yade-users team mailing list archive

[Anton Gladky <question248869@xxxxxxxxxxxxxxxxxxxxx>]

New question #248869 on Yade:
https://answers.launchpad.net/yade/+question/248869

in a very speicific conditions one of my clump is getting a very high angVel and goes to a numerical instability.
It seems, the problem is in this part of code [1] (leapfrogAsphericalRotate in NewtonIntegrator).

I could minimize the problem and there are values of all variables during 2 iterations:

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Iter:154675
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
A:  0.934319 -0.193747  0.299181
 0.188102  0.980983 0.0478483
-0.302762 0.0115711  0.952996
M:  0.934319 -0.193747  0.299181
 0.188102  0.980983 0.0478483
-0.302762 0.0115711  0.952996
state->angMom: 0 0 0
dt: 9.55301e-06
l_n: 0 0 0
l_b_n: 0 0 0
state->inertia: 6.16301e-08 1.15341e-07 1.25862e-07
angVel_b_n: 0 0 0
dotQ_n: 0 0 0 0
state->angMom: 0 0 0
l_b_half: 0 0 0
angVel_b_half: 0 0 0
dotQ_half: 0 0 0 0
state->ori: 0.9834 0.00922239 -0.153026 -0.0970739
state->angVel: 0 0 0


!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Iter:154676
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

A:  0.934319 -0.193747  0.299181
 0.188102  0.980983 0.0478483
-0.302762 0.0115711  0.952996
M:  0.934319 -0.193747  0.299181
 0.188102  0.980983 0.0478483
-0.302762 0.0115711  0.952996
state->angMom: 7.9467e-07 4.00506e-06 0.00125804
dt: 9.55301e-06
l_n: 3.97335e-07 2.00253e-06 0.000629021
l_b_n: 0.000188175 3.21368e-05 0.000599357
state->inertia: 6.16301e-08 1.15341e-07 1.25862e-07
angVel_b_n: 3053.29 278.624 4762.03
dotQ_n: 238.374 1150.47 -33.1568 2576.39
state->angMom: 7.9467e-07 4.00506e-06 0.00125804
l_b_half: 0.000376349 6.42735e-05 0.00119871
angVel_b_half: 6106.58 557.247 9524.07
dotQ_half: 401.41 2300.23 -54.5905 5160.22
state->ori: 0.987234 0.0311965 -0.153548 -0.0477782
state->angVel: 2922.13 -527.866 10932.1

Can anybody understand, whether the problem in a very high state->angMom, 
which is obtained in a Iter=154676 or something else? Reducing the timestep 
does not solve the problem

I remember, some years ago we had the same problem in clump-hopper-viscoelastic.py
But it is not reproducible any more.

Thanks ahead

[1] https://github.com/yade/trunk/blob/master/pkg/dem/NewtonIntegrator.cpp#L253

-- 
You received this question notification because you are a member of
yade-users, which is an answer contact for Yade.