yade-users team mailing list archive

[Seungcheol Yeom <question524563@xxxxxxxxxxxxxxxxxxxxx>]

Question #524563 on Yade changed:
https://answers.launchpad.net/yade/+question/524563

Description changed to:
Hello all,

I am trying to simulate a permeability of my packing using HydrodynamicsLawLBM. 
However, it seems like the lattice set up is giving me a problem.
All my walls are positioned in the positive coordinates but the width of wall appeared a negative. 
Also, all calculated dimensions of the box are incorrect. 

Here is the script:

##############################################################################################################################
#Authors: Luc Sibille  luc.sibille@xxxxxxxxxxxxxxx
#         Franck Lomine  franck.lomine@xxxxxxxxxxxxxx
#
#   Script to simulate buoyancy in a granular assembly with the DEM-LBM coupling.
#   This script was written for a practical work in the OZ ALERT school in Grenoble 2011.
#   Script updated in 2014
##############################################################################################################################


from yade import pack,timing,utils,ymport

young = 1e6 # young's modulus
dropdiam = 1e-3 # diameter of raindrop, 1mm 


print 'importing spheres...'
#import the spheres here
sp = ymport.textExt('test2.sph')
for i in sp:
	soils=O.bodies.append(i)

	
print 'done'
soils = O.materials.append(FrictMat(young=young,poisson=0.4,frictionAngle=radians(30),density=2650.0,label='soil'))

ab,cd = utils.aabbExtrema()
Xmin = ab[0]
Ymin = ab[1]
Zmin = ab[2]
Xmax = cd[0]
Ymax = cd[1]
Zmax = cd[2]
   
## Build of the engine vector
O.engines=[

	ForceResetter(),
	InsertionSortCollider([Bo1_Sphere_Aabb(),Bo1_Box_Aabb()]),
	InteractionLoop(
		[Ig2_Sphere_Sphere_ScGeom(),Ig2_Box_Sphere_ScGeom()],
		[Ip2_FrictMat_FrictMat_FrictPhys()],
		[Law2_ScGeom_FrictPhys_CundallStrack()]
	),
	HydrodynamicsLawLBM(	EngineIsActivated=False,
				WallYm_id=0,
				WallYp_id=1,
				WallXm_id=2,
				WallXp_id=3, 
				WallZp_id=5, 
				WallZm_id=4,
				useWallYm=1,
				useWallYp=1,
				YmBCType=2,
				YpBCType=2,

				useWallXm=0,
				useWallXp=0,
				XmBCType=1,
				XpBCType=1,
				LBMSavedData='spheres,velXY,rho,nodeBD',
				tau=0.6,
				dP=(0,0,0),
				IterSave=200,
				IterPrint=100,
				RadFactor=.7,  # The radius of particles seen by the LBM engine is reduced here by a factor RadFactor=0.6 to allow flow between particles in this 2D case.
				Nx=250,
				Rho=1000,
				Nu=1.0e-6,
				periodicity='',
				bc='',
				VbCutOff=0.0,
				applyForcesAndTorques=True,
				label="Elbm"
	),

	NewtonIntegrator(damping=0.2,gravity=(0.,0.,0.),label="ENewton"),
]

lowerCornerW = (Xmin,Ymin,-0.5*dropdiam)
upperCornerW = (1.2*Xmax,Ymax,0.5*dropdiam)
     
#lowerCornerW = (-0.002,0.00001,0.00001)
#upperCornerW = (0.002,5.e-3,5.e-3)

# defintiion of the material for walls
O.materials.append(FrictMat(young=young,poisson=.4,frictionAngle=radians(30),density=2650,label='walls'))

###############################################################################################################
# Creation of 6 walls at the limit of the simulation domain

## create walls around the packing
wallOversizeFactor=1.001
thickness=0.00001;

#left box
center= ((lowerCornerW[0]+upperCornerW[0])/2,lowerCornerW[1]-thickness/2.0,(lowerCornerW[2]+upperCornerW[2])/2)
halfSize= (wallOversizeFactor*fabs(lowerCornerW[0]-upperCornerW[0])/2+thickness,thickness/2.0,wallOversizeFactor*fabs(lowerCornerW[2]-upperCornerW[2])/2+thickness)
b1=utils.box(center=[center[0],center[1],center[2]],extents=[halfSize[0],halfSize[1],halfSize[2]],color=[1,0,0],fixed=True,wire=True,material='walls')
O.bodies.append(b1)

#right box
center=((lowerCornerW[0]+upperCornerW[0])/2,upperCornerW[1]+thickness/2.0,(lowerCornerW[2]+upperCornerW[2])/2)
halfSize =(wallOversizeFactor*fabs(lowerCornerW[0]-upperCornerW[0])/2+thickness,thickness/2.0,wallOversizeFactor*fabs(lowerCornerW[2]-upperCornerW[2])/2+thickness)
b2=utils.box(center=[center[0],center[1],center[2]],extents=[halfSize[0],halfSize[1],halfSize[2]],color=[0,1,0],fixed=True,wire=True,material='walls')
O.bodies.append(b2)

