Couette flow simulation with isothermal walls

Dear Lammps users,

I want to model the Couette flow for a system including solid walls and gas molecules confined between them. In my case, I want to thermostat walls also at a certain temperature. Going through the Lammps mailing list, I noticed that compute temp/partial in combination with a thermostat does what I want. In my system, I impose velocity in the x-direction. Part of my input script can be found in the following. According to the Lammps documentation for compute temp/partial: " The removal of velocity components by this fix is essentially computing the temperature after a “bias” has been removed from the velocity of the atoms. If this compute is used with a fix command that performs thermostatting then this bias will be subtracted from each atom, thermostatting of the remaining thermal velocity will be performed, and the bias will be added back in."
Therefore, I was expecting that wall temperature in the X-direction, due to the imposed drift velocity, should be always higher than the wall temperature based on removed bias center-of-mass velocity ( calculated by compute temp/com). But, as you can see in the following output file from fix ave/time the wall temperature in the x-direction (c_TBX) has being also thermostated and by the end of the simulation, it almost reaches the removed bias center of mass velocity(c_TBCOM).
I highly appreciate if someone could tell me whether my expectation here is wrong or the way that I am defining my thermostats is not correct.

#---------------------------------------Input file

compute TBCOM AuB temp/com

compute TBX AuB temp/partial 1 0 0

compute TBYZ AuB temp/partial 0 1 1

Not clear on all that you are trying to do.
But if you want the walls to move with constant velocity
based on the initial velocity you set, then you
need to use fix setforce to zero the x force on
the wall atoms. That could be the reason
that the x-vel of the wall slows down, e.g. in c_TBX.

Steve