Dear all,
I am running NVT simulation (using LAMMPS version Dec 2013) to compute diffusion of CH4 molecules in a porous medium. I have used the following commands to output X, Y, Z coordinates of the fluid molecules with their corresponding image flags in order to be able to unwrap the coordinates and to calculate MSD. In the following "sorbate” group consists of methane molecules, which are modelled as rigid 5-site bodies.
fix 1 sorbate rigid/nvt molecule temp 450.0 450.0 10.0
fix 2 sorbate ave/time 1 100 100 f_1[1] f_1[2] f_1[3] f_1[13] f_1[14] f_1[15] file msd.data mode vector
However, the image flags do not seem to be correct in the “msd.data" file , as there are non-integer numbers among them! In the same simulation I can output image flags using dump command as mentioned below, where all image flags are integer as expected:
dump 3 all custom 100 unscaled.lammpstrj id mol type x y z ix iy iz
Visual inspection of the system also doesn’t reveal any inconsistency.
I would appreciate any helpful comments in advance.
Best regards,
Amir
Dear all,
I am running NVT simulation (using LAMMPS version Dec 2013) to compute
diffusion of CH4 molecules in a porous medium. I have used the following
commands to output X, Y, Z coordinates of the fluid molecules with their
corresponding image flags in order to be able to unwrap the coordinates and
to calculate MSD. In the following "sorbate” group consists of methane
molecules, which are modelled as rigid 5-site bodies.
fix 1 sorbate rigid/nvt molecule temp 450.0 450.0 10.0
fix 2 sorbate ave/time 1 100 100 f_1[1] f_1[2] f_1[3] f_1[13]
f_1[14] f_1[15] file msd.data mode vector
However, the image flags do not seem to be correct in the “msd.data"
file , as there are non-integer numbers among them! In the same
why should there *not* be non-integer numbers, when you average the image
flags over 100 steps. some of those rigid bodies should have moved across a
periodic boundary. i don't understand what the purpose of your fix ave/time
command is. it most certainly has nothing to do with computing the MSD.
axel.
simulation I can output image flags using dump command as mentioned below,
where all image flags are integer as expected:
dump 3 all custom 100 unscaled.lammpstrj id mol type x y z ix iy iz
Visual inspection of the system also doesn’t reveal any inconsistency.
I would appreciate any helpful comments in advance.
Thank you Axel.
I guess I forgot that image flags are also averaged over time! You are right.
Btw, I only use fix ave/time to output the data. MSD will be calculated in a post processing step from the data stored in “msd.dat” file.
Best regards,
Amir