I already submitted this post some time ago and eventually ended up with a pretty interesting result modifying an already scripted version. See links
However now i have been confronted with some kinks which i am unable to solve regarding the gradient ‘release’ of the deformation. The near vertices needs a sort of gradual ‘release’ from the initial shape. I am unaware of what component to use or how to deal with the data i can locate. I have attached my file along with my ‘boundary’ objects. I also attached a screenshot which indicate the where the vertices are located
I’m quite happy but kept noticing that the vertices around the shifting directions of the sphere would act differently. So i thought that the volleyball sphere (sorry for using wrong terminology) maybe isn’t ideal for this type of data. Perhaps a globus type would handle the data better. Ive attached two screenshots to compensate for my lack of words
I thought this would be a simple fix, but I have encountered some strange errors. Perhaps you or somebody else could shine some light on how to deal with this data in this branch of the file?
is there no naked edges in this type of mesh? I am quite confused on this notion
Scanned with my Artec Leo. Postprocessing in Artec Studio 16 Professional. After scanning I figured out how to align the balloon more centered and symmetric. It’s difficult to adjust everything perfectly.
I don’t get what you mean? You could divide the mesh further to maybe get a more precise result. The more subdivisions the better the simulation, but also the more computing power it will require, and the Kangaroo settings need to get dialled in each time the count changes.
I don’t think so. The mesh sphere ex has by far the most regular topology, which greatly benefits the simulation. The regular one is denser at the poles, which makes it inflate more in those regions and less in regions with bigger quads, which might not be desirable?
It produces an American football-like shape with creases at the poles.
The outcome of the simulation greatly depends on the mesh topology!
No, both the regular mesh sphere and mesh sphere ex don’t have any naked edges, since they are closed and watertight. Only open meshes have naked edges. For a simple, flat mesh plane, all its boundary edges are naked.
Interesting. I see your point, that was also why i initially decided to use the Sphere Ex, but without proper comprehension of the problems.
But to clarify my concern it not necessarily factually proven, but only from what I can see looking at the model. In the 3D print aswell as the GH/RH model the shape tends to have a sort of bump before releasing from the shape. And since it happens around the shifting directions of the mesh lines i thought that this might be causing the problem. If you think otherwise i certainly believe you, but I do wonder how to a get more smooth transition - If you fx. look at the balloon test
adjusting all the parameters in different configuration gives me varied result, and thank you so much for that
but there seem to be an inherited factor around those point that dosn’t seem to change. Perhaps “normal” sphere isnt the right choice but would it be better with a triangulated one
That is just a mesh un-welded edge. Weld your mesh. Then run “RebuildMeshNormals” command.
Or… you can just use the “Combine&Clean” component (from kangaroo tab) on your mesh, which will have the same result.
Anyway, yes, I think it’s impossible to have a completely uniformly divided mesh sphere.
Using Weaverbird and triangle-subdividing a Icosahedron give a little better results, but still with odd vertexes visible.
You’re also right that a triangulated mesh usually works best with Kangaroo, if the triangulation is nice and even, but I usually prefer quad meshes, since they are easier to work with afterwards, without having to retopologize the mesh, which might then slightly deviate from the original.
Note the base brep is a revolved section of the ballon which is split into 4 pieces and then joined again. It is split so the mesh later on has vertices exactly on the symmetry planes. Without vertices and edges along the symmetry plane, the collision would not work as nicely.
I’ve used a rod goal for the strings.
The solid point collision is set so both solid and points are affected by the collision.
Thank you so much for your input. I was truly amazed by the investigative skills into this topic. I am terribly sorry that i havn’t been able to answer until now. I simply didn’t have the time to open up anything until today. However this look really promising, i am really impressed. I like the method of the points and solids affecting each other. I honestly disregarded that idea because i thought it would be too difficult. But from what i can understand you managed to do it
Here comes the slightly embarrassing part: I do not understand what to do with these errors:
I wish i had a better understanding of grasshopper, but my only excuse can be that i am still a beginner. I hope to get better but for the time being i simply unaware.
This one is ceramic, but the real one is about 10’ tall and made (somehow) from 1/2’ steel plate. It looked like magic to me, but I guess he knows how to do that.
Which Rhino version are you using? That error looks related to a bug which was fixed a while ago, so updating to a current service release should fix it.
Hi @Ester
Which exact version of Rhino are you running?
You can find out by choosing Help>About Rhinoceros from the Rhino menu.
In the bottom right of the window that appears there will be some text like this:
Logged in as ***********************
Version 7 SR16
(7.16.22053.15001, 2022-02-22)
Commercial
