Equal distance between points+finding support using Galapagos

Need help! I'm trying to generate an abstract structure from wooden beams, using kangaroo and karamba to test its sustainability. The project needs to be done really fast, so I'm looking fo some simple solutions. The general view is something like that

I haven't found the solutions for the following questions(in my head):

1) is there any way to generate the model from beams with defined length, for example if I want to make the entire structure from 6m-long timbers. I mean how could I achieve equal spaces between random points?

2) I like the idea of finding the location of possible supports using Galapagos to minimize the displacement of the structure, but I can't understand the principles of generating those indicies.

Could somebody possibly explain me something? Thank you!

Replies to This Discussion

Permalink Reply by peter fotiadis on December 9, 2015 at 3:25pm

Hmm (is this made via Ball Pivot on random points?)

1. Provided that you can constrain your nodes properly Kangaroo can achieve equal lengths in your axis.

2. minimize the displacement via Galapagos??? you mean achieving some reasonably minimum overhangs ??? (i.e. "min" compression forces in the overall collection of members?). If so and for this type of topology (or some "similar" anyway) the potential gains are rather minimum.

3. Is this Academic?

Permalink Reply by Maria Ostroumova on December 9, 2015 at 3:48pm

I made the connection using Nearest Neighbors(karamba) on random points

1) Yes! I'm trying to use it, but I use Gravity too and beams become not as equal as I expect, I guess I should be more consistent in my desires) I was thinking maybe there is some kind of math or geometric way of creating points with equal spacing initially...But I can't figure it out.

2) By now I agree with you, it's better to find supports intuitively.

3) What do you mean?

Permalink Reply by Maria Ostroumova on December 9, 2015 at 3:55pm

Oh, could you ,please, explain how did you measure "the potential gains"? I'm asking because I'm not experienced and maybe missing some wisdom)) Thank you for your advices, btw!

Permalink Reply by peter fotiadis on December 10, 2015 at 12:07am

Well...

"Potential gains" means a theoretical gain in profile size IF the whole thing is "balanced" with the one way or the other. On this type of "trusses" I would strongly recommend to skip that constrain for the moment.

Other than that:

1. On this type of structures equal members (although achievable) is not a consideration unless you use some type of MERO system and you have a zillion members in mind (balls must being drilled on a per ball basis, mind). But you are not after a MERO. Or ... you have "fixed" panels in mind that "fit" in your facets and make the envelope?? (but this is more theoretical than anything ... ).

2. Individually picking items in collections (your nodes for instance) although the essence of engineering ... is not achievable in GH unless you use code (I could easily provide a small demo on that matter using C# - it could change your perception on things rather drastically).

3: Academic means ... er ... academic, he he: something NOT related with a real-life project meaning that the level of detail and the whole approach is ... er ... hmm ... "indicative". Academic also means: non-profitable (in the strict sense).

Suggestions:

A. This type of stuff is 50% about designing a proper node. Provide some detailed sketches on that matter (BTW: "mobile" nodes ???). The more "accurate" they are the better. Using wood makes the structure "heavy" to the eye (and is BAD: killing trees for doing things?? No good at all - unless is fully synthetic [kitsch], that is).

B. The remaining 49.99% is the envelope: Provide ... blah, blah. Note: if the envelope "panels" are NOT engulfed in the facets (say: some planar glazing system and the likes) then the nodes MUST provide means to support the envelope as well (or at least the Primary support system). By Primary I mean "big" triangular frames that support a "grid" of smaller ones that host glass, poly carbonate sheets or some other stuff.

C: For a vast variety of reasons going from "abstract" to something "realistic" on that type of structures requires coding. Are you familiar with such freaky things? (C# for instance) OR do you have issues using a "component" that is made that way? (see 2).

Permalink Reply by peter fotiadis on December 10, 2015 at 12:51am

BTW: IF you are not afraid of C# and IF you have a certain GH experience or you want to take a plunge in the deep (he, he)... I could provide to you something [note: MERO oriented] that does all these things captured (either working on collections of points [i.e. just points] OR on meshes) - there's an indicative option for MERO type of trusses as well (but you are not after MERO):

Case points in space (connected via BPA algo):

Many had tortured that Stanford bunny ... but none like this:

Trusses out of Meshes:

And activating the MERO option (requires FAST CPU: Xeon 1630 V3 and the likes):

Permalink Reply by Maria Ostroumova on December 10, 2015 at 3:46pm

Oh,my,Peter! If you are so accurate I can tell all the pupose)) My friend asked me to help him make a structure for the concert, so it's more than REAL))) The sketch was like this(not mine)

And talking about wood, we live in Russia, so wooden beams are really cheap and affordable...Sounds awfull, but it's our reality. Addition aim is about the width of the beams - my friend is a specialist in 3d mapping, he needs the surface of tose beams to project some video content, that is all this about)))

Talking about joints, we must make them really simple, as I can see, MERO system - is some kind of thing made for shells and so on? Well my story is not a shell but a transperent structure, a decoration for a concert. It should look like totally "abstract" and freeform. And I should design some kind of simple way to make joints, like I can go to the nearest shop and buy all the materials. By this moment we desided to do it by incuting and contracting beams together, so you can see, it is hard to control the result...

