CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > OpenFOAM > OpenFOAM Running, Solving & CFD

dynamic mesh fails in parallel executation

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   May 7, 2015, 09:58
Default dynamic mesh fails in parallel executation
  #1
Member
 
Hao Chen
Join Date: Aug 2014
Posts: 66
Rep Power: 0
hchen is on a distinguished road
Dear Foamers:

I use dynamic mesh to simulate flow around oscillating cylinders in foam extend 3.1. The dynamic mesh simply update the displacement of the cylinder and call the fvMotionSolver. It works fine in serial case, but for parallel execution, the mesh is moving in a strange way in between the processor boundary. I attached the figures below. The location where the unphysical moving mesh occurs is just in between the domains for processors. I am not sure this is due to the problem of my custom dynamic mesh class (although I do not think so) or other OpenFoam problems?

Best regards
Hao Chen
Attached Images
File Type: jpg ss.jpg (94.1 KB, 124 views)
hchen is offline   Reply With Quote

Old   May 7, 2015, 13:14
Default
  #2
Member
 
Hao Chen
Join Date: Aug 2014
Posts: 66
Rep Power: 0
hchen is on a distinguished road
Hi again,,,

For some other cases, I found that the mesh in the processor boundary does not move at all. In the figure below, I impose a rigid body motion on the cylinder and also the mesh near the cylinder. The rest is moving with fvDisplacementLaplacian. However, the point in the boundary does not move at all. Any suggestions and comments are welcome.
Attached Images
File Type: jpg ss.jpg (97.0 KB, 84 views)
hchen is offline   Reply With Quote

Old   October 22, 2015, 05:01
Default
  #3
Member
 
YS
Join Date: Jan 2010
Posts: 93
Rep Power: 16
Ya_Squall2010 is on a distinguished road
I am having the same problem in running interDyMFoam using multiple CPUs. It even crashes when you run the accompanied floatingObject tutorial case in parallel. Have you identified the source of this problem? Thanks.
Ya_Squall2010 is offline   Reply With Quote

Old   January 27, 2016, 07:35
Default
  #4
New Member
 
hayder
Join Date: Dec 2013
Posts: 3
Rep Power: 12
hayder is on a distinguished road
I am having the same problem in running fluent with dynamic mesh using parallel mode . Have you identified the source of this problem?
If Yes could you please send me some tips. Thanks in advance.
hayder is offline   Reply With Quote

Old   March 15, 2016, 00:06
Default
  #5
Member
 
YS
Join Date: Jan 2010
Posts: 93
Rep Power: 16
Ya_Squall2010 is on a distinguished road
I'd like to know if anyone else is having this problem? Or is there a solution to it?
Ya_Squall2010 is offline   Reply With Quote

Old   March 16, 2016, 01:30
Default
  #6
Senior Member
 
Hrvoje Jasak
Join Date: Mar 2009
Location: London, England
Posts: 1,905
Rep Power: 33
hjasak will become famous soon enough
Please use tet FEM automatic mesh motion solvers and all will be well.

Hrvoje
__________________
Hrvoje Jasak
Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk
hjasak is offline   Reply With Quote

Old   March 16, 2016, 03:47
Default
  #7
Member
 
Hao Chen
Join Date: Aug 2014
Posts: 66
Rep Power: 0
hchen is on a distinguished road
Yes I think FEM is much robuster than FVM for foam extend versions. Otherwise if you still want to use FVM dynamic mesh, you'd better change to openfoam official versions.

Best
Hao
hchen is offline   Reply With Quote

Old   March 16, 2016, 21:50
Default
  #8
Member
 
YS
Join Date: Jan 2010
Posts: 93
Rep Power: 16
Ya_Squall2010 is on a distinguished road
The reason why I tried to switch to fvMotionSolver is that I'd like to compare the BiCGStab to GAMG, as sometimes the former is struggling in 3D cases, even 100+ subiterations are not enough for it to reach convergence.
Ya_Squall2010 is offline   Reply With Quote

Old   March 31, 2017, 03:37
Default
  #9
Member
 
chengan.wang
Join Date: Jan 2016
Location: china
Posts: 47
Rep Power: 10
wangchengan2003 is on a distinguished road
Send a message via Skype™ to wangchengan2003
Dear Jasak
I try to make parallization for the dynamic mesh problem by using the solver pimpleDyMFoam. The orginal case 'movingCone' works well under parallization process. But if "
dynamicFvMesh dynamicMotionSolverFvMesh;

motionSolverLibs ( "libfvMotionSolvers.so" );

solver velocityComponentLaplacian;

velocityComponentLaplacianCoeffs
{
component x;
diffusivity directional ( 1 200 0 );
}
" in the file 'dynamicMeshDict' is simplely changed as "
dynamicFvMesh movingConeTopoFvMesh;

motionSolverLibs ( "libfvMotionSolvers.so" );

movingConeTopoFvMeshCoeffs
{
motionVelAmplitude (1.5 0.0 0.0);
motionVelPeriod 0.003;
leftEdge 0.0;
leftObstacleEdge 0.0;
rightObstacleEdge 0.0;

right
{
minThickness 3e-5;
maxThickness 6e-5;
}

left
{
minThickness 3e-5;
maxThickness 5e-5;
}
}
", which means the mesh has topological changes, the parallization doesn't work any more. (it works fine with single processor).Of2.4.0 is used.

