yade-dev team mailing list archive

[Vincent Richefeu <richefeu@xxxxxxxxx>]

>> You're right! Some people consider granular matter as a continuous media. I don't.
>> 
> The integral is for one grain alone. Considering one grain as a
> continuum doesn't sound so absurd, does it?
> We are far above the atomic scale.

As I said, I see no problem for the computation. I just made a comment about the term "exact mean stress" which can be misleading.

>> The particles are rigid, i.e. 6 independent DOFs and that's all (Elasticity is introduced in force-laws).
>> We can always define whatever we want. The question is what is the sense of our definition.
>> As I said, the definition of a particle stress is ok, but the volume used in this definition is a matter of convenience.
> I don't really trust the "rigid particle" story. A contact point being a
> point, it can't be deformed, then it must be the particle that is
> deformable. Real silicate grains are deformable too.

Personally, I have never heard about non-strainable material, even diamond!  

>> Yade users must form their own opinions
> Thats true.
>> and given them a function that compute a particle stress is in my opinion a bit dangerous.
> It is not dangerous as long as people consider a number for what it is.
> The documentation says "exact mean stress from contour integral of
> external load". The demonstration starts with "at equilibrium", and the
> whole thing is classical elasticity. Fair enough.
>> Why don't you provide a function to compute the moment tensor and then the user is free to use the volume he want in order to compute the stress?
> Simply because I never heard about moment tensor. Not saying it would
> not be interesting to add a function for that.
>> The real aim of this proposition is to empower the user, who sometime use a function as a black box.
> Mmh... Who is to blame in that case?

Nobody.

> And, come on, those ones would
> remember they have to divide the return value by a volume?!

People are not so careless!! In any case they should not be.

Vincent