yade-users team mailing list archive

[Bruno Chareyre <question701782@xxxxxxxxxxxxxxxxxxxxx>]

Question #701782 on Yade changed:
https://answers.launchpad.net/yade/+question/701782

Bruno Chareyre proposed the following answer:
At the moment it is a bit convolved to read your changes but it seems I was able to break your algorithm.
A suspicious feature of yours is that it tries to align the period on the reference frame, which isn't helping in general.
Try a rigid body rotation (velGrad = [[0,1,0], [-1,0,0], [0,0,0]] ) and you will see the problem:
- older version leads to zero volume (wrong)
- new version does flip without "zero volume" error, but it shouldn't: it turns the cube into a more elongated thing, the opposite of what we want; it also breaks my contact model (most likely a problem in updating cellDist). 
My impression is that there must be is a much simpler fix to the issue.
I'll need to cleam my test script before showing (right now it's a mess).
Let's continue the discussion on gitlab?

B

-- 
You received this question notification because your team yade-users is
an answer contact for Yade.