peridynamics / method lps

LAMMPS LAMMPS Mailing List Mirror
1

Hello,

       I study stress on a beam with lps method. I defined a sc organisation and the studied volume is a cube on one side lattice. I applied two linear movements on each extremity.
       I get negative value for the stress and values are too low compared to usual values. After a lot of searches, i didn´t succeed in solving this problem. Do you have any explanations ?

Thanks you for your consideration

Simon

Input File :

# 3D peridynamic simulation for traction

units nano
dimension 3
boundary s s s
atom_style peri
atom_modify map array
neighbor 0.1 bin
lattice sc 0.3615

#define all the parts

region beam block 0 36.15 0 2.169 0 2.169 units box
create_box 1 beam
create_atoms 1 region beam

region leftbc block 0 1.5337 0 2.169 0 2.169 units box
group leftbc region leftbc
region rightbc block 34.6163 36.15 0 2.169 0 2.169 units box
group rightbc region rightbc

group boundary union leftbc rightbc

group middle_atoms subtract all boundary

pair_style peri/lps
pair_coeff * * 140000 48000 5 50 0.01

set group all density 8960e-6

set group all volume 0.047241

compute peratom middle_atoms stress/atom NULL

compute sigmaxx middle_atoms reduce sum c_peratom[1]

variable sigmavolxx equal c_sigmaxx/vol

fix 1 rightbc move linear 3.615 0 0 units box
fix 2 leftbc move linear -3.615 0 0 units box

timestep 0.005e-3 #0.005

thermo 100
thermo_style custom step v_sigmavolxx lx

dump 1 all custom 5000 project_6_peri_metal_res.txt id x y z
dump_modify 1 append yes

run 5000

# SIMULATION DONE

print "All done"

2

It’s more likely that your problem setup or dynamics are wrong,
than that something is wrong with LAMMPS. No one is likely
to try and figure out what is wrong with your model, other than you.

Steve

3

Maybe, you are right but I created a similar file for the method PMB and I got logical results. I just changed the parameters of peri_coeff and peri_style with usual values, but it doesn´t work. Moreover, I tried to see the influence of each parameters on the simulation but I get logical results only with bulk and shear modulus with a power 16.

Thanks you for your consideration

simon

4

Possibly Rezwan can offer an opinion.

Steve

5

make sure the units of material properties, stress etc are consistent (may be try other units like metal, real etc). also you may want to check your result in terms of force-displacement as well. timestep size is a factor too.

hope this helps.

Possibly Rezwan can offer an opinion.

Steve