Kangaroo Tensile structure facade !! Please Help me !!

I'm trying to make a facade with metal spiders holding tensile structures. The code is based on a square grid. Every cell is a four point tensile menbrane. The kangaroo component works only when there is one cell and i don't undertstand why ? When i put more than one cell it tells me: Exceptional solution: index was out of range. It doesn't have to be negative and has to be smaller than the size of the collection.

In french:1. Solution exception:L'index était hors limites. Il ne doit pas être négatif et doit être inférieur à la taille de la collection.
Nom du paramètre : index

Please help me solve the probelm, will be very grateful if you come to Paris ;)

and sorry for my bad english !!

Nenov Ivo

Attachments:

tensile structure.jpg, 626 KB
FiltresMetal3.gh, 64 KB

Replies to This Discussion

Permalink Reply by peter fotiadis on May 26, 2014 at 7:59am

You've just said the 2 wrong words (a) School/Academics (b) that AutoSomething thing (this is not CAD in the very same sense that a Harley Davidson isn't a motorcycle - but indeed has 2 wheels I confess) .

Anyway...since you look to me like a nice fella, and you are after tensile membranes ...er... I have good, bad and ugly news:

Good news: I do hope that you are a Bushido type of Architect-to-be: Obeying the Code no matter the cost (i.e. (a) Less is More, (b) Form follows Function, (c) Always serve Modesty and fight Vanity).

Bad news: Whist the component bloodshed continues (I asked some assistance from the best C# guru in my practice)...

... I reckon that this IS not the way to outline the module fixing brackets (and/or any membrane related SS 316L stuff : mini masts, tensioners blah blah).

This is:

Ugly news: But inviting real-life components like the above to the parametric chaos (and killing any argument against the so called parametric/smart thing) requires other things far and away from what Rhino/GH are designed to do (at least at present time). But we can fake things a bit (we are professional liars by trade, don't you think so?): I'll reform that chaotic pasta first and then I'll provide you with some real-life multi adjustable custom components (the variation situation that you are attempting to emulate with the attractor) suitable for the job (alas: Rhino can't understand what nested parametric driven feature modeling is) ... and that could bring your teacher(s) into a certain crisis: what are these things and how's possible being created with a parametric app? (but then again I bet that they never heard about CATIA - not to mention spending 15 years with that ugly thing, he he).

Moral: Have faith to the Code.

Permalink Reply by peter fotiadis on May 27, 2014 at 8:07am

Update :

1. In free standing cases and especially in small shading modules we usually apply the classic rule of convex/concave in membrane design for more than obvious reasons. Attempting to "stress" an almost planar tissue (your case) can cause a variety of issues up to the undo able state (metal parts/components grow in size as well for no reason). See forces estimated by FF below.

2. Therefor I strongly suggest to consider Plan B (a) mastermind a secondary "anchor" capability in order to achieve a far more stable system (b) use a mount design that can support this (and comply with the attractor concept of yours). Here's a variable mount custom system (mostly machined AND not cast) that is suitable for the scope (Rhino reads the stp file OK .... but makes a colossally big file - thus I attach here the original).

3. On first sight lot's of things in this system appear "odd". For instance: is it stable? Why these double cables are used? How far can be adjusted? (that's a classic case for feature driven parametric design - not doable with Rhino).

4. This concept (strut axis exported only) is tested in FORMFINDER and some other far more complex membrane apps that I use quite often (not RhinoMembrane). Here's is what FF tells us about:

Observe a different kind of "stress" when this is converted to radial type:

5. If you insert the stp file to the Rhino file provided (exactly as exported from FORMFINDER - no mods of mine of any kind) you'll see what goes where (and why). That way the usage of double cables is rather obvious (and a lot other things - for instance the way that the struts achieve "equilibrium", see the slots in the base mount plate.

6. If this approach is worth considering your definition requires some serious rethinking (far more simpler/manageable with the drawback that the real parts they are "static" they can adjust only as far this particular solution allows them to do - controlling them parametrically is clearly impossible with the current state of R/GH capabilities).`

All in all: this case works because the cables push the anchor points downwards and the struts push them upwards.

more in a while

Permalink Reply by peter fotiadis on May 27, 2014 at 8:08am

Here's the FF file

Attachments:

237_UAE_shaderSS316L_v4B_FROM_FORMFINDER.3dm, 991 KB

Permalink Reply by peter fotiadis on May 27, 2014 at 8:09am

Forgot the real thing

Attachments:

237_UAE_ShaderR2.zip, 1.3 MB

Permalink Reply by Nenov Ivo on May 27, 2014 at 10:23am

Thanks for sharing your file Peter and Thank you on investing your time to help me. The thing is that i don't know what to do... i can try to make a realistic project that will take me a lot of time and work that i won't invest on my architectural part or make a sexy grasshopper thing (professional liars like you said) that won't work in reality but will let me finish the boring part of the work, the plans, the autocad bullshit... I will decide this after seeing my professors tomorrow... Also another thing, my professors understand nothing about structure, they just build in concrete ... hehe.

