pair_style and integration nvt or nve

Submitted by rasoul on
moritzhoefert

Hi rasoul,

NVT ensembles are usually used for molecular dynamics and I think it does not make sense to use that for DEM simulations. In MD temperature is a directly accessible quantity: the velocity fluctuations of atoms. And the thermostat is applied to those. In DEM temperature is a per particle quantity independent of the velocities.

You could use fix nve/sphere to propagate the trajectories of DEM particles and heat/gran/conduction to track temperature.

Best,
Moritz

Tue, 27 Aug 2013 - 13:05 Permalink
rasoul

Thanks a lot for the response. So, how can I fix temperature in a target point or decrease it slowly?

Tue, 27 Aug 2013 - 22:47 Permalink
ckloss

Hi rasoul,

as Moritz said, NVT is for molecular dynamics, and temperature here is temperature in the sense of kinetic gas theory, which is not approropriate in this case

have a look at the heat transfer tutorials that come with LIGGGHTS.
with the "set" command, you can set the temperature of particles as you desire

Cheers
Christoph

Fri, 30 Aug 2013 - 15:15 Permalink

I have a question about pair_style. I'd like to simulate solid materials such as shale, clay, sand, and like this. I need to know what the appropriate pair_style for doing that is?

Another question is about integration. In my project I have to decrease temperature from X1 to X2. How can I do that?
And using "fix integer nvt/sphere temp X1 X2 DX" is true or not?

Thanks a lot for the response in advance.