Space Frame Mesh Modelling

Hi All,

I tried this for a long time in many different ways but I failed to achieve the result I am anticipating.

I need to model a Bucky space frame based on 2 triangular grid meshes (see image).

This is what I did in the definition:

1- Pick every other mesh face using David Stasuik's VB script

2-Project a pyramid mesh face on each of these

3-Get the edges.

The last step would be to connect the vertex of each of the pyramids in a triangle grid mesh that is 'slipped' from the original grid.

Any ideas?

Definition attached. Your help is appreciated.

Chap.

Attachments:

Replies to This Discussion

Permalink Reply by Chapuline on May 25, 2015 at 11:51am

Hi All,

I'm glad this thread is sparking some interesting conversations! I'm learning from following up.

Peter and Angel

All your points above are all valid and this is why this truss design is chosen

Starting with initial studies of the possible space frames that could be used, this design was the most optimal, because the tetrahedral struts allow for triangulation which make the structure a lot stiffer.
Making those connections on every other mesh face makes the structure lighter, use less material and allows for a hexagonal pattern to emerge on the second layer.
This arrangement was chosen because : Less material used, more stiffness in the shell, a less dense structure which is visually more appealing.

The second attempt to optimize is to make the pyramid strut connections grow in length from 0 height to 2.5 depth=stiffness the shell is the weakest at the least curved, and largest span, therefore the center. which brings me to my second discussion:

Dave S, you're right. I don't need to explode to mesh faces and use data trees. Not sure why I thought it wouldn't work before but now I dealt with the mesh as a whole and the normal length attractor change worked. So, no data trees.

However, I ran into another problem: Do I give the normals of every other mesh face? or the whole mesh with the attractor lenghts? I tried both, and your script would crash on both scenarios. Please see the attached GH file..

There are still 2 problems:

1-varying offset length

2-connections on all the faces not just a small patch of the mesh.

Thanks!

Chap.

Permalink Reply by peter fotiadis on May 25, 2015 at 2:13pm

Well ... it's still Monday eh?

http://www.grasshopper3d.com/forum/topics/get-points-do-something

Permalink Reply by David Stasiuk on May 26, 2015 at 1:52am

One thing I noticed with the base mesh that you sent over is that while it is a single mesh, each face has its own set of vertices...welding it fixes a lot of problems related to topological operations. [UTO]'s MeshEdit plug-in has an outstanding weld component. If you weld it, and your mesh is all valence 6, the script should correctly assign faces to two groups.

The variable offset would make more sense to have within the script itself, where face normals are derived. It depends on the attractor logic that you're using, but it's pretty straightforward.

Permalink Reply by peter fotiadis on May 23, 2015 at 12:18pm

Chap

Well ... I've postponed (a "bit") posting the def because it's code: meaning terra-firma for you: It can solve your points > mesh issue ... but code is a nasty adventure: if something goes wrong you'll probably (and rightly) blame me. If this was some kind of do-it-just-for-fun ... that would be far easier but the thing is that you intend to use it (even for Academic purposes).

BTW: Testing something to be bullet proof in any condition ... it's another animal TOTALLY DIFFERENT from writing some (actually more than "some", he he) lines of code.

So after the required clarification the good news:

1. Monday is your day: the def it could be yours.

2. 30 test domes would be included (as blocks) in the accompanying Rhino file. These tests are to blame for the delayed release, mind.

3. 3 modes to create that @^%@^ truss: the 2 that you see here and a parallax (more "dense") to the hexagon top layer.

4. 2 working modes to skin a cat: if you provide a mesh > works with that. If you don't (meaning you provide points) it makes a mesh with a ball pivot algo and keeps going (in fact it can work by making "growing" meshes until all the points are consumed by one mesh - see test animation posted).

What happened with that Skype thing? Ping me

best, Peter

Permalink Reply by Chapuline on May 23, 2015 at 12:22pm

Peter. We need more people like you in this community! Simply awesome. Thanks so much. I added you on Skype, can you accept?

Chap

Permalink Reply by peter fotiadis on May 23, 2015 at 12:32pm

Karma my friend > no request arrived here > mail me your skype ID