yade-users team mailing list archive

[Janek Kozicki <janek_listy@xxxxx>]

Václav Šmilauer said:     (by the date of Wed, 27 Oct 2010 17:48:03 +0200)

> > For preparing the sample, it's a bit complicated now because we want
> > to use sample with designated grain size distribution and porosity.
> > Can Yade create a sample which has specific grain size distribution
> > and porosity?  What i did was, i wrote a program to put particles into
> > a domain, the same size of the real sample, allowing some overlap
> > between particles.  Then in the simulation i set a higher viscosity
> > (to prevent from 'explosion' of the particles) and zero friction so
> > that the particles will move to the position.  I monitor the kinetic
> > energy till it's very low.  If this works, then very soon I can
> > calibrate the coefficient.
> 
> Hi Jesse, welcome on board! Other people might be better able to reply
> to this particular problem, but I will try at least.
> 
> In Yade, you would generate non-overlapping sphere cloud satisfying the
> desired particle size distribution (PSD); the function for that is
> SpherePack.makeCloud (for piecewise-constant probability density
> function, giving piecewise-linear distribution) or SpherePack.particleSD
> (for probability density as union of Dirac functions, giving staircase
> distribution).
> 
> The next step is to compress the packing up to the point where the
> desired porosity value is attained (this step must be currently scriped
> using a few lines of python code). You can play with friction of
> particles in the step; with zero friction, you will obtain dense
> packing, whereas higher friction will create loose packing (friction
> prevents spheres from re-arranging).
> 
> The resulting packing will have non-zero overlaps between spheres (some
> would say it is not "stress-free"). One cannot prescribe PSD, porosity
> and desired overlap at the same time, since they are not orthogonal
> (they are all geometrical characteristics).
> 
> I think Chiara has some script she could perhaps send you over, to have
> a look at the procesure described. If there is general interest in such
> a functionality, it can be factored out into a function in yade.pack as
> well.


Heh, now I need it, can you share? :) My target is void ratio=0.53,
so that would be porosity=0.347


-- 
Janek Kozicki                               http://janek.kozicki.pl/  |