Couette flow with temp controlling

Dear LAMMPS users,

I’m working on a 3d couette flow of liquid argon with lammps.
I want to let the walls at $T desired (1.5 for example (LJ units), at the same constant temperature for both walls ) while the upper wall is moved with $u=2! (couette flow) .

with what I’m doing I get the velocity profile desired but the upper wall is
heating !

I wonder if you could help me to get the same temperature at
both walls.

I’m using langvin thermostat to control the temp. could you help me
please ?

attached : velocity profile and T profile.

T_profile_u2_fluid
u_profile_u2_fit_fluid
?

If you could help I will be very grateful.

Thanks in advance.

#-------------------------------------------------------------------------------

Couette flow

#-------------------------------------------------------------------------------

velocities

velocity lower set 0.0 0.0 0.0 units box
velocity upper set ${v0} 0.0 0.0 units box

Best regards,
D. Chibouti

There is not enough information here to provide specific guidance. Could it be that you are not measuring the temperature of the walls correctly?

Thanks for replying.
here is the link to the code.
in.flow.couette.3d
I modeled 3 boxes ( fluid and 2 walls ) and I created a space under the walls to set the forces to 0 in order to let the wall stable.
Then I applied the langvin thermostat to the wall and run to verify if it works. at the first run, the temperatures are 1.5 but when I apply the velocity set to the upper wall , the temperature start to get higher witch I want to avoid.
I want to get to same temperature profile as (fig 1-a ) of the paper Viscous heating and temperature profiles of liquid water flows in copper nanochannel

but I didn’t get the same results.

If you could help, will be very grateful.

Thank you again,
D. Chibouti

Your input is extremely difficult to read because of the excessive use of variables and loads of commented out lines. It also does multiple steps and it is not clear from which step you actually took the data from. Most importantly, that input crashes immediately when I run it.

Per MPI rank memory allocation (min/avg/max) = 7.337 | 7.337 | 7.337 Mbytes
Step Temp c_temp_fluid c_temp_lower c_temp_upper PotEng KinEng TotEng 
       0   0.31540953     1.488861            0            0   -219.06999   0.47307663   -218.59691 
ERROR: Lost atoms: original 12562 current 1298 (src/thermo.cpp:439)
Last command: run         ${Niter0}

It looks the raised temperature of the “upper” wall is due to not removing the flow velocity.

If you want somebody to have a closer look, you need to prepare an input deck that is easier to debug.
In general, it is not a good idea to have all phases of a simulation (geometry creation, equilibration, production) in the same input file. So I suggest you split this into (at least) two pieces: 1) do the system setup and equilibration first and at the end write out a data file, and 2) create an input file without using variables or commented out code that only has the necessary commands to run the production simulation and compute and output only the properties that you are currently interested in and also reduce the number of simulation steps to be just enough to make the issue at hand visible sufficiently well. That will make it massively easier to debug this situation and identify problems beyond what can be gleaned from the current input (which is not much).