How to compute spatial distribution of pressure

LAMMPS users / developers,

How to correctly compute the spatial distribution of pressure for a system in which the particles do not fully occupy the domain? For example, like in the attached figure.

I just read the section in the Manual “compute stress/atom”, and it seems that the key issue is how to define a per atom volume.

So I’m wondering if there is practical workaround to get it.

Thanks in advance for sharing any comment and experience.

liquiddrop_atoms4.png

There is no simple universal answer to that.

You can look at compute voronoi/atom, which

is one solution.

Steve