← Back to team overview

yade-dev team mailing list archive

Re: time step with viscous damping

 

I think I do not have my mind in the viscous framework.

But, in a general case, some of your phrases do not correspond to my current point of view:

If more stiffness then more force for given overlap.
In my former simulations, I used this "NormalInelasticityLaw" which describes (Normal Force; overlap) states such as illustrated by attached figure. So, in this case, you can observe that the highest (tangent) stifness corresponds to lowest values of force : in unloading the stifness I used is higher than the one of loading; but unloading values of force are nevertheless lower than loading (initial loading) values of force, for given overlap.

Then, when you say:
I consider only a loading because a unloading is less the loading, so if loading is stable (not overestimate) then unloading is stable also.

Using this contact law I speak about, I had to decrease the time step. Let "K" be the ratio "unloading stifness" / "loading stifness" (= the ratio of the two slopes of the two type of lines you can see on the figure). The time step I could use with K=1 (loading = unloading), could not be used anymore if K>1 (the case of the figure).

I hope I could explain what I think ?
It's completely ok for me when you say "So, for stability you need smaller time step for more stiffness". But I tend currently to disagree if you speak only of forces.

Jerome

Attachment: FnCL1.eps
Description: image/eps


Follow ups

References