Particles deleted unexpectedly when using 'fix insert/pack' with CFDEM

Nucleophobe's picture
Submitted by Nucleophobe on Fri, 05/15/2015 - 16:42

CFDEM developers, community:

I am experiencing some strange behavior using CFD-DEM coupling with continuous particle insertion. On the timestep where a new particle is inserted, old particles are sometimes deleted erroneously.

Here are the relevant lines from my LIGGGHTS input file:
atom_style granular
atom_modify map array
communicate single vel yes
#Goal: allow particles to leave the domain when exceeding xMax:
# (using m m because STL surface for collisions extends beyond box)
boundary f m m
newton off
units si
processors 8 2 2
region reg block -0.155 0.015 -0.013 0.013 -0.013 0.013 units box
create_box 1 reg
neighbor 0.3 bin
neigh_modify delay 0 binsize 0.01 one 500000 page 5000000
#
#... more code
#
# Region for inserting particles
region bc cylinder x 0 0 0.013 -0.104 -0.096 units box # Cylinder region from x=-0.104 to x=-0.096
fix pts1 all particletemplate/sphere 1 atom_type 1 density constant 1200 radius constant 0.0015 # r=0.00025
fix pdd1 all particledistribution/discrete 1. 1 pts1 1.0
# Insert a new particle every 0.5 seconds of simulation time: (ts = 0.00005)
fix ins all insert/pack seed 11641 distributiontemplate pdd1 vel constant 0.0395 0. 0. insert_every 10000 overlapcheck yes all_in yes particles_in_region 1 region bc
#...

And here are lines from the dump file showing the disappearing particles (particle ID 1 is deleted when particle ID 3 is inserted):
ITEM: BOX BOUNDS ff mm mm
-0.155 0.015
-0.0135916 0.013
-0.013 0.013
ITEM: ATOMS id type type x y z ix iy iz vx vy vz fx fy fz omegax omegay omegaz radius
2 1 1 -0.0707589 -0.00180427 -0.00122614 0 0 0 0.0619239 0.00190548 0.00128979 5.46952e-07 -2.91851e-08 -5.11237e-09 -0.0019066 0.76076 -1.09011 0.00015
1 1 1 -0.0487447 -0.00232496 -0.00273055 0 0 0 0.0575022 0.00247728 0.00293016 5.48912e-07 -3.63463e-08 2.47315e-10 0.000969468 1.85788 -1.52325 0.00015
ITEM: TIMESTEP
20200
ITEM: NUMBER OF ATOMS
2
ITEM: BOX BOUNDS ff mm mm
-0.155 0.015
-0.0135916 0.013
-0.013 0.013
ITEM: ATOMS id type type x y z ix iy iz vx vy vz fx fy fz omegax omegay omegaz radius
3 1 1 -0.0999267 0.00563776 0.00220048 0 0 0 0.0471527 -0.000271715 -0.000118425 3.14522e-06 -4.30962e-07 -1.95865e-07 -0.00145423 -0.354587 1.05581 0.00015
2 1 1 -0.0701392 -0.00178516 -0.0012132 0 0 0 0.0620368 0.00191506 0.00129863 -3.07217e-07 6.06321e-08 2.18483e-08 0.000987506 0.763755 -1.05569 0.00015

Here's another example (particle ID 4 is deleted when particle ID 7 is inserted):
ITEM: BOX BOUNDS ff mm mm
-0.155 0.015
-0.0135916 0.013
-0.013 0.013
ITEM: ATOMS id type type x y z ix iy iz vx vy vz fx fy fz omegax omegay omegaz radius
6 1 1 -0.0786799 0.00340481 0.00389949 0 0 0 0.0410934 -0.00432919 -0.00525659 2.61628e-06 -2.41266e-07 -3.3253e-07 0.0198822 -2.25548 1.93234 0.00025
4 1 1 -0.0472689 0.000953339 0.000836285 0 0 0 0.0503288 -0.00661034 0.00473196 1.00535e-06 4.34848e-07 -3.31658e-07 -0.105532 0.589194 1.89323 0.00025
5 1 1 -0.0635287 0.00251102 0.00132621 0 0 0 0.0473274 -0.00682637 -0.00369833 1.51048e-06 -7.32165e-08 -3.44686e-09 -0.0653509 -1.19873 2.24658 0.00025
3 1 1 -0.0230975 0.00363574 0.00131126 0 0 0 0.0352516 0.00408688 0.000925187 -0.000133236 -0.000173328 9.45629e-05 -0.541975 2.72466 5.1284 0.00025
2 1 1 -0.0158195 0.00110638 0.00161044 0 0 0 0.0230949 -0.00352434 0.00778573 -5.98092e-06 1.01296e-05 2.59813e-06 -1.08076 4.96985 5.94729 0.00025
1 1 1 -0.0100704 -0.0019781 0.00101307 0 0 0 0.0120726 -0.00830111 -0.00335838 -1.16186e-05 8.13035e-07 -1.58355e-06 -0.15215 -2.61061 7.19427 0.00025
ITEM: TIMESTEP
60200
ITEM: NUMBER OF ATOMS
6
ITEM: BOX BOUNDS ff mm mm
-0.155 0.015
-0.0135916 0.013
-0.013 0.013
ITEM: ATOMS id type type x y z ix iy iz vx vy vz fx fy fz omegax omegay omegaz radius
7 1 1 -0.100432 0.00227604 -0.00762573 0 0 0 0.0402488 -4.9438e-05 0.000122548 -3.28559e-06 -1.91027e-07 7.98753e-07 -0.00929003 0.699178 0.221904 0.00025
6 1 1 -0.0782678 0.00336127 0.00384656 0 0 0 0.041335 -0.00437321 -0.00532351 1.81383e-06 -2.24594e-07 -3.22919e-07 0.019886 -2.26187 1.93835 0.00025
5 1 1 -0.0630541 0.00244276 0.00128926 0 0 0 0.0476029 -0.00682293 -0.00368379 1.63373e-06 -1.72754e-08 1.47971e-07 -0.0607807 -1.19635 2.21848 0.00025
1 1 1 -0.00995247 -0.00206118 0.00097922 0 0 0 0.0119441 -0.00830313 -0.00337976 4.24406e-06 2.22953e-06 1.19203e-06 -0.114678 -2.68875 7.00383 0.00025
2 1 1 -0.0155881 0.00107903 0.00168764 0 0 0 0.0228722 -0.00227373 0.0076712 -1.18873e-05 4.47322e-06 -2.05496e-06 -1.25972 5.17892 5.81442 0.00025
3 1 1 -0.0228135 0.00357881 0.0013771 0 0 0 0.0272345 -0.00897316 0.00860644 2.08441e-06 1.6362e-05 -1.09796e-05 -1.01763 5.80691 9.64869 0.00025

In both cases, the insertion of one particle results in the deletion of another particle (or particles, in some cases). I have not been able to identify a pattern in this behaviour.

Is this a bug?

Thank you for your attention,
-Nuc

P.S.
Some cases (rare) seem to work fine and only delete particles when they leave the x-range permitted by the boundary (xmin - xmax) as I intended. For instance, in this case, one particle (ID 5) is removed because the particle has travelled outside the limits. This occurs on the same timestep as the insertion of a new particle (ID 12):
ITEM: NUMBER OF ATOMS
11
ITEM: BOX BOUNDS ff mm mm
-0.155 0.015
-0.0135916 0.013
-0.013 0.013
ITEM: ATOMS id type type x y z ix iy iz vx vy vz fx fy fz omegax omegay omegaz radius
11 1 1 -0.0880044 -0.00547016 -0.00546559 0 0 0 0.0255435 0.00412242 0.00388754 7.26289e-06 9.9724e-07 1.03138e-06 -0.00212395 2.56564 -2.5784 0.0003
9 1 1 -0.0544477 0.00126116 0.000902297 0 0 0 0.0452507 0.0105621 0.00471717 -2.12437e-06 -1.11829e-06 -6.84461e-07 0.0319267 0.489552 -1.51031 0.0003
10 1 1 -0.078893 0.00200222 0.00521005 0 0 0 0.0325222 -0.00380874 -0.00923083 5.31206e-06 -5.7978e-07 -1.66628e-06 0.0354062 -3.141 1.16836 0.0003
8 1 1 -0.02477 0.00347819 0.00106783 0 0 0 0.0205822 -0.0079173 0.00189817 6.09786e-06 2.67063e-05 -9.55842e-06 0.0281231 1.56521 6.80862 0.0003
4 1 1 -0.0152439 -0.00284554 -0.00389426 0 0 0 -8.79267e-05 -7.88866e-06 -6.7413e-05 4.89279e-08 2.79865e-07 5.26685e-07 -0.0118238 -0.0819732 -0.00468082 0.0003
2 1 1 -0.0122527 -0.00143074 0.00109548 0 0 0 -8.29417e-05 -3.84248e-06 3.58868e-05 -1.50286e-06 3.13984e-07 8.99617e-07 -0.00146384 0.0245559 -0.0307319 0.0003
1 1 1 -0.00721803 -0.00444346 3.67808e-06 0 0 0 -2.15014e-05 -1.18537e-05 1.27733e-06 -1.52922e-07 -2.90357e-07 2.99656e-08 -0.000352378 -0.000946796 0.00802332 0.0003
7 1 1 -0.0152333 -0.00653144 0.00207487 0 0 0 -7.30332e-05 -3.39805e-05 3.43961e-05 -2.62017e-07 8.25046e-08 1.21775e-07 0.00556762 0.0344168 0.0544397 0.0003
6 1 1 0.0047888 0.00443728 -1.93021e-05 0 0 0 0.0340868 -0.00125497 0.000960791 -1.83136e-06 7.48191e-07 -1.09826e-06 0.203599 -0.900522 -1.92773 0.0003
3 1 1 -0.0176475 0.0021231 0.00248574 0 0 0 -1.00553e-05 -6.59865e-05 6.57673e-06 2.23159e-07 8.9971e-07 -5.35494e-07 -0.00621281 -0.0087654 0.02463 0.0003
5 1 1 0.0162392 0.00316116 0.000668724 0 0 0 0.0282496 -0.000578754 0.000823694 2.91496e-06 -1.02307e-06 4.12971e-07 0.342838 0.623427 -0.678642 0.0003
ITEM: TIMESTEP
110200
ITEM: NUMBER OF ATOMS
11
ITEM: BOX BOUNDS ff mm mm
-0.155 0.015
-0.0135916 0.013
-0.013 0.013
ITEM: ATOMS id type type x y z ix iy iz vx vy vz fx fy fz omegax omegay omegaz radius
12 1 1 -0.0995477 0.00700148 -0.00649198 0 0 0 0.0325532 8.5795e-05 3.5901e-05 -1.92853e-05 8.02196e-07 -7.61557e-07 0.000425956 0.228779 0.21152 0.0003
11 1 1 -0.0877476 -0.00542884 -0.00542608 0 0 0 0.0257442 0.00417355 0.00398231 7.39351e-07 8.42356e-07 4.59137e-07 -0.00200255 2.56979 -2.58085 0.0003
10 1 1 -0.078566 0.00196383 0.00511724 0 0 0 0.0329073 -0.0038598 -0.00934098 7.58346e-06 -4.72222e-07 -1.42707e-06 0.0362415 -3.12064 1.15889 0.0003
9 1 1 -0.053996 0.00136621 0.000949249 0 0 0 0.0450905 0.0104504 0.0046744 -2.83982e-07 -1.22892e-06 -5.61343e-07 0.026817 0.471132 -1.46155 0.0003
4 1 1 -0.0152447 -0.00284547 -0.00389482 0 0 0 -7.27879e-05 2.24544e-05 -5.47021e-05 2.14389e-07 2.55654e-07 -1.34634e-07 -0.0146908 -0.0746685 -0.00620721 0.0003
2 1 1 -0.0122539 -0.00143061 0.001096 0 0 0 -0.000122002 3.52847e-05 5.02257e-05 3.07092e-07 5.12211e-07 -4.32226e-07 0.000532685 0.0289826 -0.0506757 0.0003
1 1 1 -0.00721828 -0.00444363 3.70069e-06 0 0 0 -2.85908e-05 -1.72185e-05 1.77147e-06 2.2661e-09 1.2429e-07 -6.45433e-08 -0.000414651 -0.00209662 0.0159919 0.0003
7 1 1 -0.0152341 -0.00653181 0.00207527 0 0 0 -8.35488e-05 -4.11761e-05 4.33807e-05 -7.47923e-09 -1.58631e-07 3.14401e-08 0.00857 0.0539211 0.0855279 0.0003
3 1 1 -0.0176475 0.00212268 0.00248568 0 0 0 1.54842e-05 -2.79704e-05 -1.26201e-05 4.56775e-07 1.23753e-07 -3.00825e-08 -0.00571372 -0.0250766 0.049299 0.0003
8 1 1 -0.0245623 0.00340811 0.00108346 0 0 0 0.0207895 -0.00631283 0.00130862 2.64577e-06 1.34051e-05 -4.57525e-06 0.0188569 1.41147 6.42322 0.0003
6 1 1 0.00513052 0.00442508 -1.01532e-05 0 0 0 0.0342973 -0.00118293 0.000867038 1.43931e-06 8.91979e-07 -1.11891e-06 0.197354 -0.884474 -1.91146 0.0003

So, this last case seems to work, at least as far as I have let it run.

ckloss's picture

ckloss | Wed, 07/01/2015 - 14:23

Hi Nucleophobe,

which versions are you using?

Best wishes
Christoph

Nucleophobe's picture

Nucleophobe | Wed, 09/02/2015 - 18:01

Sorry for not responding earlier.

I will try this again with the latest version and let you know if I am still encountering the same behavior.

Thanks,
-Nuc

tuks123 | Mon, 08/01/2016 - 15:09

Hi Nucleophobe,

Was your problem solved?

Hello Christoph, which version is this problem fixed? We are using CFDEM_Public_2.2.0. AND We see similar problem when density of fluid (1 g/cc) and particle (1.1g/cc) are of same order. The problem doesn't occur when density of fluid is 1/100 that of particle.

Regards
Tuks

Nucleophobe's picture

Nucleophobe | Wed, 08/03/2016 - 03:21

Tuks,

I still have this problem but I discovered a workaround.

If I remember correctly, I changed the 'boundary' command from 'boundary f f f' to 'boundary m m m'.
http://www.cfdem.com/media/DEM/docu/boundary.html

This means that particles exit my simulation domain and are never deleted, which slows down the simulation over time as the total number of particles increases. However, it was a satisfactory workaround for my application.

Let me know if this doesn't work for you and I can look back at my files to make sure this is how I fixed the problem. I would look now but I don't currently have access to my files.

Nuc

ckloss's picture

ckloss | Fri, 08/26/2016 - 09:44

Hi guys,

nice to hear you found a workaround. Anyway, when reporting an issue please always do this for the latest version, it's the only one we support (we're not Microsoft ;-) )

Best wishes
Christoph