Tensile membrane on struts

My tensegrity demo works very well in kangaroo before I add this membrane fabric(16 mesh) on top of it, my goal is I want all these mesh following the movement of my tensegrity structure when I'm doing the simulation in kangaroo, can someone help me fix this definition? thank you

Siyu

Attachments:

Testing Version Peter.gh, 59 KB
5x5 Peter.3dm, 755 KB

Replies to This Discussion

Permalink Reply by SIYU GUO on December 3, 2015 at 8:17am

is it almost like create a paneling logic in grasshopper?

Permalink Reply by peter fotiadis on December 3, 2015 at 9:07am

Well ... it's made (solely via code) by dividing a List of surfaces, checking for trimming info (that's very slow) and then "combining" the quad (or triad) of points with the centroid point added to the normal vector (per quad/tri set).

Random means that the vector length can vary between a desired min/max and sin distortion means ... hmm ... something "equivalent".

Clash detection (and topology validation) means that a variety of trigonometry based checks are performed in order to assure that tubes and the other paraphernalia ... stay clear each other AND the drilling axis in the balls are doable. But this is only half the bacon: since roofing material (say VM Zink and the likes) usually imposes several restrictions (for instance: you can't "bend" a VW sheet beyond a given angle etc etc).

Then coordinate systems (Planes) are made that are used in classic Plane to Plane transformations: i.e. take an instance definition related with a part of the system (sleeve, cone etc etc) always defined at WorldXY and put it in the desired plane.This has severe limitations in Rhino (actually is undoable and thus I very rarely do this type of work in Rhino/GH). What limitations may you ask? Well ... imagine fully parametric feature driven solids (the nuts and the bits of the system) that MUST comply (NOT yielding some clash, that is) with the whole topology OR they must change OR the topology must change: stuff for CATIA and the likes.

Then the structural department evaluates the whole thingy and gives the OK - if not another variant is attempted etc etc.

Then connectivity trees are made that relate nodes to edges etc etc ... in order to assemble the thing in real-life.

Then it's roof/envelope time ... and that is a pure nightmare.

BTW: having some NVidia Kepler K4200 (and up) is a must ... other wise ... well ... dealing with 20++K "objects" is not that trivial. NEVER attempt to do that type of stuff using, say, a laptop and the likes.

Permalink Reply by peter fotiadis on December 4, 2015 at 5:04am

And here's your stuff VS a "humble" MERO solution shown in wireframe for clarity (using some "options", he he: you can even have a pregnant truss (C)(TM)(US Patent pending)).

BTW: Who said that a MERO is a heavy weight solution? Was he under the influence of spirits?

Moral: Never is too late.

Attachments: