yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

>> Going back to the dynamic case (energy.py) :
>> It can be verified that there is an unbalanced energy evolution
>> independant of friction and damping.
>> To make plastic dissipation 0, it is enough to set frictionAngle close
>> to pi/2.
>> For damping, make it 0.
>> The results are attached. ScGeom and L3 give different paterns.
> I fixed a bug in Law2_L3Geom_FrictPhys_LinPerfPl today. Can you post
> the parameters you were running those figures with?

Hi Vaclav,

The parameters are above ("close to" means *0.99999 or so, I don't
remember the exact value but it is not important).

I just ran few additional tests and found an error in kinetic energy
definition; it's fixed (r2629). With correct kinetic term and
friction=0.3pi, energy is conserved relatively well in ScGeom (but not
in L3, see curves "f03pi").

Using friction pi/2 still gives residual unbalanced energy (files
without the "f03"), but the error decreases with smaller timesteps. I
tend to see this as an elastic energy loss each time a contact is lost.

Best wishes.

Bruno

p.s. No reaction against last suggested changes in Newton, I conclude
silent agreement and will commit with undamped clumps members. Not
before a few days.

-- 
_______________
Bruno Chareyre
Associate Professor
ENSE³ - Grenoble INP
Lab. 3SR
BP 53 - 38041, Grenoble cedex 9 - France
Tél : +33 4 56 52 86 21
Fax : +33 4 76 82 70 43
________________

Attachment: energySc-f03pi.png
Description: PNG image

Attachment: energyL3-f03pi.png
Description: PNG image

Attachment: energySc.png
Description: PNG image

Attachment: energyL3.png
Description: PNG image