yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

> To me it seems as if you were saying: Oh yes, at one point I just
> removed Sinit from Ip2_FM_FM_FP, because it is arbitrary anyway. But in
> any case I don't want people to do the same.
>
>   
I didn't "remove" anything since I created this class myself. In any 
case, people can do the same, i.e. create a new class with the help of 
existing code, it will not hurt me. But I don't want somebody to put 
Sinit in Ip2_FM_FM_FP, since it is not the purpose of this class to 
handle such approach. FM_FM_FP is for granular materials (not rocks o r 
concrete) where contacts are points with no size and no length. I'll 
commit some clarifications today.

> My point was stronger. If you have particles of different radius, you
> will probably define contact-level "reference distance" by summing their
> radii. 
You can also consider the deformability of a particle under point load.
Now, with different radii, you still have a pseudo-symmetric problem, 
since the contact is a point. As a consequence, there is no effect of 
the size of particle A on the deformability of particle B. They both 
deform independently, this is the reason for harmonic average, wich 
corresponds to the sum of displacements.

> Whatever you call it, it has dimension of stiffness (even you don't
> imagine stiffness of balls, which I tend to do somehow).
>   
Oh yes, you can say it is the stiffness of the ball, no problem with 
that. It doesn't imply anything in general concerning lengths or surface.

> As soon as you define length of a particular contact (not some average
> length dimension characterizing the packing -- see above), you might
> define the area as well...?
>   
Again, I don't define any contact length.

> For that reason I would not want
> having sum of interaction volumes equal to packing volume, since that is
> in general impossible.
1/ It is possible in general. See e.g. Godard 
(http://www.springerlink.com/content/u25448v61117k1j5/).
2/ I don't know how the equations of mechanics can be written whithout a 
correct partition.

>  Reasoning about 2 spheres is simpler. Length and
> area. 
> I see it is useless to continue. Howgh.
>
>   
I don't know if it is useless. I feel like you have your head in your 
thesis now, with applications to concretes, and one specific theoretical 
background in mind (namely Hentz). I wanted to point out that you were 
perhaps overlooking other fields, and other ways of thinking (and you 
are overconfident in papers like the one you sent before, when I said 
there are no exact theories, this is because I think I know my stuff a 
little...).
Yade is a computer code, it should give results, not impose scientific 
concepts. Here, you are almost saying it is not acceptable to consider a 
contact as a point with no length and no area. It sums up with your 
previous remark on "friction should not be modified", which leads me to 
think you might mix code design and scientific positions sometimes. I 
think your framework is consistent, but a large part of researches 
around granular materials and DEM are using different frameworks, you 
need to keep that in mind.

I hope it helps. :)

Bruno