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

Moving mesh part 2

Register Blogs Community New Posts Updated Threads Search

Like Tree1Likes

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   January 15, 2008, 18:49
Default OpenFOAM/OpenFOAM-1.4.1/bin/mp
  #21
Senior Member
 
Mattijs Janssens
Join Date: Mar 2009
Posts: 1,419
Rep Power: 26
mattijs is on a distinguished road
OpenFOAM/OpenFOAM-1.4.1/bin/mpirunDebug? Or could be that I put it there.

mpirunDebug
mattijs is offline   Reply With Quote

Old   January 29, 2008, 18:02
Default Hi Frank, How are you doing
  #22
sek
Member
 
Sung-Eun Kim
Join Date: Mar 2009
Posts: 76
Rep Power: 17
sek is on a distinguished road
Hi Frank,

How are you doing, my friend? Clearly you're making progress!

Would you mind making your simple moving/deforming, FSI cases availabble to others?

SE
sek is offline   Reply With Quote

Old   January 29, 2008, 18:25
Default Hi Sung-Eun! Nice to hear f
  #23
Senior Member
 
Frank Bos
Join Date: Mar 2009
Location: The Netherlands
Posts: 340
Rep Power: 18
lr103476 is on a distinguished road
Hi Sung-Eun!

Nice to hear from you! Basically, I am preparing to share a small collection of different dynamic mesh classes, involving: 1 moving body, 2 moving bodies, with and without subsetMeshes and some applied (defined) flexibility. Unfortunately, I have done only little on FSI, since for my problem I know the wing motion / flexing, so there is no need to couple the forces to the structure. But I do have some test cases which I will share.

If you need anything on the short term, please drop me an email....

Regards, Frank

PS: maybe we can meet in Milan at the OF workshop ??
__________________
Frank Bos
lr103476 is offline   Reply With Quote

Old   January 29, 2008, 20:23
Default Hi, Frank, I am wondering i
  #24
Senior Member
 
Pei-Ying Hsieh
Join Date: Mar 2009
Posts: 317
Rep Power: 18
hsieh is on a distinguished road
Hi, Frank,

I am wondering if you can send me your dynamic mesh stuffs when you are ready?

Thanks!

Pei
email: phsieh2005@yahoo.com
hsieh is offline   Reply With Quote

Old   March 6, 2008, 15:08
Default I'm also having problems runni
  #25
Senior Member
 
Kevin Smith
Join Date: Mar 2009
Posts: 104
Rep Power: 17
kev4573 is on a distinguished road
I'm also having problems running the mesh subset motion solver in parallel. Here is the error which occurs right away when trying to run in parallel.


