Hi Aksel,
My responses below (note that I don’t want to get wrapped up in someone else’s research, so unless you have additional LAMMPS related Q’s, which most of these are not, this will probably be my last answer)
-
From your script, these are related since you used compute reduce sum - and should give the same value if they are normalized correctly. I suggest reading up on the compute stress/atom and pressure documentation - and search past postings if need be. Looking at the dimensions of the two quantities would also give an answer. There has been a lot of chatter on this subject on the mail list, so you might also look there. You may find that stress/atom (without compute reduce and proper normalization) may be more useful as you continue your studies.
-
Either give the same if normalized correctly - ideally you would make use of the whole stress or pressure tensor rather than a single component. If you don’t know what that is do some reading in any continuum mechanics book - I used Malvern’s book back in the day, though there are probably better/more accessible ones out there.
-
Note that Hookean contacts are technically impossible for linear elastic spheres. However, they are usually good enough to give physically accurate/reasonable results in granular flows. There are some methods for estimating Hookean model constants depending on what you want to keep constant, e.g. characteristic overlap or energy loss from a collision. That’s for you to decide.
I would not focus on this point however. The discrete elements in your simulation should not be confused with real spheres, they are a coarse-grained elements that you hope will act like an FEA. As such you will also find that the Poisson ratio for these spheres is unlikely to be the same as the macroscale Poisson’s ratio that you obtain from your simulations. Though the relationship between pair force coefficients and observed constitutive relationship may be of interest.
HTH.