Volumetric Meshing

Are there any plugins or methods to go about volumetric meshing? I’m curious about how the process would work…specifically volumetrically meshing with quads.

See this recent discussion


Also to add another to the ones I mentioned there-
http://www.itascacg.com/software/blockranger

Is there any equivalent plugin of tetrino for rhino 6?

Have you tried? Not that I know what the expected output is but it seems to be running fine on Rhino 6 here.
-wim

Actually you are right. It works perfectly. I can’t find out how to join the mesh between two different solids though.


The script is a bit messy but I was just testing it.

Tetrahedra mesh.gh (17.1 KB)

Try to align vertices first then join.

I’m very interested in volumetric tools as well, and I have tried Tetrino and Dendro, but experience different problems with each:

  1. Tetrino doesn’t seem to close/fill gaps between open meshes (to planes above each doesn’t get the gap between them filled)

  2. Dendro - Could be only “operator failure” by me, but I have not been able to fill open meshes without expanding the mesh. And Dendro Volume based on mesh input tends to not fill up the mesh all the way to the surface.

If I for example let Dendro build Volumes based on vertices or edges, the final mesh surface will expand the surface a radius distance larger than the vertices defining the surface.

My aim is to close open meshes and in so doing, “filling” them as to get rid of small internal meshes (typical residue from medical scans) and this is what I have failed to achieve with volumetric approaches so far. But if someone has had better experiences of some volumentric (or other) method and want to share, I’d be listening very carefully… :slight_smile:

// Rolf

Hi, @RIL, have you found a tool that works? Filling meshes to get rid of small internal meshes is something I also need to do.

I’ve wondered if you could at least begin the process of volumetric meshing with quads by first performing a quad-remesh on the desired object, then subsequently offsetting the mesh until it fills the volume.

Branch Technology has found some manner of doing this in their 3d-printed lattice structures, and I believe they use Grasshopper. It appears as though the shapes are generated by subdividing into quads.

This conversation seems to be about 2 quite different things -

  • filling a shape with many small volumetric elements (tetrahedra or hexahedra), such as you would use for a solid FEA. This is the typical meaning of the term ‘volumetric meshing’.
  • finding a single closed surface which closely wraps a shape, getting rid of holes and interior objects

For the latter, maybe something from this discussion is helpful:

1 Like

You can use Iguana Mesh, which is a pretty nice plugin:

1 Like

Materialize Magics does this in seconds really easily, but it is damn expensive (I guess justifiably as it really is amazing, many things I couldn’t do without it).

I’ve done lots of measuring of really really crappy meshes, and often even crappy meshes are actually good enough for the measurement locations I need (defined in all kinds of scientific papers). But sometimes there’s need for watertight meshes, and then crap like the below is not making me happy (notice how different colors highlights the fact that parts of surface isn’t even connected and instead form “islands” disconnected from each other. Apart from the fact that it’s full of holes… And inside it’s full of “splinters”, due to bone marrow).

Fig. Joint between Scapula/Glenoid and the Humeral head. All measurements fully automated with GH.

This means that strategies with rays from inside doesn’t work (shoots through holes, or hits an inner “double surface” representing the inside of the bone tissue at the boundary between bone and marrow), and rays from the outside (rays towards the center of a sphere) also shoots through holes, and often right through the mesh, aso.

Shrink wrapping with angle constraints on edges as to avoid sinking too deep into holes (while still allowing deep “dents” if there’s a surface underneath, is the only thing that I think would work for bone scans, since they typically never have any holes in them, with very few exceptions.

Bones that actually do have holes (like vertebras) would require mesh-parameters (again, edge angle constraints) allowing only holes bigger than a certain diameter or area to fall through (again, while still allowing for deep & steep dents which has a surface underneath which is connected to the same surface which form the “brim” of the dent.

I believe Kangaroo is capable of producing such shrink wraps, but so far I have not succeeded in tricking it to obedience…

// Rolf

1 Like