[gmx-users] A simple question

[Alex]

Hello gromacs user,

Out of a well converged NVT simulation, I calculated the "Total-Energy =
Utot" as total internal energy of the system which contain three pieces (A,
B and water).
Then, using "gmx trjconv" and "gmx mdrun -rerun", "gmx energy", I
calculated the "Total-Energy" of each piece as U_A, U_B and U_W, using the
same method, I calculate again the U_AB, U_AB, U_AW ... , which is the
internal energy due to interaction between two pieces in the soup.

I expected that summing over U_A, U_B and U_W, U_AB ...., give me the
"Total-Energy
= Utot" of the system, but no combination of U_A, U_B and U_W, U_AB ...
could be equal to the "Total-Energy = Utot" !and the difference is really
considerable.

Would you please let me know why so? or I am doing wrong somewhere?

Thanks.

Best regards,
Alex