Hi my friends, can anyone give me some suggestion to set up this tensile structure in kangaroo to stimulate?
shell.3dm (303.4 KB)
A good place to start would be Daniel’s K2 examples on tensile structures, the one with ridge beams I would presume:
1 Like
Once you get yourself acquainted with the basics as suggested above, you can try something like the attached sample. You’ll see that some of the mesh vertices “ride” along the arcs as the mesh is “stretched” over them, as in your model.
The hard part here won’t be the Kangaroo set up - it will be in the modeling of the base mesh. Lots to say there, but for now, I’d suggest you model the base mesh faces with outer edges that align in-plane with your arcs. This will make it easier to pull them together and will likely make for a more accurate looking result.
2 Likes
Thanks John
1 Like
Thanks for your suggestion, I just build one part of the model and stimulate with kangaroo,( thanks johnharding’s K2 example and his mesh convert definition ) but I get weird result, maybe, it is as you describe.
Recovered model - shell 001.3dm (580.9 KB) TensileSurface.gh (234.4 KB)
The right hand-side area should not be pull down. The left hand-side area should be push up along the ribs. Are these reason due to the incorrect setting? Your definition seems it is standard quad meshes, but in my case, it seems a bit hard to do that. 
You can build quad meshes for the taller arcs, as they seem to be parallel, or at least they don’t intersect. You can project the arcs to the CPlane, and build large, flat 4 point surfaces or mesh faces between the projected lines. These surfaces or meshes can then be converted into subdivided quad meshes.
For the lower arcs, you can use a similar process, this time using large, flat triangles. For these, you can use simple mesh faces, and subdivide them.
The process should produce combined meshes with shared edges that align with your arcs - the vertices along these edges can be pulled up to the arcs, and may find their position as you add tension.
1 Like
Thanks for you suggestion, Nick.In my case, it seems it’s hard to avoid this situation, the meshes get pull to the arc. I try to define it with a part of that arc, but it will make the topology stretch.
shell 001.3dm (646.0 KB) TensileSurface.gh (306.7 KB)
You uploaded the wrong file - can you edit your post with your file?
Sorry, that was the wrong file. Thanks for telling me that. This is a new upload, but I am not sure if this is correct way to solve it.
You have some naked vertices that are being registered as anchors. In this case, you can use Kangaroo’s Combine&Clean to fix that:
But we should also talk about your general design strategy here. It seems you’re trying to pull vertices from one mesh to multiple intersecting (by plane) arcs, but this may be impossible. Your physical model above consists of two or more pieces of fabric, correct? In which case, you may want to model the same number of meshes, so you can attach them to the corresponding arcs?
1 Like
Thanks the fix Nick. You are correct. Sorry that I should explain more clearly in the beginning and the photo is a bit confusing on displaying the model. There are two pieces of fabric. They interact with the arcs. The early one, which is the inner fabric of the model I made it wrong. It should be like the shape that is outlined, like the one below. I just barely work it out, there is issue such as the upper piece overlapping with the one below and I am not sure in the lower piece that the approach to convert patch surface to triangulated mesh is the correct one.
shell 001.3dm (862.8 KB) TensileSurface.gh (416.3 KB)
