Any way to wrap a filleted voronoi around a closed surface?

Ive been given the task to create something similar to the attached image, of a flared voronoi tube. Can anyone recommend a starting point? Im racking my brain here, All the ways that Ive tried end up as a tube with a non-filleted voronoi pattern wrapped around it, which exists as a series of curves and not as a solid. Any help would be greatly appreciated because im running out of alternative ways to attempt this

Attachments:

Replies to This Discussion

Permalink Reply by peter fotiadis on January 29, 2015 at 11:39pm

get this:

Attachments:

VOR-ON-SURF-PETER_V5.gh, 46 KB

Permalink Reply by peter fotiadis on January 29, 2015 at 11:44pm

And maybe (just maybe) that:

Attachments:

Cytoskeleton_V5.gh, 65 KB

Permalink Reply by Daniel Piker on January 30, 2015 at 11:14am

Nice. I do think in this case Cytoskeleton will give a cleaner, more predictable result than Exoskeleton, as you have a surface mesh to start with. Also the result is a quad mesh (which could even potentially be turned into NURBS). It won't work for general line networks that don't form a surface, but when you do have a surface topology it makes sense to use it.

I would also recommend to use some directional subdivision before Catmull-Clark if you want to avoid this webbing effect:

Permalink Reply by peter fotiadis on January 30, 2015 at 11:44am

Nice advices Daniel: I'll give them a spin (although my relation with meshes is ... er ... hmm ... but on the other hand this quad to NURBS possibility is enough to ignite "some" interest, he he).

Here's some test cases more: with Patrick and friends (BEFORE Cyto/Exo is involved). Obviously some proper re meshing is required (MeshLab and the likes) especially in the Tree case (that could yield some very interesting results - even for AEC usage).

Attachments:

Thickened_dual_meshV2.gh, 1004 KB
Thickened_dual_meshV2.3dm, 1.5 MB

Permalink Reply by peter fotiadis on January 30, 2015 at 3:54am

Added your object and another one for tests (load Rhino file first).

PS: That said theoretically you could achieve a nurbs "solid" geometry (Rhino polysurface to be exact) using the "wires" as input to the TSpline pipe method ... but TSplines are highly problematic and I wouldn't suggest attempting it.

Attachments:

VOR-ON-SURF-PETER_V5A.gh, 50 KB
VOR-ON-SURF-PETER_V5.3dm, 585 KB

Permalink Reply by SteveB on January 30, 2015 at 6:59am

which version of exoskeleton are you using? I tried importing your file but its not finding the plugin, even though i have that, cytoskeleton, weaverbird and topologizer all installed.

What you posted looks promising though, if i could just figure exoskeleton out

Screencap attached

Permalink Reply by peter fotiadis on January 30, 2015 at 9:05am

this

Attachments:

Exoskeleton001.gha, 24 KB

Permalink Reply by SteveB on January 30, 2015 at 9:10am

Thank you Peter, apparently I only had exoskeleton002.gha. go figure

Permalink Reply by peter fotiadis on January 30, 2015 at 9:27am

PS: the new thing (ExoWire) it doesn't work in quite a few occasions.No idea why (but then again I very rarely do things with meshes).

Permalink Reply by peter fotiadis on January 30, 2015 at 9:54am

And this ultra pathetic thing (nurbs NOT mesh) is what happens when you put the wires from your object to the TSpline Pipe thing == ultra pathetic + ultra faulty, that is.

Such a pity (McNeel should acquire TSplines long before the AutoDesk take over).

Permalink Reply by SteveB on January 30, 2015 at 10:03am

very cool peter. Ive had similar problems with tsplines in other projects, ill wait to attempt to pipe it. I think im going to spend the next couple of days tweaking the model/script, ill let you know how it goes

Permalink Reply by peter fotiadis on January 30, 2015 at 10:18am

TS Pipe it? Well .. I always admire the brave, he he.

PS: TS is like cold fusion: fantastic idea, let down by "some" minor things.

PS: And why the Exo W doesn't work? That's the 1m question.

best, Peter

RSS