Investigate and comment on huge deviation of pressure in systemtest

In `tests/system_tests/MD/LAMMPS/test_lammps_simulations.py` we have the expected results in a nice systemtest added by @TOFarmer:

```
NPT_EXPECTED = {'Atoms':(N_MOLECULES*3, 0), 'Bonds':(N_MOLECULES*2, 0),
'Angles':(N_MOLECULES, 0), 'KinEng':(384.85, 0.9),
'PotEng':(-2408.42, 6.4), 'Temp':(299.56, 0.70),
'Press':(3.87, 31.5), 'Volume':(6478.08, 30.3),
'E_bond':(0, 0), 'E_angle':(0, 0), 'E_vdwl':(455.36, 3.2),
'E_coul':(10390.32, 18.4), 'E_long':(-13254.11, 13.5),
'Nbuild':(415.84, 2.7), 'Ndanger':(0, 0)}
```

The "Press':(3.87, 31.5)" is in particular causing random failure of this systemtest, which is a pain, with values of the pressure exceeding 3.87 + 3*31.5.

I wonder if there there is a typo in "Press':(3.87, 31.5)" since 3.87 +- 31.5 implies negative pressure. @TOFarmer do you think this could be the case?

Here reconsider "Press':(3.87, 31.5)", and adjusting it to reduce the number of random build failures.