Hello,
I am attempting to create a layer of atoms within a box that will act as a “thermostat” when I run this simulation without this layer and just use fix temp/rescale it runs perfectly, but when I only apply that to a certain number of atoms, I am continually getting the lost atom error. Here is a part of my input script where the change is being made:
Before:
fix 1 atomgroup nve
fix 2 fixed1 setforce 0.0 0.0 0.0
fix 3 all temp/rescale 1 300.0 300.0 0.05 1.0
After:
(So basically I am trying to set them all to an initial temperature of 293 but only keeping a layer of atoms as the thermostat)
velocity substrate create 293 1234567 mom yes rot yes (the other atoms in the box)
velocity thermo create 293 1234567 mom yes rot yes (the thermostat layer)
velocity fixed1 set 0 0 0 (a fixed layer)
timestep 0.05
dump 1 all atom 250 mytrydump
thermo 500
fix 1 substrate nve/limit 0.05
fix 2 thermo nvt temp 293.0 293.0 .5 drag 0.5
fix 3 fixed1 setforce 0 0 0
I would greatly appreciate advice or help on this subject. Thank you.
Hello,
I am attempting to create a layer of atoms within a box that will act as a “thermostat” when I run this simulation without this layer and just use fix temp/rescale it runs perfectly, but when I only apply that to a certain number of atoms, I am continually getting the lost atom error. Here is a part of my input script where the change is being made:
if you cannot run a stable simulation without a thermostat, either your initial geometry, or your equilibration approach, or your force field settings, or your simulation settings are bad.
as mentioned before, start with a simpler system and learn how to get stable dynamics first. do not use fix temp/rescale on production simulations.
axel.
Hello,
thank you so much for your helpful and timely response.
My potential files seem sound but my geometry may be a problem as I have a problem with over lapping atoms (below I have put the start of my input file for reference) how is this problem created?
My equilibrium approach was when I was doing the velocity commands, is that not a proper way to set equilibrium?
Will you also expand on what falls into simulation settings?
Again thank you so much
echo screen
clear
units metal
boundary p p p
atom_style atomic
#creating simulation box
region box block 0 54.31 0 90 0 54.31 units box
create_box 2 box
#creating indenter
lattice diamond 3.5668
region 1 sphere 27.155 74 27.155 7.1336 units box
create_atoms 1 region 1 units box
group cutter region 1
#creating substrate (nve)
lattice fcc 5.431
region 2 block 0 54.31 10.862 54.31 0 54.31 units box
create_atoms 2 region 2 units box
group substrate region 2
#creating fixed atoms (fixed)
lattice fcc 5.431
region fixed1 block INF INF INF 5.431 INF INF units box
create_atoms 2 region fixed1 units box
group fixed1 region fixed1
#creating thermo (nvt)
lattice fcc 5.431
region thermo block INF INF INF 10.862 INF INF units box
create_atoms 2 region thermo units box
group thermo region thermo
#assign mass of atoms and indenter
mass 1 12.0107 #diamond
mass 2 28.1 #Si
set group cutter type 1
set group substrate type 2
set group fixed1 type 2
set group thermo type 2
neighbor 2.0 bin
neigh_modify delay 0 every 1 check yes
#Force Fieilds
pair_style hybrid morse 12.0 tersoff tersoff
pair_coeff * * tersoff 1 SiC.potential C NULL
pair_coeff * * tersoff 2 SiC.potential NULL Si
pair_coeff 1 2 morse .435 46.859 .19475
delete_atoms overlap .3 fixed1 thermo
delete_atoms overlap .3 thermo substrate
#min_style cg
minimize 1.0e-4 1.0e-5 10000 10000
Hello,
thank you so much for your helpful and timely response.
My potential files seem sound but my geometry may be a problem as I have a problem with over lapping atoms (below I have put the start of my input file for reference) how is this problem created?
this is your task to figure out. look at the commands, see what they are doing individually, and then it should become obvious. if you cannot figure it out, simplify and try again.
My equilibrium approach was when I was doing the velocity commands, is that not a proper way to set equilibrium?
i have no time to be your tutor. this is not what is meant with equilibration. please discuss with your adviser/tutor. the mailing list is not a place to get this kind of oversight and help.
Will you also expand on what falls into simulation settings?
everything in your input file, whether it is explicitly set or implicit as default setting. same as before, it is not the purpose of this mailing list to teach you how to do MD simulations. that is something you have to learn from your adviser/tutor/colleagues. a mailing list and e-mails in general are a very inefficient way to learn. we can give you pointers, but the rest needs to be done with personal oversight and person-to-person discussions. you have received already more than a fair share of help and pointers. at this point, you should pick up the slack and figure out how to do your work on your own.
axel.