[lammps-users] Temperature thermostat of Collision

Hi, I have read this thread, and Manual 4.6 as suggested by Steve, but I am not sure about how to set up time evolution without the use of a thermostat.

I read that to perform time integration (allowing evolution of time) some of the fixes (nvt, npt, nve or nph) must be used at least, so, how could I get time evolution if I do not use some of these fixes?.

I wonder if the suggestion is to use for example

fix FIXNVT all nvt temp 1500. 1500. 100000

so that the damping time (100000 = Tdamp = temperature damping parameter (time units)) of the temperature is much longer than the simulation time. Is that what you mean by not using thermostat for the system? or there is another way to run the Verlet algorithm without referencing any property like pressure, volume, temperature, energy … ?

Thanks

Victor

Re: [lammps-users] Temperature thermostat of CollisionFrom: Rongjun Zhang <rongjiunzhang@…24…>

Hi, all,

  I think I get it, thank you very much.

Best,

Rj Zhang

2010/1/26 Steve Plimpton <[email protected]>

So, when I use
the NVT or NPT, the velocity of nanoparticle will decrease sharply which
don’t make sense.

Thermostats change velocity. Â If the system is hot and the theromstat
says to cool it, it will reduce the velocities. Â If you don’t want that to
happen, then as Matt said, don’t use a thermostat.

Section 4.16 of the manual discusses various ways to compute temperature
and thermostat on various degrees of freedom.

Steve

Use NVE. It propagates the coordinates without a thermostat.
Matt

You’re confusing integrators with thermostats. Integrators propagate the system through time by (surprise!) integrating Newton’s laws of motion consistent with some set of constraints. A thermostat specifically controls the temperature. A thermostat must be incorporated in an integration scheme, but an integration need not involve a thermostat. For instance, the NVE integrator does not control velocities (and thus temperatures).

You should probably read an introductory text on molecular simulation (Allen and Tildesley, or Frankel and Smit, to name two), to get a better idea of how integrators and thermostats work.

–AEI

I no longer understand your question. If you
run your system with fix nve, what happens? That
is always a good place to start.

Steve