For fixes that calculate a contribution to the potential energy of the system temp/rescale

Christopher_Paredes · July 1, 2015, 7:12am

Hello,
Do not understand how this expression:
"For fixes that calculate a contribution to the potential energy of the system " ( fix_modify )
is related to the change in energy that makes the rescaling of velocities fix temp / rescale . Why the velocity rescaling contributes to the energy potential of the system ? .

I want to simulate the argon rahman 1964 , using a rescaling of temperature, however , I worry that the energy change introduced by the thermostat directly affect the results of the simulation, for that reason is that monitoring.

In my code I have the following lines:

fix control_temp all temp/rescale 10 94.4 94.4 0.05 0.1
fix integration all nve

besides the changes of cumulative energy of the temp / rescale I show with the following line

thermo_style custom step temp etotal ke pe f_control_temp

Regards

akohlmey · July 1, 2015, 10:22am

Hello,
Do not understand how this expression:
"For fixes that calculate a contribution to the potential energy of the
system " ( fix_modify )
is related to the change in energy that makes the rescaling of velocities
fix temp / rescale .

there is no relation.

Why the velocity rescaling contributes to the energy
potential of the system ? .

where does it say so. fix temp/rescale changes the kinetic energy.

I want to simulate the argon rahman 1964 , using a rescaling of temperature,
however , I worry that the energy change introduced by the thermostat
directly affect the results of the simulation, for that reason is that
monitoring.

of course, it does affect the simulation. quite significantly. every
time rescaling happens, it is as if every atom in your system is given
a kick. if you want to reproduce those very old simulations, just run
with fix nve alone.

In my code I have the following lines:

fix control_temp all temp/rescale 10 94.4 94.4 0.05 0.1
fix integration all nve

besides the changes of cumulative energy of the temp / rescale I show with
the following line

thermo_style custom step temp etotal ke pe f_control_temp

yes, but what does this have to do with fix_modify energy?

axel.

Christopher_Paredes · July 1, 2015, 3:47pm

> Hello,
> Do not understand how this expression:
> "For fixes that calculate a contribution to the potential energy of the
> system " ( fix_modify )
> is related to the change in energy that makes the rescaling of velocities
> fix temp / rescale .

there is no relation.

> Why the velocity rescaling contributes to the energy
> potential of the system ? .

where does it say so. fix temp/rescale changes the kinetic energy.

Reach that conclusion when observing the output of thermo style, when I use
fix_modify energy .
With fix control_temp all temp/rescale 1 94.4 94.4 2 1, the following
calculation:

*- When using the " fix_modify FIX-ID energy yes" command I get the
following output:*

     Step Temp TotEng KinEng PotEng
   f_control_temp
       0 120 -6.892267e-18 2.1447023e-18 -9.0369694e-18
   0
       1 94.4 -6.8922733e-18 1.6871658e-18 -8.5794391e-18
4.5561603e-19
       2 94.121879 -6.8922855e-18 1.6821951e-18 -8.5744806e-18
4.5561603e-19
       3 93.663783 -6.892305e-18 1.6740078e-18 -8.5663128e-18
4.5561603e-19
       4 93.013166 -6.8923318e-18 1.6623796e-18 -8.5547114e-18
4.5561603e-19
       5 94.4 -6.8923652e-18 1.6871658e-18 -8.579531e-18
4.1545742e-19
       6 93.294519 -6.8924064e-18 1.6674081e-18 -8.5598145e-18
4.1545742e-19
       7 94.4 -6.8924526e-18 1.6871658e-18 -8.5796185e-18
3.7126644e-19
       8 92.719102 -6.8925045e-18 1.657124e-18 -8.5496284e-18
3.7126644e-19
       9 94.4 -6.892655e-18 1.6871658e-18 -8.5798208e-18
3.053427e-19
      10 94.4 -6.8923995e-18 1.6871658e-18 -8.5795653e-18
2.6220489e-19

The energy changes ( due to thermostat ) increase the potential energy of
the system ,* maintaining a constant energy(aprox.)*.
What does the idea of adding the change of energy ( due to the
thermostat ) , to the potential of the system ?.

It is to maintain constant energy? this option has any effect on my
simulation or just to display information on the screen ?

