[philippose]

Hello everyone,

I would like to post a message I received on the Netgen Mailing list from Joachim Schoeberl (The "father" of Netgen).....

(Quote)

Dear Friends of Netgen,

I am preparing for the Netgen/NGSolve 4.5 release. Release candidate 2
is available on

http://www.mathcces.rwth-aachen.de/doku.php/netgen

It contains a compiled version for Windows, and a source package for
Linux/Unix/Windows. It compiles with gcc (on Linux/Unix), and MSVC++
2005 on Windows.

Feedback to the mailing list is welcome.

I plan to release snapshots now more often.

Netgen is moving (following myself) from JKU Linz to RWTH Aachen (at the
moment, it has two homes).

I have started a wiki - page for the documentation. I hope it will grow
now more rapidly with a joint effort.

Enjoy the new Netgen,
Joachim

(End Quote)


Now I hope that more people get to see what Netgen is capable of, and why I have been raving about it for so long :-)!

Enjoy!

Philippose

[alexandrepereira]

Hi Philippose

I am a Foam user who debates with a critical issue in CFD analysis: a properly refined mesh to capture the high gradients in boundary layer, yet coarse enough in regions where high gradients are not to be expected, so as to ease the computational burden...

I have already experimented with Netgen V 4.4 ( the default one in Ubuntu Gutsy repository ); now experimenting Harpoon v3.2 from SHARC.

So far i have several comments to make, correct me if i am wrong :

NetGen isn't as robust to geometry imperfections as Harpoon, ( I have acheived meshing STL's in Harpoon that i never got to mesh in NetGen )

Harpoon has a better boundary layer mesh control than Netgen

Netgen does not allow me to build an external flow domain, (for automotive, or aircraft CFD for instance...) as does Harpoon

Netgen is less expensive than Harpoon... :-)

So tell me, do my opinions result from a lack of knowlege of NetGen capabilities, or are they well founded...?

As a hypothetical user of NetGen i miss :

A Tutorial with-real life examples (difficult geomerties, boundary layer meshing...)

The ability to export meshes in NASTRAN format,

Besides of OpenFOAM, I have this software ADINA, which has multiphysics solver capabilities, Thermal, Radiation, FEA, CFD, Fluid Structural Interaction,... but its meshing abilities are TERRIBLE, at least for 3d unstructured mesh, but it will accept a NASTRAN mesh domain built from an external mesher...

The ability to impose boundary conditions in the exported mesh (like in Gambit, or Gridgen )

I think that NetGen was mainly targeted for FEA users... not CFD... but that is only my opinion... :-)

Am I right here or not...?

Alex

[alexandrepereira]

By The way...

Will the next release of NetGen incorporate boundary layer refinement... as well as more geometry import formats and/or mesh export formats...?

The fact that NetGen is an openSource Package is great, and it is in fact a powerfull mesher... I just wish that meshes generated would have better convergence charateristics with OpenFOAM solvers... Have better checkMesh specs... etc.

Best regards

Alex