CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (https://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Problem using AMI (https://www.cfd-online.com/Forums/openfoam-solving/95697-problem-using-ami.html)

vinz February 27, 2012 04:29

I would maybe try to make the inner part slightly larger than the outer part.

Even if they overlap by 0.1% of mesh size for example.

I gues that would ensure that during the rotation, no part of the mesh is left without cell.

liping_he April 9, 2012 23:18

Hi Stephane
I create a case to simulate two phase flowing and rotoring with the slover interDyMFoam and AMI boundary. I created a simple mesh with blockMeshDict, it runs well. I think the code is working well. But when i import my own mesh with the tool fluentMeshToFoam, the mesh is created with gambit. AMI1 and AMI2 are two same faces. it cracks. I have on idea about this problem. I with you can give me a detailed process about how to create mesh with AMI boundary. Thanks very much in advance for your advice.
liping_he@sina.cn is my email address.
May I have your email.

openfoam_user April 10, 2012 04:01

Hi,

just try to follow the steps of the propeller tutorial:
~/OpenFOAM-2.1.x/tutorials/incompressible/pimpleDyMFoam/propeller/

The first step is to redo the tutorial understanding all steps.

Then you can use your own geometry. Try to keep the same boundary names, then only you can change them.

The tutorial use snappyHexMesh (and blockMesh) for generating the mesh. The tutorial works well. But with my own geometry the computation crashes ! Now I use ICEM Hexa for mesh generation and the case doesn't crash any more.

Regards,

Stephane.

liping_he April 10, 2012 04:27

Hi Stephane

Thanks for your early reply. My mesh is created using Gambit. It has two domains, rotor and stator. AMI1 and AMI2 locate between two domains. I want to konw how you import .msh file to OF.

bestWish

liping

openfoam_user April 10, 2012 05:10

fluent3DMeshToFoam fluent.msh

with fluent.msh the mesh file in fluent V6 format.

Regards,

Stephane.

liping_he April 10, 2012 06:08

HI again Stephane

I have done the follow list.
1. I create the mesh in gambit. The mesh hase two domains
2. save every domain separately. and export the mesh using fluent V6 format
3. create separate case directories, put the .msh files there, run fluent3DMeshToFoam in every directories
4. run mergeMeshes
5.cope the new mesh, run in case.
it shows:
Quote:

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time
Create mesh for time = 0
Selecting dynamicFvMesh solidBodyMotionFvMesh
Selecting solid-body motion function rotatingMotion
Applying solid body motion to cellZone rotor
#0 Foam::error::printStack(Foam::Ostream&) in "/home/cfd/OpenFOAM/OpenFOAM-2.1.0/platforms/linuxGccDPOpt/lib/libOpenFOAM.so"
#1 Foam::sigSegv::sigHandler(int) in "/home/cfd/OpenFOAM/OpenFOAM-2.1.0/platforms/linuxGccDPOpt/lib/libOpenFOAM.so"
#2 Uninterpreted:
#3 Foam::solidBodyMotionFvMesh::solidBodyMotionFvMesh (Foam::IOobject const&) in "/home/cfd/OpenFOAM/OpenFOAM-2.1.0/platforms/linuxGccDPOpt/lib/libdynamicFvMesh.so"
#4 Foam::dynamicFvMesh::addIOobjectConstructorToTable <Foam::solidBodyMotionFvMesh>::New(Foam::IOobjec t const&) in "/home/cfd/OpenFOAM/OpenFOAM-2.1.0/platforms/linuxGccDPOpt/lib/libdynamicFvMesh.so"
#5 Foam::dynamicFvMesh::New(Foam::IOobject const&) in "/home/cfd/OpenFOAM/OpenFOAM-2.1.0/platforms/linuxGccDPOpt/lib/libdynamicFvMesh.so"
#6
in "/home/cfd/OpenFOAM/OpenFOAM-2.1.0/platforms/linuxGccDPOpt/bin/interDyMFoam"
#7 __libc_start_main in "/lib/tls/i686/cmov/libc.so.6"
#8
in "/home/cfd/OpenFOAM/OpenFOAM-2.1.0/platforms/linuxGccDPOpt/bin/interDyMFoam"
Segmentation fault
i have no idea about this error..
looking forword your help.


bestwishes

liping

openfoam_user April 10, 2012 06:20

Hi,

I haven't save the 2 domains separately.

I have used ICEMCFD to generate the mesh. Only one domain containing the 2 fluid parts and the interface (AMI1=AMI2) between the fixed and moving parts.

Stephane.

liping_he April 10, 2012 07:50

HI

I want to konw that how to set the rotating cellzone..


best

liping

openfoam_user April 10, 2012 08:12

Hi,

ALL IS EXPLAINED IN THE PROPELLER TUTORIAL.

topoSet -dict system/createAMIFaces.topoSetDict

Please redo the propeller tutorial again and try to understand all steps !!!

Regards,

Stephane

wangqiangele April 11, 2012 12:10

Hi Stephane,

I have also encountered this problem, and I have tried a lot of methods. Thanks for your kind advice.

It seems you have created the background mesh with ICEM and used it for snappyHexMesh. So there is no need for the innerCylinderSmall any more. Am I right?

I will try it right now.

Best regards,
Qiang Wang

wangqiangele April 11, 2012 12:22

Hi Stephane,

I am sorry for the question, but how could I draw a background mesh like yours. It seems a little difficult for me to make the mesh follow a circle.

Do you use a cylindrical block(with small radius) in the middle to make the mesh outside the block follow a circle?
How do you draw the Hexa mesh in the center of the cylinder?

Thanks in advance,
Qiang Wang

openfoam_user April 12, 2012 03:53

Hi,

I have tried 2 ways:

1. Background mesh with ICEM and sHM for the propeller. The case runs well at the beginning and then crashes suddently.

2. Whole mesh with ICEM (hard work and time consuming). The case runs well. I have done 5 complete revolutions.

I hope that AMI will be improve in the next release of OF.

Regards,
Stephane.

wangqiangele April 13, 2012 01:21

Hi Stephane,

Thanks for your advice. Building the whole mesh with ICEM seems quite difficult for me. I will try to use the background mesh with sHM first.

Best regards,
Qiang Wang

vinz April 13, 2012 02:58

I managed to use SHM to create the grid, so you should be able to do the same.
However as I said in this post, I had to try different meshes to find one working.
Good luck.

wangqiangele April 13, 2012 07:54

Hi,

I managed to use mesh created only by sHM. The case has run over 2 revolutions.

At the same time, I have managed to use mesh created by sHM while the background mesh created by ICEM. The case is running, and seems OK.

Thanks for all of you.
Qiang Wang

openfoam_user April 13, 2012 08:04

Hi,

happy to hear that users are able to run case using grid generated by sMH.

Mine is also running for some time then it crashes suddenly (despite nice convergence of residuals and forces).

With the ICEM mesh it runs well (5 revolutions). It require more time to use ICEM hexa, but sharp edges are perfectly represented.

Best regards,

Stephane.

kid April 17, 2012 08:23

Developers please improve AMI implemetation.
Most of us are stuck, not that propeller case is not working. Propeller works fine, but when we try to implement our own case like rotating cube, things go really difficult.
Also some documentation .

If someone guides me implementing AMI using a generic case. I will do that documentation my-self and place it on forum.

Also, if people are interested , let us do it together ans start a new thread.
regards

linnemann April 17, 2012 17:13

I use AMI every day and I almost never have issues.

Also the meshes are almost all SHM to SHM meshes.

In general you just have to be a little more careful generating the mesh at the AMI's. Using the same and constant refinement level for both sides of the AMI works almost all the time.

Code:

        ami1
        {
            level (3 3);
        }

and also for
Code:

        ami2
        {
            level (3 3);
        }

I although agree that a little better output from the AMI's whilst running the code could be beneficial in order to track down where in the mesh there is an issue. A possibility to output the affected faces to a faceZone that could be watched in Paraview would be nice.

EDIT: Using OF21 and OF21x

wyldckat April 17, 2012 17:18

Quote:

Originally Posted by linnemann (Post 355235)
I use AMI every day and I almost never have issues.

You forgot to mention if this was with OpenFOAM 2.1.0 or 2.1.x :D

kid April 18, 2012 01:45

@linnemann
yes you must be correct in stating so, We people need better understanding of AMI.

Regards


All times are GMT -4. The time now is 15:35.