hybrid tip4p water with aribo and LJ

LAMMPS LAMMPS Mailing List Mirror
1

Hi,

I have some problem on tip4p water. The simulation always stopped at step 0 or 1. It reported error like " Shake atoms 29 69 109 missing on proc 7". I reduced timestep and distances between water molecules, but it doesn't work. Could you please help look at my input file:

units metal
atom_style full
dimension 3
boundary p p p
neighbor 2.0 bin
neigh_modify delay 5

read_data tmp.lmp

pair_style hybrid airebo 2.5 lj/cut 8.50 lj/cut/coul/long/tip4p 6 7 1 1 0.125 12.0 10.0
pair_coeff * * airebo CH.airebo C C C C C NULL NULL
pair_coeff 1*2 3*4 none
pair_coeff * 5 none
pair_coeff 1*2 3*4 lj/cut 0.00384 3.40
pair_coeff * 5 lj/cut 0.0 0.0
pair_coeff * 6*7 none
pair_coeff 6 6 lj/cut/coul/long/tip4p 0 0
pair_coeff 6 7 lj/cut/coul/long/tip4p 0 0
pair_coeff 7 7 lj/cut/coul/long/tip4p 0.16275 3.16435
pair_coeff 1*2 6 lj/cut 0.006499 3.19
pair_coeff 3*4 6 lj/cut 0.006499 3.19
pair_coeff 1*2 7 lj/cut 0 0
pair_coeff 3*4 7 lj/cut 0 0

kspace_style pppm/tip4p 0.00001
bond_style harmonic
angle_style harmonic
bond_coeff 1 80.0 0.9572
angle_coeff 1 200.0 104.52

group oatoms type 6
group hatoms type 7
group water union oatoms hatoms
group virtual type 5
group low_m type 3
group low_f type 4
group high_f type 1
group high_m type 2
group lo union low_f low_m
group hi union high_f high_m virtual
group mobile union low_m high_m
group mobileall union mobile water

displace_atoms hi move 12 0 0 units box

velocity mobile create 10 77685 dist gaussian mom yes rot yes
velocity virtual set 0.04 NULL NULL units box

#fix 12 water shake 0.0001 20 2000 b 1 a 1
fix 1 mobileall nvt temp 10 10 0.1 drag 0.2
fix 2 low_f setforce 0.0 0.0 0.0
fix 8 high_f spring couple virtual 1.0 0.001 0.0 NULL 0.0
fix 9 virtual setforce 0.0 0.0 0.0
fix 10 virtual nve
fix 7 high_f rigid single torque 1 off off off
fix 5 high_f addforce 0.0 0.0 0.0
fix 11 high_f viscous 0.0005
fix 6 low_f nve

compute 1 high_f reduce sum z
compute 2 low_m reduce sum x
compute 3 lo group/group hi
compute 4 mobile temp

timestep 0.000001
thermo 100
thermo_style custom step temp c_3[1] c_3[2] c_3[3] f_8[1] f_8[2] c_1
thermo_modify lost warn

dump 1 all atom 4000 FL1L.file
dump_modify 1 scale no

run 40000

timestep 0.002

run 600000

velocity virtual set -0.04 0.0 0.0 units box

timestep 0.002
thermo 100
thermo_style custom step temp c_3[1] c_3[2] c_3[3] f_8[1] f_8[2] c_1
thermo_modify lost warn

run 600000

Thanks a lot.

Best,

Justin

2

Hi,

I have some problem on tip4p water. The simulation always stopped at step 0
or 1. It reported error like " Shake atoms 29 69 109 missing on proc 7". I
reduced timestep and distances between water molecules, but it doesn't work.
Could you please help look at my input file:

this is a horribly complex input file.
it would require some divine intervention
to be able to tell what is wrong.

based on previous experiences, however,
my guess would be that it is not the settings
in the input that are giving you troubles, but
rather that your starting configuration has some
close contacts and the resulting large forces
generate some instability in the integration
resulting in the error.

there are two general strategies:

1) replace the shake on the water bond and angle interactions
with very stiff harmonic potentials and first do a minimization
to reduce the close contact forces

2) split your data file into pieces which will also lead to a
simpler input and then check out each piece until you find
the problematic one and then narrow down the issue.
sometimes it can be as simple as making the simulation
box a little bit too small.

axel.

3

SHAKE atoms missing - means the atoms in a water
molecule have become too far apart - that almost always
means bad dynamics.

Steve

4

Hi Axel,

Thanks a lot. I will try them.

Best,

Justin