# tension test of a bulk

Hello,

I want to simulate the tension test of a material which has periodic boundary condition in three dimensions. From the documentation, I have found the fix “ deform “ command probably useful for the purpose. So I have written:

fix 2 all deform 1 z erate 0.01 units box remap x
run 1000000

But there is a problem with it. The box expands unlimitedly throughout the running without any necking and fracture.
Please guide me how I can resolve the problem. Is there any other useful command for tensile loading in periodic dimension?

Thanks,
Ezel

I imagine LAMMPS is doing what you ask it to. Maybe you
need non-periodic dimensions in the transverse x,y.

Steve

I have done what you said and checked it for non-periodic dimensions in x,y but it didn’t work again. The box length expands in z direction and the problem is that the atoms change their positions in order to fill the whole box during the length expansion. So after some time steps, the same-sized gaps appear between all the atomic layers. It seems there is a force for atoms that obliges them to change their positions into any box size.
The input script comes below. Please tell me where the problem comes from. I have tried a lot to solve the problem. I really need your help.

#tension test

units metal
boundary s s p
lattice bcc 3.015

variable lx equal 39
variable ly equal 39
variable lz equal 108
variable lz2 equal {lz}-10 variable lx2 equal {lx}-10
variable ly2 equal \${ly}-10

region box block 0 {lx} 0 {ly} 0 \${lz} units box
region box2 block 4 34 4 34 1 107 units box

create_box 2 box
create_atoms 1 region box basis 2 2

group box2 region box2

#Potential
pair_style eam/fs
pair_coeff * * NiTi.fs Ni Ti

thermo_style custom step temp pe press lx ly lz
thermo 100

# initial velocities

velocity all create 300.0 5812775 units box
fix 1 all npt temp 300 300 0.01 z 0.0 0.0 10 drag 0.2

dump a box2 atom 1000 dump.indent
dump t box2 atom 5000 dump-box2-total.indent

# equilibrium

run 20000

unfix 1
undump a

dump 2 box2 atom 1000 dump-l.indent

# Store final cell length for strain calculations

variable tmp equal “lx”
variable L0 equal {tmp} print "Initial Length, L0: {L0}"

Fix deform has a remap option which controls whether the atoms
are continuously remapped to the changing box or not. If you pull
slowly in one periodic dimension, you probably do want to remap.
If the other 2 dims are non-periodic, then you should get necking
and eventual fracture. if you pull too fast, all bets are off.

Steve

Dear Steve,

I have really got confused. The commands are correct. I have also checked the boundary s s p with low strain rate ( fix 2 all deform 1 z erate 0.001 units box remap x ). Everything in input script seems to be alright. But no necking and fracture!

Ezel

I don't know. My guess is that it is not a problem
with LAMMPS, but that the simulation is simply
not doing what you expect.

Steve

Hi Ezel,
There could be a problem with the use of the potential; make sure that the
EAM potential "NiTi.fs", which does not ship with LAMMPS, you are using is
suitable for a bcc lattice. Since Ni and Ti are closed packed fcc and
hcp, so the potential may not be well suited for studying necking and
cracking of a bcc.
Best,
Ray

Hi,

1). A box can not expand unlimitedly . The fracture or necking depends
on the srtress-strain curve. Once you indentiy the yield strain i.e
where the elastic plastic transition occours , then you can use a
Centro Symmetry or CNA to look at the defect structure...

Oscar G.

Hi,

Ray, I have tested other potentials and even eam/fs for its main structure which is a monoclinic. But I have got the same problem and result. I believe that potential is an important factor here. But I think, fracture is not something so dependent on potential. In my view, when the box length is expanding by tensile loading, at least there should be an eventual fracture. Necking is an issue that depends on several things. So at this stage, I do not look for necking. I just expect the material to fracture after about 20% elongation. While the stress-strain curve of the simulation shows some odd numbers of 150%-200% elongation without fracture!!!

Oscar, what is important to me is drawing the stress-strain curve. I know a box cannot expand unlimitedly, but in my case it does!! The only thing that happens during simulation by “fix deform” is creating same-sized gaps between atomic layers which gradually become larger. I see no fracture by visualization. As you talked about identifying the yield strain, I think it needs a reliable simulation. What I have seen as the results of my simulation is not reliable.
I should appreciate again for your attention. Let me know your opinions and suggestions.

Regards,
Ezel

Hi,

@Starry, I'm not sure what kind of structure your trying to simulate.
I have run simulations to predict the mechanical stability in
nanowires and i have found that they dont break or fracture even after
300% of enlogation. Neverthless you should make a plot of stress- vs
strain and observe at what percentage of enlogation the stress
decays.. Also try to run simulation using few atoms at first, and then
try to incrase the size of your system ....

Oscar G.

Hi Ezel,

In addition to Oscar's nice comment below, I would suggest you try with a potential that is well known to reproduce fracturing. For example, you can use your same script, with necessary modifications, to run some tensile test of a fcc Ni structure with EAM Ni_u3.eam potential. This preliminary test will allow you to verify that your script works as expected and get a feel what sizes and erate you need. Then move on to a less known potential NiTi.fs that you are using. Also try fix deform without remap and a slower erate.

Best,
Ray

Dear Oscar and Ray,

Thank you so much for your very rich comments. I will work on your suggestions and report the results to the lammps mailing list.

Best,
Ezel