hello, its been a while since I have used grasshopper and Kangaroo etc. I was going back through old projects and wanted to have a go at this project. The proper and mathematical way.
I was just wondering where would be a good starting point to begin to develop this type of structure. I as s going through the forums and couldn’t exactly find what i needed. There has been so many plugins, tutorials, etc that is hard to find a starting point for research.
My main goal is to
create and optimize the form
run structural analysis of the form
create different structure studies (waffle, Steel tube, space frame). Which i vaguely remember how to do
Full analysis with structure.
Any help to where to begin with the form would be awsome! thank you everyone.
I quickly got a sketch of four arcs and a base surface mesh but being an amateur with Kangaroo, flailed for hours on that part (white group). Finally got something looking decent though, and understand much better how it all works. Balancing ‘Strength’ with damping and tolerance values was a big part, once I got the right(?) ‘Solver’ goals defined.
Well … having triangles in mind (quads require planarization: avoid) in case that you want a LBS with no trusses (say via self supporting modules as in the images below) the very first task (after some relaxation) is to achieve torsion free beams via K2 (I think that Daniel has posted an example or two on that matter - but I can hardly recall the post):
Green is planar (where planar means that the min diagonal distance is less than tolerance) and then … the more the red the more the deviation . That said avoid “waffle” type of LBS (nobody does that in real-life … with the rarest of exceptions).
In case that you want a W truss (recommended) … you’ll need to compute connectivity and via that to compute axis to axis angles in order to deal with clash events between, say, MERO members (connecting bolts, sleeves, cones, tubes and the likes). For what means clash see these:
Obviously you should treat all the “solid” items as Instance Definitions otherwise you would wait R a week (or two) to finish (and your file would be a 500Mb pointless thingy).
Then comes the envelope: depending on the solution this may become far more challenging than the LBS.
But … the most elegant way is to do a big U-Turn … go after tensile membranes … and forget all these “old times” shapes/methods/restrictions:
Thank you very much, yes this does help and will lead in the right direction. And it does help me better understand where to begin, in the end i thikn i was leaning more toward tensile structures, which i do recall doing in the past. But the structure and LBS side is intersting and will give it a shot when i get that far into it. Thank you
Remember: nothing in real-life can be achieved without code. Other than that - if you are not after tensiles - I would suggest to start with stand-alone modules that are easily assempled on site (with the envelope fixed in place as well) instead of designing a complex node that could handle a variety of levels of freedom (also this is impossible without the likes of CATIA or Siemens/NX).
Here’s a topology “near” the one that you posted above (tosion free beams as explained above etc etc).
With regard tensiles … well … the K2 relaxation is the 0.0001% of the task.The rest are tricky things related with designing the real-life parts of the system (most notably “hinged” anchor plates from a given anchor plate angle apperture and up). For this you’ll need feature driven solid MCAD apps and a certain experience. Not even think to attempt this with surface modellers like Rhino.
Yea that is very true, I think the goal is to play with the form on our own arch department. if we where to build it , we would send the initial form to a engineering company that would then guide the design into the feasible. Most of the similar projects we have seen have been tube steel, with tensile membrane over. or A structure like you posted above.
Er … we are on the opposite side of the fence: I’m after the form follows function thingy meaning that for me the form … well … is just a by product (tensiles excluded).
Anyway NEVER do that (i.e outline a form and then instruct “others” to finish the job) before fully resolving ALL the nuts and bits of the structure … for a very simple reason: structural engineering is about FEA stuff and the likes and has nothing to do with the core aesthetics of the machine (that’s the reason why Honda would never do a Ducati no matter how many zillions are available).
For instance imagine another class (so to speak) of envelopes: tensegrity truss systems - push by tubes, pull by cables (very expensive stuff, mind). These are the main 3 (abstract) types that we use in real-life (the double tetra is my favorite):
If the above (3 out of a zillion variants on that matter) are pig ugly … then the whole is pig ugly as well. Just visit Louvre and observe what the Master (I.M.Pei) did .