#back box
center=(lowerCornerW[0]-thickness/2.0,(lowerCornerW[1]+upperCornerW[1])/2,(lowerCornerW[2]+upperCornerW[2])/2)
halfSize=(thickness/2.0,wallOversizeFactor*fabs(lowerCornerW[1]-upperCornerW[1])/2+thickness,wallOversizeFactor*fabs(lowerCornerW[2]-upperCornerW[2])/2+thickness)
b3=utils.box(center=[center[0],center[1],center[2]],extents=[halfSize[0],halfSize[1],halfSize[2]],color=[0,0,1],fixed=True,wire=True,material='walls')
O.bodies.append(b3)

#front box
center=(upperCornerW[0]+thickness/2.0,(lowerCornerW[1]+upperCornerW[1])/2,(lowerCornerW[2]+upperCornerW[2])/2)
halfSize=(thickness/2.0,wallOversizeFactor*fabs(lowerCornerW[1]-upperCornerW[1])/2+thickness,wallOversizeFactor*fabs(lowerCornerW[2]-upperCornerW[2])/2+thickness)
b4=utils.box(center=[center[0],center[1],center[2]],extents=[halfSize[0],halfSize[1],halfSize[2]],color=[1,1,0],fixed=True,wire=True,material='walls')
O.bodies.append(b4)

#bottom box
center=((lowerCornerW[0]+upperCornerW[0])/2,(lowerCornerW[1]+upperCornerW[1])/2,lowerCornerW[2]-thickness/2.0)
halfSize=(wallOversizeFactor*fabs(lowerCornerW[0]-upperCornerW[0])/2+thickness,wallOversizeFactor*fabs(lowerCornerW[1]-upperCornerW[1])/2+thickness,thickness/2.0)
b5=utils.box(center=[center[0],center[1],center[2]],extents=[halfSize[0],halfSize[1],halfSize[2]],color=[0,1,1],fixed=True,wire=True,material='walls')
O.bodies.append(b5)

#top box
center=((lowerCornerW[0]+upperCornerW[0])/2,(lowerCornerW[1]+upperCornerW[1])/2,upperCornerW[2]+thickness/2.0)
halfSize=(wallOversizeFactor*fabs(lowerCornerW[0]-upperCornerW[0])/2+thickness,wallOversizeFactor*fabs(lowerCornerW[1]-upperCornerW[1])/2+thickness,thickness/2.0);	 	
b6=utils.box(center=[center[0],center[1],center[2]],extents=[halfSize[0],halfSize[1],halfSize[2]],color=[1,0,1],fixed=True,wire=True,material='walls')
O.bodies.append(b6)


# loop over bodies to change their 3D inertia into 2D inertia (by default in YADE particles are 3D spheres, here we want to cylinder with a length=1).
for s in O.bodies:
	if isinstance(s.shape,Box): continue
	r=s.shape.radius
	oldm=s.state.mass
	oldI=s.state.inertia

	m=oldm*3./4./r
	s.state.mass=m

	s.state.inertia[0] = 15./16./r*oldI[0]	#inertia with respect to x and y axes are not used and the computation here is wrong
	s.state.inertia[1] = 15./16./r*oldI[1]  #inertia with respect to x and y axes are not used and the computation here is wrong
	s.state.inertia[2] = 15./16./r*oldI[2]  #only inertia with respect to z axis is usefull
	
O.dt=1e-7 #use a fix value of time step, it is better to not use a time stepper computing a new step at each iteration since DEM time step is eventually fixed by the LBM engine.


yade.qt.View() #to see what is happening 


print"___________________"
print"PHASE 1"
print "Settlement of particle under gravity only, fluid flow is not activated at the present time... please wait"
print"___________________"

O.run(60000,1)  #settlement of particle under gravity only, no action of
the fluid flow.


#definition of a runnable python function to increase progressively the fluid pressure gradient.
def IncrDP():
	currentDP=Elbm.dP
	Elbm.dP = (1.0/1.e10,0,0)
#	Elbm.dP=(currentDP[0]+1.0/1.e10,0,0)
#	O.engines[6].dP=currentDP+1.0/100.0	
#	if not(O.iter%10):
#		print "dP=", Elbm.dP[0]

O.engines=O.engines+[PyRunner(iterPeriod=1,command='IncrDP()')]

#Necessary to initiate correctly the LBM engine
O.resetTime()

#Activation of the LBM engine
Elbm.EngineIsActivated=True

#Cundall damping for solid particles removed (the fluid viscosity should be enough to damp the solid particles ... in general, not always...)
ENewton.damping=0.0

# Now you just need to run the simulation as long as you want to ...
note that if you wait to long the pressure gradient will become very
large inducing great fluid velocities, and the fluid simulation may
become crazy!! Look at the Mach number (M) printed in the console, it
should not be too high...

print"___________________"
print"PHASE 2"
print "Fluid flow has been activated now with a gradual increase of the pressure gradient."
print "Run the simulation, until the pressure gradient is strong enough to compensate the gravity forces, then particles will mouve to the right."
print"..."
print"..."
print "Velocity and pressure field of the fluid can be plotted with the matlab script yadeLBMvisu.m"
print "Terzaghi critical hydraulic gradient can be compared with the one deduced from the simulation with the matlab script Buoyancy_Gradient.m"
print"___________________"

here is  the external data in sph in the following link:
https://www.dropbox.com/s/mgmhoi7wjrlq49k/test2.sph?dl=0

Thank you.

-- 
You received this question notification because your team yade-users is
an answer contact for Yade.