relaxed space frame

Hi!

Im working on developing this project that you see on the image below. It is created using a 3d voronoi algorithm that does not allow me precise control of where nodes and bridges are. Therefore I need to build another algorithm that creates a visually similar output but gives me more control. form will also need to be different with more complexity near the top.

I found a minimal surface algorithm which seems to work well, though i have only tried it using very simple geometry. I like it because it gives me ok contoll of thickness of different parts. see example below. it originates from this thread

What i now have to do is create a sort of space frame for use with this definition. simple enough job but i play around with advanced scripts though i still dont know the basics of the game.. :/

I have built a test point cloud and a jumble of lines that connect the points. Now i need to find a way to boolean intersect all the pipes i create.

How do i do that?

Or is this not the best way to attack this issue to start with? Should i go about it differently?

Attachments:

bygga kub i grasshopper.3dm, 472 KB
grundad på stretchedfabric.ghx, 304 KB

Replies to This Discussion

Permalink Reply by Ludvig Haav on March 26, 2014 at 5:03

I suspect node smoothing is what i am after. This is a thread about it that is not encouraging at all.

It is strange to me that it should be a hard thing to do, seems so basic.

Permalink Reply by Fredrik Skåtar on March 26, 2014 at 5:04

It seems you'll have to construct the initial structure as a continuous unified polysurface which you turn into a mesh in order to distort/smoothen the joint areas. -I guess, that is what you mean by making a boolean union.

If yes, there is a geometric problem with joining pipes, - the very same problem that occurs if you would weld pipes together for real. That is, there will be a gap.

The solution, also when welding for real, can be to insert a sphere which is slightly larger (see image).

The resulting mesh, you could add to the gh-file you attached previously, but it takes quite much computer power.

The question is, why do you really need to control the thickness of the pipes?

I mean, you could insert stairs in those shapes and if you do that, you work with the Voronoi script which turns your project into a continuous concept.

Working with pipes, like in my image, will give you a rather uniform expression unless you start randomizing radii, which will be a huzzle.

What do you say?

best regards

Fredrik

Attachments:

pipe_boolean_union.jpg, 42 KB
Ludvig____pipe_boolean_union_from_Skatar.gh, 9 KB

Permalink Reply by Fredrik Skåtar on March 26, 2014 at 7:59

So, I found a solution.

It is all about reducing the mesh complexity.

(images correspond to digits)

1) Bake the result from the previous file (Ludvig____pipe_boolean_union_from_Skatar.gh)

2) Turn the baked result into a mesh in Rhino

3) Select the mesh, in Rhino and type "ReduceMesh".

4) Set the number of meshes as simple as possible
(i.e. visually - click preview) without destroying the main structure.
You'll see that the spheres dissapear. 500 seems to be good when
handling the entire structure

5) insert the resulting, reduced, mesh into Grasshopper and subsequently into
the script. The image only shows a section of the structure since the operation takes a while.

Exploding the mesh takes out the bulgy node effect we're after.

But, experiment with it.

Attachments:

1.jpg, 215 KB
2.jpg, 233 KB
3.jpg, 233 KB

Permalink Reply by Fredrik Skåtar on March 26, 2014 at 7:59

rest of images

Attachments:

4.jpg, 241 KB
5.jpg, 257 KB

Permalink Reply by Fredrik Skåtar on March 26, 2014 at 8:01

and the models.

Attachments:

grundadpstretchedfabric___Skatar.ghx, 334 KB
byggakubigrasshopper___Skatar.3dm, 688 KB

Permalink Reply by Ludvig Haav on March 27, 2014 at 3:52

Thanks a lot for your help Fredrik, no matter how simple i made the structure subjecting it to the physics script made my CPU depressed.

instead im now working with T-splines. John Ross showed me some basic t-splines modelling yesterday night and i have now just managed to output my original voronoi pattern into a very editable t-splines model. things are moving forward, for now. i really hope i have settled with a modus operandi with this one.

see image of input model and output tspline. and how easily tweakable it is.