Constrained MD using andersen thermostat

[sutapa.dutta]

I am trying to perform constrained Md simulation of a single water molecule using andersen thermostat by fixing bond length and bond angle of the molecule. Initially the O atom was at the corner ofthe box. During the simulation it is not leaving the corner. It seems that the O atom is not moving properly. Following are the details of my input files. What modification I need to do to run MD in a right way?

POSCAR
H2O
0.52918 ! scaling parameter
20 0 0
0 20 0
0 0 20
O H
1 2
cart
0.00 0.00 0.00
1.10 -1.43 0.00
1.10 1.43 0.00

INCAR

ALGO = V
MAXMIX = 40
IBRION = 0
NSW = 1000000
NBLOCK = 100
POTIM = 3.0
TEBEG = 0; TEEND = 0
ISYM = 0
MDALGO = 1
ANDERSEN PROB = 0.0
LREAL =.TRUE.
LCHARG =.FALSE.
NELMIN = 4
PREC = LOW
ISTART = 0
ISMEAR = 0; SIGMA=0.1

[admin]

1. You need to define temperature (e.g. TEBEG=300). At 0K (TEBEG=0) atomic positions are
practically frozen.
2. The vibrational frequency of O-H is app. 4000cm-1, i.e one oscillation takes app. 9 fs.
To have a decent description of the dynamics you need at least 10 steps per the vibration amplitude.
This means your timestep must be 0.9 fs, or smaller (e.g. POTIM=0.5).
3. In your constrained dynamics you are probably interested in details of the sampling.
Writing of the detailed report is switched on using LBLUEOUT=.TRUE.
This produces the file REPORT.