Hi Axel,
I understand that, thanks. I'm just wondering if there's any other reason for it. I see that I might have to modify the test so that LAMMPS can determine which side of the wall it is on. In general, it is conceivable that one might have a wall potential for selective atom types and periodic boundaries for other types, for example.
sorry, but with periodic boundaries it doesn't make
any sense unless you have a "double wall" or "barrier".
for a single wall, there is no "side". that is a principal
problem. whether the wall applies only to some atoms
and not others doesn't matter.
since LAMMPS cannot know without going through a
lot of hoops whether an acceptable "double wall" scenario
exists, it is not supported. i would recommend to create
a new fix "barrier" that would combine two walls with
a suitable distance between them.
axel.
axel.