variable dt

hi every one
i want to simulate cu-water.and compute thermal conductivity.i have compared results with the others.the results are good but there is a articles result is about 7/5… but my answer is 7.5e-014.thus i changed variable dt to 20e+14(it was 02) and the answer became 7.5*e-005…is changing the variable dt resonable?? does anybody know what should i do to give a resonable answer?
best wishes

variable T equal 298
variable V equal vol
variable dt equal 200000000000 # .00000000002

variable x equal 23.41
variable y equal 23.41

variable rho equal 0.6
variable t equal 20
variable rc equal 2.5

variable p equal 200 # correlation length
variable s equal 2 # sample interval
variable d equal $p*$s # dump interval

convert from LAMMPS real units to SI

variable kB equal 1.3806504e-23 # [J/K] Boltzmann
variable kCal2J equal 4186.0/6.02214e23
variable atm2Pa equal 101325.0
variable A2m equal 1.0e-10
variable fs2s equal 1.0e-15
variable convert equal {kCal2J}*{kCal2J}/{fs2s}/{A2m} #thermal conductivity
variable convert equal {atm2Pa}*{atm2Pa}{fs2s}*{A2m}{A2m}*{A2m} #*${kCal2J}

#set up problem

dimension 3
echo screen
boundary p p p

atom_style full
bond_style harmonic #hybrid harmonic
angle_style harmonic #hybrid harmonic
kspace_style pppm 1.0e-5

group hydrogen type 1
group water type 1 2
group cu type 3
group oxygen type 2

lattice fcc 3.615 #Cu lattice constant
region Cu sphere 0 0 0 3 units box
create_atoms 3 region Cu

#set group oxygen charge -1.040 #???
#set group hydrogen charge .520 #???
#set group cu charge 0.000

pair_style hybrid lj/cut/coul/long 0.1521 3.157 eam lj/cut 5 # .583 #2.8 # 3.157 # 7.5
pair_coeff 1 1 lj/cut/coul/long 0.0460 0.4000 #H-H epsilon sigm # 108.0e-21 32.0e-11
pair_coeff 1 2 lj/cut/coul/long 0.0836 1.7753 #O-H epsilon sigma
pair_coeff 1 3 lj/cut 0.6589 0.2117 #H-Cu epsilon sigma
pair_coeff 2 2 lj/cut/coul/long 0.1521 3.157 #O-O epsilon sigma # 0 0
pair_coeff 2 3 lj/cut 1.198 1.587 #O-Cu epsilon sigma
pair_coeff 3 3 eam cu.eam #Cu-Cu

for cu-cu bond sigma=.227 epsilon(Lj)=.583 ev # sigma=2.34 epsilon=9.4512 kcal/mol … cu eam cut off= 4.95 Ang

bond_coeff 1 450 0.9572 #O-H
angle_coeff 1 55 104.52 #H-O-H


neighbor 2.0 bin
neigh_modify delay 0 every 1 check yes

min_modify dmax 0.01
minimize 1.0e-8 10.e-4 100000 300000

timestep .0000000000000002 # ${dt}
thermo $d

velocity all create 298 4928459 rot yes dist gaussian #23482341

fix 1 hydrogen shake 1e-6 500 0 m 1.0 a 1 #for hydrogen
fix 12 water npt temp 298 298 100.0 iso 0.0 0.0 1000.0

---------- Relaxation -----------------------------------------

minimization : avoid atoms overlapping

#min_style fire

#thermo_style custom step etotal enthalpy pe press ke
#thermo_modify flush yes

run 1200
reset_timestep 0


#dump 1 all custom 10000 dump.equilibrium. id type x y z vx vy vz


Green-Kubo viscosity calculation

Define distinct components of symmetric traceless stress tensor

variable pxy equal pxy
variable pxz equal pxz #-press
variable pyz equal pyz

fix SS all ave/correlate $s $p $d &
v_pxy v_pxz v_pyz type auto file S0St.dat ave running

variable scale equal 1.0/$tvols*dt variable scale equal {convert}/(${kB}$T)$V*s*{dt}

variable v11 equal trap(f_SS[3]){scale} variable v22 equal trap(f_SS[4])*{scale}
variable v33 equal trap(f_SS[5])
#*******************************#thermal conductivity
compute myKE all ke/atom
compute myPE all pe/atom
compute myStress all stress/atom virial
compute flux all heat/flux myKE myPE myStress
variable Jx equal c_flux[1]/vol
variable Jy equal c_flux[2]/vol
variable Jz equal c_flux[3]/vol
fix JJ all ave/correlate $s $p d & c_flux[1] c_flux[2] c_flux[3] type auto file J0Jt.dat ave running variable scale equal {convert}/${kB}/$T/$T/$V
variable k11 equal trap(f_JJ[3])
{scale} variable k22 equal trap(f_JJ[4])*{scale}
variable k33 equal trap(f_JJ[5])

thermo_style custom step temp press v_Jx v_Jy v_Jz v_k11 v_k22 v_k33 v_pxy v_pxz v_pyz v_v11 v_v22 v_v33

run 400
variable k equal (v_k11+v_k22+v_k33)/3.0
variable ndens equal count(all)/vol
print “average conductivity: $k[W/mK] @ T K, {ndens} /A^3”

hi every one
i want to simulate cu-water.and compute thermal conductivity.i have
compared results with the others.the results are good but there is a articles result is about 7/5... but my answer is
7.5*e-014.thus i changed variable dt to 20*e+14(it was 02) and the answer
became 7.5* changing the variable dt resonable??

Of course not. Please read a MD book and try to understand what dt (time
step size) means. Time step sizes of 200000000000 and 0000000000000002 do
not make sense at all for any kind of unit style.

does anybody know what should i do to give a resonable answer?

Here are some of my last pieces of advices. There only way you are ever
going to get a reasonable answer is to follow them.

1. Throw away your input script. It is not correct at all. Just delete
it, and you will not lose anything.

2. From one of the examples directories, find an EAM potential input
script. Start with pure Cu, and make sure you can equilibrate it and
obtain good lattice properties. Make sure you use metal units.

3. Use equilibrium or non-equilibrium methods to estimate the thermal
conductivity of pure Cu. Again, find relevant input scripts to start
with. This requires running for many many MD steps, far more than your
"400" steps.

4. Switch to water. Start from LJ or real units. repeat steps 2 and 3 for

5. Then convert your water script to metal units. repeat steps 2 and 3

6. Finally, you can combine Cu and water together (under metal units). The
reason of choosing metal units is it is hard to convert EAM to other unit
styles, but relatively easier to convert lj/cut/coul/long to metal units.

Unless you follow these steps, you will never get reasonable answers and
likely no one will ever respond to you again.


Thanks for another addition to my horror cabinet. We all had a really good laugh at the scientific computing workshop we are doing right now.

I’m actually surprised people are still answering your messages given how much you have spammed this list. But at this point you need to realize that you have become a buffoon around here. C’mon, is your English that bad that you cannot notice?