Errors ( "-nan" in thermo / killed by signal:9) was encountered when using PERI package in PD-LAMMPS

Dear LAMMPS users,
I am a novice user of the LAMMPS, currently trying to simulate the stretch of an elastoplastic system with initial crack using PERI package.However, I have encountered some errors when setting the velocity of the top several layers of particles using the “fix move” or “velocity set” command. The simulation terminated with a “killed by signal:9” error or results in “-nan” values in thermo recorded in the out.log file. (the dump output file show orderly crack between the top and other particles)
Similary, when I switched to using the “fix defoem” command and set periodic boundary conditions (p p p), a similar error “-nan” happened. (these error happened even when I tested a PMB brittle model.)

When “killed by signal:9” occurs, the simulation task is killed, and when “-nan” occurs in thermo, the simulation continues, but particles are missing from the dump output structure, leaving only the box. (error messages in attached figures)

The objective I aim to achieve is calculating crack growth for elastoplastic or hyperelastic ductile materials under stretch conditions and obtaining a stress-strain curve. Any assistance provided would be greatly appreciated.

Best Regards.

Jichun Zhao

①here is the dump file:
image
②here is the errors:


③here is the input:

3D Peridynamic simulation"

units si
dimension 3
boundary s s s
atom_style peri
atom_modify map array
neighbor 0.001 bin
#neigh_modify one 40000

lattice sc 0.0004

Create desired target 32x30x12.4mm and add atoms

region specimen block -32e-3 0.0 -30e-3 0.0 -12.4e-3 0.0 units box
create_box 5 specimen # 5-types atoms
create_atoms 1 region specimen

region boxA block -15.91e-3 0.0 -17.3e-3 -15e-3 -12.4e-3 0.0 units box
region boxB block -15.91e-3 0.0 -15e-3 -12.7e-3 -12.4e-3 0.0 units box
region boxC block -18.5e-3 -15.9e-3 -15e-3 -12.7e-3 -12.4e-3 0.0 units box
region boxD block -18.5e-3 -15.9e-3 -17.3e-3 -15e-3 -12.4e-3 0.0 units box

group a region boxA
group b region boxB
group c region boxC
group d region boxD

set group a type 2
set group b type 3
set group c type 4
set group d type 5

Peridynamic material settings for the target

pair_style peri/eps

pair_coeff * * 210E9 70e9 2.0001e-3 0.10 0.2 30e8
pair_coeff 2 3 210E6 70e6 2.0001e-3 0.01 0.2 30e8
pair_coeff 2 4 210E6 70e6 2.0001e-3 0.01 0.2 30e8
pair_coeff 3 5 210E6 70e6 2.0001e-3 0.01 0.2 30e8

Set mass density 800 kg/m^3

set group all density 800

volume = lattice constant^3 every particle volume

set group all volume 6.4e-11

Zero out velocities of particles

velocity all set 0.0 0.0 0.0 sum no units box

Use velocity-Verlet time integrator

fix F1 all nve

Compute damage for each particle

compute C1 all damage/atom
compute sigma all stress/atom NULL
compute pe_per all pe/atom # potential energy per atom
compute ke_per all ke/atom # kinetic energy per atom
write_data geo.data

stretch along y-axis

region top block -32e-3 0.0 -30e-3 0.0 -12.4e-3 0.0 units box # top boundary
group top_atoms region top
fix stretch_top top_atoms move linear 0.0 1 0.0 units box # stretch velocity
#fix stretch_box all deform 10 y final -20e-3 20e-3 units box # deform

Output total energy and kinetic energy

timestep 1.0e-7
thermo_style custom step atoms temp press pe ke vol
thermo 100
dump D1 all custom 4000 dump.peri id type x y z c_C1 c_pe_per c_ke_per #c_sigma[1] c_sigma[2] c_sigma[3] c_sigma[4] c_sigma[5] c_sigma[6]
run 20000

Have you noticed this warning in your output?

WARNING: One or more atoms are time integrated more than once (src/modify.cpp:296)

This indicates a serious problem in your input. Everything else you observe as problems is likely a consequence of that.

Thank you very much for your advice. My negligence did not take into account the warning information here. I will try to correct this problem.

To avoid “integrated more than once” ,I replace “fix move linear” with “velocity set” command, there is no WARNING in my log file, but same error hannpened:

Thank you very much.

Your simulation is not stable. Look at your kinetic energy and pressure. That is usually a sign of bad simulation settings.