Issues with extruding a tiled pattern

[ohmarinus (Marinus)]

Hi, I’ve been working on tiling of a flexible system I’m designing.

After some effort I managed to get it to tile in a way (not perfect, but getting there) and now I want to try out if I can even extrude my patter and this is where I get stuck.

This is the pattern type:

Is it generated with the offset tool from the Clipper components from a base curve like this one:

The issue that happens is that the Clipper Offset component creates a polyline, otherwise known as a segmented curve, this will hinder making an extrusion by a lot:

So instead I looked at generating a better type of curve, which I managed to do with Offset Curve:

The result is a nice curve instead of a polyline, though this offset is missing the nice round cap that the Clipper offset has:

So I went back to Clipper to see how else I could extrude the offset curves and then I ran into the issue that the Clipper result doesn’t respect the boundaries:

So then I started trying things out with Unions, Mesh Unions and other ways to subtract the shapes and manipulated lists to see if I could subtract the smalled holes from the large outline surface:

This has unexpected results:

I hope my screenshot journey is still able to catch your focus, anyway, I’m at this point where I’m stuck where I don’t know how to make a better offset and a better intersection of the shapes I’m generating. My goal is to be able to make even bigger and more complex shapes, so if this smaller and simpler approach is already bugging out or just taking too long to generate, I’m becoming slightly pessimistic for the future of my project. However I do not believe there is not a solution.

Would it help speed things up if I start using plain values between 0.0 and 1.0 instead of a mix of integer and floating point?

Is there a way to generate the desired offsets and extruding them, without slowing down the script too much?

Here is the file with internalized data, I hope someone has a piece of advice:
curve_offset_extrusion.gh (736.0 KB)

This is what the file looks like:

[maje90 (Riccardo Majewski)]

Your curves have duplicates.
Also, please attach the single, central path.

[ohmarinus (Marinus)]

Hey, that’s strange, thanks for pointing that out, I see that I made a mistake somewhere with the tiling.

curve_offset_centerline.gh (136.8 KB)

Here’s an updated version with something that looks more like something I’m trying to achieve. Also fixed the multiples.

It also looks like the solid union leaves some undesired artifacts:

[Ryan14 (Ryan)]

MergeCo-plainerFaces.

[inno]

@ohmarinus

It also looks like the solid union leaves some undesired artifacts

I think you can try to remove those with Merge Faces:

also there is the secret Pipe intersecting XY Plane technique that might work to create the offsets faster?

curve_offset_centerline_Re.gh (151.9 KB)

[Quan_Li]

curve_offset_centerline Edited V0.gh (29.6 KB)
How did you get the curves, look beautiful;

[ohmarinus (Marinus)]

Hi, thanks, I saw this pattern over on thingiverse: Mesostructured Cellular Materials: Early Prototypes by AndreasBastian - Thingiverse

And I was simply wondering how I would go about and make that So it was for simple practice.
However, as with many things in life, doing this type of stuff will also bring about new ideas and now I came up with a new project that I can utilize this for.

Sadly the links in the creators description are all dead links now, but the idea by itself is already very inspiring to me.

Making a small piece is not that hard and the compute time is between 5 to 10 seconds, but when I start to enlarge this model this is where the issues start to arise. Sometimes hijacking my computer for over a minute and the compute time seems to increase exponentially even though I have 32Gb of RAM, a dedicated GPU and a CPU from 2020.

The search for a faster compute time goes on

[ohmarinus (Marinus)]

Ah yes, of course, how could I forget about Merge Coplanar… Thanks for reminding me.

I tried out the pipe solution which is a nice workaround, however this still doesn’t give round caps, as when I set the cap to round, the Boolean Fails and I can’t for the life of me find why. Maybe it’s having issues where the edges of the round caps are exactly on top of each other?

Ah, fixed it, what I did was remove the boolean after the Flip and introduce a Region Union in between Plane Section and Surface:

This way the shapes have a rounded cap but then it messed up boundaries when extruding:

Here’s the changed file:
curve_offset_centerline_Re_Re.gh (27.5 KB)

Update:
Wait, I fixed it by adding Boundary Surfaces:

New fixed file:
curve_offset_centerline_Re_Re_fixed.gh (27.4 KB)

This also sped up the extrusion!

[Quan_Li]

[Quan_Li]

Perfect for viewport background

[ohmarinus (Marinus)]

Ha! That’s actually not something I thought about
Yeah, looks like it works well for something like that.