Hello everyone,
I am trying to obtain the stress-strain curve of tobermorite 14A with LAMMPS using CLAYFF potential. For non-bonded energy, lj/cut/coul/long was used. I am running into several issues here. I am using fix deform to deform the simulation box. For equilibration I am using fix nve and fix npt. After the equilibration, I am deforming the box in x-direction. Two things happened here: 1) If I commented out the fix/nvt part, the stress kept on increasing, 2) If I keep the fix/nvt part, the stress oscillates around 0. But ideally, the stress should increase and then fracture after a particular strain.
Because tobermorite consists water, should I use fix/shake? If I used fix/shake, before the equilibration, I am running into "non-numeric positions" errors. Can anyone help me with the solution? Here's a part of my input file.
# Bonded Interaction
bond_style morse
angle_style harmonic
dihedral_style harmonic
improper_style cvff
minimize 1.0e-4 1.0e-6 100 1000
min_style cg
min_modify dmax 0.1
fix 5 water shake 0.0001 20 10 b 5 a 8 t 6 7 m 15.999400 1.007970
fix 1 all nve
fix 2 all langevin ${temp} ${temp} 5.0 2341456
compute peratom all stress/atom mobile_temp virial
compute p all reduce sum c_peratom[1] c_peratom[2] c_peratom[3] c_peratom[4] c_peratom[5] c_peratom[6]
variable multiplication_factor equal 101.325*1e-6 # stress converstion from atm to GPa
variable f equal ${multiplication_factor}
variable sigmax equal (v_f)*(c_p[1]/${Tvol}) # Volume of tobermorite crystal
fix 10 all langevin ${temp} ${temp} 5.0 2341456
fix 9 all deform 1 x erate ${erate1} units box remap x
timestep 1
thermo 100
thermo_style custom temp v_strainx v_sigmax
run 30000
Regards,
Baig
I didn't see how you defined your stress.
But it is normal to see fluctuating stress as atoms are moving in all directions and atomic force will be changing all the time. You may want to try averaging more atoms or averaging longer time to get stabler results.
Hope this helps.
Zheng
Hello Zheng,
I updated my question. Now, I hope you can see how I have defined stress computation. The atomic stress will fluctuate with timestep but when I am deforming the box, they should increase. And they are increasing if I don't use fix NVT with fix deform. But, if I use fix NVT with fix deform, the stress is oscillating. My concern is, because I am using morse potential for bond_style, the tobermorite should fracture at some point, but it is not.
I will apply your suggestion and let you know. When you are saying, averaging more atoms, you mean increasing the neighbor list, am I correct?
I am not an expert with water simulation, but as far as i know, this is not how they usually calculate strain stress curve.
1. they generally use "fix npt".
2. they use overall pressure as stress instead of stress/atom.
refer to: https://icme.hpc.msstate.edu/mediawiki/index.php/Uniaxial_Tension
and http://lammps.sandia.gov/threads/msg12420.html
- Badges
- Science topic