I’m currently working on a project using bionic structures to reduce the weight of parts made by 3D printing.
Since my university doesn’t teach Rhino or Grasshopper, I tried to learn the Software all by myself. Therefore, please don’t be surprised about some basic stuff I don’t know yet
After baking the cube (with the bionic structure inside) I receive a 95% solid part. But by having a closer look at the surface of the component, some kind of graphical errors (kind of holes in the outter surface) can be recognized. For this reason, I’m not able to export the component into an STL. file.
Do you have any idea why these mistakes occure and how to fix it in the grasshopper algorithm?
Furthermore, I think about transforming this “shell-model” into a solid component. Is it enough to just bake it and export it as a STL file?
by itself, I am not sure that “solid intersection” should give any specific result, given that a mesh is open. However, seeing that it does, it seems that it does not do it accurately. It’s unfortunate that mesh Booleans are not very robust in Rhino, and we are trying to improve this for future versions of Rhino.
For now, I have a suggestion. You could try to increase the amount of faces of the “box” surface. This way, the intersector would not have to deal with too many faces at once, and it will be more likely to succeed. I’ve tried this with Weaverbird’s Split Triangles subdivision, and that gives a good result. Check it out… (tested in Rhino 6 BETA)
Thank you very much for your quick response. Because of your editing I now fixed my problems with the cube.
Additionally to the cube, I was also trying to apply the algorithm at more rounded parts. But in this case, almost the entire outter structure is missing. Unfortunately, the Weaverbird Split Triangles doesn’t work here.
Do you have any idea how to solve this problem?
I also added photos of the part, but the algorithm is still “test-patched”.
sorry I have no idea regarding how to fix Rhino’s mesh Booleans for this. Please note that your method, as it stands, does not ensure that with curvy objects the results are still coincident with the surface from which they were created. One alternative option (maybe where you could increase tolerance?) might be to try with CARVE_GH: http://www.grasshopper3d.com/profiles/blogs/carverhino-carvegh-a-wrapper-for-the-carve-mesh-boolean-library
A further, maybe better, alternative would be to use a more constructive method from the beginning, but it would involve creating the geometry manually. I am showing here a method that works that way: test-patched-build-exterior.gh (28.7 KB)
It also uses MeshEdit to weld with tolerance (Boolean operations introduce a lot of tolerance in the model).
I think this should be able to put you back on track. You might have to fix your previous logic a little with curvy objects, but it’s mostly working.
Giulio
–
Giulio Piacentino
for Robert McNeel & Associates giulio@mcneel.com
thank you for all your effort and time you put into the test-patched build exterior algorithm. It was a great example for me to improve my skills especially with Weaverbird.
Over the weekend I tried to work with Carve_GH but it wasn’t really successful.
By having a closer look at your test algorithm, I figured out that in my opinion the algorithm is not the problem. I think the problem is more related to the lack of points at the outside of the component. By reducing the amount of points, the component becomes more angular because the software builds the outer structure by connecting points with straight lines.
I assume by increasing the amount of points in the “populate geometry” the component becomes more rounded. Buy first, my computer is not powerful enough to use more than 80 points and additionally, more points also lead to a buggy inner structure due to the in
Therefore, I’m currently trying to figure out how to use more points to build the outside by keeping the amount of points on the inside of the component low.
As I told you (and the community) that I’m not an expert in GH, I really hope for additional help because until now, I couldn’t solve the problem…
You need to figure out better what is going on in the interior structure (the definition that was already there before I showed you how to approach the exterior). That is because there was a problem that is still present now.
I’ve given this a go, fixing a further problem in the logic. I’m thinking this could be made easier if we only used the opposite, that means, only the interior and capped it in the ends. test-patched-2.gh (29.1 KB)
Thank you for your reply. Unfortunately, I just recognised your first reply that day, therefore the late response.
I checked our your test-patched-opposite file.
Yes, I got your point. The opposite-algo is distinctively easier but not exactly what i was looking for. It just gives me an interior-like structure with capped ends. A solid exterior structure is missing.
At the moment, I try to figure out how to keep the interior structure but adapt the way how the exterior structure is built to be able to get a rounded exterior. Do you know how to additionally add data from the rounded outer structure to the exterior mesh construction?
Enclosed you can find the two different algorithms applied to the same curvy object.
A word of caution, though: at this point this project seems to be on the road to become an “infinite” assignment to always solve better. If you are up to that task, that’s fine and I could suggest some strategies, but this is something that will require some rather large time working on your side. I have too many other projects to work on.
Alright, i got your point. I know it’s a more problematic task and I can understand you. I’m now trying to figure out other ways how to get along with my first idea because there is still my project that has to get done.
Something that would definitely help the Boolean process is the fact that the inner structure exceeds the “shell”. But probably the most stable method would be one that does not require any booleaning. Maybe lofting a surface between exterior face and inner structure, and then meshing that with minimal points.
