delaunay meshes from (poli)surfaces

hello,
i need delaunay meshes from (poli)surfaces with compound curvature for simulation cloth in max or maya.
can anybody think, if and how this could be solves with grasshopper?

any comment would be appreciated.

thanks in advance roberto

Replies to This Discussion

Permalink Reply by Suryansh Chandra on August 26, 2009 at 4:47am

Roberto,

I see two ways to go about it:

1. Do your cloth simulations in Rhino + GH itself. Take a look at these:

http://en.wiki.mcneel.com/default.aspx/McNeel/ClothSimulation.html
http://liquidtectonics.blogspot.com/2009/03/cloth-systems-next-iter...

2. Secondly, if you are going to be using a delaunay to export to max/maya, curvature doesn't matter, no? delaunay will always triangulated meshes, which will always be flat.

If you do a simple UV division of your polysurface to get the points, and then use them to generate to a delaunay, wouldn't that work for you? Delaunay is now a 'factory fitted' function in GH .60018 onwards. Find it under the Mesh tab.

Permalink Reply by roberto gagliano on August 26, 2009 at 5:31am

hi suryansh,
thanks for the quick reply and the interesting input.
@1
the video links look promising, i will check out asap.
i do real-world-inflatables, so i need to simulate pressure and physical correctness is a must. i am at the beginning and i am not sure if 3dsmax cloth/garment maker or maya ncloth is the way to go under this aspect, but at least there are many parameters to adjust and fine tune. moreover (as i am not a programmer) i need to work with figures and parameters that reflect cloth properties on as close as possible to a real-world-basis. do you know, if this ispossible with david´s scripts?

for cloth simulation delaunays are always better than quads, as quads tend to fold in one of the two uv-directions, so also when simulating with gh it should be a deaunay.

@2
there is kind of a misunderstanding here:
the single tri-mesh faces are flat, as you point out.
what i need is a delaunay to be derived from a curved surface. i get good delaunays with garment maker in max, but there the input has to be a flat closed curve (just as the tailor does when producing his layout), so the whole surface has to be flat, before mesh generation.

when i make a delaunay from uv-division-points the quads will be divided into tris but it will still be kind of regular.

i will check out the delaunay-function in gh .60018 (at the moment i run .60012) to see if it accepts a surface input or if it works the same as in "pset recon tools" (which i tried without good results)

thanks again roberto

Permalink Reply by roberto gagliano on August 26, 2009 at 6:21am

hi again,
just checked the delaunay-function in current gh.
it accepts only point input, which doesnt fit for me (see explanation screenshot and 3dm).
one benefit is: obviously the point input does not have to pe planar.

still hope, i am missing something and there is an easy solution.

roberto

Attachments:

delaunay_mesh.jpg, 583 KB
delaunay_vs_quad.zip, 172 KB

Permalink Reply by Evert Amador on August 26, 2009 at 6:44am

Hi Roberto,

I saw your post this morning on the Rhino NewsGroup.

Sorry I've haven't been on the loop, pretty busy at work trying to finish things up to go on holiday.

I've opened the file you attached and I'm assuming the "desired result" comes from the max garment. I've just downloaded the latest Panelling tools and grasshopper releases to try the delaunay triangulation, and point grid creation but I still think that the problem is finding out the delaunay vertices in the first place. It seems that GH only find the closest 3 points suplied and creates a facet, it seems it does not allows to create extra points where required so the result is not 100% delaunay I would say.

I'll continue researching on the subject to see what happens.

In the menatime have alook at the enclosed file, you can see the original surface in red to which I applied a point grid creation in Panelling Tools and the GH delaunay.

I did not know you work in inflatables, have you seen rhino membrane?

Cheers!

Attachments:

NoisedSurfaceTriangulated.rar, 1.5 MB

Permalink Reply by roberto gagliano on August 26, 2009 at 7:25am

hi evert,
nice to meet you here!

> I've opened the file you attached and I'm assuming the "desired result" comes from the max garment.

yes. and pulled to the surface with applyUVN.

unfortunately i cant download your rar. it says:
Access Denied
You don't have permission to access "http://storage.ning.com/topology/rest/1.0/file/get/2768589584?profile=original" on this server.

Reference #18.4c447b5c.1251292821.34edb881

> have you seen rhino membrane?
never heared about it, but looks interesting.
do you have any experiences with it?
do you know the price?

cheers roberto

Permalink Reply by Evert Amador on August 26, 2009 at 8:55am

Hello!

I haven't tried it yet. Have a look at this page for rhino membrane and prices.

http://www.ixcube.com/dettagli.asp?n=20

I'm enclosing the file again I hope it works this time.

regards

Attachments:

NoisedSurfaceTriangulated.3dm.zip, 1.6 MB

Permalink Reply by roberto gagliano on August 26, 2009 at 1:17pm

thanks for the rhinomembrane link.
didnt see the prices on my first visit, the site was too long ;)
saw your file, but didnt really get what you intended to do and how you got there...

cheers roberto

Permalink Reply by Vicente Soler on August 26, 2009 at 6:08pm

You can create a delaunay mesh in GH from a polysurface made out of untrimmed surfaces using only components rather easily. Just get the uv coordinates of random points on each face and delaunay these, then get the faces of the resulting mesh and create a new mesh with the corresponding 3d points as vertices instead of the uv coordinates. To have a similar density on each face, make the number of random points proportional to the surface's area.
To make the edges match, just create a rectangle that matches the surfaces uv size, divide it in many points and add them to the rest of the random points before creating the delaunay mesh.

This is obviously not the best way of doing it, but sort of works.

Permalink Reply by Vicente Soler on August 26, 2009 at 8:08pm

Something like this:

Permalink Reply by Evert Amador on August 27, 2009 at 2:52pm

Hi Vicente,

Very interesting, could you post the definition?

Regards

Evert

Permalink Reply by Vicente Soler on August 27, 2009 at 3:11pm

Sure. To get an even point distribution reparametrize the brep's surfaces in rhino first, select 'automatic'. I'd try to find a better way of placing points that's not just completely random. You could also create a grid of points and move them randomly by a small amount. The better methods would probably require some sort of recursion.

Attachments: