I keep getting an error during bond creation:
creating bond btw atoms 1501 and 1503 (i has now 6 bonds) at step 1
creating bond btw atoms 1502 and 1501 (i has now 1 bonds) at step 1
creating bond btw atoms 1502 and 1503 (i has now 2 bonds) at step 1
creating bond btw atoms 1502 and 1504 (i has now 3 bonds) at step 1
ERROR: Internal fix bond/create error (../fix_bond_create_gran.cpp:465)
it occasionally goes away if I tweak the number of atoms in the simulation, reduce the bond length, or various other small adjustments to parameters. I haven't been able to find anything as to what this error actually means, has anyone encountered this error before?
richti83 | Wed, 02/15/2017 - 17:50
are you running in serial or
are you running in serial or parallel mode ?
In parallel mode I have seen this error when one proc bonds the particle but the other one doesnt. Can you provide the full setup ? (input skript, stl geometry) I can have a closer look.
tjleps | Wed, 02/15/2017 - 21:42
Thanks Richti83, your work on
Thanks Richti83, your work on getting the granular bonding built into LIGGGHTS has been enormously useful to me.
I'm running in parallel compiled on a macbook pro.
Here's the script, it's based off of the "sphere" example in the bonds package:
#bond
atom_style hybrid granular bond/gran n_bondtypes 1 bonds_per_atom 6
#atom_style granular
atom_modify map array
boundary f f f
newton off
units si
communicate single vel yes
processors 2 2 1
region reg block -.1 .1 -.1 .1 -.1 .1 units box
create_box 2 reg
neighbor 0.0005 bin
neigh_modify delay 0 check yes
#New pair style
pair_style gran model hertz tangential history
pair_coeff * *
#bond
bond_style gran
bond_coeff 1 1 5e10 5e10 1 1e18 1e18
# ==============================================================
# create particles IN A DEFINED REGION
# ==============================================================
region membrane block -.025 .025 -.025 .025 0 .0005 units box
region bigbox block -.0255 .0255 -.0255 .0255 0 .0005 units box
region outer block -.025 .025 -.025 .025 0 .0005 units box side out
region anchor intersect 2 bigbox outer
region TallBox block -.025 .025 -.025 .025 0 .1 units box
fix m1 all property/global youngsModulus peratomtype 1.e7 1.e7
fix m2 all property/global poissonsRatio peratomtype 0.3 0.3
fix m3 all property/global coefficientRestitution peratomtypepair 2 0.5 0.5 0.5 0.5
fix m4 all property/global coefficientFriction peratomtypepair 2 0.2 0.2 0.2 0.2
lattice bcc 0.00025
create_atoms 1 region bigbox
set type 1 density 8500 diameter 0.0001
group anchored region outer
fix bondcr all bond/create/gran 1 1 1 0.00025 1 6 prob 1 67867979
#Material properties required for new pair styles
fix boxwall1 all wall/gran model hertz tangential history primitive type 1 zplane -.1
fix boxwall2 all wall/gran model hertz tangential history primitive type 1 zplane .1
fix boxwall3 all wall/gran model hertz tangential history primitive type 1 xplane -.1
fix boxwall4 all wall/gran model hertz tangential history primitive type 1 xplane .1
fix boxwall5 all wall/gran model hertz tangential history primitive type 1 yplane -.1
fix boxwall6 all wall/gran model hertz tangential history primitive type 1 yplane .1
timestep 0.0000005
#region and insertion
#insert the first particles so that dump is not empty
shell mkdir post
run 1
dump dmp all custom 2000 post/dump*.liggghts id type x y z vx vy vz fx fy fz
fix_modify bondcr every 0
#insert ball
create_atoms 2 single 0 0 .0036
set type 2 density 153800 diameter .005
#group NVE region TallBox
fix gravi all gravity 9.81 vector 0.0 0.0 -1.0
fix freeze anchored setforce 0.0 0.0 0.0
#apply nve integration
fix integrate all nve/sphere
fix ts all check/timestep/gran 1000 0.1 0.1
compute 1 all erotate/sphere
#run
run 150000 upto
apple | Sat, 04/22/2017 - 17:03
How to resolve this problem
When I simulate your script, there are something wrong. The feedback is :
ERROR: All masses are not set (../atom.cpp:1176)
ccc@ubuntu:~/LIGGGHTS/LIGGGHTS-WITH-BONDS-master/examples/bondspackage/sphere$
Could you help me ? Your script is very helpful for me. Could you give me the correct script? Thanks very much!
tjleps | Tue, 08/15/2017 - 19:41
Sorry
I completely missed this question sorry I didn't get back to you. I'm not sure why it's giving you that error, did you copy and paste it directly?
richti83 | Thu, 02/23/2017 - 10:00
First try to increase bonds
First try to increase bonds_per_atom to 26 (theoretical a 3D package could have max 3*3*3-1contacts) and set
fix bondcr all bond/create/gran 1 1 1 0.00025 1 26 prob 1 67867979Does it work in single-core operation ? As far as I remember there is a problem when bonds crossing the domain border, you could try to generate the bonds in single core mode, run 1, write_restart and than use this restart file in your main - simulation on many cores.
tjleps | Tue, 08/15/2017 - 19:49
multicore
I'm attempting to do as you suggested and write a restart file after creating bonds in single processor mode, then restart with multiple processors. When I attempt to restart I get segmentation faults. Have you successfully restarted a bonded granular simulation with a new 3D processor grid?
tjleps | Tue, 08/15/2017 - 20:01
Restart
Actually I seem to get segmentation faults even when restarting with a single processor grid and bonds.
tjleps | Thu, 03/02/2017 - 18:24
Thanks, I'll give that a shot
Thanks, I'll give that a shot, I didn't try running in single core operation, but I set the simulation box partitions so the initial conditions were all on one CPU after your first response and I didn't get any errors, so I suspect that it, is as you suspect, an error across CPU boundaries.
rob1 | Sat, 08/19/2017 - 14:53
Hello richti83,
Hello richti83,
I have similar problems in using bond/gran.
Please can you upload a description of the bond_coeff for this bond style? Because there is no documentation online and I do not know what all the coefficients mean.
Thank you,
rob1
richti83 | Wed, 08/23/2017 - 10:55
Hey rob1,
Hey rob1,
have a look in D.O.Potyondya, P.A.Cundall "A bonded-particle model for rock". Sorry I've no time to write a manual page for this style.
best,
Christian