Anyone tackled this method before on Grasshopper ?
I have a variety of C# stuff that does tensegrity trusses on BrepFaces (in 4 types, shown a double tetra and a x frames options). Implementing some abilities nore couldn’t be nuclear science … but the big real-life challenges on tensegrity trusses are related with clash free node design (Impossible even to think to do it with Rhino or GH). That said tensegrity for real-life AEC costs MANY times more than any, say, MERO type of truss … so you would be surprised if you assume that light means “rational” as well.
8 Likes
Thank you for the info.
“Rational” is not a problem as this is for an ART piece work 
the goal would be to create a morphed abstract general shape but to get truss and cable point algorithm to fill it up and make it hold shape, tensegrity style.
like if this (attached) is possible in Rhino+GH ?
Yes it is … but I speak having a C# solution in mind.
That said I have no idea if it’s doable with native components (but several aspects of the whole case - even for artistic purposes - are more or less impossible).
That said … if the structure has some size (and you are after high class parts aesthetics) you’ll encounter several detail puzzles with regard the node design. The smaller the node size … the more challenging the task is.
For instance … er … hmm … this has missed the target by some miles (too big stuff in too little space plus some questionable detailing)
This Monte Carlo method is interesting, but to me it doesn’t seem very practical or very fast. The paper is quite well documented and implementing it in Grasshopper shouldn’t be too bad. I would look at scripting most of it as you’ll otherwise quickly end up with a mess of components that are essentially doing simple math. I would also look at separating the potential energy component and the moving of points. That way, you could compare the MCFF method to other optimizers in Grasshopper like Galapagos.
Have a look at this thread, https://www.grasshopper3d.com/forum/topics/tensegrity-structures?groupUrl=kangaroo&groupId=2985220%3AGroup%3A120977&id=2985220%3ATopic%3A164904&page=2#comments @DanielPiker Mentions the very article you’ve posted here. I would have a stab at your problem with Kangaroo first before trying to code something up from scratch. You may very well be able to make a tensegrity system that looks roughly like your starting shape.
However, as @PeterFotiadis mentioned, finding a stable geometry and finding a stable geometry you can actually build are two different things. You may need to figure out how you’ll be building this thing and force those limitations onto your form finding steps.
1 Like
Leaving apart the art part of the story … sooner or later at some time reality dictates that all in all you have 2 options:
Do a “regular” T truss of some type using “simple” non adjustable nodes like this one (easy to get the gist of the tension method):
OR do an irregular one and deal with the challenging task to design a node with several levels of freedom.
2 Likes
Here’s a classic thingy as well:
http://rsif.royalsocietypublishing.org/content/11/98/20140520
that proves that Academics know a thing (or two) about Maths but nothing about real-life engineering.
Plan Z: Do a T Truss like this and be a happy bunny (the ultimate complexity is simplicity). Plus I could provide a working multi adaptable node design (ultra slim, using mini ball pivot joins and Ronstan hardware) for similar node types (SS 316 L + dollars aplenty).
PS: The kings of tensile hardware stuff:
Also : (good topology aesthetics crude nodes):
BTW: I was talking about your case with my top IT Guru at the practice: Me? I have no idea about coding, C#, D#, … , Z#, computers and the likes (I’m just the businessman).
Guru (of gurus (of gurus)) said:
See this? It’s a platonic (or not) tetrahedron that grows randomly (shown: the 2nd generation) on a per face “expand” basis. Solid struts are “pushed” (NOT connected in the classic sence) against a mini sphere (at tetra CoG) because cables (white) force them to do so. If you remove just one cable the whole thing does bananas.
Guru said: Is this thing stable? I said: I have no idea nor I care to know (I’m just the businessman) . Guru (of gurus) said: If we do some recursion on that this is what we get:
Guru said: got the plus points (all struts/nodes are the same for instance, not to mention a zillion other things) of a similar solution? I said: don’t ask me (I’m just the businessman). Guru said: see the underlying mesh in (random grow) action:
1 Like
I wonder what does the Psychiatrist say after listening to the mono(dia)log, above? 
You got it 100% wrongly: that was a real dialog since there’s the guru (and the Lord) talking serious scientific matters (In the Name of Science, what else?). Here’s the guru in a demo with regard latest and greatest add-ons (node connectivity, angles, this and that) on that single tetra LOL T truss variant.
Just what the Guru (of gurus (of gurus)) is out after…
1 Like
Indeed … I suspect that’s the reason for going from 80 ms to double+ that … when the Guru (of Gurus (of Gurus)) attempts to compute a simple node to node connectivity tree: a clear proof that he is focused on other things. I mean that even the most stupid Guru should know the cost of the IndexOf thingy.
