About Dynamic simulation using fix nve

Dear lammps users,

I am running a dynamic simulation with constant volume and constant energy. With the free molecular motion the pressure build up should’t be excessive but I am getting high pressure fluctuation and also the temperature of system also have to remain somehow constant but the temperature is increasing.

I have attached my input file along with this email. Please suggest me, whether I made an error on some of the commands.

Thanks,
Aaron

1tube.txt (133 KB)

script.in (586 Bytes)

I am running a dynamic simulation with constant volume and constant energy.

That’s an unusual choice. But okay.

With the free molecular motion the pressure build up should’t be excessive
but I am getting high pressure fluctuation

The magnitude of fluctuations in pressure and temperature should decay as 1/sqrt(N) (N = number of particles), so if your pressure fluctuations are large, it could just be a reflection of your small system size
(unless you are near some kind of phase transition).

and also the temperature of
system also have to remain somehow constant but the temperature is
increasing.

Gradual increases in kinetic energy can indicates your timestep is too large. Try decreasing it. You might also want to watch the total energy of the system (“etotal”) to see that it is not increasing. I use:

thermo_style custom step temp pe etotal press

(As a reference point. In my simulations at constant temperature, I sometimes allow for a 0.5% increase in kinetic energy every 100,000 timesteps, or so. I don’t know if this is considered acceptable by the community. To compensate for that I measure the temperature of my simulation instead of assuming it was what I requested. In your case, for NVE simulations you will have to choose a timestep that does not cause an unacceptable increase in total energy over the entire duration of your simulation.)

Also, If you haven’t visualized your system yet, to see what it is doing, do that. Some instructions (using VMD) here.


Check out the vibrant tech community on one of the world’s most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

(slashdot advertises here?)

Hi Andrew,

Thanks, the simulation looks strange because I want to find out the thermostating time from this fix nve setting. I have found out the parameter for thermostating. But still the system is blowing up with Nose hover thermostat. Is this because my thermostat time is too short.

Thanks

Hi Andrew,

Thanks, the simulation looks strange because I want to find out the
thermostating time from this fix nve setting. I have found out the
parameter for thermostating. But still the system is blowing up with Nose
hover thermostat. Is this because my thermostat time is too short.

​first of all, your time step is too small by about a factor of 100.0. you
can easily do a 1fs timestep with an all carbon system.

what you are seeing is what is generally referred to as "equilibration".
doing this with fix nve only will take a long time, since there is no
solvent in your system.
it would be a good idea to enable a dissipative thermostat, e.g. fix
langevin to speed up the equipartitioning of kinetic energy. you also need
to look at your periodic box dimensions. dumping an .xyz trajectory loses
that information. but it looks to me, as you have interactions across PBC
in z-direction, while there is plenty space in x and y. is this what you
want?

i think after about 20-50ps with fix nve + fix langevin, you should be able
to switch to fix nvt, or simply stick with fix nve.

axel.