Permalink Reply by Maria Ostroumova on December 10, 2015 at 3:55pm

And talking about equal spacing - its all about the speed of installation, so I desided to make it all from those beams I can buy in hardware store, without any fabrication. All about speed)))

Permalink Reply by Maria Ostroumova on December 10, 2015 at 4:13pm

I really like your animal-like stuctures, so nice! I dream about future projects and hope to use all this experience, mabe C# is a method. I took the "advanced" course of GH with some addons like kangaroo, I learned karamba by myself for this project, I like all this logic a lot. It's a pity I haven't learned maths at school as much as I need today( I am all about humanities...

Permalink Reply by peter fotiadis on December 10, 2015 at 11:52pm

Hmm ... nice aesthetics I must say: bravo. It's a rare sign these days were the kitsch Holy Blob reigns supreme.

1. Still I don't see any reason to go after equal axis on that one. The pic that you've posted rather testifies that. Randomness is the virtue here NOT equality. BTW: forget completely "optimization" of the mount points et all.

2. The thing in pic is airy (quality Numero Uno on that type of stuff, especially for things with no real load bearing capability) meaning minimum profiles ... meaning FORGET wood. Use aluminum tubes (rather cheaper than wood) as follows: screw the Captain Hook "node" in some kind of machined tube end (a humble massif cylinder that is screwed or clued [Araldide 2 part Epoxy] to the tube AND machined with threads for the hook).

NOTE: I could make a simple tube "adjustment" system that could allow you to build that on-the-fly WITHOUT any GH/K or anything: just start connecting variable length tubes ... er ... hmm ... randomly. This is the recommended way to do it anyway: we can't emulate art with software and even if we could: it's the art of pointless.

3. Additionally the whole conceptual aesthetics BEG for some kind of metal instead of wood. The fact that wood is aplenty in Russia doesn't justify killing trees (for any scope), anyway.

4. Using rings to "attach" the hooks ... well that could yield a highly unstable structure for more than obvious reasons. I could provide to you dozens of highly sophisticated bespoke solutions on that matter ... but they are unsuitable for this DIY occasion: I must think on a zero basis on that puzzle - allow me some time to propose the best "adapter" (easy, cheap, stable and allowing some liberties).

5. A Connectivity tree (see for instance Sandbox) can resolve with easy the equal axis worry of yours (thus: FORGET equality, just buy a hack-saw [ aluminum is very easy to "cut", he he]).

All in all: I like that a lot. I'll post soon some examples related with the all overall approach (including the node, he he). You don't need Kangaroo for that (and dare I say no structural analysis IF the structure to be is "similar" in size with the one pictured).

more soon, best, Lord of Darkness

Permalink Reply by peter fotiadis on December 11, 2015 at 12:20am

BTW: Is there any good Samaritan who could ... er ... machine threads on things? (inner threads for the tubes and classic ones for the "adapter").

BTW: For the thing that you've posted (or "similar"): can you find aluminum tubes with R 100 and thickness 6-10? (mm). Steel is very heavy for DIY assembly mind. What about Carbon Fiber? (hit some bank, get the millions and I'll provide the ultimate solution upon how to spend them, he he).

BTW: Is this type of performance repeatable? (think U2 going around the world and making that ultra kitsch "scenery" thingy).

BTW: The only way to use rings with the Captain Hook stuff is to apply NEGATIVE tension (pull) to the tubes AFTER assembling the thing ... but for that you'll need a Norseman type of tensioner (rather unavoidable using that one ON EACH member end I'm afraid).

Permalink Reply by peter fotiadis on December 11, 2015 at 12:34am

I mean (pull AFTER assembly) replacing the Captain Hook stuff with these (or some variant of these anyway):

Anyway get the Bible on these matters (but since they are SS 304/316 L stuff they are not exactly cheap).

Attachments:

Navtec norseman gibb catalogus.pdf, 2.3 MB

Permalink Reply by peter fotiadis on December 11, 2015 at 1:32am

BTW: Here's a sequel of pics related with that matter (BUT with a MERO type of stuff in mind although we can achieve "similar" rigidity with the negative tension on members - God MAY help us [not to mention the Dark Side, he he]).

BTW: If you are interested on that type of methodology > Skype is required (approx 1 hour about this, this and that).

Enjoy (using TSplines, C#, Y#, Z#, the Dark Side, the White Side (not really), stuff from Daniel, stuff from planet Zorg ... and in general ... "stuff", he he).