Instantiating a new FlowEngine after removing does not work

Hello,

> According to Questions #452277 and #670889

I'm not sure what these questions are, can you please link to them?

Two things to try:

1) Don't delete the old FlowEngine, it should beenough to simply set flow.dead = 1
2) Don't reuse labels in python, here both of your flowengines have the same label 'flow.' It seems unnecessarily risky. Why not just label the new one 'flow2?'

Cheers,

Robert

The links are as follows:
https://answers.launchpad.net/yade/+question/452277
https://answers.launchpad.net/yade/+question/670889

Can anyone help? Thank you very much!

Hello,

Please see my suggestions.

Cheers,

Robert

Dear Robert,

I'm doing erosion experiments, so I need to delete particles many times, maybe 10000 times, maybe more. According to your method, I need to add 10000 FlowEngines. Will this cause the problem of insufficient memory

New information has entered the fold ! ;)

Probably yes. So you will need to implement a reset in FlowEngine source like Bruno suggests in [1].

[1]https://answers.launchpad.net/yade/+question/452277

Thank you. I'll try your advice. It helps me a lot. Thank you very much

Dear Robert,
I reinstalled Yade from the source code，the version is yade-2021-06-08.git-4852d56, and the version of Ubuntu is 18.04.I tried to delete the particles. There is no problem for the FlowEngine, but the PeriodicFlowEngine will give the following warning.
CHOLMOD warning: matrix not positive definite. file: ../Supernodal/t_cholmod_super_numeric.c line: 911
something went wrong in Cholesky factorization, use LDLt as fallback this time1
Does it matter to the simulation? Here is my code:
##______________ First section, generate sample_________

from __future__ import print_function
from yade import pack, qt, plot
from math import *

nRead=readParamsFromTable(
        ## model parameters
        num_spheres=1000,
        targetPorosity= .4,
        confiningPressure=-100000,
        ## material parameters
        compFricDegree=15,#contact friction during the confining phase
        finalFricDegree=30,#contact friction during the deviatoric loading
        young=2e8,
        poisson=.2,
        density=2600,
        alphaKr=7.5,
        alphaKtw=0,
 competaRoll=.22,
        finaletaRoll=.22,
        etaTwist=0,
        normalCohesion=0,
        shearCohesion=0,
        ## fluid parameters
        fluidDensity=1000,
        dynamicViscosity=.001,
        ## control parameters
        damp=0,
        stabilityThreshold=.001,
        ## output specifications
        filename='suffusion',
        unknowOk=True
)

from yade.params.table import *

O.periodic=True
O.cell.hSize=Matrix3(1,0,0, 0,1,0, 0,0,1)
# create materials for spheres
#shear strength is the sum of friction and adhesion, so the momentRotationLaw=True
O.materials.append(CohFrictMat(alphaKr=alphaKr,alphaKtw=alphaKtw,density=density,etaRoll=competaRoll,etaTwist=etaTwist,frictionAngle=radians(compFricDegree),momentRotationLaw=True,normalCohesion=normalCohesion,poisson=poisson,shearCohesion=shearCohesion,young=young,label='spheres'))

# generate particles packing
sp=pack.SpherePack()
sp.makeCloud((0,0,0),(1,1,1),-1,0.3333,num_spheres,False, 0.95,seed=1)
sp.toSimulation(material='spheres')

O.engines=[
        ForceResetter(),
        InsertionSortCollider([Bo1_Sphere_Aabb()]),
        InteractionLoop(
            [Ig2_Sphere_Sphere_ScGeom6D()],
            [Ip2_CohFrictMat_CohFrictMat_CohFrictPhys(label='contact',setCohesionNow=False,setCohesionOnNewContacts=False)],
            [Law2_ScGeom6D_CohFrictPhys_CohesionMoment(useIncrementalForm=True,always_use_moment_law=True)],
 ),
        PeriodicFlowEngine(dead=1,label="flow"),#introduced as a dead engine for the moment, see 2nd section
        GlobalStiffnessTimeStepper(active=1,timeStepUpdateInterval=100,timestepSafetyCoefficient=0.8),
        PeriTriaxController(label='triax',
            # specify target values and whether they are strains or stresses
            goal=(confiningPressure,confiningPressure,confiningPressure), stressMask=7,
            # type of servo-control, the strain rate isn't determined, it shloud check the unbalanced force
            dynCell=True,maxStrainRate=(10,10,10),
            # wait until the unbalanced force goes below this value
            maxUnbalanced=stabilityThreshold,relStressTol=1e-3,
            doneHook='compactionFinished()'
            ),
        NewtonIntegrator(damping=0)
]

import sys
def compactionFinished():
 # after sample preparation, save the state
        O.save('compactedState'+filename+'.yade.gz')
        print('Compacted state saved', 'porosity', porosity())
        # next time, called python command
        triax.doneHook=''
        O.pause()
O.run()
O.wait()

#B. Activate flow engine and set boundary conditions in order to get permeability
flow.dead=0
flow.defTolerance=-1
flow.meshUpdateInterval=-1
flow.useSolver=3
flow.permeabilityFactor=1
flow.viscosity=10
flow.bndCondIsPressure=[0,0,1,1,0,0]
flow.bndCondValue=[0,0,1,0,0,0]
flow.boundaryUseMaxMin=[0,0,0,0,0,0]

O.run(1,1)
csdList=flow.getConstrictionsFull()
print(len(O.bodies),len(csdList),'finished')
flow.dead=1
print(O.bodies[10].shape.radius)

for i in range(10,50):
    O.bodies.erase(i)
print(len(O.bodies))
O.run(1000,True)

flow.dead=0
flow.updateTriangulation=True

O.run(1,1)
csdList1=flow.getConstrictionsFull()
print(len(csdList1))

Thanks Robert Caulk, that solved my question.