Selecting dynamicFvMesh subsetMotionSolverFvMesh
Selecting motion solver: laplaceFaceDecomposition
Selecting motion diffusivity: quadratic
*** glibc detected *** malloc(): memory corruption: 0x0000000000791140 ***
*** glibc detected *** malloc(): memory corruption: 0x0000000000811220 ***
[compute-2-1:13295] *** Process received signal ***
[compute-2-1:13295] Signal: Aborted (6)
[compute-2-1:13295] Signal code: (-6)
[compute-2-1:13295] [ 0] /lib64/tls/libc.so.6 [0x35b802e2b0]
[compute-2-1:13295] [ 1] /lib64/tls/libc.so.6(gsignal+0x3d) [0x35b802e21d]
[compute-2-1:13295] [ 2] /lib64/tls/libc.so.6(abort+0xfe) [0x35b802fa1e]
[compute-2-1:13295] [ 3] /lib64/tls/libc.so.6 [0x35b8063291]
[compute-2-1:13295] [ 4] /lib64/tls/libc.so.6 [0x35b8069881]
[compute-2-1:13295] [ 5] /lib64/tls/libc.so.6(malloc+0x92) [0x35b806b272]
[compute-2-1:13295] [ 6] /usr/lib64/libstdc++.so.6(_Znwm+0x2a) [0x35b91af50a]
[compute-2-1:13295] [ 7] /usr/lib64/libstdc++.so.6(_ZNSs4_Rep9_S_createEmmRKSaIcE+0x7e ) [0x35b919024e]
[compute-2-1:13295] [ 8] /usr/lib64/libstdc++.so.6 [0x35b919260b]
[compute-2-1:13295] [ 9] /usr/lib64/libstdc++.so.6(_ZNSsC2EPKcRKSaIcE+0x43) [0x35b9192783]
[compute-2-1:13295] [10] /home/krs289/OpenFOAM/OpenFOAM-1.4.1-dev/lib/linux64GccDPOpt/libfiniteVolume.so( _ZN4Foam12fvMeshS
ubset18setLargeCellSubsetERKNS_4ListIiEEiib+0x605) [0x2a9688b1f5]
[compute-2-1:13295] [11] /home/krs289/OpenFOAM/OpenFOAM-1.4.1-dev/lib/linux64GccDPOpt/libfiniteVolume.so( _ZN4Foam12fvMeshS
ubset18setLargeCellSubsetERKNS_7HashSetIiNS_4HashI iEEEEib+0x141) [0x2a9688d451]
[compute-2-1:13295] [12] /home/krs289/OpenFOAM/OpenFOAM-1.4.1-dev/lib/linux64GccDPOpt/libdynamicFvMesh.so (_ZN4Foam24subset
MotionSolverFvMeshC1ERKNS_8IOobjectE+0xd64) [0x2a9557bd34]
[compute-2-1:13295] [13] /home/krs289/OpenFOAM/OpenFOAM-1.4.1-dev/lib/linux64GccDPOpt/libdynamicFvMesh.so (_ZN4Foam13dynami
cFvMesh29addIOobjectConstructorToTableINS_24subset MotionSolverFvMeshEE3NewERKNS_ 8IOobjectE+0x31) [0x2a9557ed91]
[compute-2-1:13295] [14] /home/krs289/OpenFOAM/OpenFOAM-1.4.1-dev/lib/linux64GccDPOpt/libdynamicFvMesh.so (_ZN4Foam13dynami
cFvMesh3NewERKNS_8IOobjectE+0xaa9) [0x2a9556a2f9]
[compute-2-1:13295] [15] moveDynamicMesh [0x402127]
[compute-2-1:13295] [16] /lib64/tls/libc.so.6(__libc_start_main+0xdb) *** glibc detected *** malloc(): memory corruption:
0x0000000000794d60 ***
[compute-2-1:13296] *** Process received signal ***
[compute-2-1:13296] Signal: Aborted (6)
[compute-2-1:13296] Signal code: (-6)


It should look something like this -

Create mesh for time = 0

Selecting dynamicFvMesh subsetMotionSolverFvMesh
Selecting motion solver: laplaceFaceDecomposition
Selecting motion diffusivity: quadratic
Number of cells in new mesh : 2772
Number of faces in new mesh : 11286
Number of points in new mesh: 5940
oldInternalFaces : 198
Selecting motion solver: laplaceFaceDecomposition
Selecting motion diffusivity: quadratic


Running on one processor works great and I can move a subset just fine. However when decomposing the '0/setSubset' and '0/motionSubset' directories do not end up in any of the processor directories (should they?). Here's my case and modified subsetMotionSolverFvMesh code respectively -

http://www.box.net/shared/ph43019sss

http://www.box.net/shared/rllamuaok0

Thanks for any advice,
Kevin
kev4573 is offline   Reply With Quote

Old   March 6, 2008, 18:49
Default The subsets should be decompos
  #26
Senior Member
 
Frank Bos
Join Date: Mar 2009
Location: The Netherlands
Posts: 340
Rep Power: 18
lr103476 is on a distinguished road
The subsets should be decomposed as well. Somewhere on this message board their lives a decomposeParWithSets, or something like that. That utility should do the job, for me it worked fine....

Regards, Frank
__________________
Frank Bos
lr103476 is offline   Reply With Quote

Old   March 10, 2008, 19:25
Default Frank, could you post your
  #27
pbo
Member
 
Patrick Bourdin
Join Date: Mar 2009
Posts: 40
Rep Power: 17
pbo is on a distinguished road
Frank,

could you post your icoFsiFoam case with the cylinder and trailing flat plate. I try to understand why the fore part of my airfoil, which is clamped (fixedValue (0 0 0) for the displacement), gets deformed after the mesh motion in icoFsiFoam (only the rear part and the whole fluid region are coupled in the couplingProperties dictionary).

