wall contact not detected

Submitted by richti83 on
Attachment Size
script.tar_.gz 1.19 KB
aaigner

Thx Christian.

I opened an issue in the master repository... somebody will care about it ;-)

cheers,
Andreas

Wed, 25 Mar 2015 - 18:35 Permalink
ckloss

Hi Christian,

thanks for the info - we'll have a closer look in the frame of the next release!

Best wishes
Christoph

Fri, 3 Apr 2015 - 09:35 Permalink
ckloss

Hi Christian,

thanks for posting this test case. I'll post a patch that improves the handling of such meshes in the release next week

Cheers
Christoph

Sat, 25 Apr 2015 - 15:26 Permalink
richti83

Hi Christoph,

could you please be so kind and tell me what in detail has changed in the code to solve this issue ? It's impossible to find the right place in the commit because of 4848 files has been changed.

Thanks,
Christian.

Wed, 29 Apr 2015 - 15:44 Permalink

I have encountered a strange behaviour in very rare cases when a particle hits exactly a common node of two triangles.
Attached a simple testcase with a single particle and a mesh with two triangles.
When the particle is at
0.075 0.0750001 0.425
everything works fine,
but if it is at
0.0750001 0.075 0.425
the contact is not detected.
I looked deep in the code but don't understand everything what happens in tri_mesh_I resolveEdgeContactBary.
The function everytime ends up in

....
else{
double closestPoint[3];
if(!edgeActive(iTri)[iEdge])
return LARGE_TRIMESH;

It seems to me that for both triangles the edge is not active, and in this case the distance is not calculated by any triangle.

Cheers,
Christian.