I am trying to create a contour of equidistant curves on top of a mesh with double curvature. On a planar mesh this is easy to achieve with the Contour component. However on a mesh with double curvature the resulting curves are not at equal distance from each other. Is there any trick to achieve such a geometric result.
Equidistant curves on a mesh.gh (10.7 KB)
You could try the compass method which was once introduced by Frei Otto. There were some Plug-Ins, when I searched some months ago, but I ended up scripting a very buggy one myself.
I think David is asking about sets of curves going all the way across the surface in each direction at equal distance from the adjacent curves, not a grid of equal length segments(aka Chebyshev net) like you might use the compass method for.
I think that for a general mesh, finding these equidistant curves is an unavoidably iterative problem, so will require looping in some fashion, either with scripting or plugins (or just chaining repeated components together if you don’t need a huge number).
A simple approach can be to pipe the starting curve, find the intersections, pipe the intersection curves etc
The simple approach makes sense and I believe will work wonders. I thought that it will be iterative too.
Yes, I think that should work too. The source would need to be all the vertices of the starting curve.
It ought to be something like this. I just don’t know how to loop it.
Equidistant curves on a mesh_v2.gh (17.0 KB)
Geodesic distance is what you are after
Can you upload it? I think that you are correct on the geodesic part.
Here it is
I wanted to test it with a subdivision at level 2 and it didn’t finish !!
Heat distance method with Heat component
My tool using Geodesci WaveFront coding
Aren’t you trying to do this only from the edge of the surface not a point on it?
Kind of but with a linear distribution (with a guide vector as a direction) not radial as this example.
Like for my tool this implementation of Heat Method takes points, so like said by @DanielPiker
it is what I have done, it works quite well with 200 points, not good with 50 points.
I hit a roadblock. The Nautilus is for Rhino 7. I am still stuck at 6. Will upgrade straight to 8.
There must be some mesh iso splitting somewhere
You’ll have to extract the edges and then find the good ones !!
That heat method discussion does go into vectors later on.
See this link
There is a “Mesh Contour in Field” Millepede
I have managed to create a working loop and the results are close to perfect.
I get small deviations when it comes to the length of the individual quad edges.
What I notice is that the quads that result from this method progressively lose their orthogonality.
Do you know of a method to keep the quads as square as possible? They don´t need to be planar.
Or maybe it is impossible to achieve this as well?
I upload a result here.
Equidistant Mesh.3dm (151.3 KB)
Do you mean the quads formed by the green curves here?
Can I ask what’s the end result you’re aiming for with this?
I’m guessing it’s maybe something to do with cutting patterns for a fabric structure?
If so, depending on whether you’re aiming to minimise twist within each strip in the assembled form, or to keep the flat pieces as straight as possible, or with least width variation, or some trade-off between these,
you might want different curves, eg the 2 principal curvatures (which will intersect orthogonally), geodesics, or contours of geodesic distance.
Yes these quads.
I wonder if it is possible to create an ortogonal cable net and then fill the quads with simple fabric squares.
Such a thing is very simple to do with plane geometry. Troubles arise when you try to do it with doubly curved shape.
