Per-type masses

[Dorian_Thompson]

Hello,

So with the molecule file, I found the “unknown section”, as it was the lines of comments after the script; there were some #s missing so it ruined the file.

Now, I got it to run and am getting the error message that not all per-type masses are set, but I am fairly certain I set all the masses. Below is the input script.

#Variables

#variable fname index data.UIO-66-2.ddec

#variable latparam equal

# Initialization

log log.UIO-66 append

units real

boundary p p p

atom_style full

# Dreiding Potential Info

neighbor 2.0 bin

neigh_modify every 1 delay 0 check yes

bond_style harmonic

angle_style hybrid cosine/periodic fourier

dihedral_style harmonic

pair_style lj/cut/coul/long 12.500

improper_style fourier

kspace_style ewald 1e-5

special_bonds lj/coul 0 0 1.0

dielectric 1

pair_modify tail yes mix arithmetic

read_data data.UIO-66-2.ddec extra/atom/types 12 extra/bond/types 11 extra/angle/types 19 extra/dihedral/types 6

#Adding in DMMP

#lattice fcc 0.8

#region box_DMMP block 0 41.4008 0 41.4008 0 41.4008

molecule DMMP data.dmmpMolecule offset 5 7 14 3 1

#group MyDMMP DMMP

#fix 1 all deposit 1 0 1 9123654 region box_DMMP near 2 attempt 100 mol DMMP

create_atoms 0 single 10 10 10 mol DMMP 754901 units box

mass 6 15.0345

mass 7 30.97699

mass 8 15.9994

mass 9 15.9994

mass 10 15.9994

mass 11 15.0345

mass 12 15.0345

#Simulation

velocity all create 500.0 1234

timestep 1

dump 1 all custom 100 DUMP_F.lammpstrj id mol type x y z ix iy iz

fix 2 all nve temp 500.0 500.0 100.0

fix 3 all langevin 500.0 500.0 10.0 853910

thermo 100

thermo_style custom step temp pe lx ly lz press pxx pyy pzz

run 100000

print “All done”

Sorry if I am bothering you alot, I am just really tired of hitting these brick walls and learning this program is a challenge. I greatly appreciate the help you are giving!

Thanks,
Dorian Thompson
University of Pittsburgh
Department of Chemical & Petroleum Engineering
Class of 2019

[akohlmey]

you ask for 12 extra atom types in your read data command, but you
only set 6 masses in your input. that cannot work.

axel.

[akohlmey]

Hello,

So with the molecule file, I found the “unknown section”, as it was the lines of comments after the script; there were some #s missing so it ruined the file.

so PEPCAK as suspected. it still won't hurt to update your LAMMPS
version, as your version has known issues, specifically related to
using molecule files.

Now, I got it to run and am getting the error message that not all per-type masses are set, but I am fairly certain I set all the masses. Below is the input script.

you are making the mistake of many beginners in that you are
overconfident in doing things correctly. there are many more ways to
mess up things, than you may imagine. i am answering to people using
different simulation software packages for over 15 years now and this
is a very common pattern among beginners and people with a background
in engineering seem to be particularly impatient and vulnerable.
(remember, i have been a beginner myself, so i've been through the
same process). for a software as flexible and complex as LAMMPS it is
easy to overlook details. also, there cannot be many "safety nets" as
something that may be "wrong" in one context, could be "correct" in
another.

[...]

Sorry if I am bothering you alot, I am just really tired of hitting these brick walls and learning this program is a challenge.

part of the problem is that you don't take the time to build your
input gradually and thus learn things at a more manageable pace.
molecule files are definitely an advanced topic in LAMMPS and
something that hasn't been planned from the beginning, so there are
more rough edges to expect and more possibilities for inconsistent
behavior, so you should be *much* more paranoid about what you are
doing and look for more ways to build your input incrementally (e.g.
start with something simpler, e.g. without depositing, then do a
simple system without molecules, then use molecule files with only a
single atom, and so on). also, you should look for ways to confirm,
that what you are doing is actually happening. for example, the info
command can be extremely helpful in debugging inputs and verifying,
that what you think is happening does happen. similarly, using
write_dump or write_data to get a snapshot of the current system and
its properties and then inspect or visualize the files can be
extremely helpful. rather than getting frustrated with mistakes, you
have to embrace them as a way to learn. any time you are running your
head against the wall is a sign, that you have not yet properly
understood what you are trying to do. the trick with errors is to try
and make them in a way, that it is easier to figure out what the
mistake is, and that is best done by breaking down problems in smaller
pieces and solve them incrementally.

axel.