Reciprocal structures

Hi everyone,

I am trying to design a timber pavillion for a school project, and I am having huge trouble about the structure.

The structure that I need is similar to the Kreod Pavillion in London. The problem is that it I need to apply it on a free form surface, but the connections resemble a reciprocal structure, making it very hard to parameterize.

for a more basic form of this system: csm_20140530_Zollinger_9c4bc539aa.jpg

I have been looking for Kangaroo scripts for reciprocal structures everywhere, but nothing helped so far.

I would appreciate any thought.

Cheers!

Replies to This Discussion

Permalink Reply by Nik Willmore on June 15, 2015 at 4:44pm

You haven't defined the problem well. What I see in your image is just a nearly uniform hexagonal grid with a likely few pentagons to really curve the plane into a dome. What is "reciprocal" here? The triangles I see are just slippage of ideal struts past ideal hubs, a simple operation.

This brings up one question in my mind of whether it's topologically possible to make an all-distorted-hexagon spherical polyhedron, since usually 12 pentagons are required, even for highly distorted domes, that suggests that topological connectivity ruins it for all-hexagons.

Kangaroo merely relaxes and reforms meshes or series of lines as springs, etc.

Even a pencil on paper sketch about the "free form surface" and the result would help. Otherwise, this is all way too generic of a question.

Is it the specific joints that are bothering you or the overall constructive geometry?

So I looked up reciprocal structure and it's like Fuller's tensengrity but just uses gravity or beam tension to hold it together?

https://en.wikipedia.org/wiki/Reciprocal_frame

If it's the connections that are your sticking point, what's the simplest model of the problem that might define it well?

Permalink Reply by Amir Ojalvo on June 16, 2015 at 5:25am

Thank you so much for your reply. You are right, the question wasn't clear enough.

I can show a few of my failed attempts:

On a planar surface, rotating the lines of a triangular grid seems to do the trick. The more you rotate, the bigger the hexagons become. And since all the lines are on the same plane, it's possible to trim the extensions.

On a non-planar surface, however (I looks planar from top view, but actually it's not) each hexagon is different. You can see that some lines are too short and some are too long... And the problem is that the lines are all on different planes, making it impossible to trim or extend them.