which are the force field parametres we can try changing to get the experimental density.Currently I am doing simulation of water,started with a density of 0.5 after 500 steps nvt and ten lakh npt steps not yet reached the 1g/cm^3 .plz suggest.
You keep misusing this forum. Please have a look at this page before you post.
Concerning your question, force fields for water have been tested and compared in countless publications, and most of these publications provide the density.
Btw 500 steps, assuming that you are using a timestep of 2 fs, is 1 ps. This is not an acceptable duration for a box of water to relax.
Edit: this is my personal favorite force field for liquid water at ambient temperature and pressure. As you can see from the abstract, the density matches the experiments.
Please keep in mind that you are posting for an international audience, most of which have no clue that “ten lakh” is a 10*100000 = 1000000. Why not just say “a million”?
This is a self-inflicted problem. Going from low density to high density with an equilibrium simulation is a notoriously difficult and time consuming process. Much more than going the other direction. This critically also depends on the system size, but is also particularly prominent with long molecules (e.g. polymers) or molecules with strong directional forces (e.g. water). The explanation of this comes from statistical mechanics: as the density increases, you need “collective rearrangements” of molecules so that they can all fit together well, but these require significant energy fluctuations which are rare and their magnitude depends on the system size. As a simple macroscopic example, think of a box of different size building blocks that you want to fit densely without holes into a box. If you just put them all in the box and shake it a bit (which is more drastic than what you do with an equilibrium NPT MD run), you will rarely achieve a good and dense packing.
Taking all of this into account, a less time consuming approach would be to first “overshoot” by using elevated pressure and temperature until you reach a density beyond your expected target, and then slowly and carefully relax toward the target temperature and pressure. You will still have a likely less ordered state unless you use excessively long simulation times, same as you would have a too ordered system when starting from a perfect crystal and trying to turn it into a liquid, but since density is not a very sensitive properly, it will be close enough. The systematic error from the water potential will be larger. For more complex systems or polymers, the only way to get a dense equilibrium with a reasonable computational effort is to use a combination of MD and MC simulation, since MD alone will get “stuck” too much, a phenomenom referred to a “jamming”.