NastranIO support for reading Nastran meshes/subdomains/boundaries.

MASTmultiphysics / mast-multiphysics

Multidisciplinary-design Adaptation and Sensitivity Toolkit (MAST) - Sensitivity-enabled multiphysics FEA for design

https://www.mast-multiphysics.com

GNU Lesser General Public License v2.1

44 stars 25 forks source link

NastranIO support for reading Nastran meshes/subdomains/boundaries. #50

Closed JohnDN90 closed 4 years ago

JohnDN90 commented 4 years ago

This adds the feature of importing meshes from a Nastran input file by utilizing an interface to pyNastran.

jdeaton commented 4 years ago

A number of updates in this pull request now:

Obviously have the work related to NastranIO capability that John started and I refactored to make it better fit within the existing MAST source
Converted docs/assets over to a git submodule where the files are now located at https://github.com/MASTmultiphysics/mast-assets.
Added example with simple Nastran mesh input to demonstrate NastranIO capability as well as setup for structural modal analysis. This example also shows how you can use pure HTML elements in our tutorial files. I used it to organize result pictures in an HTML table.

Since we have the big change of the submodule in this, might only want to do the "Merge pull request" option and not the "Squash and merge" to keep those commits separated for clarity.

jdeaton commented 4 years ago

Looks like the build is currently only failing with libMesh v1.3.1. I think we should just drop support for that since its working with the newer libMesh 1.4.1 and 1.5.0.

jdeaton commented 4 years ago

Looks like the build is currently only failing with libMesh v1.3.1. I think we should just drop support for that since its working with the newer libMesh 1.4.1 and 1.5.0.

Just fixed this. Looks like it was related to the refactoring that the libMesh team did. Question remains whether we should drop support since they did clean up the includes in the more recent versions.

manavbhatia commented 4 years ago

I would recommend to create a mast-multiphysics/contrib folder for non-MAST codes, instead of mast-multiphysics/contrib/external. This way, libfort can reside as a clone of the original repository in `mast-multiphysics/contrib/libfort'.

jdeaton commented 4 years ago

I just went ahead and reorganized as you suggested and moved the external source into ROOT/contrib/libfort. This required a little bit of CMake refactoring.

The reason I originally set it up with ROOT/src/external was to try to simplify CMake, but I think it actually made things a bit more confusing and probably only made sense to me.

Once the last changes clear Travis-CI I'm going to go ahead and merge this PR in.

manavbhatia commented 4 years ago

Thanks! Yes, I think we can merge after this change.

manavbhatia commented 4 years ago

Just a sidenote: I pushed a commit #54 yesterday fixing the mass matrix issue for beam element that John had raised. I would recommend that you verify the results of your Nastran example with this update.