Applying normal strain rates in PeriTriaxController engine

Hi Mehdi,

My guess is that the simultaneous use of PeriTriaxController and velGrad assignements may be not possible.

Because, unlike TriaxialStressController again, it seems it is not possible to deactivate PeriTriaxController along certain axes: PeriTriaxController is missing something like e.g. TriaxialStressController.wall_back_activated

Thus, there seems to be no means to avoid the conflict you described (?), at the moment.

Maybe it could be possible to perform yourself the servo-loop on O.Cell.velGrad to maintain the constant stresses in the directions you would like ?

Hi,
The diagonal terms of the velocity gradient are controlled by PeriTriaxController.maxStrainRate.
You you change them directly it will be replaced by values from PeriTriaxController.
The off diagonal terms are free, you can change them independently of the controller (e.g. simple shear with imposed normal stress).

@Jérôme
"it is not possible to deactivate PeriTriaxController along certain axes: PeriTriaxController is missing something"

Sorry but no, I don't see anything missing. What would it be to "deactivate" along one axis? Rate=0?
Well, it is straightforward: goal=current.

My complete sentence was "PeriTriaxController is missing something like e.g. TriaxialStressController.wall_back_activated"....

Deactivate along one axis would be to cancel the control by the Engine of stress/strain along this axis.

I think this is possible to do with TriaxialStressController, setting e.g. TriaxialStressController.wall_bottom_activated = TriaxialStressController.wall_top_activated = false

PS : my comment still holds even though I know there are no walls in periodic boundary conditions ;-)

What should happen when it is «not controlled»? Whatever you imagine here
can be obtained with current behavior, i think.
B

Le jeu. 19 juil. 2018 09:33, Jérôme Duriez <
<email address hidden>> a écrit :

> Question #670862 on Yade changed:
> https://answers.launchpad.net/yade/+question/670862
>
> Jérôme Duriez posted a new comment:
> My complete sentence was "PeriTriaxController is missing something like
> e.g. TriaxialStressController.wall_back_activated"....
>
> Deactivate along one axis would be to cancel the control by the Engine
> of stress/strain along this axis.
>
> I think this is possible to do with TriaxialStressController, setting
> e.g. TriaxialStressController.wall_bottom_activated =
> TriaxialStressController.wall_top_activated = false
>
> PS : my comment still holds even though I know there are no walls in
> periodic boundary conditions ;-)
>
> --
> You received this question notification because your team yade-users is
> an answer contact for Yade.
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-users
> Post to : <email address hidden>
> Unsubscribe : https://launchpad.net/~yade-users
> More help : https://help.launchpad.net/ListHelp
>

Thanks for your response Bruno.

I did try changing PeriTriaxController.maxStrainRate. But obviously, this is only the maximum rate, not the current rate. It means that as long as the current rate is smaller than the maximum rate, the normal strain rate would keep changing itself.

So unless I am missing something, in the general case, the normal strain rate in PeriTriaxController cannot be constant. (?)

Mehdi

Is current different from max? Any example script?
B

Le sam. 28 juil. 2018 01:22, Mehdi Pouragha <
<email address hidden>> a écrit :

> Question #670862 on Yade changed:
> https://answers.launchpad.net/yade/+question/670862
>
> Mehdi Pouragha posted a new comment:
> Thanks for your response Bruno.
>
> I did try changing PeriTriaxController.maxStrainRate. But obviously,
> this is only the maximum rate, not the current rate. It means that as
> long as the current rate is smaller than the maximum rate, the normal
> strain rate would keep changing itself.
>
> So unless I am missing something, in the general case, the normal strain
> rate in PeriTriaxController cannot be constant. (?)
>
> Mehdi
>
> --
> You received this question notification because your team yade-users is
> an answer contact for Yade.
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-users
> Post to : <email address hidden>
> Unsubscribe : https://launchpad.net/~yade-users
> More help : https://help.launchpad.net/ListHelp
>
>

Here is a simple example. 2D though. As long as the maxStrain rate is 0.5, it is maintained. But if I increase it to 1000 then the current rate changes.

Of course these rates are meaningless when we have such a small sample. As the sample becomes larger, the maximum rate at which the strain rate follows the maxStrainRate drastically decreases.

