TriaxialStressController

Hello everyone:

Here are my codes..I try different max velocity of wall during consolidation..but the initial normal stress on three directions is still very large..Hope you can help check this..

from yade import plot
from yade import qt
from yade.pack import *

Centroid_X=[-0.092447, -0.069173, -0.27585, -0.043333, -0.311997, -0.030057, 0.309978, 0.368362, 0.132105, 0.403417, 0.398405, 0.394845, 0.115757, 0.177633, -0.354947, -0.366737, -0.113036, -0.350854, -0.188196, 0.224647, -0.195432, 0.228807, 0.390975, -0.18349, 0.163522, -0.321283, -0.098903, 0.129782, -0.010825, -0.367059, -0.382682, -0.353875, -0.405821, -0.402318, 0.3965, 0.256225, 0.403996, 0.390108, 0.097005, -0.062893, 0.392252]
Centroid_Y=[0.5999, -0.096868, 0.166983, -0.304628, 0.32389, 0.37278, -0.132354, 0.37821, -0.311325, -0.415205, -0.057077, 0.5006, 0.31917, 0.089013, -0.178724, 0.04522, 0.526631, 0.42096, 0.58706, 0.54971, 0.48997, -0.414162, 0.12819, -0.247131, 0.01415, -0.342179, -0.006503, 0.031844, 0.34549, 0.02605, -0.393463, -0.391638, -0.409767, 0.50673, -0.408674, -0.212701, -0.408868, -0.385581, -0.416159, 0.15939, 0.154113]
Centroid_Z=[-0.337109, 0.035291, 0.274027, -0.300666, -0.28512, -0.36685, -0.007652, -0.163221, 0.289719, 0.202803, -0.281833, 0.039395, 0.093978, -0.137465, -0.216228, -0.376156, -0.014532, 0.009707, 0.167293, -0.349804, 0.376997, -0.229445, 0.377528, 0.364034, -0.36308, 0.072812, -0.390037, 0.344641, 0.398662, -0.040234, 0.352715, -0.384614, -0.174679, 0.272686, 0.39796, -0.302071, -0.046035, -0.360983, 0.003109, -0.180008, -0.089793]
Radius=[0.1177, 0.2097, 0.2484, 0.2195, 0.1807, 0.1495, 0.2126, 0.1376, 0.2126, 0.1016, 0.1137, 0.1242, 0.2426, 0.1048, 0.1619, 0.1458, 0.1172, 0.1657, 0.1088, 0.1915, 0.142, 0.09959, 0.1277, 0.1498, 0.1561, 0.1845, 0.1329, 0.1802, 0.1191, 0.1391, 0.1281, 0.1293, 0.1069, 0.1135, 0.1094, 0.1224, 0.1092, 0.1293, 0.1001, 0.1647, 0.125]

mn,mx=Vector3(-0.50,-0.50,-0.50),Vector3(0.50,0.725,0.50)

O.materials.append(FrictMat(young=5e11,poisson=0.,frictionAngle=0.,density=2650,label='spheres'))#Wei- Density=2650*1e4
O.materials.append(FrictMat(young=5e11,poisson=0.,frictionAngle=0.,density=0.,label='walls'))

walls=aabbWalls([mn,mx],thickness=0,material='walls')
wallIds=O.bodies.append(walls)

O.bodies.append([
 sphere(Vector3(Centroid_X[i],Centroid_Y[i],Centroid_Z[i]),Radius[i],material='spheres')
    for i in range(41)
])

print(len(O.bodies))
stabilityThreshold=2e-3

young=5e11
triax=TriaxialStressController(
 maxMultiplier=1.0+2e5/young,
 finalMaxMultiplier=1.0+2e4/young,
 thickness=0,
 stressMask=7,
 internalCompaction=False,
 max_vel=0.01,
        label='triax'
)

O.engines=[
 ForceResetter(),
 InsertionSortCollider([Bo1_Sphere_Aabb(),Bo1_Box_Aabb()]),
 InteractionLoop(
  [Ig2_Sphere_Sphere_ScGeom(),Ig2_Box_Sphere_ScGeom()],
  [Ip2_FrictMat_FrictMat_MindlinPhys()],
  [Law2_ScGeom_MindlinPhys_Mindlin(includeMoment=False)]
 ),
 triax,
        GlobalStiffnessTimeStepper(active=1,timeStepUpdateInterval=10,timestepSafetyCoefficient=0.8),
 TriaxialStateRecorder(iterPeriod=100,file='WallStresses'),
 NewtonIntegrator(damping=.2),
 PyRunner(iterPeriod=100,command='chenkshear()',label='recorder')

]
qt.View()
triax.goal1=triax.goal2=triax.goal3=-500000
O.dt=.5*PWaveTimeStep()

def chenkshear():
 unb=unbalancedForce()
 print ('unbalanced force: ',unb,'mean stress: ',triax.meanStress)
 if unb<stabilityThreshold and abs(-500000-triax.meanStress)/500000<0.001:
  O.save('Phase1.xml.bz2')
  O.pause()

Hello,

Please do not mix 2 questions into 1. Some remarks considering your second one:

DEM is a simulation method being dynamic by nature (Cundall & Strack, 1979). As such, having a stable structure at every moment (i.e. a quasistatic evolution) requires extra care from the user.

If you're compressing an initial cloud with very few particles, it's probably even more difficult to ensure permanent quasi-staticity and "smooth" macro-curves:
- going from a cloud to a solid-like packing is a quite dramatic process between two very different states
- with very few particles, macro-curves can certainly be much more affected by discrete events on 1 or 2 (or ..) particles