steady-state shear flow in LAMMPS

Hi Steve,

I am trying to do the steady-state shear flow in LAMMPS for nanoparticles. I follow the example for the NEMD simulation by in.nemd with the LAMMPS package. The input script is also attached.
However, when I monitor the velocity and shear stress profile of the atoms along the shear direction, I find out some strange behaviors:

  1. Under the high shear strain rate (0.1), the velocity profile almost immediately becomes linear during the simulation. However, the shear stress dose not approach the linear behavior after a long time.
  2. Under the low shear strain rate (0.0001), both the velocity and shear stress profiles do not approach the linear behaviors after a long time simulation.

I do not know why. Do you have any suggestions/comments?

All my best,

In general, these kids of system can take a long time
to equilibrate and then even longer to generate good
statistics, esp for something like the off diagonal
pressure tensor components. You also
need to insure in a NEMD model that both the
NPs and the solvent are shearing together and the
system is remaining disperse. There is no
guarantee that the toy problem in the examples
file satisfies any of these issues.

You might ask Flint Pierce (fpierce at sandia.gov) for
advice. He has done a lot of modeling with these
kinds of systems.

Steve

Dear all,

I submitted an example input script of sheared LJ chains on lammps
website, Steve promised he would put it online soon :wink:

Best,
Laurent