yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

> The volume must obviously be greater or equal to the particle-volume.
The volume is slightly less, according to elasticity, or strictly equal
if Poisson's ratio is 0.5.
I thank you for this discussion. I will include it somehow in the
documentation to make more clear what this function gives and gives not.
It is not giving any sort of equivalent mesoscale stress, averaged in a
volume comprising voids. It is the per-particle solid phase stress in a
packing of elastic and initialy spherical beads ("not-an-equivalent" ->
the true/real/exact one, what is the best naming?) . Emanuele needs this.

It would be interesting to define more virtual stress averages on a
variety of volumes, I agree on that point, but again it needs to change
the position x in the integral accordingly, which can't be simply done
by the same function dividing by a different volume. For Voronoi for
instance, it would need to find intersections between normal axis and
radical planes. It would be interesting to implement this (eh, I need to
fix TesselationWrapper first).

For now, if somebody really needs the same thing not divided by sphere
volume, I'm afraid he will have to _multiply_ the returned value by
sphere volume on his side. Since it is very unlikely to happen, and
would be inconsistent from my point of view, keeping the current
equation is on the user friendly side, and Emanuele will not have to
divide by spheres volumes in his scripts. :)
 
Bruno

> A function that provide the moment tensor (\sum_k x_i^k f_j^k without division by a volume) for one particle 
> can be used with any shape (avoid the volume computation by means of radii). 
> It can also be summed and then divided by the volume you want...
>
>
> _____________