yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

> Now, whether to use the (anti)symmetric split vs difference of angular 
> velocities... What is the reason to use the second one, as you do in ScGeom6D 
> for instance, is there something physical reasoning behind, or is it just an 
> arbitrary choice?

How would you define the rotation of one body relative to a fixed coordinate system?
Then what if the coordinate system is not fixed and you want the rotation of the object in
this moving system? You would probably define it as BodyRotation*SystemRotation^(-1), no?
That is what Janek explained already (btw, Janek, it is the code you wrote that I moved
into ScGeom6D).
For me, it is a bit as if you were asking "why relative displacement of two points is
u2-u1? Is it arbitrary?".
Would anything else be objective (see below)? Why do you think it is not consistent?

> How about other DEM codes?

I know you have PFC "theory and background" somewhere (section 2.3.2) ;-).

Going back to the last formula you sent (nice thdb plugin!), it seems to me that :

1 - a rigid body rotation perpendicular to x1x2 will generate bending rate (think discs in
2D) proportional to (r1-r2)/(r1+r2). If they have the same radii, your a safe, else...

2 - a rigid body rotation along x1x2 will give a bending rate proportional to
2(u1+u2)/(x1x2), which is not related (seems to me) to the RB rotation applied. A correct
evolution would be more like bending rate = RBspin.cross(bending) (or, if updating the
orientation is done independently, then this rate should be null).

Bruno