Mehdi
----------------------------------------------------
from yade import pack,qt
from yade import plot
import numpy as np

## computing sample size
num_spheres=500# number of spheres
key='_shearTest_'
compFricDegree = 10
finalFricDegree = 30
stabilityThreshold=0.01
young=1e6
max_Strainrate = 0.5

rMin,rMax = 0.5e-3,1.0e-3
vSol = num_spheres * pi * ((rMin + rMax)/2.)**2
volTot = vSol / (1-0.9)
size = volTot**(1./2.)

O.materials.append(FrictMat(young=young,poisson=1,frictionAngle=radians(compFricDegree),density=2600,label='spheres'))

O.periodic=True
#O.cell.setBox(size,size,1)
O.cell.hSize=Matrix3(size,0,0, 0,size,0, 0,0,1)
sp=pack.SpherePack()
radius=(rMin + rMax)/2.
num=sp.makeCloud((0,0,0.5),(size,size,0.5),radius,.2,num_spheres,periodic=True) # min,max,radius,rRelFuzz,spheresInCell,periodic
#O.bodies.append([sphere(s[0],s[1]) for s in sp])
sp.toSimulation(material='spheres')

for k in O.bodies:
 if isinstance(k.shape, Sphere): k.state.blockedDOFs='zXY'

O.engines=[
 ForceResetter(),
 InsertionSortCollider([Bo1_Sphere_Aabb()],verletDist=.05*radius),
 InteractionLoop(
  [Ig2_Sphere_Sphere_ScGeom()],
  [Ip2_FrictMat_FrictMat_FrictPhys()],
  [Law2_ScGeom_FrictPhys_CundallStrack()]
 ),
 PeriTriaxController(dynCell=True,mass=0.2,maxUnbalanced=0.01,relStressTol=0.001,globUpdate=5,maxStrainRate=[max_Strainrate,max_Strainrate,max_Strainrate],label='triax'),
 NewtonIntegrator(damping=.2),
]

O.dt=PWaveTimeStep()
## compaction
conf = -5 #kPa
triax.stressMask=3
triax.goal = [conf,conf,0]
stabilityThreshold = 1e-2

while 1:
  O.run(1000, True)
  #the global unbalanced force on dynamic bodies, thus excluding boundaries, which are not at equilibrium
  unb=unbalancedForce()
  print 'unbalanced force:',unb,' xx stress: ',triax.stressTensor[0,0]
  if unb<stabilityThreshold and abs((conf-triax.stressTensor[0,0])/conf)<0.01 and abs((conf-triax.stressTensor[1,1])/conf)<0.01:
    break

triax.stressMask=4 ## stress controlled along z
deps = -0.1
eps_xx, eps_yy = triax.strain[0], triax.strain[1]
triax.goal=[eps_xx+deps, eps_yy+deps ,0]
O.run(2,True)
print 'current maxStrainRate=', triax.maxStrainRate
print 'current strain rates: xx=', O.cell.velGrad[0,0], ' yy=',O.cell.velGrad[1,1], ' zz=', O.cell.velGrad[2,2]

ps_xx, eps_yy = triax.strain[0], triax.strain[1]
triax.goal=[eps_xx+deps, eps_yy+deps ,0]
print 'setting triax.maxStrainRate=1000'
triax.maxStrainRate = [1000,1000,1000]
O.run(2,True)
print 'current maxStrainRate=', triax.maxStrainRate
print 'current strain rates: xx=', O.cell.velGrad[0,0], ' yy=',O.cell.velGrad[1,1], ' zz=', O.cell.velGrad[2,2]

Would 1000/sec for one iter not simply exceed the target 0.1 strain?
Why this would be related to sample size is unclear to me.
B

Le sam. 28 juil. 2018 02:22, Mehdi Pouragha <
<email address hidden>> a écrit :

