My problem is to find a way to add some time-dependent value to wall atoms’ velocity while their temperature is controlled by another fix such as fix nvt or temp/berendsen+fix nve.
In fact I want to have an oscillating heat source wall.
I tried “velocity set” command with “sum yes” keyword but the error message says that you can not add a time-dependent value. On the other hand I can not use “fix move” command because I want to control the temperature with another appropriate fix command.
My problem is to find a way to add some time-dependent value to wall atoms'
velocity while their temperature is controlled by another fix such as fix
nvt or temp/berendsen+fix nve.
In fact I want to have an oscillating heat source wall.
I tried "velocity set" command with "sum yes" keyword but the error message
says that you can not add a time-dependent value. On the other hand I can
not use "fix move" command because I want to control the temperature with
another appropriate fix command.
what frequency and amplitude of oscillation of the wall are we talking
about here?
Thanks for your response.
I did not get your purpose of asking this question.
Though, this is variable. You suppose a period of 2 ps and an
amplitude of 2 angstroms
Thanks for your response.
I did not get your purpose of asking this question.
Though, this is variable. You suppose a period of 2 ps and an
amplitude of 2 angstroms
the purpose of that question is to identify the time scales of the
thermalization and the wall motion. if they differ sufficiently, there
is no need to control both.
with a 2ps period with a 2 angstrom oscillation, for example, it seems
to me, that the energy asserted by the wall motion itself will
dominate, so requiring that the wall is thermostatted as well, would
not really make that much of a difference. you can simply quantify
this by setting up a simulation of a immobile, but thermostatted wall
with an equilibrated medium next to it, but then running the wall
adjacent medium with fix nve and observe the energy transferred and
compare that result to the same setup with a wall that is controlled
by fix move.
i am also wondering what kind of process this is modeling. after all,
a 500GHz vibration corresponds to lower infrared and the average
velocity of the wall approaches the speed of sound in dense media.
More generally, you can define a wall consisting
of atoms which move in a prescribed oscillating
manner via fix move. But make those atoms
be type A atoms which are “invisible” to the fluid.
Then define harmonic bonds between one type B atom
and one type A atom with bond-length 0.0.
B atoms do influence the fluid.
And you can thermostat the B atoms, e.g. via
fix langevin. The bond strength K will determine
how far the mobile B atoms move from
their associated A sites.