yade-users team mailing list archive

[Robert Caulk <question670765@xxxxxxxxxxxxxxxxxxxxx>]

Question #670765 on Yade changed:
https://answers.launchpad.net/yade/+question/670765

    Status: Open => Answered

Robert Caulk proposed the following answer:
>>  I wonder whether there is a method which can adjust dt
automatically.

Probably! If you determine such a method I encourage you to share with
the community.

>> Can I just ignore the warning? Will it have any negative effect on
the simulation?

Here is the origin of the warning [1], looks like the author is using a
triple product to evaluate each cell's volume [2] so the pseudoscalar
output needs to be signed [3]. In other words, the volume calc may yield
a negative value despite being perfectly safe. Hence the "should still
be safe". However, in your case you are likely seeing this as a result
of the large particle size mismatch in your simulation and the fast-
moving particles. I think this combination results in very acute
tetrahedral angles and if these particles move faster than the
triangulation updates, they end up easily inverting themselves,
resulting in a negative triple product. You are still getting the
correct volume of a pore represented by the locations of those
particles, but the forces are probably pretty inaccurate since the
triangulation is outdated. If you are scared of this you could try
updating the triangulation more frequently with meshUpdateInterval.

[1]https://github.com/yade/trunk/blob/master/pkg/pfv/FlowEngine.ipp.in#L629
[2]https://github.com/yade/trunk/blob/master/pkg/pfv/FlowEngine.ipp.in#L628
[3]https://github.com/yade/trunk/blob/master/pkg/pfv/FlowEngine.ipp.in#L630

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