Does anyone know how to simulate 3D still water on a flat substrate by SPH ?

LAMMPS LAMMPS Mailing List Mirror
1

Hello, All.

I want to simulate 3D liquid system by SPH method.
To test SPH in LAMMPS, I simulated 3D systems consisted of water and a flat substrate, but calculations always end with error.

Does anyone know how to simulate still water on a substrate by SPH ?

Initial condition are following:
velocity of each particle = 0.
gravity = (0,0,-9.81)
The motion of substrate particles is fixed by “meso/stationary”. The thickness of the substrate is at least 4 times lattice.
Boundary is “p p p” or “p p m” or “p p f” or “p p s”.

The error messages are following:
When the boundary is “p p p” or “p p f”, ERROR: Lost atoms.
In other boundaries,
if neighbor is 4e-6 bin, ERROR on proc 0: Too many neighbor bins
if neighbor is 5e-6 bin, ERROR: Domain too large for neighbor bins

Thanks.

S. Yoshida

2

If you are losing atoms even with periodic boundaries,
it is likely b/c they are moving too far, too fast. You have
bad dynamics or need a smaller timestep. The too many
bins problem is typically b/c your domain has exploded
in size (e.g. b/c a particle flies away). Either way, you
need to visualize and dump out your atom coords and
figure out what is going wrong.

Steve