yade-users team mailing list archive

[liukeqi <question404284@xxxxxxxxxxxxxxxxxxxxx>]

Question #404284 on Yade changed:
https://answers.launchpad.net/yade/+question/404284

    Status: Answered => Open

liukeqi is still having a problem:
Hi Jan,

  I read your paper "OPEN SOURCE FEM-DEM COUPLING"[1] and my work is
similar like "2.5Multiscale coupling" in the paper.

  We know the advantages of mutiscale coupling are as following.
  Firstly, it is more precise than FEM because its constitutive law straightforwardly derive from DEM.
  Secondly, it is more efficient than DEM because it does not need to compute the large sample entirely consisted of DEM. It just need to use a lot of small DEM samples to compute as RVE.

  I have some confusion about verifying the case that mutiscale coupling is more efficient than pure DEM.
  Normally, should I use a pure and large DEM sample which use common boundary to compute (it may cost many days) and compare the result that obtained from using mutiscale coupling which use FEM to mesh the sample with the same size as pure DEM? Or maybe I can just use a pure and small RVE size DEM sample which use periodic boundary to compute and compare the result obtained from using mutiscale coupling.

  If my understanding of periodic boundary is right, then comparing the
results between RVE size DEM and mutiscale coupling is enough to verify
the efficient? If so, how to select the shape of periodic boundary, cube
or rectangular prism? And how to value the geometric dimensioning of
periodic boundary.

Liu


[1]http://www.engmech.cz/2012/proceedings/pdf/018_Stransky_J-FT.pdf

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