If you could give me some suggestions?

Best regards.
wangchengan2003 is offline   Reply With Quote

Old   March 31, 2017, 03:44
Default
  #10
Member
 
chengan.wang
Join Date: Jan 2016
Location: china
Posts: 47
Rep Power: 10
wangchengan2003 is on a distinguished road
Send a message via Skype™ to wangchengan2003
Dear Jasak,

I try to make parallization for the dynamic mesh problem by using the solver pimpleDyMFoam. The orginal case 'movingCone' works well under parallization process. But if

///////////////////////////////////////////////////
dynamicFvMesh dynamicMotionSolverFvMesh;

motionSolverLibs ( "libfvMotionSolvers.so" );

solver velocityComponentLaplacian;

velocityComponentLaplacianCoeffs
{
component x;
diffusivity directional ( 1 200 0 );
}
///////////////////////////////////////////////////

in the file 'dynamicMeshDict' is simplely changed as


///////////////////////////////////////////////////
dynamicFvMesh movingConeTopoFvMesh;

motionSolverLibs ( "libfvMotionSolvers.so" );

movingConeTopoFvMeshCoeffs
{
motionVelAmplitude (1.5 0.0 0.0);
motionVelPeriod 0.003;
leftEdge 0.0;
leftObstacleEdge 0.0;
rightObstacleEdge 0.0;

right
{
minThickness 3e-5;
maxThickness 6e-5;
}

left
{
minThickness 3e-5;
maxThickness 5e-5;
}
}
///////////////////////////////////////////////////
, which means the mesh has topological changes, the parallization doesn't work any more. (it works fine with single processor). And Of2.4.0 is used.

I think the problem is similar. If you could give me some suggestions?

Best regards,
wangchengan2003 is offline   Reply With Quote

Old   September 26, 2017, 12:22
Default Problem solved in FE 4.0
  #11
New Member
 
alia's Avatar
 
Ali Aghaei
Join Date: Oct 2014
Posts: 12
Rep Power: 11
alia is on a distinguished road
Quote:
Originally Posted by hchen View Post
Hi again,,,

For some other cases, I found that the mesh in the processor boundary does not move at all. In the figure below, I impose a rigid body motion on the cylinder and also the mesh near the cylinder. The rest is moving with fvDisplacementLaplacian. However, the point in the boundary does not move at all. Any suggestions and comments are welcome.
This bug has been fixed in foam-extend 4.0

Cheers.
alia is offline   Reply With Quote

Old   June 14, 2018, 00:30
Default
  #12
Member
 
Join Date: Jan 2018
Location: Malaysia
Posts: 58
Rep Power: 8
jiahui_93 is on a distinguished road
Quote:
Originally Posted by wangchengan2003 View Post
Dear Jasak,

I try to make parallization for the dynamic mesh problem by using the solver pimpleDyMFoam. The orginal case 'movingCone' works well under parallization process. But if

///////////////////////////////////////////////////
dynamicFvMesh dynamicMotionSolverFvMesh;

motionSolverLibs ( "libfvMotionSolvers.so" );

solver velocityComponentLaplacian;

velocityComponentLaplacianCoeffs
{
component x;
diffusivity directional ( 1 200 0 );
}
///////////////////////////////////////////////////

in the file 'dynamicMeshDict' is simplely changed as


///////////////////////////////////////////////////
dynamicFvMesh movingConeTopoFvMesh;

motionSolverLibs ( "libfvMotionSolvers.so" );

movingConeTopoFvMeshCoeffs
{
motionVelAmplitude (1.5 0.0 0.0);
motionVelPeriod 0.003;
leftEdge 0.0;
leftObstacleEdge 0.0;
rightObstacleEdge 0.0;

right
{
minThickness 3e-5;
maxThickness 6e-5;
}

left
{
minThickness 3e-5;
maxThickness 5e-5;
}
}
///////////////////////////////////////////////////
, which means the mesh has topological changes, the parallization doesn't work any more. (it works fine with single processor). And Of2.4.0 is used.

I think the problem is similar. If you could give me some suggestions?

Best regards,
Hi, i am facing exactly the same problem as yours. Had you found the solution? If yes, would u mind n could you please share with me? Thanks a lot

Sincerely,
Jia Hui
jiahui_93 is offline   Reply With Quote

Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
how to set periodic boundary conditions Ganesh FLUENT 15 November 18, 2020 06:09
Dynamic Mesh on Pintle type injector. herntan FLUENT 16 September 4, 2020 08:27
[snappyHexMesh] No layers in a small gap bobburnquist OpenFOAM Meshing & Mesh Conversion 6 August 26, 2015 09:38
[Other] dynamic Mesh with angular velocity not parallel to x axis donQi OpenFOAM Meshing & Mesh Conversion 5 October 2, 2014 23:19
InterDyMFoam dynamic meshing in parallel fails under nonquiescent conditions adona058 OpenFOAM Bugs 7 November 18, 2008 14:58


All times are GMT -4. The time now is 14:24.