Containment Error

Hey guys,

I got this error while working with containment component, "Curve point relationship cannot be determined" ... At first i thought this was a plane problem as some of the points were in up in z-axis, but the error was still there when I turned all z- values to 0... although all the curves are polyline curve, yet majority shows result after containment and some give this error... the white planar surfaces in the image are the one that shows result and empty curves with point in the center are the one which gave the error... the image shows one branch of the tree... the remaining error curves are in the other branches hence the long list of error...

What might be causing this problem?

Files are attached...

Thanks in advance

Attachments:

containment error.3dm, 4.7 MB
containment error.gh, 26 KB

Replies to This Discussion

Permalink Reply by Ching on January 7, 2012 at 7:34am

I *think* it might be a bug in either rhino or grasshopper in dealing with lots of decimal places or some other issues.

I have changed the units absolute tolerance settings in Rhino from 0.1 to 0.0001 and 0.000001 to see what happens. And I found out that the grasshopper is giving some different random Z values for the points inside the Containment component when in theory they should all be 0. Nothing else is changed, just the unit tolerance settings.

Weird.

Anyone knows why?

Permalink Reply by David Rutten on January 7, 2012 at 1:03pm

It's not a random value, or rather, it's not any random value. 7.0040e-23 is very, very close to zero. Almost all mathematical operations in the computer that deal with floating point values introduce some inaccuracy in the least significant digits. You should start getting worried when the noise gets larger than 1e-12.

David Rutten

david@mcneel.com

Poprad, Slovakia

Permalink Reply by Yoann Mescam (Systemiq) on January 7, 2012 at 10:21am

Some curves are not closed.

Permalink Reply by Khizer Ishtiaq on January 7, 2012 at 10:32am

Hmm... Then how can Brep/Plane intersection component is creating open curves... since the geometry is closed and upright? the section should be all squares

I'll try moving plane a bit up ....

Permalink Reply by Khizer Ishtiaq on January 7, 2012 at 10:58am

Moving up the plane 0.1 units resolves many errors, but there are still 21 remaining... and when i move 0.2 or 0.3 units, the number of error goes upto 35, the absolute tolerance of the file is 0.1 and changing it to 0.01 doesn't change the number of errors, after re-plugging the wires.... I guess the issue was not containment but the intersection component.... that explains the change in values when Ching changed the absolute tolerance values...

Permalink Reply by Khizer Ishtiaq on January 7, 2012 at 11:12am

It works when the plane provided for intersection is moved at 0.13 units in z-axis .... No errors... Thanks guys ...

Permalink Reply by Yoann Mescam (Systemiq) on January 7, 2012 at 11:03am

Well some Breps have volume=0, extruded surfaces with a vector with 0 amplitude ?

Permalink Reply by Khizer Ishtiaq on January 7, 2012 at 11:13am

But here every brep has some volume...

RSS