Dear all,
I met some questions about fix langevin when I used the nemd method to compute the thermal conductivity of a polymer system.I used the command:
fix hot hotatom langevin {tHot} {tHot} {tDamp} 698043 tally yes zero yes and
fix cold coldatom langevin {tCold} {tCold} {tDamp} 698043 tally yes zero yes
to control the temperature at the two ends of the system,and used fix 1 all nve command to perform time integration to the whole system.
When i checked the output,i found a peculiar thing that thought f_hot+f_cold is not a constant,the total energy was conserved. And here i list some output informations:
Step c_temphigh c_templow Temp Press KinEng PotEng TotEng f_hot f_cold v_tdiff
0 298.85936 294.09574 221.89138 170.13883 4232.4014 -4581.1087 -348.70731 -0 -0 0
1000 340.79683 267.87193 221.39554 -70.535765 4222.9437 -4571.8651 -348.92142 -135.71943 116.02388 -81.982755
2000 315.14474 272.55546 223.06478 -250.81996 4254.783 -4603.5599 -348.77681 -159.34009 125.70528 -87.411916
3000 351.50841 248.06747 222.909 5.3265744 4251.8116 -4601.0364 -349.22481 -214.81823 166.30118 -90.976747
4000 351.1287 246.61775 222.32826 64.183234 4240.7345 -4589.4403 -348.7058 -230.24551 213.39552 -89.296911
5000 355.22314 246.3202 219.83776 -575.83652 4193.2301 -4541.8974 -348.66732 -226.39558 210.18953 -91.084328
I think just as what NVE ensemble means, the system did not exchange energy to the environment , so only the fix langevin command could affect the total energy of the system.But the total energy is not changed even thought the accumulative energy of f_cold not equals to f_hot. So how does the system manage to conserve it’s total energy? Are there any other ways to add or extract energy from the system?
Best regards,
XiaonengRan