Changing simulation size causing issues

I am trying to simulate two graphene sheets (spaced a certain distance apart) in the vicinity of a nanotube and water. I am trying to understand how surface asperities affect heat transfer in water (with the nanotubes representing surface asperities). In order to do this, I first simulated a smaller box size (graphene sheets 4 nm by 4 nm, spaced 4 nm apart) than the one that I will be working with for the actual run (8 nm by 8 nm, spaced 8 nm apart) in order to make sure that all of my parameters were correct and the system was set up correctly. These simulations ran, and, after equilibration to 300 K and 1 atmosphere, gave me a density of 1.03 g/cm^3, fairly close to the actual density of water in that situation. So I assumed that the system was working, and the parameters were correct and proceeded to perform the actual simulation. Unfortunately, even though I used the same code to generate both sets of input files (with the only change being the change in size of the box volume by a factor of 8, and a corresponding increase in the number of water molecules by a factor of 8, from 1000 to 8000), and used the same LAMMPS input script, the new set of structures does not make it past the first step of equilibration. Specifically, during the first step of equilibration, I get nan pressure even though the last step of minimization has a pressure of -455 atmospheres. I wrote my own code to solvate the structure file once I had them, and I ensured that all oxygen atoms in the water molecules were placed at least 3 Angstroms away from any other atom in the system. I have the old (test case, named gCNT10,10_l2_2by2_water.
lammps) structure file (unfortunately, the new structure file is over 5 MB, and, as such, the user list will not let me attach it), the input script used to run both systems (named nve.lammps), and the log file for the new system (named log.lammps). Any input into this issue would be greatly appreciated.

gCNT10,10_l2_2by2_water.lammps (534 Bytes)

log.lammps (23.2 KB)

nve.lammps (1.54 KB)

I get nan pressure

If you get this on step 0, it means your forces are bad,
i.e. infinite, typically due to high overlaps in your initial
configuration. Fix your input and the problem should go away.

Steve