yade-users team mailing list archive

[Nasibeh Moradi <nasibeh.moradi@xxxxxxxxx>]

I did uniaxial test  with contact law in YADE,  the graph is attached.  the
graph is very bad, I forget anything?
Also, the crossSection from utils.uniaxialTestFeatures() is 2*Area(Area is
calculated by hand). it is possible?!

*The code:*

import itertools, time
from yade import utils, pack,timing,plot

planeSize=Vector3(0.08,0.08,0.
006)
rSphere=0.001
strainRate=0.001

#Add material
O.materials.append(FrictMat(young=74.1e9,frictionAngle=0.179,poisson=0.2,density=2500,
                       label='PlaneMaterial'))

#generate a FCC Pack
a=2*rSphere;
Spheres=[]
h=a*sqrt(2)/2
ii,jj,kk=[range(0,10),range(0,20),range(0,30)]

for i,j,k in itertools.product(ii,jj,kk):
    x,y,z=0+rSphere+i*sqrt(2)*a,0+rSphere+j*h,0+rSphere+k*h
    if j%2==0: x+= sqrt(2)*a/2. if k%2==0 else -sqrt(2)*a/2.
    if k%2!=0: x+=sqrt(2)*a/2.

Spheres+=[utils.sphere((x,y,z),radius=rSphere,color=(1,0,1),material='PlaneMaterial')]

PlaneIds=O.bodies.append(Spheres)

bb=utils.uniaxialTestFeatures()
negIds,posIds,axis,crossSectionArea=bb['negIds'],bb['posIds'],bb['axis'],bb['area']
O.dt=0.8*utils.PWaveTimeStep()

#O.initializers=[
#       BoundDispatcher([Bo1_Sphere_Aabb()])
#]
O.engines=[
        ForceResetter(),

        BoundDispatcher([
                Bo1_Sphere_Aabb(aabbEnlargeFactor=1.07)
        ]),

        InsertionSortCollider(),

        InteractionDispatchers(
                [Ig2_Sphere_Sphere_Dem3DofGeom()],
                [Ip2_FrictMat_FrictMat_FrictPhys()],
                [Law2_Dem3DofGeom_FrictPhys_Basic()]
                 ),

        #GravityEngine(gravity=[0,0,-9.81]),
        NewtonIntegrator(),
        UniaxialStrainer(strainRate=-strainRate,axis=axis,asymmetry=0,

posIds=posIds,negIds=negIds,crossSectionArea=crossSectionArea,
                    blockDisplacements=False,blockRotations=False,
                    setSpeeds=True,label='strainer'),
        PeriodicPythonRunner(virtPeriod=3e-7/strainRate,realPeriod=1,

command='addPlotData()',label='plotDataCollector',initRun=True)
]

plot.plots={'eps':('sigma',)}

O.saveTmp('initial');
def addPlotData():

yade.plot.addData({'t':O.time,'i':O.iter,'eps':strainer.strain,'sigma':strainer.avgStress})

Nasibeh

2010/4/18 Václav Šmilauer <eudoxos@xxxxxxxx>

> > 1- Uniaxial Strain Test is not a good test for verification of contact
> > law?
> No, not for the very first check. First you need to verify that the
> contact law really obeys formulas you have on paper and which you can
> solve (plot, whatever) by hand. Triaxial will give you macroscopic
> numbers, which are made up from interaction of hundreds or thousands
> particles and interaction. How do you check that by hand??
> > 2- why the uniaxial starin test give bad results (even with contact
> > law in YADE)?
> What do you mean, bad? The program does exactly what you told it to do.
> (BTW if you have the non-equality of Material::young and macroscopic
> stiffness in mind: they are not the same, and that is OK. If you wrote
> your own Ip2 functor, you should understand how E maps to contact
> stiffness, which in turn influences (but is not equal to) macroscopis
> stiffness.
> > 3- In my law (written for a continuum case and its fracture) there are
> > tention, cohesion, compression, shear ,
> >       there are two kinds of stiffness for the contact, before
> > fracture and after fracture,
> >       with a simple case (a sphere-sphere interaction) I can check it?
> Yes you can, of course. Put 2 spheres in interaction and make them move
> in such way that each feature of your contact law is isolated (pure
> tension, pure shear etc); check with hand-computed results.
>
> v
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-users<https://launchpad.net/%7Eyade-users>
> Post to     : yade-users@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~yade-users<https://launchpad.net/%7Eyade-users>
> More help   : https://help.launchpad.net/ListHelp
>

Attachment: UniaxialStrainTest.pdf
Description: Adobe PDF document