Creating external volume around 3D mesh model

[Anonymous]

We’re working with a customized developing software which can export our 3D models as NASTRAN.dat file format (inherently a 3d format). During the imported with Autodesk Simulation CFD (ASC) the following message pops up:

The 2D model is either not planar, or it is not in the Z = 0 plane

                       

Seems that ASC interpret the file as 2D model even if I have later a 3D model in ASC:

 

The problem that occours now: I can’t set up an external volume. Both possibilities to access External Volume Creation don’t work:

Setup (tab) > Setup Tasks (panel) > Geometry Tools, and click the Ext. Volume tab.

Alternatively, right click on the Geometry branch of the Design Study bar, and click Edit...

I exported also a few 3D models with merged surfaces as *.obj file and imported them into Inventor Fusion 2013. The import wors without any error message but afterwards I’ve the same problem: I can’t create the external volume:

I checked the tutorials and it works fine with the car models etc. but not with our 3D model. Any suggestions?

Thx

[Royce_adsk]

Could you upload your 3D model for us to checkout?

Thanks,

Royce

[Anonymous]

Following the models (see attachment):

TestWing_1.dat (NASTRAN File)

TestWing_2.obj (WaveFront OBJ File)

Beside this I can export the 3D model as:

Please let me know if you need some other files.

Thanks for your help!

[Royce_adsk]

Interesting list. Try saving it as a Solidworks file and upload that.  Haven't had a chance to look at the other 2 files you posted, but maybe someone else will get to it before I do.

-Royce

[Anonymous]

"Good" choice - solid works is the only file format which I can't create. For the export solid works must be installed on the computer but I'm an AutoCAD 'aficionado' 🙂 So no solid works on my computer...

[Royce_adsk]

Well, that is a bummer!  Too bad it didn't output a better Autodesk format.

-Royce

[Anonymous]

Hi, I'm the developer of the software making the original export model.  I can get the model into Solidworks format, although the files are often around 50Mb in size and take a long time to generate through Solidworks, and this format is not supported by Autodesk Simulation CFD.

We would really like to work with the NASTRAN format files, as these are generated quickly and easily, and are supported by Autodesk Simulation CFD.  Please let me know what is going wrong with the import of those files so that we can get these working.  The file is definitely a 3d format, and we can see the 3d model in the CFD program, but for some reason it will not allow analysis of this as a 3d model for external flow, as mr_siegel has described.

[Royce_adsk]

1) We can read Solidworks files using Inventor Fusion which is included in the CFD installer.  Once it is in the SolidWorks format, we might be able to read it into Fusion and convert to a 3D geometry.  Depends on the quality of the geometry.

2) The Nastran import needs to be a 3D mesh and what you actually have is a grid mapping of surfaces (CQUAD4 Element). That is why you see so many errors when you load the model.  The nastran import does not support the geometry tools because the assumption is that you have already meshed your model for simulation.  If you wanted to use only the CQUAD4 elements it would have to be located on the XY plane for a 2D analysis.

You are seeing the surfaces in CFD, but that is after you went through many error messages indicating that something was not correct.  The model will not work as is.

3) The .obj and Nastran fails to open in both Inventor Fusion and Inventor Pro.

[Anonymous]

Yes, I have used the CQUAD4 elements to define the surfaces in the NASTRAN file.  The NASTRAN format does not support a "MESH" as such.  Do I need to use CHEXA, CPENTA and/or CTETRA to define my model as a solid?  I could convert each CQUAD into a CHEXA with a small thickness if this would help.

Does Simulation CFD support analysis of surface-based models and membranes?  Our models are constructed from cloth, and for some models (not so much this particular one), it is important to analyse the flow on both sides of a single layer of cloth, so a membrane based model would be perfect.  However, I know that many CFD packages cannot analyse surface models, and require the model to be given a solid thickness before analysis can be done.  Is this the case for Simulation CFD?  If so, is there a minimum thickness required?  (e.g. Some PARSOL analysis software requires at least 1-2mm thick surfaces, otherwise flow may occur through the surface.) 

Do you have a sample NASTRAN file that does work in Simulation CFD?  Even a simple geometry would be helpful as a reference, so that I can mimic the model definition method used in that file - even just a cube.

[Anonymous]

I have made a solidworks model to test with.  Getting this file working will not solve the main problem of interfacing between programs, as mr_siegel does not have Solidworks to use, but will be an interesting model for mr_siegel to test with.

www.surfplan.com.au/temp/Part1.zip

[Anonymous]

I am having a similar problem.  I just downloaded SimCFD 2015 (way earlier in the release cycle than I normally would) , in part because it can natively import rhino files.  I imported a file with a bunch of surfaces, and no solids.  The program seems to think this is a 2D file - it gave me the same error message above, and my geometry options say (Cartesian 2D) and don't give the option to switch to 3D. 

I would like to model this as surfaces instead of solids because there are hundreds of thin parts, and I don't want the meshing to be unmanageable (though I will miss the adaptive mesh ability).  I too was planning to make the external volume (wind tunnel) around it in CFD.  This is because Inventor is horrible at importing complex surface files from neutral formats.

I tried using a .stp file and had the same issue.

Do I need to trick it by adding a dummy solid to the file before saving from Rhino?  I'm going to try this and report back but I thought I'd flag that this is still an issue.

And no, I can't upload this model, but maybe a test part if this drags on.

[Royce_adsk · Accepted solution]

When you only import surface parts the software will think you are trying to perform a 2D simulation.  The software probably gave you an error that the surfaces were not on the XY plane.

Also, when using surface parts they must be 100% embedded in a volume when importing into CFD.  That means, as you said, you will need to create a volume in Rhino.

When working with surface parts in Simulation CFD I would highly recommend using the following flag and turn on Surface Refinement.  This flag and refinement will help with incomplete surface mesh warning messages and build a better mesh around the free edges of the surface.

mesh_boundarylayer_blend with a value of 1.

[Anonymous]

Thanks, Royce, I will try this, including the flag.

Before your note above, I made a small dummy volume, rather than making the whole wind tunnel in Rhino, and then just made that solid into an air part inside of the CFD-generated fluid volume.  This succeded in convincing it to run 3D, but the mesh was terrible (all surfaces had sawtooth edges, even with surface refinement) so I will see if embedding it in the volume helps.

[Royce_adsk]

The flag & Surface Refinement are critical in my opinion when it comes to using surface parts that are in the middle of the flow domain.

[Anonymous]

Does the thickness assigned to a surface part in the material dialog box make a difference when meshing?  In other words, will I be able to create a smaller/faster/more robust mesh with 0 thickness, actual thickness, or extra thickness on the surfaces parts?  Or is this only used for thermal?

For example, I have many thin sheets that are <0.12 inch thick, but will happily assign them to 0 thickness or 0.5 inch thickness if it will let me mesh the model more efficiently.

Does the refinement length or any other meshing parameters have to be somehow matched to the part thickness (more/less/same/doesn't matter)?  Again, for surface parts, not solids.

Thanks.