MDS water flow in nanochannel

hi dear users , excuse in my simulation for water flow in 300K for water and 300 for walls , with 100000 timestep and timestep=3.5fs ,
the temprature doesnot stay on 300 and goes up , in timestep about 60000 , temperature arrives to 1400K and lammps error that bond 1250 1252 are missing , i want to have a steady temperature (300K) in my simulation, how can i do for this?i fixed wall nvt and fixed the water flow nve and fix temp/rescale 300… , ? my velocity profile is very good but temperature doesnot stay on 300 and goes up , why? thank you for proposes ,

Look at compute temp/profile and the discussion in
section 6.16 of the manual on how to apply biasing
to thermostats. You must also be careful to
measure the temperature correctly when
the system is flowing.

Steve