crack with peridynamic theory

LAMMPS LAMMPS Mailing List Mirror
1

Hi,

I am trying to use peridynamics for crack simulation in lammps. I have
a questions:
I use the acrylic glass mechanical material for simulation, but the cracks do not grow from pre-crack. I got the force from the test results.
Is the potential not correct or there is another code error?
Thanks for your help

leila abbasiThis is my file:
# 2d crack
units si
dimension 2
boundary s s p
atom_style peri
atom_modify map array
neighbor 0.0002 bin

# create geometry

lattice sq 0.0001
region plate block -0.0065 0.0065 -0.0285 0.0285 -0.004 0.004 units box
create_box 7 plate
create_atoms 1 region plate
region crack block -0.0065 0 0 0.000001 -0.004 0.004 units box
group crack region crack
delete_atoms group crack
pair_style peri/pmb
pair_coeff * * 2.94138e24 0.000301 0.038614 0.22
pair_coeff 4 5 2.94138e24 0 0.038614 0
pair_coeff 2 * 2.94138e24 0.0003001 1 0
pair_coeff 7 * 2.94138e24 0.0003001 1 0

set group all density 1180
set group all volume 25e-8
velocity all set 0.0 0.0 0.0 sum no units box
region 2 block INF INF INF -0.028199 INF INF units box
group A region 2
region 3 block INF INF -0.028199 -0.027898 INF INF units box
group B region 3
region 4 block INF INF -0.027898 0 INF INF units box
group low region 4
region 5 block INF INF 0.000001 0.027898 INF INF units box
group up region 5
region 6 block INF INF 0.027898 0.028199 INF INF units box
group E region 6
region 7 block INF INF 0.028199 0.0285 INF INF units box
group F region 7

set group A type 2
set group B type 3
set group low type 4
set group up type 5
set group E type 6
set group F type 7

fix 1 all nve
fix 7 F addforce 0 2.943e15 0
fix 2 A addforce 0 -2.943e15 0
compute 1 all damage/atom
timestep 3e-8
thermo 20
dump 1 all custom 20 dump.lammpstrj id x y z c_1
dump 2 all xyz 20 dump.xyz
run 200

2

This is a Q for a Peridynamics expert. Probably
won’t get help from the LAMMPS mail list.

Steve

3

even as a non-expert, here are some observations, that are worrying
and should be checked or adjusted to give meaningful results:

- the input creates a very large number of particles. could there be a
typo somewhere with a length or density somewhere?
- despite the high particle density there is no interaction energy at
all, and hence no forces. this, again, suggests a typo or conceptual
error somewhere
- nothing much will happen in only 200 MD steps (of course in this
specific input that is irrelevant, since there are no interactions
computed).

in general, it is common sense, to start with a tiny (even if
unphysical) test system to get the basics right before starting a
(complex) production simulation.

axel.

4

Hi Leila,
I cannot run your code right now given the fact that the amount of particles is huge, but at first glance it looks like you are trying to model pre-crack using different values of the critical strain s00 in pair_coeff. From my understanding, this cannot be done in the present version of LAMMPS.
You might be interested in taking a look at https://github.com/lammps/lammps/issues/984 since yours should be the same problem.
If you have a smaller test system I can try to run it so to gain a better understanding of what is going on.

Riccardo Galbiati