Dear All,
Do you have any suggestion (or example) of how to create a good rough wall (made by particles)?
I have start following the LAMMPS manual: Create a square lattice (group 1) and particles (group 2) and read it via read-data command.
The the particles of the wall interacting itself via Lennard-Jones potential ( is this a good choose?) and with a weak repulsive potential with particles of group 2.
The wall can be move freely on the x-direction, but I’ve used to set the force constant on y and z direction by { setforce NULL 0.0 0.0 } .
Is this the better way to do the rough wall? For the other walls could be reflective or Periodic boundary condition - but when is reflective the wall falls.
Thanks in advance,
Márcio Sampaio
Dear All,
Do you have any suggestion (or example) of how to create a good rough wall (made by particles)?
there is too little specific information here to give specific advice. below are some questions that come into my mind.
what makes a rough wall “good”?
what degree of roughness do you need?
what kind of wall material do you want to study at what resolution?
what kind of material should the wall interact with? using what kind of potential?
i also recall a previous discussion about this topic, so a search through the mailing list archive may provide additional pointers. apparently, there are algorithms to construct rough surfaces in the published literature. those could be used to construct a suitable data file. another option would be to generate a block of particles at the proper (ideal) lattice positions and then define a region that leaves 1-2 outmost layers of atom untouched and apply delete_atoms with the porosity keyword on that region to remove atoms at random positions (> 50%) followed by minimization and possibly some annealing (again leaving the outmost atom layers unchanged through using fix setforce 0 0 0 on them).
I have start following the LAMMPS manual: Create a square lattice (group 1) and particles (group 2) and read it via read-data command.
The the particles of the wall interacting itself via Lennard-Jones potential ( is this a good choose?) and with a weak repulsive potential with particles of group 2.
The wall can be move freely on the x-direction, but I’ve used to set the force constant on y and z direction by { setforce NULL 0.0 0.0 } .
i don’t think that this is very physical. i prefer having 1-2 “carrier” layer of crystal lattice atoms plus add-ons atoms, where the “carrier” is either immobile or moves as a whole block using fix aveforce, and the “rough” atoms can move freely with fix nve. but that is just my personal preference, which worked for cases where i needed this. there are likely other options and approaches that are better suited for other purposes. there is likely a lot to be learned from descriptions in the published literature.
axel.