Cheers,

Patrick
pbo is offline   Reply With Quote

Old   March 11, 2008, 06:07
Default Hello I'm trying to model a
  #28
Member
 
Juho Peltola
Join Date: Mar 2009
Location: Finland
Posts: 89
Rep Power: 17
juho is on a distinguished road
Hello

I'm trying to model a wing, or a plate, in flow suspended on an axle. I would greatly appreciate any sample cases that are related ie. on solving the forces acting on the body and updating the mesh accordingly.

If you could email me at juho.peltola@tut.fi it would be great!

Juho
juho is offline   Reply With Quote

Old   March 13, 2008, 15:25
Default Frank - Thanks for pointing me
  #29
Senior Member
 
Kevin Smith
Join Date: Mar 2009
Posts: 104
Rep Power: 17
kev4573 is on a distinguished road
Frank - Thanks for pointing me to that decomposition utility, it seems to work ok but I could only get the patched decomposePar to compile. After decomposing my case (decomposePar . cylSimple) I try running 'moveDynamicMesh' in parallel but it fails with this message:

[1] --> FOAM FATAL IO ERROR : cannot open file
[1]
[1] file: ../cylSimple/processor1/system/motionSubset/fvSchemes at line 0.
[1]
[1] From function regIOobject::readStream(const word&)
[1] in file db/regIOobject/regIOobjectRead.C at line 66.
[1]
FOAM parallel run exiting


If I'm not mistaken, the processor directories don't need their own system directory - Did you add any parameters or do anything different when decomposing or running?

When I manually copy these files over to each processor directory it eventually gives an error :

[1] --> FOAM FATAL ERROR : Cannot find file "faces" in directory "constant/motionSubset/polyMesh"

Did you end up with the case structure as openFOAM expects it?

Kevin
kev4573 is offline   Reply With Quote

Old   March 13, 2008, 15:35
Default I did just copy the files and
  #30
Senior Member
 
Frank Bos
Join Date: Mar 2009
Location: The Netherlands
Posts: 340
Rep Power: 18
lr103476 is on a distinguished road
I did just copy the files and directories which is being asked for.......Indeed you'll end up with some dublicate stuff, but for me, it worked.

Are you sure that there is a polyMesh inside the motionSubset. Maybe you should also create the subsets using subsetMeshes utility and put the resulting polyMesh where needed.....In my case I had constant/motionSubset1 and constant/motionSubset2, so I had to create both polyMesh dirs using subsetMeshes.....

Cheers, Frank
__________________
Frank Bos
lr103476 is offline   Reply With Quote

Old   March 14, 2008, 16:14
Default Ok, I've made some progress an
  #31
Senior Member
 
Kevin Smith
Join Date: Mar 2009
Posts: 104
Rep Power: 17
kev4573 is on a distinguished road
Ok, I've made some progress and can now at least run in parallel but now the mesh moves strangely - seems the processor interface right around the 'oldInternalFaces' is not moving correctly.

Here is the original mesh at t = 0



And after 0.02 seconds I get this



Everywhere else the motion seems fine, but those rogue points at the top of cylinder seem to be causing a problem. Anyone have ideas?

Kevin
kev4573 is offline   Reply With Quote

Old   March 22, 2008, 13:46
Default Hello I'm trying to set up
  #32
Member
 
Juho Peltola
Join Date: Mar 2009
Location: Finland
Posts: 89
Rep Power: 17
juho is on a distinguished road
Hello

I'm trying to set up a case with icoDyMFoam. Mesh motion works neatly when a set the movingWall in cellMotionUy and pointMotionUy to fixedValue.

