freegsnke.jtor_refinement module

class freegsnke.jtor_refinement.Jtor_refiner(eq, nnx, nny)

Bases: object

Class to allow for the refinement of the toroidal plasma current Jtor. Currently applied to the Lao85 profile family when ‘refine_flag=True’. Only grid cells that are crossed by the separatrix are refined.

__init__(eq, nnx, nny)

Instantiates the object and prepares necessary quantities.

Parameters:

• eq (freegs4e Equilibrium object) – Specifies the domain properties

• nnx (even integer) – refinement factor in the R direction

• nny (even integer) – refinement factor in the Z direction

build_LCFS_mask(core_mask)

Builds a mask composed of all gridpoints connected to edges crossed by the LCFS. These trigger refinement.

Parameters:

core_mask (np.array) – Plasma core mask on the standard domain (self.nx, self.ny)

build_bilinear_psi_interp(psi, core_mask, unrefined_jtor, thresholds)

Builds the mask of cells on which to operate refinement. Cells that are crossed by the separatrix and cells with large gradient on jtor are considered. Refines psi in the same cells.

Parameters:

• psi (np.array) – Psi on the standard domain (self.nx, self.ny)

• core_mask (np.array) – Plasma core mask on the standard domain (self.nx, self.ny)

• unrefined_jtor (np.array) – The jtor distribution

• thresholds (tuple (threshold for jtor criterion, threshold for gradient criterion)) – tuple of values used to identify where to apply refinement, by default None

build_from_refined_jtor(unrefined_jtor, refined_jtor)

Averages the refined maps to the (nx, ny) domain grid.

Parameters:

• unrefined_jtor (np.array) – (nx, ny) jtor map from unresolved method

• refined_jtor (np.array) – maps of the refined jtor, dimension = (no cells to refine, nnx, nny)

Return type:

Refined jtor on the (nx, ny) domain grid

build_jtor_gradient_mask(unrefined_jtor, threshold, quantiles=(0.5, 0.9))

Selects the cells that need to be refined based on their value of the gradient of jtor. Selection is such that it includes cells where the norm of the gradient exceeds the value calculated based on the quantiles and threshold[1].

Parameters:

• unrefined_jtor (np.array) – The jtor distribution

• thresholds (float) – the relevant value (in the tuple) used to identify where to apply refinement

build_jtor_value_mask(unrefined_jtor, threshold, quantiles=(0.5, 0.9))

Selects the cells that need to be refined based on their value of jtor. Selection is such that it includes cells where jtor exceeds the value calculated based on the quantiles and threshold[0].

Parameters:

• unrefined_jtor (np.array) – The jtor distribution

• thresholds (float) – the relevant value (in the tuple) used to identify where to apply refinement

build_mask_to_refine(unrefined_jtor, core_mask, thresholds)

Selects the cells that need to be refined, using the user-defined thresholds

Parameters:

• unrefined_jtor (np.array) – The jtor distribution

• core_mask (np.array) – Plasma core mask on the standard domain (self.nx, self.ny)

• thresholds (tuple (threshold for jtor criterion, threshold for gradient criterion)) – tuple of values used to identify where to apply refinement, by default None

get_indexes_for_refinement(mask_to_refine)

Generates the indexes of psi values to be used for bilinear interpolation.

Parameters:

mask_to_refine (np.array) – Mask of all domain cells to be refined

Returns:

indexes of psi values to be used for bilinear interpolation 4 points per cell to refine, already set in 2-by-2 matrix for vectorised interpolation dimensions = (no of cells to refine, 2, 2)

Return type:

np.array

prepare_for_refinement()

Prepares necessary quantities to operate refinement.