yade-users team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

I think I will have to remove the call to cell->integrateAndUpdate() 
from scene code and put it in PeriCompressor.
I don't see how to get relevant values otherwise (you, Vaclav, 
mentionned some time ago that we were using transformation of the step 
before, which is true and problematic).
Cell must be updated between contact law and Newtons law.

> the name "velocity gradient" which I still don't get really
>   
Oh, really? What is not clear?
> (why not just transformation rate? 
In continuum mechanics it is common to introduce the "gradient of 
displacement tensor" (http://en.wikipedia.org/wiki/Finite_strain_theory).
You could say it is better to call it "gradient of transformation". It 
is equivalent, I don't see why one would be better.
If you take the symmetric part of gradDisp, you have the infinitesimal 
strain tensor.
If you take the symmetric part of gradVel, you have strain rate tensor.
> Of what _velocity_ is velGrad a
> gradient? 
Define a homogeneous (*) vector field describing the velocity at each 
point in a continuum. Now consider the gradient of it. How can you name 
it better than the "gradient of the velocity field"?
> rate==velocity? etc etc)
>   
No, rate is not velocity. You can move the solid at high speed without 
loading it. The loading rate (kinematic) is really described using the 
velocity gradient.
Honnestly, I feel like explaining obvious things. I don't understand 
what is not clear for you.

Bruno

(*) In "french" continuum mechanics, a "homogeneous" disp/velocity field 
is not uniform, it is something like Vx=a.x in 1D, leading to uniform 
strain. Not sure it translates correctly.