Yade

Too high unbalanced forces in Yade (even just aplying only gravity)

Asked by Criss Zanelli

Hello everybody.
I modeled a wall composed by polyhedra blocks. I does not matter what force or aceleration I aply, only with the gravity It has unbalanced forces of infinite values at the beginning, then a short period of 1e10 values an then is almost constant with the value o 1e6.
My aim then is to make a pseudo static simulation under out of plane forces, but I will never reach the static state if unbalanced forces are in that magnitudes. Please recommend me some alternatives to solve it. Thanks very much.

Question information

Language:
English Edit question
Status:
Solved
For:
Yade Edit question
Assignee:
No assignee Edit question
Solved by:
Bruno Chareyre
Solved:
Last query:
Last reply:
Revision history for this message
Bruno Chareyre (bruno-chareyre) said : 		#1

Hello,
It seems something wrong happens and the unbalanced force is only revealing it.
I would suggest to inspect the evolution in the 3D view and by other means.
It is not possible to help more based on your description, you may like to visit [1].
Regards
Bruno

https://yade-dem.org/wiki/Howtoask

Revision history for this message
Criss Zanelli (criss.zanelli) said : 		#2

Thanks for answering, Bruno.

 My script by now is really simple. I only create the geometry (the wall and the floor), then in the engine only acts the gravity. Please if you can run my script in order to check the the magnitud of the unbal force (ploted) I would really appreciate it.

#Script

from yade import qt
from numpy import arange
import numpy as np
import random
from yade import polyhedra_utils
from yade import plot

aceleration=0 #m/s2
damping=0.80
#O.dt=1e-9

m = PolyhedraMat()
m.density = 2500 #kg/m3
m.kn = 1e7 #Pa
m.ks = 1e7 #Pa
m.young = 6.02e11 #Pa
m.frictionAngle = 0.610 #rad35
O.materials.append(m)

mfloor = PolyhedraMat()
mfloor.density = 1e10 #kg/m3
mfloor.kn = 1e15 #Pa
mfloor.ks = 1e15 #Pa
mfloor.young = 6.02e11 #Pa
mfloor.frictionAngle = 0.873 #rad50
O.materials.append(mfloor)

# dimenssions of the pirca ****************************************************
Length=6 #m
Heigth=1.0 #m
Width=0.35

# dimensions of the rock blocks ****************************************************
ancho = Width #m
largo = 0.3 #m
alto = .25 #m

N=int(Heigth/alto)
M=int(Length/largo)
print 'Numero de filas', N
print 'Altura de pirca', Heigth,'m'
print 'Aceleracion total', aceleration,'m/s2'

# composition of the wall************************************************************
vertices1 = [[0,0,0],[ancho,0,0],[ancho,largo,0],[ancho,largo,alto],[0,largo,0],[0,largo,alto],[0,0,alto],[ancho,0,alto]]
for i in range(0,M):
 for j in range(0,N):
  b = polyhedra_utils.polyhedra(m,v=vertices1)
                b.state.pos = (0,(i+1)*largo,(j+1)*alto)
                O.bodies.append(b)

# restrinction of the degree of freedom lateral ****************************************

vertices2 = [[0,0,0],[ancho,0,0],[ancho,0.1,0],[ancho,0.1,Heigth],[0,0.1,0],[0,0.1,Heigth],[0,0,Heigth],[ancho,0,Heigth]]

rigth = polyhedra_utils.polyhedra(m,v=vertices2, color=(.6,.4,.3))
rigth.state.pos = (0,0.1,Heigth/2+alto/2)
O.bodies.append(rigth)

left = polyhedra_utils.polyhedra(m,v=vertices2,color=(.6,.4,.3))
left.state.pos = (0,Length+.2,Heigth/2+alto/2)
O.bodies.append(left)

# floor ***********************************************************

vertices3 = [[-1,-1,0],[-1,7,0],[4,-1,0],[4,7,0],[-1,-1,-0.1],[-1,7,-0.1],[4,-1,-0.1],[4,7,-0.1]]
floor = polyhedra_utils.polyhedra(mfloor,v=vertices3, color=(.6,.4,.3))
floor.state.pos = (1.5,3,0.075)
O.bodies.append(floor)

O.bodies[80].state.blockedDOFs='xyz'
O.bodies[81].state.blockedDOFs='xyz'
O.bodies[82].state.blockedDOFs='xyz'

