DPD Thermostat does not keep the temperature constant

Dear Lammps Users,

I am using dpd thermostat along Lennard-Jones pair_style in a simulation in which I shear a system of particles with rough walls movements. I am willing to keep my temperature constant at 0.2 (LJ units) but this does not happens and what I do get is something fluctuating around 0.188. When I do use Nose-Hover Thermostat, this is kept fluctuating around 0.2. Temperature is an important parameter in my simulation, so I need it to keep it at 0.2.

I used the dissipative force constant of 1 as recommended by Soddemann et al ( PRE, 68 2003). I am using cut-off radius of 2.254 the same as my LJ pair_style.

I wonder if anyone have similar experience with dpd thermostat and have an idea how to fix this issue? I have also attached my input script.

Best regards.

Hassani

in.shear (2.88 KB)

Dear Lammps Users,

I am using dpd thermostat along Lennard-Jones pair_style in a simulation in
which I shear a system of particles with rough walls movements. I am willing
to keep my temperature constant at 0.2 (LJ units) but this does not happens
and what I do get is something fluctuating around 0.188. When I do use
Nose-Hover Thermostat, this is kept fluctuating around 0.2. Temperature is
an important parameter in my simulation, so I need it to keep it at 0.2.

please note that an "exactly constant" temperature is only possible
with an infinitely large system. the instantaneous temperature will
*always* fluctuate. how much it fluctuates depends on the size of the
system. consider the extreme case of a single harmonic oscillator. its
temperature will fluctuate from 0K to 2xT, if T is the (average)
temperature.

I used the dissipative force constant of 1 as recommended by Soddemann et al
( PRE, 68 2003). I am using cut-off radius of 2.254 the same as my LJ
pair_style.

I wonder if anyone have similar experience with dpd thermostat and have an
idea how to fix this issue? I have also attached my input script.

this is a far too complex input to make a simple statement. all one
can say is that it looks to me that your way of determining the
temperature may be flawed. also, i don't think it is a good idea to
apply the dpd thermostat to particles that are time integrated via fix
move.

axel.