[lammps-users] Some modification for compute heat/flux


I modified the compute heat/flux command, and now it can output the autocorrelation file “ac.dat” in dimensionless units and thermal conductivity file “tc.dat” in dimensional units, no need to again deal with the flux.log file to get the autocorrelation and thermal conductivity. And i also add three parameters in the compute heat/flux command, which is mbig, ntimestep/nsteps, and tc_ntimestep_start(valid for ntimestep)
mbig is the parameter M in the discrete expression for green-kubo of thermal conductivity; ntimestep means output a thermal conductivity file every 100000 steps; nsteps means output only one thermal condutivity file at the end of the simulation; tc_ntimestep_start is only for ntimestep instead of nsteps, which means the value from which step the thermal conductivity starts output, whose units is 10000 timesteps.

In the attached the modificated file is there.

|

compute_heat_flux.cpp (9.99 KB)

compute_heat_flux.h (1.11 KB)

atom.h (5.99 KB)

run.cpp (6.85 KB)

Thanks - I'll let Reese take a look at these changes. The
changes you made to run.cpp (and atom.h) are not appropriate for
that file - so you and Reese will have to see if that
can be done in the compute heat flux or someplace
else.

Steve