Relax Surface

grasshopper
rhino5
unhandled

(Ana) #1

Hello, Is there a way to create organic routes via rhino or grasshopper? See sketch, red lines indicate proposed connection between opening and organic routes connecting them.
Thank you a lot. Best,
I am not expert in grasshopper, just learning something new for myself for educational purpose.

wqunnamed.3dm (167.9 KB)


(Pfotiad0) #2

Easy with Kangaroo2 . Notify if you want a demo (but better have some first hand K2 experience )


(Ana) #3

Hi I tried doing same as per relax tension tutorial, but they create the mesh boxes first, didnt work within the script. wonder if it is the right way for my purpose, please advise with some tutorial maybe or starting script. thank you a lot:)


(Pfotiad0) #4

Provide something to work: say a def that does the structure captured (I guess some sort of artistic frame). Be very explicit with regard the rules desired for defining the openings (the rectangles) to bridge.


(Ana) #5

The main part that i dont know how to do is how specify a a geometry linked between the curves(to bridge them) all i am looking for is to have adjustable rhino curves and automatically generated bridges between them( routes). hint? :slight_smile: sorry trying to find some logic in here.hard
ana.gh (12.9 KB)
ana.3dm (474.7 KB)


(Pfotiad0) #6

First things first:

Get a dual (K1 099 or K2 231) def that does the elastic inwards effect.

We’'ll see later how to define the mesh out of various openings (rectangles) .

Warp_K1orK2_V1.gh (16.9 KB)


(Ana) #7

Hi thank you, it has worked somehow. wonder if there a way to automate mesh being connected to the curves, so you drag change the curve opening and it adjusts the surface.
Sorry not sure i understood your comment, but tried to use it for this geometry purpose. Thank you Peter.
ana.gh (22.5 KB)
ana.3dm (654.4 KB)


(Pfotiad0) #8

Since nobody else answers to that: Of course there’s several ways to do it … but in my case several C# defs that (I suspect) would be useless to you (unless you want one just for fun). See a similar random mesh generator (any configuration imaginable) that feeds K2 231 for the final elastic effect.


(Ana) #9

Can I have files rhino/grasshopper. ps. yes the C#…coding are too complicated, never even tried to understand it, not that much into grasshopper yet, just trying something new for myself.
so, in the pictures i see you create mesh boxes but what if i need to done based on openings, not manually drawn mesh boxed routes?


(Pfotiad0) #10

The pictures show a random mesh generator that makes the ideal (so to speak) background in real-time (zero delays) for creating any elastic solution with any complexity having that type/class of topology in mind (one cental spine of variable N of edges/size/hight/etc + any random protrusions in any position).

I could very easily modify some lines for doing exactly what you want (using openings in some facades, so to speak) … but for what? It could be a freaky black box for you.

Anyway get the attached (a static demo related with the above + native components ) as well and I’ll try to mastermind some solution that uses native components on facade openings.

Warp_K2_staticDemo_V1.gh (25.3 KB)


(Pfotiad0) #11

Find a couple of minutes and added 2 sorts or things:
Warp_K2_staticDemo_V1A.gh (48.8 KB)

1.The first gets a closed curve measures the segments lengths and makes variable divisions (trying to create “even” width results) on the extruded surfaces (out of the segments) and then makes randomly frames (or profiles if you like) on the surfaces. It can make the full combo (no) or some (yes) depending on the eliminationFactor variable:

2.The second is 1200% Academic: creates a random “spine”, feeds the 1 with that and then creates curves (kinda Bezier spans) - NOTE: if no spine no connecting curves are made - that could serve as paths for extruding the frames (sweep1), then unite all with a sweep1 along the spine (the union takes a year and half to finish), then use MeshMachine (that’s fast) then make the Mesh and then relax the Mesh with K2. But … in real-life this VERY slow approach violates the rule N1 in parametric (the interactive thingy) … meaning that is just a naive theoretical attempt to solve a challenging problem.

What’s the real-time way to do that? Only via voxels (creating a 3d grid of “boxes” and taking care that all have a face to face contact with their neighbor boxes). But this requires code to do it.


(Pfotiad0) #12

Switching entirely to C# and using all the tricks known to man. Despite all that …here’s a history of a misarable failure (good news: up to MM/K2 phase response is bettered by a factor of 100, bad news: miserable final K2 results unless you want to wait a day/week/month/year). Using i5 (no), i7 (maybe) and i9 (hope dies last).

  1. Changed policy with regard the final (to MM/K2) tube like Brep. Using flat regions instead of tube like Breps (plus a lot of checks for dealing with non manifold and/or self intersecting curves etc etc - that’s real-time) :

.

  1. Using the flat Breps thickened tubes are made (that’s very tricky) and combined with the central spine tube yield the final Brep for the MeshMachine and K2 (that’s almost real-time):

  1. Control Alt Defeat: MM/K2 take a very long time (for good reason) for a decent result … meaning that the whole approach is not interactive and thus is naive and stupid. Here’'s bad results by limiting the N of loops:

Moral: a misarable failure, what else?