Poiseuille flow - High temperature on wall atoms

LAMMPS LAMMPS Mailing List Mirror
1

Dear Lammps users,

I am new to lammps. I am trying to model the classic example of poiseuille flow in 3D. So I tired to modify the example problem given in the Lammps package mainly for a real wall model ( velocities are not set to zero for wall particles and a berendsen thermostat is used on wall and fluid particles separately).

In current simulation Argon gas is confined in two parallel Platinum plates. Now the problem I am facing while running the code is the very high temperature in the wall atoms with a time step less than 0.0001 fs or atoms are lost with a time step greater than 0.0001 fs.

A part of the code related to the potentials are given below:

pair_style lj/cut 10.215
pair_coeff 1 1 3.45 0.238 10.215
pair_coeff 2 2 2.47 7.495 6
pair_coeff 1 2 2.94 0.157 9

[ 1 for argon particle, 2 for platinum particle]

How can we reduce the high velocity/energy on wall atoms due to the potential?

Thanking you for your time.

Regards

Abdul Gafoor

2

Hard to say. In principle your model should work. I suggest

you build up the model one step at a time. E.g. can

you run stable dynamics with no applied force to

the flow atoms?

Steve

3

Thank you sir for your kind reply

As You suggested I tried to run without any external force on flow atoms and applying some energy drain commands ( fix viscous and nve/limit) , I am able to run the simulation with larger time-step (0.1 fs) and I am getting the wall temperature damped to the required value ( In my problem I set it to 350 K). But when I am trying to visualize the flow by VMD, the wall atoms are not maintaining its position, i.e., its moving into the flow region.

Any suggestion on this will be helpful for me.

Thanks in advance.

4

Your questions are too vague. I suggest you debug your

script by starting simple, and adding one feature at a time.

Steve