[lammps-users] 回复: 回复: atoms fly in nanocutting simulation

please always keep the list in cc.

thanks.

Dear Steve:
    Thank you very much for replying my question.
    I have tried the methods you told me. FIrst, I use
relaxation(minimization,nvt) for 2000 timesteps. And I monitor the
temperature to be about 300K, and the force on tools (thermo_style costum

300K temperature?

your logfile indicates a temperature of 40,000K!

fx) to about 0. The I add volecity on my tools, and my tool begin to move.

yes but have to computed how fast your tool would
be moving in reality with a velocity of 300 AA/ps??

that is 30,000 m/s or almost mach 90, i 90 times the
speed of sound. if you hit something _that_ hard,
what do you expect to happen?

The volume begins to shrink for the tool moving near the substrate. When the
tool touch the substrate, the atoms fly out. If I use a long
timestep(0.01fs.for instance), the atoms would fly quickly, and the program
would terminate. If I use a short timestep(0.0001fs), the atoms would fly
slow in front of the tool

the time step does not change the speed in reality,
it only changes the resolution in your trajectory.
.

   I have my log file and the picture of atom flying attached to this
letter, I didn't use relaxation in this log file because i find that no
matter i relaxed it or not , the atoms would always flying out.
  Would you please help me finding out what is wrong with my simulation.

you have a PEPCAC problem (or GI-GO if you prefer that abbreviation).
a simulation can only simulate the system that you feed it.
the image you see makes perfect sense and the simulation
seems to have run correctly. that you don't see what you expect
to see is only because you didn't put in the input that you
expected to put there.

there is not much else that you can do other than sit down,
have a cup of tea and ponder what you have done.

cheers,
     axel.

Dear Alex:
Thank you for telling the right thing to do.
I change the speed of my tool, and look closely at my simulation. I find that the flying is due to the rise of temperature too quickly and too enormously. I tried to control the temperature but cannot find the way. But when I change my units to LJ(i used metal ), the rise of temperature turns small, and the atoms would not fly. I don’t quite understand why is this, could you please helping me again? Thank you so much!
Sincerely yours
Chen

Dear Alex:
    Thank you for telling the right thing to do.
    I change the speed of my tool, and look closely at my simulation. I find
that the flying is due to the rise of temperature too quickly and too

in case you have not noticed,
the rise of temperature is *due* to the movement of your tool!

enormously. I tried to control the temperature but cannot find the way. But
when I change my units to LJ(i used metal ), the rise of temperature turns
small, and the atoms would not fly. I don't quite understand why is this,

this happens, *because* you are changing your units
and in particular the relation between the time unit the
timestep command uses and the time implied by the
mass, length and energy units.

could you please helping me again? Thank you so much!

there really is not much else that i can do. *you* have to
understand what it is exactly that you are modeling and
what the right parameters for that are. i.e. what is a proper
velocity of the tool to move? in meter per second?
if you translate this to angstrom per picosecond, you have
the value for the velocity that you need to give your tool.

these are simple physics (i.e. classical mechanis) issues
that i had to learn in middle school. perhaps you can find
your old school books and have a look at it again.

cheers,
    axel.