Dear developers,
in the documentation it is mentioned that the Tsuji damping model is not compatible with cohesive normal models such as JKR or DMT . What does ‘not compatible’ mean there?
I understand that using DMT plus a Tsuji-like term, due to cohesion the actual restitution coefficient will not correspond to the one specfied as an input parameter to tsuji’s damping. Is that the reason for non compatibility?
I’m asking that because one could be interested in using Tsuji’s scaling, even after having understood that the specified restitution coefficient does not correspond to the actual one.
Thank you,
Kind regards
Riccardo
That is a question for an expert in DEM. So perhaps @jtclemm can help out.
I am speculating here, but it could just be a technical one, i.e. that it has not been implemented.
We just released a new version of LAMMPS Release Feature release 28 March 2023 · lammps/lammps · GitHub
that contains a significant refactor of the granular pair style, which is supposed to make things more modular and consistent. Perhaps that allows combinations of models previously not available.
Hello Riccardo, I believe you are correct that the only reason mixing Tsuji damping with cohesive normal models is not allowed in PairGranular is the fact that Tsuji’s calculation of the restitution coefficient doesn’t hold for non-Hertzian normal forces.
If one understands the mathematical inconsistency, there’s no reason it wouldn’t work in LAMMPS if the error was turned off. This is pretty trivial to do in the latest development version of LAMMPS. There’s just a flag that needs to be set: https://github.com/lammps/lammps/blob/35121a2ed8a3621fed6cba7b93919caec64bf30b/src/GRANULAR/gran_sub_mod_damping.cpp#L102.