simple 2D tensile test simulation FIX settings

Dear LAMMPS-users,

I am a beginner in LAMMPS and MD. My question is regarding whether NVE or NVT ensemble to use in the simulation.

I am simulating a simple 2D tensile test with atomic style, LJ pair potential and hex lattice in order to get stress-strain curve. I modified the crack example problem by adding relaxation step before straining using the following commands.

###### Relaxation ###############
fix f1 all nvt temp 0.01 0.01 .01
run 10000
unfix f1

then, the tensile straining of the model is done by fixing the bottom layer and giving the upper layer a certain velocity. This simulation settings are as follows.

###### Tensile Settings-1 ###############

fix 1 all nve
fix 2 boundary setforce 0.0 0.0 0.0 # boundary is union of upper and bottom layers
run 1000000

But, in literature, i am seeing different settings like,

###### Tensile Settings-2 ###############

fix 1 boundary nve
fix 2 boundary setforce 0.0 0.0 0.0
fix 3 mobile nvt temp 0.01 0.01 .01 # mobile is the middle region
run 100000

If i do not intend to maintain the temperature of the model, can i use the settings-1? Or is it always necessary to maintain the temperature in the tensile test as in settings-2?
Please suggest me. Your help will be very useful in my analysis. Thanks in Advance.

Regards,
Mathiazhagan S

Hi,

if you want Isothermal tensile deformation , then use NVT . Else you could try using NVE, just make sure energy is conserved (fluctuate around a value)

Oscar G.

Thanks Oscar.

During Relaxation the TotEng is converged close to -1.952.

But, during straining of the model the TotEng is increasing from -1.9364 to -1.8698. Temperature does not change significantly (~0.025) till stress reaches maximum. After the fracture initiated, the temperature increase from 0.025 to 0.065 with high rate.

I think that the total energy has to increase due to the increase in potential energy during straining. Is this acceptable?

I have one more question:
Is it possible to simulate the air/water environment with constant temperature around the solid where the solid material is undergoing tensile straining? So that the heat exchange can happen with the environment due to contact of solid and gas/liquid particle. This situation is similar to the macrolevel tensile test. Please help me whether this is possible in LAMMPS. what fixes are necessary?

Note: total atoms in the simulation is 167,700. strain rate 2E-4, timestep 0.001, in reduced LJ units.

Thanks
Mathi

Your posted script is using both NVE and NVT. You cannot do this.
You are integrating all the atoms twice. You should have gotten a
LAMMPS warning about this.

Steve

Steve, I used in my tensile test simulation following fix commands.

Relaxation

fix f1 all nvt temp 0.01 0.01 .01
run 10000
unfix f1

Tensile Settings-1

fix 1 all nve
fix 2 boundary setforce 0.0 0.0 0.0
run 1000000

My question in my first mail was: Do i need to use settings-1 OR settings-2 which i mentioned in my first mail.

When i used the above mentioned settings, the following things are observed:

1)During Relaxation the TotEng converged close to -1.952.
2a)During tensile straining of the model the TotEng is increasing from -1.9364 to -1.8698.
2b)Temperature does not change significantly (~0.025) untill stress reaches maximum value.
2c)After the maximum stress reached, the temperature increased from 0.025 to 0.065 with high rate.

I think that the total energy has to increase due to the increase in potential energy during straining. Is this acceptable during NVE?

my one more question was:
Is it possible to simulate the air/water environment with constant temperature around the solid where the solid material is undergoing tensile straining? So that the heat exchange can happen with the environment due to contact of solid and gas/liquid inter-phase. This situation is similar to the macro-level tensile test. Please help me whether this is possible in LAMMPS. what fixes are necessary?

Thanks,
Mathi

Hi,

Mathi wrote:
1)During Relaxation the TotEng converged close to -1.952. 2a)During tensile straining of the model the TotEng is increasing from -1.9364 to -1.8698.2b)Temperature does not change significantly (~0.025) untill stress reaches maximum value. 2c)After the maximum stress reached, the temperature increased from 0.025 to 0.065 with high rate.

Yeah that’s fine, a change of -1.952 to -1.8698 its acceptalbe (I guess) , and yeah the temperature seem to be ok . … I have a hard time when trying to get a meaning from numbers . Its more easy to visualise, and tell if the simulation is correct if a graph of energy vs timesteps , or temperature vs timesteps is provided =>

Mathi wrote:
I think that the total energy has to increase due to the increase in potential energy during straining. Is this acceptable during NVE?

Yes, but , then it should stabilize

Mathi wrote:
Is it possible to simulate the air/water environment with constant temperature around the solid where the solid material is undergoing tensile straining? So that the heat exchange can happen with the environment due to contact of solid and gas/liquid particle. This situation is similar to the macrolevel tensile test. Please help me whether this is possible in LAMMPS. what fixes are necessary?

No idea… And not enough brain power to come with some ideas of how to do it

Oscar G.

Comments below.

Steve

Steve, I used in my tensile test simulation following fix commands.

###### Relaxation ###############

fix f1 all nvt temp 0.01 0.01 .01
run 10000
unfix f1

###### Tensile Settings-1 ###############

fix 1 all nve
fix 2 boundary setforce 0.0 0.0 0.0
run 1000000

My question in my first mail was: Do i need to use settings-1 OR settings-2
which i mentioned in my first mail.

When i used the above mentioned settings, the following things are observed:

1)During Relaxation the TotEng converged close to -1.952.
2a)During tensile straining of the model the TotEng is increasing from
-1.9364 to -1.8698.
2b)Temperature does not change significantly (~0.025) untill stress reaches
maximum value.
2c)After the maximum stress reached, the temperature increased from 0.025 to
0.065 with high rate.

I think that the total energy has to increase due to the increase in
potential energy during straining. Is this acceptable during NVE?

All of these results sound fine to me. If you're stretching the
material you are
adding energy to the system.

my one more question was:
    Is it possible to simulate the air/water environment with constant
temperature around the solid where the solid material is undergoing tensile
straining? So that the heat exchange can happen with the environment due to
contact of solid and gas/liquid inter-phase. This situation is similar to
the macro-level tensile test. Please help me whether this is possible in
LAMMPS. what fixes are necessary?

You can thermostat a liquid or gas surrounding a solid. Whether enough
heat will transfer from the solid to the surroundings to effectively
thermostat the solid is up to you to experiment with.

Thanks Steve and Oscar.
Your comments are useful for me to understand the problem better.

Regards,
Mathi