question about lammps

Hello
I am working on a project about age hardening. i want to quench my system. it means that the temperature of my system muss decrease in short period of time. so I wrote a following code for this part of my project :

#…Quench…#
units metal
dimension 3
boundary p p p
atom_style atomic
neighbor 2 bin

read_data data3

mass 1 63.546
mass 2 107.868

pair_style eam/alloy
pair_coeff * * cuag.eam.alloy.alloy Cu Ag

timestep 1
thermo 1
thermo_style custom step temp
dump 1 all atom 1000 mydump2.in
fix 1 all nve
fix 2 all heat 1 -0.000000000823
write_data data4
run 2

even though i put a negative number as rate of subtracting , the temperature was increased.I can’t fix this problem.
.
.
run 2
Memory usage per processor = 13.9745 Mbytes
Step Temp
0 1073
1 265612.81
2 1056553.9
Loop time of 0.084645 on 1 procs for 2 steps with 32000 atoms

.
.
.
please help me to fix this problem.
thank you for your consideration

Hello
I am working on a project about age hardening. i want to quench my system.
it means that the temperature of my system muss decrease in short period of
time. so I wrote a following code for this part of my project :
#.....................Quench..................#
units metal
dimension 3
boundary p p p
atom_style atomic
neighbor 2 bin

read_data data3

mass 1 63.546
mass 2 107.868

pair_style eam/alloy
pair_coeff * * cuag.eam.alloy.alloy Cu Ag

timestep 1
thermo 1
thermo_style custom step temp
dump 1 all atom 1000 mydump2.in
fix 1 all nve
fix 2 all heat 1 -0.000000000823
write_data data4
run 2

even though i put a negative number as rate of subtracting , the temperature
was increased.I can't fix this problem.

nothing can really happen in two MD steps. if you see a big change,
then this is a very bad sign. that increase in temperature is usually
an indication of a problem with either your starting geometry or the
potentials you are using or the boundary conditions and your MD
simulation parameters.

all it requires to set the temperature to 0K is:

velocity all set 0.0 0.0 0.0

however, if your system has high potential energy - as it seems to be
having - then temperature will increase rapidly.

the most suspicious setting in your input is the size of the timestep
of 1 picosecond. that is *far* too large for the kind of atoms you
have.
a safe choice should be in the range of 1 to 10 femtoseconds

axel.