Hi everyone !
I am doing molecular dynamic simulations of organic phase with extracted metallic cations. In my lab we use amber for MD. I would like calculate the viscosity of my simulated solutions but it’s not easy in amber. I saw that with lammps (in 6.21. Calculating viscosity) there are some methods. I am interesting by the one using the Green-Kubo formula. No one uses lammps in my lab so I would like to ask you some information.
So I have trajectory from amber simulations. Can I convert it in lammps format and use the script to calculate viscosity or I have to perform the entire simulation on lammps ? In this case it would be more difficult for me because I would have to convert the amber force field for using it in lammps.
We have already lammps installed in our cluster (07/30/16 version)
Thank you
Amaury Paquet
PhD student
CEA/DEN/DRCP/SMCS/LILA
amaury.paquet@…819…
00 33 4 66 79 55 17
France
Hi everyone !
I am doing molecular dynamic simulations of organic phase with extracted
metallic cations. In my lab we use amber for MD. I would like calculate the
viscosity of my simulated solutions but it’s not easy in amber. I saw that
with lammps (in 6.21. Calculating viscosity) there are some methods. I am
interesting by the one using the Green-Kubo formula. No one uses lammps in
my lab so I would like to ask you some information.
So I have trajectory from amber simulations. Can I convert it in lammps
format and use the script to calculate viscosity or I have to perform the
entire simulation on lammps ? In this case it would be more difficult for me
because I would have to convert the amber force field for using it in
lammps.
green kubo requires the computation of the autocorrelation of the
pressure tensor, which in turn requires forces and velocities.
theoretically, you could reuse an existing trajectory, but it would
have to contain velocities and you would still need to port the force
field to LAMMPS to compute the pressure tensor.
on the other hand, you could just record a "pressure tensor
trajectory" from amber (which may need inserting a few lines of code)
and then write a small script/program to compute the autocorrelation
of that and get the same result. the computation done by fix
ave/correlate in LAMMPS is not very complicated.
axel.
Thank you for your answer.
So using my trajectory from amber in lammps should be complicated for me because we use a special force field with polarization. I know better amber than lammps so I should try to get all the necessary data for pressure tensor calculation from amber. It's easy to print out velocities for each step so I should find a way to do that.
-----Message d'origine-----
Thank you for your answer.
So using my trajectory from amber in lammps should be complicated for me because we use a special force field with polarization. I know better amber than lammps so I should try to get all the necessary data for pressure tensor calculation from amber. It's easy to print out velocities for each step so I should find a way to do that.
if you have the pressure tensor, you should not need the velocities.
the kinetic contributions are usually included already. i only
mentioned velocities in case of using the rerun feature in LAMMPS,
where you would need the velocities matching the positions (and the
forces) in order to correctly reconstruct the pressure tensor
elements.
axel.