Documentation/UserGuide/Thermal Conduction

Both isotropic and anisotropic thermal conduction can be added. In either case, configure the code with

    % configure --enable-conduction


Both are added at first-order via operator splitting.

The update is explicit in time, so that a very restrictive CFL constraint on the time step will be used.

# Isotropic thermal conduction

Enable by setting a value for the coefficient of thermal diffusion kappa_iso in the problem generator. For example, to read a value from the <problem> block in the input file, or to set a default value of zero if a value is not specified in the input file, add the line

    kappa_iso = par_getd_def("problem","kappa_iso",0.0);


anywhere in the problem generator. If the code detects kappa_iso > 0, then thermal conduction will be applied. Currently kappa_iso must be a constant.

If both kappa_iso and kappa_aniso (see below) are zero, the code assumes an error was made in initializing these constants, and aborts.

# Anisotropic thermal conduction

In this case, the heat flux is parallel to the magnetic field lines. MHD must be enabled to use anisotropic thermal conduction.

Enable by setting a value for the coefficient of parallel thermal diffusion kappa_aniso in the problem generator. For example, to read a value from the <problem> block in the input file, or to set a default value of zero if a value is not specified in the input file, add the line

    kappa_aniso = par_getd_def("problem",kappa_aniso",0.0);


anywhere in the problem generator. If the code detects kappa_aniso > 0, then thermal conduction will be added. Currently kappa_aniso must be a constant.

To add both non-zero parallel and isotropic thermal conduction, simply set values for both kappa_iso and kappa_aniso in the problem generator.

If both kappa_iso and kappa_aniso are zero, and conduction is enabled, the code will terminate with an error message.

# Units

Note the kappa’s are diffusivities, not conductivities. Also note the current implementation uses “dimensionless units” in that the factor (mbar/k_B) is not included in calculating the temperature (instead, T=P/d is used). For cgs units, kappa must be entered in units of (cm^2/s), and the heat fluxes calculated in the counduction functions would need to be multiplied by (k_B/mbar).