OpenFOAM Feasibility...
 

The company I currently work for is planning to start adding CFD capabilities in our analysis function. We have looked into several commercial packages and OpenFOAM. While I was still in school, I was very briefly exposed to OpenFOAM, and became very excited about how powerful it is and the support available here on the forum.

Initially most of the analysis will consist of a structure with a cavity and inlet/outlet ports. Our main concerns at this time are pressure drops from inlet to outlet. Further down the line, analysis may include pressure/velocity profiles through a passage that has a surface with thermal expansion. Pressures will typically fall in the range of 1,000 psi - 20,000 psi, and the fluid will be hydraulic fluid.

We currently utilize Ansys 11.0 Classic (with occasional Workbench) for all structural analysis, and do not have any licenses for Fluent/CFX/CFD-Flo. If we could utilize node/element information generated from classic or workbench mesher, it would certainly be beneficial.

I would greatly appreciate any feedback as to whether or not OpenFOAM should be able to meet the needs specified. If it seems like a viable option, I would likely attend the training session in the US either the March or June date.

Thanks!

Hi,

I think OpenFOAM will meet your requirements, since your simulation case don't seem to include any "specialties", so I would recommend you to have a look at the OpenFOAM User Guide, to check if one of the standard solvers will already fit perfectly.

Regarding the mesh export from ANSYS to OpenFOAM: I'm in the same situation needing such a transfer procedure from time to time. Although OpenFOAM includes useful meshing tools, there are special cases where I find it more convenient to create the mesh in ANSYS.

You can export the ANSYS mesh as Fluent binary but none of the existing fluent-to-foam mesh converters can read binary files. Fortunately, in my company we also a have a Fluent license, so opening the mesh in Fluent and saving it as ASCII .cas file is no big deal. But sorry, that's no help for you at all.

I would also prefer to export an ASCII fluent mesh out of ANSYS, since that would speed up my workflow. Maybe something could be done with APDL scripts?
Since I'm mainly a Workbench user and not very familiar with APDL, I'll post in the ANSYS forum, maybe we get some help there.

Best regards,
Johannes

Hi Johannes,

Thank you for the suggestions; I have printed off the user manual and other documentation, and will begin working my way through them. I believe the standard solvers appear promising from first glance.

After talking with the IT department at my work, I was only able to get the unofficial windows port installed on my workstation :(. I believe if I decide this is looking more and more promising, I would be able to get the official version installed on a Linux machine. A friend has referred me to a meshing utility called enGrid, which sounded as if it would allow me to convert the Ansys mesh and apply my boundary conditions (If I recall correctly).

I wonder what the output file from Fluent looks like. If I knew what format was read in, I could write a macro/APDL to convert the mesh data from regular Ansys or Workbench.

Thanks,

Jason

Hi Jason,

Another possibility of testing OpenFOAM under a Windows environment would be to install an arbitrary linux distribution on an external HDD and use it as a boot device. I tried this myself and found it quite useful, especially because there are no changes to be applied to the windows machine (maybe aside of the boot device priority).

That should not be a problem. If you'd like, I can provide you with a simple mesh mesh example both as binary .msh and ASCII .cas, so you can compare them and write such an export macro for ANSYS.

Best regards,
Johannes

Hi,

Yes the external HDD does seem like a good solution. I have used vituralbox with a linux distribition on my personal laptop while in school, and then ran OpenFOAM through that. Everything but post-processing seemed to run fine, which was alleviated by running the windows version of Paraview outside of the emulator.

I am certainly interested in a generic mesh. It is no rush though, as I have some other projects that are more pressing at the moment; so,it will be a little while before I can start devoting time to OpenFOAM.

Thanks again

Hi Jason,

that's completely ok, just drop me a line or respond to this thread when you're going to start.

Best regards,
Johannes

I still can't seem to figure out an "easy" way to get Workbench to export an ascii fluent file instead of the binary. I have some ideas of how to make the *.inp file exported from Ansys match the template needed for a *.msh file, but I really hope I do not have to resort to that.

Hi Jason,

I saw your reply in my "Fluent mesh export as ASCII" thread. Using the environment variable as suggested there works fine for me with ANSYS 12.x even if Fluent is not loaded.

I did a quick test with ANSYS V11 and unfortunately it seems that the variable is not recognized and the output format is still binary.

Although there may be a workaround: Could you try if can save your mesh as Fluent ASCII case file (*.cas)? I can't test because I've no Fluent 6.x ready. At least this works with Fluent 12 and the resulting case file is readable by the existing Fluent to foam mesh converters without any problems.

Best regards,
Johannes

That is great news! I have been looking for a reason to get IT to bump up my version to 12.x, and this looks like it :).