As i see your replies, i think you're an engineer, am i right ? So may i show one of my work that will maybe, almost sure be built, i send you the renders right now in private messages.

Permalink Reply by peter fotiadis on May 27, 2014 at 12:17pm

i think you're an engineer

Er...I'm Architect. What exactly do you mean by "engineer"? (maybe a structural engineer?)

my professors understand nothing about structure

Well...why I'm not surprised at all?

they just build in concrete

So did T-Rex and Brontosaurus Gorgeous.

Permalink Reply by Nenov Ivo on May 27, 2014 at 10:25am

I didn't know we have to be friends to send a private message...

Permalink Reply by peter fotiadis on May 28, 2014 at 1:05am

Warning: I'm not a good person (avoid friendship with me when possible).

he he

Here's the latest (and greatest)

1. This started life (the critical bits, that is) far far and away from Rhino and GH. But ... have a close look on these weird "slots" in the base mount plate - allow the struts to "follow" some base "auto" arrangement (up to a point).

2. After various ... er ... hmm... "communications" with a variety of apps.(some of them are not for public eyes) ...here's a concept demo about what could be done and fool the academics (that's the bit that I like the most)

In plain English (work in GH):

1. Create some wires that represent the struts and PAY attention on their limits of adjustability.

2. Create a nurbs curve through the points indicated with "balls" in the demo. Patch the nurbs.

3. Trim the nurbs surface with some "indicative" profiles OR use Kangaroo by applying a minimum possible relax state (if the latter add the rhomboid cables as well - they deform by pulling the membrane downwards).

4. Optionally put the real things in place (quite GPU taxing that one - do some Viz control).

best, Peter

Attachments:

237_UAE_shaderSS316L_v4B_CONCEPT_GUIDE.3dm, 2.2 MB

Permalink Reply by SA on May 28, 2014 at 2:48am

Great work !!! And thanks for sharing.

As a newbie in kangaroo had a little success till the point I could develop my desired tensile shape but struggling with the fabrication of the non develop-able surfaces. Any suggestions are more than welcome.

Attachments:

cone.jpg, 57 KB

Permalink Reply by peter fotiadis on May 28, 2014 at 7:19am

You mean that you intend to work solely with GH/Rhino? (or you use apps the likes of NDN, Aeronaut, Forten4000, K3, Mehler and you don't know how to do it?).

Permalink Reply by SA on May 28, 2014 at 7:40am

Hi again,

No I have no clue about these but you think its hard (by which i mean not so precise) to do it in Rhino/Grasshopper or Kangaroo for that matter or may be Python? I am open for trying new softwares, hopefully they are not rocket science unlike the ones mentioned above.

Permalink Reply by peter fotiadis on May 28, 2014 at 8:51am

Here we are ...

BEFORE READING ANYTHING: I have a great respect for stuff the likes of GH/Kangaroo/Rhino etc etc.

(1) Membranes ARE NOT like anything else you know. You can't design (or if you are student your scholars should NEVER steer you into that approach) them by just finding nice shapes and that's all. A myriad of "auxiliary" parts are required (either ready or custom for bespoke projects), special software is required for force/load calculations and/or INTERACTIVE/LIVE formfinding (especially in nested complex cases - the norm these days), wind behavior, thermal behavior (closed envelopes), fabric choice, cost estimation, longevity estimation and ... unavoidably some close collaboration (on a daily basis for projects worth the name) with specialists the likes of Birdair. Membranes are a classic case where the bottom-to-top design mentality is King.

Moral: these things require a totally holistic approach: do it to the end or do something else.

(2) Membranes comply with the 2 axioms of architecture best than anything else. Bad news: It's like dating Charlize Theron: most unlikely but worth trying.

Moral: long is the path (and hilly).

(3) Therefor ... some form finding process/approach IS BY NO MEANS enough for the job. This is the reason that relatively few people in the world (statistically speaking) know how to deal with these animals - don't forget to add high class aesthetics (bad news: designers are born, not made).

Moral: buy a mirror, what tells you?.

(4) Membranes are like smoking: most unlikely to quit. Don't start playing with them ... they'll change the way that you see things for ever (a dangerous path that one).

Moral: handle with care.

All in all: I would strongly suggest to have a look on membrane design/fabrication software (the more the better) first and then ... return to Rhino/GH ... but only after having the big picture in mind. Bad news: the best membrane software known to man ... it's not for sale.

may the Force (dark option) be with us all