← Back to team overview

yade-users team mailing list archive

Re: Sample uniformity



Maybe you should wait for more experienced advices, but here is "humble"

First, have you used the utils.plotDir on the entire volume of the assembly?
If yes, I think that it is quite normal that you have preferential
directions due to the action of boundary walls. Try to use
utils.plotDirections(utils.aabbExtrema(.2)). It will compute contact
orientation distribution, but inside 0.8 of the whole volume (.2 or anything
else BTW).

Actually, this issue makes me think about the use of "walls" to apply
loading on granular packings. For instance, when I use the triaxial test on
cohesive pre-generated assemblies (extreme case here), I can see that walls
only apply forces on few particles (often less than 10%) of the packing
surface (unfortunately, I don't have it here, but I have an explicit picture
showing this "pb". Will try to include it Monday). So here is a question for

Saying that boundary particles can be identified (actually I tried with
Python and it is quite easy), would it be possible/relevant/easy to replace
each wall element by a list of particles in the Triaxial files without
causing big changes?

Besides the consequences in terms of "homogeneity", this "trick" would also
allow to get rid of the ScGeom need for performing triaxial tests...

Any advice on that?



2010/7/10 chiara modenese <c.modenese@xxxxxxxxx>

> Hi all,
> I would have a question about the uniformity of the packing in the
> triaxial. There is a nice function utils.plotDirections() which returns the
> orientation of the interactions with respect to the three planes of the
> global system. I was using this function at the end of the initial phase of
> the triaxial, both trying the compaction and the expansion method
> alternatively. I am getting preferential directions along the vertical and
> horizontal global axes. I was expecting a better solution adopting the
> expansion method but apparently it does not help too much.
> Did/Do you have the same problem? What would you suggest to avoid it?
> Thanks.
> Chiara
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-users<https://launchpad.net/%7Eyade-users>
> Post to     : yade-users@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~yade-users<https://launchpad.net/%7Eyade-users>
> More help   : https://help.launchpad.net/ListHelp

Follow ups