Hi people, was just wondering how you could do something like this through grasshopper to test different options a lot faster than manually doing it?
Would love some tips or starting point for me to investigate
Cheers
Hi people, was just wondering how you could do something like this through grasshopper to test different options a lot faster than manually doing it?
Would love some tips or starting point for me to investigate
Cheers
It’s very difficult to reverse engineer any logic from an image! What you can tell first though is the outer shape is a constraint (it looks identical each time, just rotated a bit for reasons unknown). It might be that you can create some inner form in a more orthogonal way, and then trim this using a Solid Intersection or something similar.
The way the forms inside are configured seems to suggest some sort of stepping of volumes, so you would have to build this logic inside your grasshopper definition, with parameters for each level up which can adjust the size of each element, then Solid Union them together. Use the Box Mapping component on the result to get it aligned with the target shape, and then a Solid Intersection might be one way to go. It all depends how much logic is built into the form generation and how much is essentially random. Just having random shapes would give you something like this (note a much simpler target BRep shape here, but essentially you could make one with a courtyard similar to that above just in Rhino):
Essentially, if you’re searching through many options then you are searching through the parameter space (genotype) that generates the design (phenotype). It’s usually this kind of loop:
Constraints + Parameters > Modelling > Analysis > Evaluation > Optimisation (loop back to parameters).
So you need to work out what logic you want to build into the model, and exactly what it is you would like to search for (or just pick things that feel right, I will plug my own tool ‘biomorpher’ for this).
As always, the best place to start is not actually in Grasshopper, but with a pen and paper, trying to work out a) what the constraints are, b) what the adjustable parameters are and what they control c) what the goal(s) are, d) are they measurable. Then try and build something and be prepared to change the assumptions as you progress.
John.
Here for example is a very very simple model that creates 3 stacked boxes with 3 size parameters (9 in total), with the corner for each box based on one of those of the last. Finally a parameter for rotating the shape before mapping to the target BRep. Something similar but requiring much more effort could be made to recreate something similar to the image above, but as I say we really need more information as to what the geometry means otherwise it’s just mindless form-making (and we got through that phase in the 1990s). Hopefully this will give you some pointers however.
Thanks for the reply John, sorry for not being very clear, I am referring to this building https://architizer.com/projects/urbain-villa-4-in-1/ where the units wrap around the facade, so thediagram above is showing how each unit occupies the form. Cheers for your help though without much context I think this will definitely be beneficial.