Add unstructured mesh capability for R2S workflow

[gonuke]

Modify R2S workflow for unstructured mesh capability. This will greatly simplify the material definition part of the workflow, at the expense of requiring every region to be meshed with a tet-mesh.

[erelson]

Attempting to list the various details that need to be taken care of for this. Some of these will be their own issues.

mmGridGen/mmgrid.py #38

• Ray tracing of tet mesh for material fractions
• Use voxel center point material

ALARA

• Feed tet volumes into ALARA input Calculate volumes for arbitrary tetrahedra and hexahedra #40
• Feed unstructured mesh tally neutron fluxes into ALARA input Generate ALARA's fluxin file from unstructured mesh tally data #39

Additions to source.F90 to read MOAB mesh directly:

• Look for and read in bias/custom energies tags
• Figure out how to pass mesh file name (defaulting to n_fluxes_and_materials.h5m currently) Pass arbitrary mesh file (.h5m, .vtk) to MCNP's source.F90 #41
• Code for tets/hexahedra Read MOAB mesh directly with source.F90 #42
  • TBD: how much of the code can be the same for both voxel types
  • Arbitrary hexahedra probably supported
  • Will support meshes that mix tets/hexs (though a use case doesn't really exist yet?)
• Normalize photon source data within source.F90. Re-implement Photon Source Normalization Factor in source.F90 #43
  • We already tag the mesh with data taken from phtn_src. write_gammas.py does some normalization before writing the tagged source info to the gammas file. Since having the information from phtn_src tagged to the mesh is useful for visualization, it is probably better to not replace this data with normalized values.
  • However, if a calculation is re-run, this normalization ends up being re-done. Not sure how costly this is.
• Calculate volumes for arbitrary tetrahedra (and hexahedra) Calculate volumes for arbitrary tetrahedra and hexahedra #40

Compiling DAGMCNP with iMesh API calls in Fortran #44

• Understand/implement changes to Makefile; work in progress for distribution
  • gfortran call definitely requires adding -fcray-pointer -I/filespace/groups/cnerg/opt/MOAB/opt-cubit-c12/include -liMesh -L/filespace/groups/cnerg/opt/MOAB/opt-nocgm/lib
  • may also require -lstdc++ -lMOAB -lhdf5 -lnetcdf
  • I need to better understand how current DAGMC functionality uses MOAB (if it does?)

gammas file: #45

• Decide whether to replace the gammas file completely with the mesh. gammas file is unneeded for unstructured workflow
  • Keeping the file makes it easier to quickly change parameters (e.g. source particle debugging, void rejection options)

[gonuke]

Admin note: You could make this a milestone and each of the major bullets could be issues attached to that milestone.

[erelson]

Admin reply: That works nicely.

[erelson]

Capability exists. Todo list above has been updated. Related enhancements are now tracked in #50.