[gmx-users] Potential energy positive. Can be?

[Mark Abraham]

Eudes Fileti wrote:
> Hi gmx users,
> I have tried to simulate a box with 1000 molecules of toluene
> and I always have obtained a positive value for the potential energy.
> I have used three different models of potential (with and without 14 
> interactions)
> and all of them provides a positive value for energy. I must be falling in
> some error or making some wrong analysis but I do not know what.
> Below I put the average values that I got from the simulations that I made.
> If somebody will be able to help me, I will be grateful!

How are you equilibrating these structures? For how long are you doing so?

> Two questions about these results:
> # Why are so larger the values for LJ-14?

Because they're from atoms near each other. You don't just turn them on 
and off to suit the phase of the moon, you need to choose whether 
they're included according to how the force field was developed and is 
intended to be used. Search GROMACS manual for OPLS to see what is 
recommended, and/or read the original OPLS paper (reference in the 
GROMACS manual).

> # How can I compose the potential energy from the components?
> Would not be U = LJ + Coul + bond + ang + dihed ? And LJ-14?

You add them up. Every energy term that isn't kinetic or total is a 
potential energy.

Mark