[lammps-users] some problems about thermal shock simulation

LAMMPS LAMMPS Mailing List Mirror
1

Dear all,
I have utilized lammps simulate thermal shock passing through the iron system, however, there are some phenomenon which I am very confused:

  1. The higher temperature of the NVT ensemble is only 500.0K, but the result is much higher than 500.0K, how does this happen?
  2. I want to get displacement-time distribution during the thermal shock, so I use the lammps commands “compute displace/atom” and “fix ave/time” get the result. But the result is incredible. I think the displacement-time distribution curve should be one peak after another peak, so we can estimate the velocity of the thermal shock wave. Such information couldn’t be output by the command ‘fix ave/time’, isn’t it? what should I do?
    Could anyone should give me some advice?
    Sincerely,
    Maosheng

simulation process:

  1. I establish a system about 1001010, and the system is uniformly divided into 10 regions along direction x— in the 1st region a hot bath-NVT ensemble is imposed (T=500.0K), in the 10th region a cold bath-NVT ensemble is imposed (T=0.01K), and the rest regions are imposed NVE ensemble.
  2. free boundary is imposed on the left outermost surface and a fixed boundary condition is imposed on the right outermost surface in direction x, and periodic boundary conditions are imposed on outermost surface in direction y and z

In the simulation, the timestep is 1fs, and the simulation lasts 6ps.

The input script:

Initialization*************************************************************

units metal
dimension 3
newton on
boundary sf p p
atom_style atomic
neighbor 0.3 bin
neigh_modify check yes
lattice bcc 2.87
region box block 0 100 0 10 0 10 units lattice
create_box 1 box
create_atoms 1 box

2

Your damping constant on the thermostat is 100 in metal units
which is 100 psec, which is 100,000 timesteps. And you
are running for 6000 steps, so the thermostat will do
nearly nothing. In general you will have trouble maintaining
temperature when you invoke a shock, which be definition
is a high temperature event, so I'm not clear on why
you would try to thermostat in this way on a boundary.

You might look at the fix msst command.

Steve

3

Dear Steve,
Thank you for your letters and advice. Because I think when the ensemble is imposed, the temperature evolution process before equilibrium is the same as thermal shock, so I do like this.
I have confirmed that the MSST is developed for mechanical shock waves passing over the system, it isn’t suitable for simulating this kind of shock.
Best wishes,
Maosheng