Determining the compressive force lines (Guastavino Tiling Arch system)

[Liana1 (Liana)]

Hi everyone,

I am trying to create a parabolic vault made out of ceramic tiles inspired from Guastavino’s Tiling Arch System. I have attached my current model’s snapshot but the gridsystem of the polysurface is random. I am trying to find the lines of compression experienced by this shell so that i can arrange my ceramic tiles in that order to design further.

any help or direction would be great, thank you!

[PeterFotiadis (Pfotiad0)]

Some tips:

1. Be rational: AVOID double curved envelope modules (tiles/cats/dogs/whatever).
2. Get the gist of K2 (the most important GH add-on by a million miles [add some more]).
3. Do a flat layout (i.e. some BrepFace) and some Mesh (the more “even” it is the better). Like:
   
   
   .
4. Relax the Mesh (that does the K2 for this case). The shown C# is using a variety of Anchor Policies (6 in fact) plus a variety of vars for the K2 Goals required for the job.
   
   
   
   
   
5. As an option do some Truss. Shown a W type with the bottom layer (hexagons) upside or not: always use the triangles for the envelope - avoid quads and the likes (planarity in real-life is a chimera ).
   
   
   
   
   
   
   .
   

[Liana1 (Liana)]

Hey thanks for getting back so promptly. I understood the anchor points although I am not at all familiar with C#. Do you mind sharing an easier way to approach or send me this script so that I can fiddle around to get my understanding in it?

[PeterFotiadis (Pfotiad0)]

@Liana1

send me this script

Well … this is 100% C# code thus at present time it would be a big black box to you. It could cut the mustard quite easily … but only if you want THAT type of approach/envelope design. I mean it could do various other “variations” as well … but for that you should do things that you can’t (unless you want to walk the C# walk).

Let’s reset the whole situation: First - given that you want to exploit the merits of K2 (i.e. particles, IGoals and the likes) - you should get the gist of that freaky thing. Time required 1 - 2 months.

BTW: There’s various tutorials (i.e posted cases etc) on the subject: how to relax a Mesh and/or a collection of Geometry (say lines - see pics below). For the latter see some other vault option: get lines via some sort of logic, do the Ccx all VS all part, get the points/segments and then use K2 to relax the collection of segments (out of the Ccx Events). As usual various Anchor policies are possible (spot some “inner” Anchors used).

OK the last 3 images are not for engineers (unless a giant 3d printer is available) … but you can imagine the possibilities.

That said the last 2 are just K2 VS a flat plain humble ortho grid … meaning that you can start from something that nobody rates high … and then … blah, blah. For you THIS is the way to start to walk the K2 walk.

But all the vaults are a bit pathetic (Topology wise) … thus I hate’m. Even a geodesic dome (also a bit pathetic) is better - at least is rational.