[wyldckat]

Hi Reza,

Sorry about that. It's so much information I'm looking at here in the forum, that I'm failing to notice some details.

I don't have access to Pointwise, so I have to ask:

1. How are you exporting the mesh to OpenFOAM format?
2. Is it able to preserve cell zones?
3. What kind of cells does the mesh have, at least in the interface between rotor and stator?

Best regards,
Bruno

[reza1980]

Brunu,
Would you please specify what you mean clearly. We have run this mesh as self-propelled case, a propeller installed on a ship with hull and rudder. It seems OK and done by my supervisor.

[wyldckat]

Hi Reza,

OK, I'll try to be clearer.

You have a mesh generated with Pointwise, correct? But how did you get that mesh into a format that OpenFOAM can use?

Best regards,
Bruno

[reza1980]

You can export the mesh into 3D-OPenFoam.(Please look at the attachment)

[wyldckat]

Ah, OK, I didn't know that.
What export options does it give you?

[reza1980]

For OF just OF3D.

[wyldckat]

OK, I think I'm getting the idea.
Can you please generate a very small example case in Pointwise, such as having an internal cylinder and an external cylinder, so that the internal one should rotate inside the external one?
Then share the two meshes in OpenFOAM format here on the forum or at Dropbox or something like that?
Because this way I could see for myself what we're working with.

[reza1980]

Bruno
I can send my major case includes stator+rotor,Ok?

[wyldckat]

If you are allowed to share the case and if the mesh is not too big, I think you can.

[reza1980]

please give your account in dropB

[wyldckat]

Hi Reza,

Only for the last 20 minutes have I managed to look into this with enough attention.
I have a question for you... How exactly did you do this step :

Update Boundary (Add cyclicAMi and tolerance)

Best regards,
Bruno

[reza1980]

The interface between Stator and Rotor must be in patch type before merging and then convert to cyclicAMI, and define a tolerance based on tutorial, that is 10-4 , and neighbor for AMI1 and AMI2. The files I sent you must be implemented based on AMI.

[wyldckat]

Then do you mean that you used "createBafflesDict" and createBaffles?

[reza1980]

no not at all, please take a look here,
http://143.248.98.6/ExCFD/dynamicMesh.pdf

[wyldckat]

Hi Reza,

Mmm... I suspected as much. Well, the trick of editing the file "constant/polyMesh/boundary" sometimes works, other times it doesn't. Or at least, I think it might not work as intended.

The reason I say this is because creating baffles requires an indication of "which side is up", or in other words, a reference cell zone for figuring out which side is on the baffle side master and which is on baffle side slave. In theory, incorrectly defining these patches can lead to wrong oriented faces, which could in turn lead to crazy values on the baffle surfaces.

I'm going to see if I can still test this today, namely to use createBaffles for producing a proper surface connection... but only after confirming first the quality of the resulting mesh, after editing the "boundary" file directly, namely by running:

checkMesh -allTopology -allGeometry -constant

Best regards,
Bruno

[reza1980]

Thanks so much Bruno ,I am online If you have any question or needs to sent final case that is ready to run.
Furthermore,I run this case with different Inflow . I can sent some pictures from the results.
One odd thing is to see sometimes the non-overlapped cells .
REGARDS
Reza

[reza1980]

Originally Posted by wyldckat

Hi Reza,

Mmm... I suspected as much. Well, the trick of editing the file "constant/polyMesh/boundary" sometimes works, other times it doesn't. Or at least, I think it might not work as intended.

The reason I say this is because creating baffles requires an indication of "which side is up", or in other words, a reference cell zone for figuring out which side is on the baffle side master and which is on baffle side slave. In theory, incorrectly defining these patches can lead to wrong oriented faces, which could in turn lead to crazy values on the baffle surfaces.

I'm going to see if I can still test this today, namely to use createBaffles for producing a proper surface connection... but only after confirming first the quality of the resulting mesh, after editing the "boundary" file directly, namely by running:

checkMesh -allTopology -allGeometry -constant

Best regards,
Bruno

Attachment is the log file from checkMesh -allTopology -allGeometry -constant

[wyldckat]

From your log file, this looks like a problem:

<<Writing 81 cells with two non-boundary faces to set twoInternalFacesCells

It's the only thing that seems to be bad, when compared to the tutorial "incompressible/pimpleDyMFoam/propeller".

OpenFOAM's 2.x paraFoam is able to represent these sets, namely the set in question "twoInternalFacesCells"; search for the "sets" option on the "Object Inspector".

[reza1980]

i access to OF2.1x 2.2x in Cluster network .would you please specify what you mean

[wyldckat]

Sorry, I meant 2.x, with x=0,1,2.
In other words, as of OpenFOAM 2.0.0 and beyond, you can use this feature in paraFoam. Therefore, both 2.2.x and 2.1.x should have it.