Variable as a function of temperature

Submitted by toru68 on Sat, 09/07/2019 - 09:29

I want to use thermal conductivity as a function of temperature. For instance, as simplest form,

variable kt equal temp
or
variable kt equal f_Temp

and

fix ftco all property/global thermalConductivity peratomtype v_kt

But this gives “ERROR: Could not find thermo custom fix ID”

LIGGGHTS manual is a bit confusing about thermo_style. Unlike LAMMPS, liggghts thermo_style command does not take “temp” as an argument. I can dump f_Temp values by just using f_heattransfer argument in thermo_style, when I use a constant conductivity.

Would anyone teach me how to solve this?
--------------------------
I now understand my silly mistake. The right thing to do seems to be as follows.

variable tk atom f_Temp
variable kt atom "some function of v_tk"
fix at0 all property/atom thermalConductivity scalar no no no 7.0
run 1
set atom * property/atom thermalConductivity v_kt

It seemed to work all right.

Forums:

40044600's picture

40044600 | Mon, 12/09/2019 - 22:58

Hey toru86,

Did you successfully implement temperature dependant variables?
I tried to use your set command above but returns an error.

annjeffers | Mon, 04/17/2023 - 01:48

For anyone else trying to implement temperature dependent conductivity, I ran into a segmentation fault using the approach given here in LIGGGHTS-PUBLIC. I believe this is because LIGGGHTS-PUBLIC is set up only for conductivity and specific heat to be property/global fixes. LIGGGHTS-INL however does allow conductivity and specific heat to be inputted as property/atom fixes.

annjeffers | Wed, 04/19/2023 - 20:03

Another thing I discovered about this approach is that (1.) the conductivity calculation needs to be changed from "property/global" to "property/atom" in the physicsheatconduction.cpp, (2.) set.cpp needs to be modified so that it includes thermal conductivity as an atom property to be modified in the set command, (3.) I had to introduce a loop in the input script to continuously call in the set command every n steps during the heat transfer analysis. I'm documenting all this here so that others know how to handle this.