When I try to use oscillatingFixedValue or timeVaryingFixedValue it gives me an error message:

================================================== =

[0] --> FOAM FATAL ERROR : Not implemented#0 Foam::error::printStack(Foam:stream&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
#1 Foam::error::abort() in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
#2 Foam::defaultFvPatchField<double>::defaultFvPatchF ield(Foam::fvPatch const&, Foam::DimensionedField<double,> const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libfiniteVolume.so"
#3 Foam::fvPatchField<double>::addpatchConstructorToT able<foam::defaultfvpatchfield <double> >::New(Foam::fvPatch const&, Foam::DimensionedField<double,> const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libfiniteVolume.so"
#4 Foam::fvPatchField<double>::New(Foam::word const&, Foam::fvPatch const&, Foam::DimensionedField<double,> const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/applications/bin/linuxGccDPOpt/icoDyMFoam"
#5 Foam::GeometricField<double,>::GeometricBoundaryFi eld::GeometricBoundaryField(Fo am::fvBoundaryMesh const&, Foam::DimensionedField<double,> const&, Foam::List<foam::word> const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libfvMotionSolvers.so"
#6 Foam::GeometricField<double,>::GeometricField(Foam ::IOobject const&, Foam::fvMesh const&, Foam::dimensioned<double> const&, Foam::List<foam::word> const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libfvMotionSolvers.so"
#7 Foam::velocityComponentLaplacianFvMotionSolver::ve locityComponentLaplacianFvMoti onSolver(Foam::polyMesh const&, Foam::Istream&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libfvMotionSolvers.so"
#8 Foam::motionSolver::adddictionaryConstructorToTabl e<foam::velocitycomponentlapla cianfvmotionsolver>::New(Foam::polyMesh const&, Foam::Istream&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libfvMotionSolvers.so"
#9 Foam::motionSolver::New(Foam::polyMesh const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libdynamicMesh.so"
#10 Foam::dynamicMotionSolverFvMesh::dynamicMotionSolv erFvMesh(Foam::IOobject const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libdynamicFvMesh.so"
#11 Foam::dynamicFvMesh::addIOobjectConstructorToTable <foam::dynamicmotionsolverfvme sh>::New(Foam::IOobject const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libdynamicFvMesh.so"
#12 Foam::dynamicFvMesh::New(Foam::IOobject const&) in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libdynamicFvMesh.so"
#13 main in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/applications/bin/linuxGccDPOpt/icoDyMFoam"
#14 __libc_start_main in "/lib/tls/i686/cmov/libc.so.6"
#15 Foam::regIOobject::readIfModified() in "/home/juho/OpenFOAM/OpenFOAM-1.4.1/applications/bin/linuxGccDPOpt/icoDyMFoam"
[0]
[0]
[0] From function defaultFvPatchField<type>::defaultFvPatchField(con st fvPatch& p, const DimensionedField<type,>& iF)
[0] in file fields/fvPatchFields/basic/default/defaultFvPatchField.C at line 50.
[0]
FOAM parallel run aborting
[0]

=================================================

I'm using 1.4.1 with the precompiled binaries. The timeVaryingFixedValue works perfectly in turbFoam and oodles.

Any tips how to make it work?
juho is offline   Reply With Quote

Old   March 22, 2008, 18:06
Default Hi Juho, I've always put the '
  #33
Senior Member
 
Kevin Smith
Join Date: Mar 2009
Posts: 104
Rep Power: 17
kev4573 is on a distinguished road
Hi Juho, I've always put the 'oscillatingFixedValue' mesh motion implementation in the file 'motionU', never had to touch the other files 'cellMotionUy' and 'pointMotionUy'. I'm using 1.4.1-dev off of the svn but I think you should be able to do this 1.4.1 .
kev4573 is offline   Reply With Quote

Old   March 22, 2008, 19:21
Default When using tetDecomposition mo
  #34
Senior Member
 
Frank Bos
Join Date: Mar 2009
Location: The Netherlands
Posts: 340
Rep Power: 18
lr103476 is on a distinguished road
When using tetDecomposition motion solvers, the user only has to specify the motionU file. For your moving wall, you either specify oscillatingFixedValue or just fixedValue if your motion is defined by a (custom) dynamicFvMesh library.

This tetDecomp motion solver is only available in the OF-1.4.1-dev version. When using Finite Volume based motion solvers, you only have to specify pointMotion (the interpolation to cellMotion is done accordingly in the code).

Enjoy the mesh motion, it will lead to a lot of fun!!!

Frank
__________________
Frank Bos
lr103476 is offline   Reply With Quote

Old   March 23, 2008, 02:59
Default Thanks for the replies! I g
  #35
Member
 
Juho Peltola
Join Date: Mar 2009
Location: Finland
Posts: 89
Rep Power: 17
juho is on a distinguished road
Thanks for the replies!

