yade-users team mailing list archive

[ehsan benabbas <question688681@xxxxxxxxxxxxxxxxxxxxx>]

Question #688681 on Yade changed:
https://answers.launchpad.net/yade/+question/688681

    Status: Answered => Solved

ehsan benabbas confirmed that the question is solved:
Yes, you are right Jan. 
Just to mention, based on the formulation for Kn, we have Kn=E*r (if we assume E and r are constant for all particles)
While, based on the timestep calculation provided in the DEM formulation of the YADE website, in the section "Estimation of Δtcr
by wave propagation speed", this formula is Kn=E*D because π′ is equal to 2 in the classical DEM model (Ip2_FrictMat_FrictMat_FrictPhys) as implemented in Yade. 

My conclusion is that as the particle mass, young, and size changes the
timestep, it's not avoidable to have such a long runtime and the best
solution is multiprocessing using as many cores as we can provide.
Therefore, we can use E=Kn/D and need to think of a high-end PC or a
supercomputer.

Thank you for your comments,

Bests,
Ehsan

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