> Question #670862 on Yade changed:
> https://answers.launchpad.net/yade/+question/670862
>
> Mehdi Pouragha posted a new comment:
> Here is a simple example. 2D though. As long as the maxStrain rate is
> 0.5, it is maintained. But if I increase it to 1000 then the current
> rate changes.
>
> Of course these rates are meaningless when we have such a small sample.
> As the sample becomes larger, the maximum rate at which the strain rate
> follows the maxStrainRate drastically decreases.
>
>
> Mehdi
> ----------------------------------------------------
> from yade import pack,qt
> from yade import plot
> import numpy as np
>
>
> ## computing sample size
> num_spheres=500# number of spheres
> key='_shearTest_'
> compFricDegree = 10
> finalFricDegree = 30
> stabilityThreshold=0.01
> young=1e6
> max_Strainrate = 0.5
>
> rMin,rMax = 0.5e-3,1.0e-3
> vSol = num_spheres * pi * ((rMin + rMax)/2.)**2
> volTot = vSol / (1-0.9)
> size = volTot**(1./2.)
>
>
>
> O.materials.append(FrictMat(young=young,poisson=1,frictionAngle=radians(compFricDegree),density=2600,label='spheres'))
>
> O.periodic=True
> #O.cell.setBox(size,size,1)
> O.cell.hSize=Matrix3(size,0,0, 0,size,0, 0,0,1)
> sp=pack.SpherePack()
> radius=(rMin + rMax)/2.
> num=sp.makeCloud((0,0,0.5),(size,size,0.5),radius,.2,num_spheres,periodic=True)
> # min,max,radius,rRelFuzz,spheresInCell,periodic
> #O.bodies.append([sphere(s[0],s[1]) for s in sp])
> sp.toSimulation(material='spheres')
>
> for k in O.bodies:
> if isinstance(k.shape, Sphere): k.state.blockedDOFs='zXY'
>
>
> O.engines=[
> ForceResetter(),
> InsertionSortCollider([Bo1_Sphere_Aabb()],verletDist=.05*radius),
> InteractionLoop(
> [Ig2_Sphere_Sphere_ScGeom()],
> [Ip2_FrictMat_FrictMat_FrictPhys()],
> [Law2_ScGeom_FrictPhys_CundallStrack()]
> ),
>
> PeriTriaxController(dynCell=True,mass=0.2,maxUnbalanced=0.01,relStressTol=0.001,globUpdate=5,maxStrainRate=[max_Strainrate,max_Strainrate,max_Strainrate],label='triax'),
> NewtonIntegrator(damping=.2),
> ]
>
> O.dt=PWaveTimeStep()
> ## compaction
> conf = -5 #kPa
> triax.stressMask=3
> triax.goal = [conf,conf,0]
> stabilityThreshold = 1e-2
>
> while 1:
> O.run(1000, True)
> #the global unbalanced force on dynamic bodies, thus excluding
> boundaries, which are not at equilibrium
> unb=unbalancedForce()
> print 'unbalanced force:',unb,' xx stress: ',triax.stressTensor[0,0]
> if unb<stabilityThreshold and
> abs((conf-triax.stressTensor[0,0])/conf)<0.01 and
> abs((conf-triax.stressTensor[1,1])/conf)<0.01:
> break
>
>
> triax.stressMask=4 ## stress controlled along z
> deps = -0.1
> eps_xx, eps_yy = triax.strain[0], triax.strain[1]
> triax.goal=[eps_xx+deps, eps_yy+deps ,0]
> O.run(2,True)
> print 'current maxStrainRate=', triax.maxStrainRate
> print 'current strain rates: xx=', O.cell.velGrad[0,0], '
> yy=',O.cell.velGrad[1,1], ' zz=', O.cell.velGrad[2,2]
>
> ps_xx, eps_yy = triax.strain[0], triax.strain[1]
> triax.goal=[eps_xx+deps, eps_yy+deps ,0]
> print 'setting triax.maxStrainRate=1000'
> triax.maxStrainRate = [1000,1000,1000]
> O.run(2,True)
> print 'current maxStrainRate=', triax.maxStrainRate
> print 'current strain rates: xx=', O.cell.velGrad[0,0], '
> yy=',O.cell.velGrad[1,1], ' zz=', O.cell.velGrad[2,2]
>
> --
> You received this question notification because your team yade-users is
> an answer contact for Yade.
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-users
> Post to : <email address hidden>
> Unsubscribe : https://launchpad.net/~yade-users
> More help : https://help.launchpad.net/ListHelp
>
>

Bruno:

The sample size affects the rate chosen by the engine.

But, anyways, I think I have found a way to solve the problem;

As far as I understand, in PeriTriaxController engine, when the goal is strain, the strain rate will initially be equal to the maxStrainRate. The rate starts to decrease only when the target strain is being reached (?). This final ramping stage is probably not a big deal in most cases, but for cases involving small strain probing, it quickly shows up.

The solution I came up with, is that I make sure the strain goal is always far from the current stage (but on the right side, i.e. larger or smaller than current state). This way, the final ramping stage never takes place. I then stop the straining manually.

Thanks for all the help.

erratum: the sequence will be {0.03, 0.03, 0.03, 0.01, 0}

Thanks a lot Bruno for the ample explanation. It clarifies everything.

Mehdi

Thanks Bruno Chareyre, that solved my question.