I guess I should switch to 1.4.1-dev then. Shouldn't the oscillatingFixedValue and timeVaryingFixedValue work with Finite Volume motion solvers? At the moment I've used the movingCone tutorial as an example.

Another question:

I want to move only a part of a continous wall. How do I make the transition between the moving and fixed par smooth? Like bending the wall.
juho is offline   Reply With Quote

Old   March 23, 2008, 09:11
Default All boundary conditions should
  #36
Senior Member
 
Frank Bos
Join Date: Mar 2009
Location: The Netherlands
Posts: 340
Rep Power: 18
lr103476 is on a distinguished road
All boundary conditions should work fine with both tetDecomp and FV motion solvers......

Frank
__________________
Frank Bos
lr103476 is offline   Reply With Quote

Old   March 23, 2008, 09:36
Default Any ideas what might cause the
  #37
Member
 
Juho Peltola
Join Date: Mar 2009
Location: Finland
Posts: 89
Rep Power: 17
juho is on a distinguished road
Any ideas what might cause the error message I get if I use other boundary conditions than fixedValue for the moving wall? Message in my earlier post.
juho is offline   Reply With Quote

Old   March 23, 2008, 09:48
Default You should only provide a poin
  #38
Senior Member
 
Frank Bos
Join Date: Mar 2009
Location: The Netherlands
Posts: 340
Rep Power: 18
lr103476 is on a distinguished road
You should only provide a pointMotionU file with oscillatingFixedValue i.e. for your moving wall. (The cellMotionU is derived from that, you could delete that file).....Just try this. Additionally, you should set U for your moving wall to movingWallVelocity.
__________________
Frank Bos
lr103476 is offline   Reply With Quote

Old   March 23, 2008, 10:53
Default I deleted the cellMotionU.
  #39
Member
 
Juho Peltola
Join Date: Mar 2009
Location: Finland
Posts: 89
Rep Power: 17
juho is on a distinguished road
I deleted the cellMotionU.

Currently I have:

=================================================
0/pointMotionU:
=================================================


boundaryField
{
farField
{
type fixedValue;
value uniform (0 0 0);
}
movingWall
{
type oscillatingFixedValue;
refValue uniform (0 0 0);
amplitude uniform (0 2 0);
frequency 50;
value uniform (0 0 0);
}

outlet
{
type slip;
}
inlet
{
type slip;
}
frontAndBackPlanes
{
type empty;
}
}

=================================================
0/U:
=================================================

movingWall
{
type movingWallVelocity;
value uniform (0 0 0);
}

=================================================
constant/dynamicMeshDict:
=================================================

dynamicFvMesh dynamicMotionSolverFvMesh;

motionSolverLibs ("libfvMotionSolvers.so");

twoDMotion yes;

solver velocityLaplacian;

diffusivity directional (50 500 0);

=================================================

Same error message as before.

Is it possible something is wrong with my installation? The timeVariedFixedValue gives the same error message but works perfectly in turbFoam and oodles.

Also the mesh motion works fine when I set a fixedValue to the pointMotionU.

Thank you for your time!
juho is offline   Reply With Quote

Old   March 24, 2008, 04:08
Default The boundary conditions seem t
  #40
Member
 
Juho Peltola
Join Date: Mar 2009
Location: Finland
Posts: 89
Rep Power: 17
juho is on a distinguished road
The boundary conditions seem to work fine with the -dev version's tet decomposition.

And now towards the next problem...
juho is offline   Reply With Quote

Reply


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
Dynamic Mesh tutorial Part I attachDetach hsieh OpenFOAM Running, Solving & CFD 5 October 11, 2012 16:00
Can we add extra solid mesh part to the analsis? Farhath Alam FLUENT 0 December 22, 2006 03:41
Designating a part of mesh as wall megan FLUENT 0 October 15, 2006 22:04
Problem with rotational mesh deformation: Part II NymphadoraTonks CFX 2 November 4, 2004 04:05
Moving part in a fluid E. BOINOT FLUENT 0 April 24, 2002 10:34


All times are GMT -4. The time now is 15:46.