yade-dev team mailing list archive

["Vasileios Angelidakis (PGR)" <V.Angelidakis2@xxxxxxxxxxxxxxx>]

Hi,


Jan, thank you for even trying to give a solution to this! :) It is much appreciated!


I am using the latest version of YADE downloaded from github (I don't have a number of the version, as it appears as yade-Unknown) and my operating system is Ubuntu 16.04.5 LTS. I have compiled the code after enabling the "PotentialBlock" and the "PotentialParticle" classes.


I found something interesting. Using the PotentialBlocks class, contact detection works smoothly, but the periodic boundaries are not respected, while if I use the PotentialParticles class, the contact detection does not work, but the generated particles interact with the periodic box. This is quite bizarre, since these two classes are practically twins (same mathematical formulation of the particles, using different contact detection algorithms each).


Kind Regards,


Vasileios


Vasileios Angelidakis

Post-Graduate Researcher in Geotechnical Engineering

School of Engineering, Newcastle University

Room 3.04, Drummond Building

Devonshire Terrace, Newcastle upon Tyne, NE1 7RU, UK

E: v.angelidakis2@xxxxxxxxx<mailto:v.angelidakis2@xxxxxxxxx> T: +44 (0)7380317986 W: Personal Page<https://www.students.ncl.ac.uk/vangelidakis2/>

________________________________
From: Yade-dev <yade-dev-bounces+v.angelidakis2=ncl.ac.uk@xxxxxxxxxxxxxxxxxxx> on behalf of Jan Stránský <honzik.stransky@xxxxxxxxx>
Sent: 01 November 2018 08:38:42
To: Vasileios Angelidakis (PGR)
Cc: Yade developers
Subject: Re: [Yade-dev] Potential Blocks in a Periodic Box

Hello,
what version of Yade and what operating system do you use?
PotentialBlocks are not compiled by default. I can **try** your code and to help you (without any guarantee to succeed :-)

In general, bodies should have no problems with O.periodic=True, but interaction evaluation should implement something extra..

cheers
Jan


st 31. 10. 2018 v 17:11 odesílatel Vasileios Angelidakis <b7063391@xxxxxxxxxxxxxxx<mailto:b7063391@xxxxxxxxxxxxxxx>> napsal:

Hi,

I have started working on the "PotentialBlock" code in YADE for the generation of polyhedra using the Potential Particles approach. I want to use these particles in a periodic cell, but it seems the PotentialBlock class is not compatible with periodic boundaries. Would be grateful to get any advice on whether this is the case and where I should focus to implement it myself? (FYI I am still a rookie in C++ development).



In the following lines I paste a minimal working script to demonstrate a potential block falling through the periodic cell.
Visualisation of the simulation is available only in a VTK format (not in qt).

Cheers,



Vasileios



Vasileios Angelidakis

Post-Graduate Researcher in Geotechnical Engineering

School of Engineering, Newcastle University



The script:

from yade import pack
import math

# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
# Clear the directory where the VTK results will be written of all files (but not subdirectories).
# The directory is cleared, so files from previous runs do not interfere with files from new runs.
# If the directory does not exist, it is created.

import os, shutil
folder = './vtk_ele'
try:
    os.mkdir(folder[2:])
except:
    for the_file in os.listdir(folder):
        file_path = os.path.join(folder, the_file)
        try:
        if os.path.isfile(file_path):
            os.unlink(file_path)
    #        elif os.path.isdir(file_path): shutil.rmtree(file_path) #uncomment to also delete the subdirectories
        except Exception as e:
        print(e)
# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #



# Engines
O.engines=[
    ForceResetter(),
    InsertionSortCollider([PotentialBlock2AABB()],verletDist=0.01),
    InteractionLoop(
        [Ig2_PB_PB_ScGeom()],
        [Ip2_FrictMat_FrictMat_KnKsPBPhys(kn_i=1e8, ks_i=1e8, Knormal = 1e8, Kshear = 1e8, useFaceProperties=False, calJointLength=False, twoDimension=False, unitWidth2D=1.0, viscousDamping=0.05)],
        [Law2_SCG_KnKsPBPhys_KnKsPBLaw(label='law',neverErase=False)]
    ),
    NewtonIntegrator(damping=0.0,exactAsphericalRot=False,gravity=[0,-10,0]),
    VTKRecorder(fileName=folder+'/vtkPeriodicCell-VTKRecorder-',recorders=['pericell'],iterPeriod=200)
]

# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #



# Basic dimensions used in the model
distanceToCentre= 0.5
meanSize = 1.0
heightOfBase = 5.0*meanSize

# Material definition
powderDensity = 2000
O.materials.append(FrictMat(young=150e6,poisson=.4,frictionAngle=radians(0.0),density=powderDensity,label='frictionless'))

# Creation of a spheres packing, following a desired PSD
sp=pack.SpherePack()

mn,mx=(Vector3(0,0,0),
       Vector3(4*heightOfBase, 4*heightOfBase, 4*heightOfBase))

sphereRad = sqrt(3.0)*0.5*meanSize
sp.makeCloud(mn,mx,sphereRad,0,10,periodic=True)

# Replacement of the spheres with cuboids
count= 0
rPP=0.05*meanSize
for s in sp:
    b=Body()
    dynamic=True
    wire=False
    color=[0,0,255.0]
    highlight=False
    b.shape=PotentialBlock(
k=0.2, r=0.05*meanSize, R=1.02*sphereRad,
a=[1.0,-1.0,0.0,0.0,0.0,0.0],
b=[0.0,0.0,1.0,-1.0,0.0,0.0],
c=[0.0,0.0,0.0,0.0,1.0,-1.0],
d=[distanceToCentre-rPP,distanceToCentre-rPP,distanceToCentre-rPP,distanceToCentre-rPP,distanceToCentre-rPP,distanceToCentre-rPP], isBoundary=False, color=color, wire=wire, highlight=highlight,
minAabb=Vector3(1.0*sphereRad,1.0*sphereRad,1.0*sphereRad),
maxAabb=Vector3(1.0*sphereRad,1.0*sphereRad,1.0*sphereRad),
maxAabbRotated=Vector3(1.0*sphereRad,1.0*sphereRad,1.0*sphereRad),
minAabbRotated=Vector3(1.0*sphereRad,1.0*sphereRad,1.0*sphereRad),
AabbMinMax=True,fixedNormal=False)

    length=meanSize
    V= 1.0
    geomInert=(2./5.)*powderDensity*V*sphereRad**2
    utils._commonBodySetup(b,V,Vector3(geomInert,geomInert,geomInert), material='frictionless',pos=s[0],  dynamic=True, fixed=False)
    b.state.pos = s[0] #s[0] stores center
    b.state.ori = Quaternion((random.random(),random.random(),random.random()),random.random()) #s[2]
    b.state.mass =V*powderDensity
    O.bodies.append(b)
    b.dynamic = True
    count =count+1

# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #

# Set up Periodic Boundaries
O.periodic = True
O.cell.setBox(4*heightOfBase, 4*heightOfBase, 4*heightOfBase)

# PotentialBlockVTKRecorder recorder
sampleQuality=50 # Increase to improve visual quality of the particles in Paraview
O.engines=O.engines+[PotentialBlockVTKRecorder(fileName=folder+'/eleTest_', iterPeriod=200, twoDimension=False, sampleX=sampleQuality,  sampleY=sampleQuality,  sampleZ=sampleQuality, maxDimension=0.2, label='PBvtkRecorder')]

# Time step
O.dt = 0.2*sqrt(0.3*O.bodies[0].state.mass/1.0e9)
import yade.timing
O.timingEnabled = True
yade.timing.reset()

_______________________________________________
Mailing list: https://launchpad.net/~yade-dev
Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx<mailto:yade-dev@xxxxxxxxxxxxxxxxxxx>
Unsubscribe : https://launchpad.net/~yade-dev
More help   : https://help.launchpad.net/ListHelp