CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CD-adapco (https://www.cfd-online.com/Forums/cd-adapco/)
-   -   v4 & AMG warning msg (https://www.cfd-online.com/Forums/cd-adapco/56393-v4-amg-warning-msg.html)

more work December 19, 2007 05:56

v4 & AMG warning msg
 
Hi! I'm now running v4 but using trim mesh created in v3. It ran quite well earlier without porous media. However, now in a model containing porous media I got this warning msg below but only after 45 iterations onwards. Kindly someone let me what this warning message means.

45 2.75E-01 1.00E+00 7.00E-01 -- -- -- -- 1.03E+05-6.73E+13 4.40E-01 -- -- -- -- FINISH ITERATION NO. 45 CPU time is 306.86 Elapsed time is 307.11

46 3.82E-01 1.00E+00 2.81E-01 -- -- -- -- 3.04E+05-1.05E+14 6.09E-01 -- -- -- -- FINISH ITERATION NO. 46 CPU time is 311.89 Elapsed time is 312.14 WARNING: Ap = 0 on multigrid level 1, nRows = 49184, blockSize = 1

AMG coarsening halted. WARNING: Ap = 0 on multigrid level 1, nRows = 49974, blockSize = 1

AMG coarsening halted. WARNING: Ap = 0 on multigrid level 1, nRows = 50823, blockSize = 1

AMG coarsening halted.

47 2.92E-01 1.00E+00 6.90E-01 -- -- -- -- 2.73E+05-1.06E+14 1.67E+00 -- -- -- -- FINISH ITERATION NO. 47 CPU time is 400.92 Elapsed time is 401.17 WARNING: Ap = 0 on multigrid level 1, nRows = 49011, blockSize = 1

AMG coarsening halted. WARNING: Ap = 0 on multigrid level 1, nRows = 50353, blockSize = 1

AMG coarsening halted. WARNING: Ap = 0 on multigrid level 1, nRows = 49831, blockSize = 1

AMG coarsening halted.


rcm March 7, 2012 00:34

This is due to small un conected surfaces in mesh
To rectify this right click and select repair surface in Representation >> remeshed surface, which you can find below plots

A new window called Surface Prepration Options will pop up. Click OK

In the side pane, select mode as "Edit conected surfaces" from the drop down and click start diagonosis

it will display more than one line of surface groups in a table.
Highlight and delete the small unconnected surfaces and close the window.

Select generate volume mesh to remesh

fshak92 March 28, 2012 08:42

Quote:

Originally Posted by rcm (Post 348025)
This is due to small un conected surfaces in mesh
To rectify this right click and select repair surface in Representation >> remeshed surface, which you can find below plots

A new window called Surface Prepration Options will pop up. Click OK

In the side pane, select mode as "Edit conected surfaces" from the drop down and click start diagonosis

it will display more than one line of surface groups in a table.
Highlight and delete the small unconnected surfaces and close the window.

Select generate volume mesh to remesh

Do we have to regenerate a volume mesh after using 'remove invalid cells in region' ? or it is compulsory just after using surface repair?
Thank you in advance.

rcm March 28, 2012 09:56

yes. generate volume mesh and switch on cell quality remediation

fshak92 March 29, 2012 04:47

Quote:

Originally Posted by rcm (Post 351939)
yes. generate volume mesh and switch on cell quality remediation

Thank you for your answer. :)

Just a question.After removing some cells by 'remove invalid cells' ,first i did not regenerate volume mesh.So we should expect the residual isn't changed.
but it was changed a lot.
What does it mean?

Thank you in advance.

Sprgopack December 28, 2016 23:19

I have tried clicking on Automated Mesh.Remesh underneath Representations >> Geometry, but I do not get any option to select surface repair.

singlav January 30, 2017 14:12

Floating point error :coupled flow.coupled implicit solver
 
Hi

I am running a DEM simulation for mixing powders. Simulation was running for more than 2 months on a high computing cluster when today this error popped up-

' A floating point exception has occurred: floating point exception [Invalid operation]. The specific cause cannot be identified. Please refer to the troubleshooting section of the User's Guide.
Context: star.coupledflow.CoupledImplicitSolver'

In the output window it says :
Bi-conjugate Gradient Stabilized solver didnot converge !
*CFL 50 ->5....

WARNING: Insufficient precision on multigrid level 6, nRows=6
AMG Coarsening halted.

I got floating point exception error back in October when I was setting up simulation. At that time I resolved it by making the mesh finer. But now since the simulation is running for more than 2 months I can't go back.


Someone please help urgently, what should I do now?

arjun February 8, 2017 21:24

Quote:

Originally Posted by singlav (Post 635185)
Hi

I am running a DEM simulation for mixing powders. Simulation was running for more than 2 months on a high computing cluster when today this error popped up-

' A floating point exception has occurred: floating point exception [Invalid operation]. The specific cause cannot be identified. Please refer to the troubleshooting section of the User's Guide.
Context: star.coupledflow.CoupledImplicitSolver'

In the output window it says :
Bi-conjugate Gradient Stabilized solver didnot converge !
*CFL 50 ->5....

WARNING: Insufficient precision on multigrid level 6, nRows=6
AMG Coarsening halted.

I got floating point exception error back in October when I was setting up simulation. At that time I resolved it by making the mesh finer. But now since the simulation is running for more than 2 months I can't go back.


Someone please help urgently, what should I do now?

Mesh change is only thing you can do here.

When I was at cd adpaco i proposed a fix with that this issue will never crop up because AMG then detect Ap 0 and replace it with something appropiate.

Since AMG was not my area of work they decided they will not put the fix in.

I looked into this problem because another customer was not able to run his simulation and this cropped up any time during simulation.

So fix exist and it still exist as review code at adapco but i do not think they will ever put it in.

singlav February 10, 2017 00:17

This is sad, this simulation is my senior year project and I can't go back to reset the mesh and start the whole simulation again. I wish cd-adapco could have done something for this.


All times are GMT -4. The time now is 14:21.