How to create lattice shell from 3D surface?

I want to split free surface into many pices of triangle edges and create lattice shell.
I also want to let lattice triangle edges avoid holes.
Please tell me how to solve the problem.

shell.3dm (1.3 MB)

if you need this

You have to use a mesher. I choose Meshmachinestatic from Daniel Piker.

  1. Join (or not) your Polysurfaces
  2. Extract the borders of the polysurface
  3. Choose the good length of the lattice edge length. Plug that data first in MeshMachine, by default it is one and it will take ages to mesh …
  4. connect the curves, points, number of itterations …
  5. connect a mesh component to to Plankton Mesh and you will get triangles. (258.1 KB)
1 Like

Thanks for the reply !

However, I still have 3 errors…

(This eroor window appears when I load grasshopper.)

(This error window appears when I try to open grasshopper document you created.)

(MeshMachine desappear.)

Please tell me what I have to do.I have no experience in using…

Well … IF this is some sort of AEC thingy here’s some advises:

  1. NEVER attempt to do an overcomplicated LBS (Load Bearing Structure). You gonna pay a big price (design wise) and when time to do it on site arrives > mama mia.

  2. ALWAYS separate LBS and envelope (in your case appears that you fancy irregural openings : do them on the envelope and let the LBS do the nice and clean job).

  3. Due to 1 > using nice and clean Curves (otherwise: garbage in, garbage out) do a sweep 2, get a SINGLE surface (if Rhino reports the result as polysurface > try again). Then do a Mesh (or subdivide the Surface) and do the LBS (if facets are triangles you’ll get max rigidity).

  4. A skin deep LBS requires a lot of material more (struts and nodes) and has (more or less) questionable seismic and thermal expansion behaviour (that’s critical). Do a W truss and use struts from Carbon Fibre (LOL) … you’ll be barely able to see the members in real-life.

Thus I would strongly suggest a concept like this one (W positive for clarity, but avoid doing that in real-life … meaning that the LBS should deploy under the envelope):


Well you have to install the components Plankton Meshmachine static … The most is to drop gha file on Grasshopper canvas.

1 Like

Thank you for your affable reply.
I study more.

BTW: Less is more (I’m not talking about the amount of study).

BTW: You can obviously “skip” some LBS modules (in the proximity of envelope’s openings) but that yields kitsch results in the majority of cases. On the other hand … if the top (truss inwards) LBS layer supports some false ceiling … then do it.

1 Like

BTW: Given the opportunity here’s a very expensive way to skin the cat (totally oriented towards engineering [aesthetics: just a by product]).

Imagine a tetrahedron where the apex to the base nodes are connected via struts and the base nodes are connected via cables. Imagine another (rotated) terahedron - with the same struts/cables concept - upside down where the new apex is at a point on the old tertahedron height. Connect old apex to all new base nodes (and new apex to all old base nodes) with cables and get what is known as a double tetrahedron tensegrity truss (pull via rigid members, push via cables). Do the ultra freaky thing (on a surface like the other concepts as above) that way and create openings only in the bases of the tetrahedrons (meaning: forget the shape of your openings … because the openings are just a derivant out of the whole topology). To the untrained eye a thing like this is impossible to stay in 3d space (but it does).

Shown the concept without the openings:

1 Like