I'm confused

*- on the other hand " FIX- ID fix_modify energy no" shows the following:*

     Step Temp TotEng KinEng PotEng
   f_control_temp
       0 120 -6.892267e-18 2.1447023e-18 -9.0369694e-18
     0
       1 94.4 -7.3478893e-18 1.6871658e-18 -9.0350552e-18
4.5561603e-19
       2 94.121879 -7.3479015e-18 1.6821951e-18 -9.0300966e-18
4.5561603e-19
       3 93.663783 -7.3479211e-18 1.6740078e-18 -9.0219288e-18
4.5561603e-19
       4 93.013166 -7.3479478e-18 1.6623796e-18 -9.0103274e-18
4.5561603e-19
       5 94.4 -7.3078226e-18 1.6871658e-18 -8.9949885e-18
4.1545742e-19
       6 93.294519 -7.3078638e-18 1.6674081e-18 -8.9752719e-18
4.1545742e-19
       7 94.4 -7.2637191e-18 1.6871658e-18 -8.9508849e-18
3.7126644e-19
       8 92.719102 -7.2637709e-18 1.657124e-18 -8.9208949e-18
3.7126644e-19
       9 94.4 -7.1979977e-18 1.6871658e-18 -8.8851635e-18
3.053427e-19
      10 94.4 -7.1546044e-18 1.6871658e-18 -8.8417702e-18
2.6220489e-19

in this case the energy of the system varies

> I want to simulate the argon rahman 1964 , using a rescaling of
temperature,
> however , I worry that the energy change introduced by the thermostat
> directly affect the results of the simulation, for that reason is that
> monitoring.

of course, it does affect the simulation. quite significantly. every
time rescaling happens, it is as if every atom in your system is given
a kick. if you want to reproduce those very old simulations, just run
with fix nve alone.

If I run only NVE , as you would expect the temperature fluctuates a lot.
Rahman (1964 ) controlled the temperature using a rescaling of velocities ,
which is why I'm using temp / rescale .

sjplimp · July 1, 2015, 4:22pm

The purpose and operation of the fix modify energy option is the

same for all the thermostats in LAMMPS. Just think of fix temp/rescale

as a simple, bad thermostat.

Any thermostat should keep the temperature roughly constant

over time, hence the kinetic energy is constant.

The thermostat fixes (including temp/rescale) have an output

you can print with thermo output which is the cummulative

energy the thermostat has added/subtracted. Note that

it may be per-atom or total, depending on the thermo output

normalization. This is simply a way to monitor the thermostat.

If something else is draining energy from the system, then

the thermostat has to continually add energy.

Fix modify energy is just a way to alter the bookkeeping and

add the thermostat energy to the potential energy. Then

the total energy of the system (kinetic + potential) should

stay (roughly) constant, e.g. if you did thermo output of etotal.
For more sophisticated thermostats,

the energy conservation can then be quite good.

Steve

Christopher_Paredes · July 1, 2015, 8:37pm

Excuse me for being so insistent. For example, consider a fictitious out system and examine the next steps of interaction

case 1:
step Temp Etotal Echange Ethermostat
0 120 1000 0 0
1 94 800 -200(drining) +200(thermostat add 200,to correct energy change due to rescaling and maintain constant energy.)

akohlmey · July 2, 2015, 2:01am

Excuse me for being so insistent. For example, consider a fictitious out
system and examine the next steps of interaction

case 1:
step Temp Etotal Echange Ethermostat
0 120 1000 0 0
1 94 800 -200(drining) +200(thermostat add 200,to
correct energy change due to rescaling and maintain constant energy.)

case 2:
step Temp Etotal Echange Ethermostat
0 94 800 0 0
1 120 1000 200 -200(thermostat sustract 200, to
correct energy change due to rescaling and maintain constant energy.)

Ethermostat is the cummulative energy the thermostat has added/subtracted.

Looking at the two fictional examples. Is this correct way I think it works
the cumulative energy of the thermostat?.

why don't you just set up a test system and see if your hypothesis is confirmed.

The purpose and operation of the fix modify energy option is the
same for all the thermostats in LAMMPS. Just think of fix temp/rescale
as a simple, bad thermostat.

Any thermostat should keep the temperature roughly constant
over time, hence the kinetic energy is constant.

The thermostat fixes (including temp/rescale) have an output
you can print with thermo output which is the cummulative
energy the thermostat has added/subtracted. Note that
it may be per-atom or total, depending on the thermo output
normalization. This is simply a way to monitor the thermostat.
If something else is draining energy from the system, then
the thermostat has to continually add energy.

Fix modify energy is just a way to alter the bookkeeping and
add the thermostat energy to the potential energy. Then
the total energy of the system (kinetic + potential) should
stay (roughly) constant, e.g. if you did thermo output of etotal.

Then with fix_modify energy potential = epair+emol+etc...+ Ethermostats ?
which is the physical justification ?, adding it to potential energy is
conserved artificially .

as steve already said, there is no physical justification. only the
practical effect hat this way etotal is the conserved quantity and you
don't have to allocate another variable.
fix_modify energy yes is most important for minimizations..

personally, i don't use it for thermostats. if i want to monitor the
conserved quantity, i do define a variable that adds the scalar output
of the thermostat fix to etotal and output that.