#FORCE ***********************************************
#def addF():
# for a in range (0,M):
# O.forces.addF(N*a, (523.43,0,0), permanent=True)
# O.forces.addF(N*a+1, (373.88,0,0), permanent=True)
# O.forces.addF(N*a+2, (224.33,0,0), permanent=True)
# O.forces.addF(N*a+3, (74.78,0,0), permanent=True)

# ENGINE ********************************************************************************

def checkUnbalanced():
   if unbalancedForce()<.01: # to ensure stability
     O.pause()
     plot.saveDataTxt('bbb.txt.bz2')

if O.iter>0:
   print O.bodies[36].state.pos[0], O.bodies[37].state.pos[0], O.bodies[38].state.pos[0],O.bodies[39].state.pos[0]

def addPlotData():
#plot the displacement of each block of the column located in the midle of the wall
    plot.addData(iteracion_distrib1=O.iter,Unbalanced=unbalancedForce())

plot.plots={'iteracion_distrib1':('Unbalanced')}

O.engines=[
   ForceResetter(),
   InsertionSortCollider([Bo1_Polyhedra_Aabb()],label='collider'),
   InteractionLoop(
      [Ig2_Polyhedra_Polyhedra_PolyhedraGeom(),Ig2_Facet_Polyhedra_PolyhedraGeom()],
      [Ip2_PolyhedraMat_PolyhedraMat_PolyhedraPhys()], # collision "physics"
      [Law2_PolyhedraGeom_PolyhedraPhys_Volumetric()] # contact law -- apply forces
   ),
   NewtonIntegrator(damping=damping,gravity=(aceleration,0,-9.81)), # constant aceleration in x
   PyRunner(command='checkUnbalanced()',iterPeriod=50,label='checker'),
   PyRunner(command='addPlotData()',iterPeriod=50),
# PyRunner(command='addF()',iterPeriod=100),
   ]

collider.verletDist=.1
plot.live=True
plot.plot(subPlots=True)

#print ('D_36 %d D_37 %d D_38 %d D_39 d%'%(O.bodies[36].state.pos[0], O.bodies[37].state.pos[0], #O.bodies[38].state.pos[0],O.bodies[39].state.pos[0]))

qt.Controller()
qt.View()
R=yade.qt.Renderer()
R.bgColor=(1.,1.,1.)

# end *******************************************

Revision history for this message
Best Bruno Chareyre (bruno-chareyre) said : 		#3

The wall is stable.
The apparent problem is because bodies 80-82 have only translations blocked (not rotations as in xyzXYZ), hence they are taken into account in the unbalanced force (and the total force on the bottom plate is not supposed to be zero of course). With full fixities unbalanced force goes to 0.
It also appears that you are using an extremely small timestep, which makes the stabilization very slow.
Bruno

Revision history for this message
Criss Zanelli (criss.zanelli) said : 		#4

Thanks Bruno Chareyre, that solved my question.

Revision history for this message
Robert Caulk (rcaulk) said : 		#5

Hey Yade user, I am an automated bot :-). After reading your question, I decided to find some related threads and summarize them for you here. Please do not hesitate to follow up on my summarizations by following the links provided below.

Title: "I don not understand how to simulate triaxial-test in Yade."
Thread summary: PeriTriaxController is used for prescribing deformation of space based on user-defined targets, such as stress or strain. The code in the Periodic Triaxial Test tutorial has been understood almost entirely, except for the purpose of specifying triax.maxUnbalanced to 10 after calling the function triaxFinished(). There is a guess that there might be an increase in unbalanced forces when switching from static to deforming regime, but further clarification is needed. Thanks Bruno Chareyre, that solved my question.
https://answers.launchpad.net/yade/+question/683427

Title: "[Triaxial test] Kinetic Energy is way too big, how to reduce it?"
Thread summary: The user experiences high kinetic energy in a triaxial test simulation with Yade for a quasi-static situation. Chiara suggests comparing total kinetic energy with elastic potential energy and ensuring it's a small fraction of the latter. Despite this, the user remains confused about why the value of Kinetic Energy at a specific iteration where unbalancedForce is <0.001 is 2000.
https://answers.launchpad.net/yade/+question/228886

Title: "Too high unbalanced forces in Yade (even just aplying only gravity)"
Thread summary: The user encounters unbalanced forces in a Yade simulation of a polyhedral wall, causing infinite values at first and constant values around 1e6. The issue is resolved by setting full fixities for bodies 80-82 and reducing the timestep.
https://answers.launchpad.net/yade/+question/663620