Dear Lammps Users,
I have constructed a minimal working example (MWE) to illustrate the difficulty that I have towards comprehending these boundaries.
To describe the problem:
I have a simulation box that is a cube of dimension 30 (LJ units). I create a random set of 28 atoms in the lower half of the cube (the dimensions of the lower half 0 ,30, 0, 30, 0, 15).
The boundary conditions are of the form p p s. It is shrink wrapped on the z axis, periodic along x, y.
I move the atoms with a constant x velocity of 1.0 (LJ units)
Timestep is 0.0005 and I step through 20000 timesteps
When I focus on atom at (14.7725,1.4443,2.3487) this atom’s:
x coordinate moves from X=14.77 to X= 22.37 and shifts to X = 10.0 and increases
y coordinate jumps from Y=1.44 to Y=22.89 when this shift occurs and otherwise remains constant
z coordinate jumps from Z=2.349 to Z = 8.971 when this shift occurs and otherwise remains constant
When I focus on atom at ( 7.3471,2.6826,13.2146) its x component uniformly increases till 17.347 and its other coordinates remain constant.
This behavior is quite confusing as I was always under the impression that if the atom moves along the x axis that needs to hit the edge before it gets remapped and the y/z coordinate never changes. Can anyone please point me to the right direction? I have pasted the input file of MWE (my zipped attachment did not get through to the list I suspect).
Choice of units - relative units
3 dimensional problem
We are simulating the interface
so it is periodic on two sides and not on the third side
boundary p p s
Type of atoms involved
atom_modify map array
region sim block 0.000 30.000 0.000 30.000 0.000 30.000
region lower block 0.000 30.000 0.000 30.000 0.000 15.000
create simulation box
create_box 1 sim
Fill up atoms
create_atoms 1 random 28 10231 lower
mass 1 1.0
Neighbor set up
neigh_modify delay 1 check yes
pair_style lj/cut 2.250
pair_coeff 1 1 1 1
fix 1 all move linear 1.0 0.0 0.0
dump 101 all custom 1 config.dat id x y z