Find the curvature of a segment in the geometry

Hi all,

I have a curved geometry and from this geometry I want to section panels that do not exceed a certain deviation. I want to cut out the panels as flat as possible.
I first made a gradient of the curvature to see where the flat (white) parts are and then I made a box around two points and calculated the angle, which is the parameter that refers to the deviation.

Now this box does (obviously) not box the other points on the surface, while eventually these points within the box range have to be taken into account for sectioning
the panels.
Further, the box is not in the relative u,v,w orientation.

Therefore my approach does not work yet.

Does anybody have an advice on how to extract panels with a certain curvature deviation from a curved geometry?


@simonnpro Are you interested in deviation of curvature, direction of normal vector, or position?

Hi David,

Thank you for your reply.

I’m interested in the deviation of the curvature.
Eventually I want to section panels as flat as possible, with a certain allowed deviation (for example: allowed deviation of 16mm per meter).

Does this make it more clear?


Which curvature? Curvature of a surface in general is not a scalar but two scalars plus a direction. (See my response in your other thread.) Why is curvature important?

You mentioned creating panels. Do you you want to make panels which are “close enough” to planar? If so then Gaussian curvature may be the appropriate metric.

What you are suggesting is right, I want to make panels which are close enough to planar indeed. I will dive into the Gaussian curvature. If you have an advice on which approach might be suitable, I’d love to hear it!

Thanks in advance


After some more thought I’m not sure Gaussian curvature or any other curvature metric will be of much use to you.

Divide surfaces into flat rectangular panels has been extensively studied and has been the topic of previous threads here.