I am applying gravity to a hydrostatic problem. When I apply gravity, operating in LAMMPS “real” units, I apply the equivalent 1g acceleration of gravity as -2.31e-4 (kcal/anstrom-gram), when converted to real units, which I think is correct. My issue is that I am generating a pressure gradient that is many orders of magnitude higher than it should be. Has anyone run into this issue before? If so, can you see where I am making a conceptual error?
From your description it is not clear what kind particles you are simulating, but please note that for atomic scale simulations and the corresponding time scales, gravity forces are negligible and tiny compared to the forces from atomic interactions.
Thank you for the response. I am using fictional particles that are essentially noble gas (no angular torques or particle attractions). I am building models that are being used as analogs for much larger systems, and I am mostly doing this as a learning exercise. I just want to make sure I am using the right numbers. Perhaps I will switch to SI units, but there is no happy median as far as unit systems, since some parameters are very dependent on significant figures required for a stable solution.
Please note that units real has masses in atomic mass units (or Daltons, or g/mol). Redo your calculations and you will find that the gravitational acceleration due to Earth’s mass at its surface, in units real, is exceedingly negligible.
Thank you for that clarification on units. I will double check for atomic mass units. I am scaling up the gravity to a level where it will be measurable in these models, in order to understand the nature of its contribution to the dynamics of much larger models which are UN-modelable due to computation limitations. What I am making are tiny proxy models that I will draw inferences for behavior at true scale by projecting these behaviors to scale. The LAMMPS mathematical solver is a useful tool for me to build analogs because of its computational efficiency.
Note that LAMMPS has unit styles micro, cgs, and si for directly modelling macroscopic systems.
I will experiment with some unit changes. That should clarify things for me. Thanks for your help.