Hello all,
I have used fix nvt/sllod in conjunction with fix deform to run a NEMD simulation of Couette flow. My system is composed of water between two parallel plates and a spherical particle being solved in water. I want to apply fix nvt/sllod command on all particles in the system while thermostatting only water molecules. To do so, I have defined a compute and used fix_modify to exclude other molecules from thermostatting. When I run the input file, I get the following error. Could anyone please help me find the solution to this problem?
ERROR: Temperature for fix nvt/sllod does not have a bias (…/fix_nvt_sllod.cpp:76)
I have brought just part of my input file to save space.
group lower type 1 2 # lower wall
group upper type 3 4 # upper wall
group water type 8 9
group cluster type 5 6 7
group boundary union lower upper
fix 4 boundary setforce 0.0 0.0 0.0
reset_timestep 0
timestep 1.0
compute temp_water water temp
compute temp_wall boundary temp
thermo 1
thermo_style custom step time pe etotal press temp c_temp_water c_temp_wall
thermo_modify norm no
fix 1 all nvt/sllod temp 298.0 298.0 100.0 tchain 1
fix_modify 1 temp temp_water
fix 2 all deform 1 x erate 0.02 remap v
fix 3 water shake 1e-5 20 0 b 1 a 1
run 50000
The error message is telling you the problem:
fix 1 all nvt/sllod temp 298.0 298.0 100.0 tchain 1
fix_modify 1 temp temp_water
You specifically assigned a temperature to fix nvt/sllod
that does not account for the deforming box, b/c it
has no bias. You can read about biases for compute temperatures
in Section 6.16 of the manual.
Steve
Steve,
Thanks so much for your answer. In fact, I wanted to use SLLOD algorithm to simulate Couette flow. My system model consists of rigid walls and fluid between walls. I wonder if I can apply SLLOD to this system or not. If I want to use SLLOD, should I remove rigid walls from my system and let SLLOD itself apply shear flow in the fluid?
Thanks.
Farshad
The error message is telling you the problem:
fix 1 all nvt/sllod temp 298.0 298.0 100.0 tchain 1
fix_modify 1 temp temp_water
You specifically assigned a temperature to fix nvt/sllod
that does not account for the deforming box, b/c it
has no bias. You can read about biases for compute temperatures
in Section 6.16 of the manual.
Steve
It’s more typical to not have walls and deform the box.
That’s what fix nvt/sllod requires b/c it uses the
box deformation tensor to apply the SLLOD eqs.
If you want to use walls, you could use some other
thermostat, e.g. fix nvt biased by a compute temp/profile.
That’s not SLLOD per se, but probably pretty close
to the same, since the SLLOD correction is typically small.
Steve