I am trying to do a shear flow NEMD simulation similar to this paper. In short, the simulation was set up with polymer melt confined between two walls, one stationary wall at the bottom and one moving wall at the top of the box. A simple shear flow was imposed by the moving upper wall and the fixing bottom wall. Furthermore, p-SLLOD algorithm was implemented in the polymer melt. In LAMMPS, I noticed that fix nvt/sllod is performed together with fix deform to achieve the NEMD simulation of shear. However, the simple shear simulation in the paper did not change the box.
My question is how to execute the shear flow using the method above with the p-SLLOD algorithm using LAMMPS.
I set up the shear simulation without the p-SLLOD algorithm. The input file and the data file should be uploaded here. All the simulation tests should be finished within 1 min with mpirun -n 4 lmp -i *.in on my AMD 5700X with WSL and LAMMPS (2 Aug 2023).
Any guidance or suggestions would be greatly appreciated!
Please note that this is different from a “true” SLLOD run, because the walls impose an orientational discontinuity on whatever they are confining. This is appropriate if you are simulating an actual nanoconfined fluid, and may introduce insurmountable error if you are trying to retrieve the bulk shear response of that fluid.
In any case, you need to introduce a temperature compute to your simulation that subtracts out the velocity profile, such as compute temp/ramp or compute temp/profile. You then need to “connect” the temperature compute to your thermostatting (and possibly integrating) fix, such as fix nvt, using the fix_modify ... temp option.
Please consult the documentation for further details.
Thank you for your valuable guidance! Your reply is so enlightening!
I am focusing on the molecular motion of polymer melts subjected to shear at solid interfaces, which could be different to the bulk shear response. Thus, the problem became quite tricky.
As far as I know, the SLLOD algorithm subtracts the “streaming” velocity of the atoms to produce the correct velocity and temperature. In my case, I need to subtract the wall effect and the drift of the center of mass of the melt to reasonably integrate the velocity to get the system temperature. I spent the last few hours reading this part of the manual and did some simulation tests(still working on it).
