Support for brep export with imprinting, merge and material tagging to allow h5m DAGMC geometry

[shimwell]

DAGMC allow use of faceted based geometry to simulate particle transport when used in conjunction with a suitable neutronics code like OpenMC.

Mainly neutrons and photons are transported using ray tracing like algorithms but there is support for charged particles in some codes. We use these codes for fusion reactor design. The faceted geometry has a few particular aspects:

• First any surfaces that are touching can be imprinted and merged. This effectively removes one surface and has the impact of speeding up particle transport through the geometry. We normally do this in the CAD stage prior to faceting.

• Additionally the geometry is also surrounded by a thin container geometry that terminates transport of particles when they react the edge of the universe, often referred to as a graveyard. However I think we already have plans to automated the production of this so nothing to worry about here.

• Finally the geometry allows material tagging for each volume. Neutrons and photons interact differently with different materials so this material tagging is essential.

h5 files are the a great format partly due to the speed they can be read into the code. The geometries can get quite large so this really helps.

There is one neutronics code (serpent II) that transport particles on STL geometry, hence my earlier interest in exporting STL format.

Not many formats support the virtual surfaces needed for imprinting and merging but I believe brep does. It is also tricky keeping track of the materials and volume numbers while they go through this process of removing surfaces. The input order in is not necessarily the volume order in the brep file.

There is an open source tool that takes the imprinted, merged, materiel tagged brep file and facets it and exports the h5m file

So log story short, supporting brep (or some other format) exporting with imprinted, merged and material geometry would naturally be of interest to our fusion neutronics community.

[shimwell]

I think this might be a duplicated of #449 which has a bit more detail so I should close this one and put comments over on #449