
Vaze

mvee · June 30, 2015, 23:56

Number of times I faced same issue, that when I changed UDS for all zones from liquid zones, I have not observed the UDS diffusivity for solids in material panel. You first make all the scalars to zero and then make proper changes.

June 30, 2015, 03:23		#3
Dranzer New Member Join Date: Feb 2015 Posts: 13 Rep Power: 11	Hi Vaze, Thanks for the reply. The problem for me is that even though I am assigning the Scalars in liquid phase, Fluent is still solving it for the regions where there is no liquid. Thats why I returned 0 for that regions, I know my approach is wrong. Can you suggest any other alternatives. I tried return m_lg but it is not working Last edited by Dranzer; June 30, 2015 at 07:49.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Error: Divergence detected in AMG solver	siri	FLUENT	8	March 12, 2018 17:08
Divergence detected in AMG solver	msatrustegui	FLUENT	16	March 20, 2017 09:20
Fluidized Bed: Error: Divergence detected in AMG solver: pressure correction Error Ob	Mole89	FLUENT	5	April 12, 2014 09:32
Divergence detected in AMG solver	tiankongasd	FLUENT	2	March 28, 2013 12:34
AMG solver and floating point divergence	ndabir	FLUENT	1	July 13, 2012 16:20

June 30, 2015, 00:52		#2
mvee Senior Member Vaze Join Date: Jun 2009 Posts: 172 Rep Power: 16	Dear Dranzer I think the problem is not due to Source1, Source2 and Source3. You can check it by putting them in off mode. UDF code for source 1, 2 and 3 are correct. The problem is due to return 0. It should be return m_lg. Best wishes Vaze

June 30, 2015, 09:02		#4
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 26	If you initialized UDS0, UDS1 and UDS2 all to zero, the calculated sources are indetermined (0/0). So if that is the case, initialize them at a non-zero value, and your problem might go away.

June 30, 2015, 09:09		#5
Dranzer New Member Join Date: Feb 2015 Posts: 13 Rep Power: 11	I am first developing the flow field and then introducing the scalars and then I am patching the zones to a non-zero value for the scalars. Should I initialize them?

June 30, 2015, 23:26		#7
mvee Senior Member Vaze Join Date: Jun 2009 Posts: 172 Rep Power: 16	Dear Drazner m_lg will transfer the correct value to mass equation. It may not solve your problem. Can you please tell me How do you assign particular scalar to liquid phase? Best wishes Vaze

June 30, 2015, 23:56		#8
mvee Senior Member Vaze Join Date: Jun 2009 Posts: 172 Rep Power: 16	Number of times I faced same issue, that when I changed UDS for all zones from liquid zones, I have not observed the UDS diffusivity for solids in material panel. You first make all the scalars to zero and then make proper changes.