FW: temperature gradient destory by SHAKE

Hi Lammps users,

I see that http://lammps.sandia.gov/doc/fix_thermal_conductivity.html (restriction part). SHAKE algorithm will alter the amount of transferred momentum. So it can only be used with flexible molecules. And I also see SHAKE method destroy temperature gradient which is established by command fix thermal/conductivity at later period of simulation. Is this another reason make it only be valid for flexible molecule?

My input files are attached.

in.gene will generate water.equil

in.test will read water.equil, and activate fix thermal/conductivity to do NEMD.

water.267 is water molecule file with 267 molecules.

At about 15000 timestep, temperature gradient is becoming unstable, and finally disappear.

All reply will be appreciated!

Thanks.

Best regards,

in.gene (1.89 KB)

in.test (2 KB)

water.267 (43.6 KB)

water.equil (124 KB)

I don't understand your Q. Your input script
is using fix shake with fix thermal/cond which
the doc pages tell you not to do. it's not working
like you want. Why would you expect a different
outcome?

Steve