[lammps-users] nvt and relax time

Dear all

i have had a puzzle on relax time.now i want simulation the
solidification of the metal melt droplet, i used fix nvt command like
this :

units metal
.................
velocity all creat 2300 32545
.................
fix 3 all nvt 2000 300 100
......................
timestep 0.1
run 1000

Then my question is

a. In order to get the melt of metal , i set the temperature 2300
above the melt point of metal in the velocity command ,then
equilibrated in this temperature for a special time by fix nve (run
1000 timestep),is it right or valid?

b. how shoud the random seed "32545" be set? is it arbitrary ? what is
considered when it did.

c in simulation solidification process (fix 3 all nvt 2300 300
100), if i can get the temperature each timestep is a ramped value
during the run (1000) from (2300)to (300), that is we got
2298,2296,2294,2292,2290........306,304,302,300

d what is the mean of the Tdamp(100), is it relax time? what relation
is between the relax time and the timestep? some people say the relax
time should be set in a order with timestep and <= the timestep ,why?

e Is the relax time the real time? about the timestep? then in my
inout file ,the relaxtime is 100 time untis,that is 100ps, and the
timestep is 0.1ps

f. the decrease temperature rete is (Tstart -Tend)/(run * timestep)

thanks
your sincerity

Dear all

i have had a puzzle on relax time.now i want simulation the
solidification of the metal melt droplet, i used fix nvt command like
this :

units metal
.................
velocity all creat 2300 32545
.................
fix 3 all nvt 2000 300 100
......................
timestep 0.1
run 1000

Then my question is

a. In order to get the melt of metal , i set the temperature 2300
above the melt point of metal in the velocity command ,then
equilibrated in this temperature for a special time by fix nve (run
1000 timestep),is it right or valid?

correct

b. how shoud the random seed "32545" be set? is it arbitrary ? what is
considered when it did.

arbitrary

c in simulation solidification process (fix 3 all nvt 2300 300
100), if i can get the temperature each timestep is a ramped value
during the run (1000) from (2300)to (300), that is we got
2298,2296,2294,2292,2290........306,304,302,300

this looks correct

d what is the mean of the Tdamp(100), is it relax time? what relation
is between the relax time and the timestep? some people say the relax
time should be set in a order with timestep and <= the timestep ,why?

the doc for NVT describes what Tdamp is

e Is the relax time the real time? about the timestep? then in my
inout file ,the relaxtime is 100 time untis,that is 100ps, and the
timestep is 0.1ps

Tdamp is real time (for whatever units you are in), not timesteps. You
want to set it so your simulation equilibrates after many timesteps, not 1 or 2.

f. the decrease temperature rete is (Tstart -Tend)/(run * timestep)

You can see the formula in fix_nvt.cpp for the target temperature - it;s
just a linearly ramped value between T1 and T2

thanks
your sincerity
--
Road

Steve