ADAS integration

To evolve the ion densities and temperatures in DREAM simulations, we solve ion rate equations which use rate coefficients from the Atomic Data and Analysis Structure (ADAS). The necessary coefficients are downloaded from Open-ADAS at build-time and compiled into the source file src/ADAS/adasdata.cpp, which is then subsequently compiled and built into DREAM. Rate coefficients are only downloaded for a select set of elements which are regulated through the file tools/elements.json.

Currently, DREAM implements support for four different rate coefficients:

Signature	Description
ACD	Effective recombination coefficients
PLT	Line power driven by excitation of dominant ions
PRB	Continuum and line power driven by recombination and Bremsstrahlung of dominant ions
SCD	Effective ionisation coefficients

Handling of coefficients in DREAM

In DREAM, the coefficients are stored numerically in static arrays declared in src/ADAS/adasdata.cpp (which is automatically generated by tools/get_adas.py during build time). During initialization of a Simulation, an ADAS object is constructed. This ADAS object in turn constructs separate ADASRateInterpolator objects for every element in the DREAM ADAS subset database and for every type of rate coefficient. The job of the ADASRateInterpolator is to evaluate a particular ADAS rate coefficient, for a particular element, at any given charge state, background density and temperature. Thus, the ADAS object (and all its ADASRateInterpolator objects) are accessible to, for example, the IonRateEquation equation term during a simulation so that rate coefficients can be evaluated.

The ADASRateInterpolator takes the ion charge state, background plasma density (in \(\mathrm{m}^{-3}\)) and temperature (in \(\mathrm{eV}\)) as input and interpolates in the tabulated values. Prior to interpolation, the input density and temperature are logarithmized as this is how the ADAS values are tabulated. The coefficient tables also store the 10-logarithm of coefficients, but the ADASRateInterpolator appropriately exponentiates the coefficient before returning it.

Adding new elements

Data for the rate coefficients are not stored in the DREAM git repository. Instead, the user will have download the necessary ADAS data while building DREAM. This is done using the Python script tools/get_adas.py which should automatically be invoked when running CMake. The script downloads data for the elements listed in tools/elements.json and generates the file src/ADAS/adasdata.cpp which contains the actual coefficient data.

Adding support for new elements in the DREAM ADAS databased is fairly straightforward and only requires you to add another entry in the file tools/elements.json. This file (as its name indicates) contains data in the JSON format and generally looks something like the following:

{
    "H": "96",
    "He": "96",
    "Ne": "96",
    "Ar": "89"
}

The key in each key-value pair gives the name of the element to download data for, while the value is a two-digit integer giving the year in which the dataset was published. The Open-ADAS database contains several sets of data for most elements and we can only pick one. Which dataset to pick can be difficult to decide, but the ADAS manual can sometimes provide some guidance. Generally, we want to pick the dataset which has the highest indicated quality and contains all the rate coefficients we need (see the table above).

Warning

In general, adding new elements to DREAM is more complicated than simply adding them to the ADAS database. In addition to rate coefficients, data for the mean excitation energies of ions are also needed by the SlowingDownFrequency object. Such data is not available in the literature for Z>18, and is therefore not yet supported by DREAM.

The script tools/get_adas.py can also be run manually. It has a number of configuration options which can be listed by running ./get_adas.py --help.