I want to create this parametric bubble pattern shown below with grasshopper which will enable me to to use it in multiple applications and resize it accordingly.
In the image above you can see where I am in grasshopper at the moment (left) and where I would like to be in the future (right). I hit a wall in grasshopper when attempting to copy this piece over and shift it to continue the pattern.
To create the pattern on the right I baked my current geometry, and modeled the rest of it in rhino. As mentioned I would like this to be used in multiple instances and as projects arise I would like to be able to parametrically adjust it to fit the context of the project.
The end goal is to create vertical contours of this geometric pattern, CNC mill the pieces (3 axis), and also somehow embed planter boxes into each of the bubbles, making a green feature wall. But for starters.. some help just modeling the wall in the way which will allow me to parametrically manipulate it: size of the wall, size and proportions of the bubbles, etc.
Also, from the point of aesthetics, I would like if the seams between bubbles werenât so harsh, so any advice on that would be helpful as well.
Perhaps for this youâd have to reconsider the starting segment (arc) thatâs getting lofted. These perfect half-circles will contribute to that âharshâ seam. You could attempt to create your own nurbs profile to smooth the ends of the curves:
The fillet produced weird results once baked, so I created this workaround solution that is point based rather than using arcs. I used the prior solutions in this thread as a base file and built off that. Does anyone know why loft isnt working here?
Also any advice on cleaning up this file? It feels redundant and im sure there is a simpler solution.
Seems to me that âplanter boxesâ is the whole point of a âgreen wallâ, yet I havenât seen them yet? Also, there is a structural component of a wall of planter boxes? What holds it up? How much dirt and what kind of plants will it hold?
I havenât addressed structural issues myself yet, though I can imagine stackable boxes, like concreate blocks. OrâŠ
I substantially rewrote the code I posted, with a novel solution to staggering the planter shapes and making the shapes more accurately. Starting at the bottom, odd numbered rows have wider shapes at each end. There is a switch (blue group) between all separate shapes (perhaps best for stackable boxes?) or each row as a joined brep.
P.S. I very hastily added stackable boxes (white group). Sorry but donât have time now to add back surfaces, structural details or check the code for errors.
I agree with Joe that some explanation or decision making has to be made on the constructive side of things, but following up on the displeasure you voiced over the the filleting of the the semi-circles:
This script gives alot of control over the a 2nd degree curve created as a fillet between two Bezier spans set out of phase and of different amplitudes, in the 1st and 2nd halves of a period.
I used MetaHopper to set the domains of the sliders. Unfortunately, the Get Connected Objects needs to be connected directly to a slider, so I couldnât place it inside the cluster expanding the number of cluster inputs.
You can use the resulting output geometry and flip and array it to get your sequence of joined curves. Here, what the cluster looks like on the inside:
