[lammps-users] error: Attempting to rescale a 0.0 temperature

Dear all:
The below is my input file,I try to simulate water in carbon tube:

units real
neigh_modify every 1 check yes

boundary p p p
atom_style full
bond_style harmonic
angle_style harmonic

read_data data.in
bond_coeff 1 450 0.9572
angle_coeff 1 55 104.52

group water type 2 3
group carbon type 1

pair_style hybrid/overlay lj/cut 12.0 lj/cut/coul/long/tip4p 2 3 1 1 0.1250 12.0 10.0 tersoff

kspace_style pppm/tip4p 1e-4
#kspace_modify slab 3.0 #caution

pair_coeff 1 2 lj/cut 0.4785 0.3275 12 # C-O
pair_coeff 1 3 lj/cut 0.15968 0.31597 12 # C-H

pair_coeff 2 2 lj/cut/coul/long/tip4p 0.16275 3.16435 12.0
pair_coeff 3 3 lj/cut/coul/long/tip4p 0 0 3.0
pair_coeff 2 3 lj/cut/coul/long/tip4p 0 0 3.0

pair_coeff * * tersoff C.tersoff C NULL NULL

#compute
compute p all stress/atom
#compute T water temp # set water’s T as system’s T

write_restart data.restart
dump 1 all custom 50 dump.tube id type x y z vx vy vz c_p[1] c_p[2] c_p[3]

#mini if neccesary

#velocity
velocity all create 277.0 1234567 units box
#fix 2 carbon setforce 0.0 0.0 0.0
fix 1 all npt 277 277 100 xyz 1 1 1000
fix 3 water shake 0.00001 50 10 t 1 a 1
timestep 0.5
run 40000

I don’t know how the error “Attempting to rescale a 0.0 temperature” occurs.Thank you all.

When does it happen? At setup or during dynamics?
Normally, the error means the temp is 0.0 and you are attempting
to rescale to a finite value, which can't be done. Rather you
need to create a new velocity distribution.

Steve