Hi all,
I am attempting to use the GK auto-correlation method to calculate the thermal conductivity for my system of SPCE vapor water molecules. It works well with large volumes but as I try to reduce the volume of which I am analyzing the calculation becomes less accurate. I believe this is expected as when the volume decreases the number of molecules within the space decreases as well, thus the given molecules do not create a good PDF in state space representative of the desired bulk temperature and pressure. Of the other methods in LAMMPS used to calculate k, are there any which have less of a dependence on volume? Or is this unavoidable because of the low density of molecules within the volume.
Regards,
Peter
I think you have to just try the other methods. See the examples/KAPPA
directory for toy scripts that illustrate the other methods.
Steve