About the estimation of the error in the calculation of thermal conductivity by the Green-Kubo method

Dear LAMMPS community,

I am doing thermal conductivity calculations for a polymeric system, I have obtained the HCACF and it decays to zero satisfactorily, now I want to know how to estimate the error associated with the thermal conductivity calculation (Δk), does anyone have experience on this? thanks for your time.

Hi @julianslipk,

This is a methodology problem rather independent of the use of LAMMPS. Do not hesitate to discuss this with people around you.

There are issues to be careful with when using GK methods, mainly convergence. So there are numerous papers in the literature dealing with error estimations of the measured values. A small search made this one pop up in my browser but this is not the first I see. It might be a first lead to help you find something that suits your needs.

As a rule of thumb, you can also make several independent simulations and compute an error-bar as with any measurement.

1 Like