SD4SOLPS.jl
- SD4SOLPS.jl
This repository serves as the top most workflow manager with helpful utilities to use other repositories in this project.
Documentation of other repositories in this project
GGDUtils.jl
SOLPS2IMAS.jl
SynthDiag.jl
Installation
Using make:
After cloning this repo, check the make menu:
SD4SOLPS.jl % make help
+Help Menu
+
+make env_with_cloned_repo (or make r): Creates a Julia environment with the cloned repositories
+make env_with_git_url (or make u): Creates a Julia environment with the git urls without creating local clones
+make clean: Deletes Project.toml and Manifest.toml for a fresh startmake r
This option creates local copies of required private repositories at the same level as current repository and uses them in develop mode to create a Manifest.toml
make u
This option uses url of required private repositories to create a static Manifest.toml attached to current master branches of these repositories.
make clean
Deletes Manifest.toml so that environment can be recreated, to update or change the last used method.
Using Julia REPL and installing using Github url
Or, in julia REPL:
julia> using Pkg;
+julia> Pkg.add(; url="https://github.com/ProjectTorreyPines/IMASDD.jl.git");
+julia> Pkg.add(; url="https://github.com/ProjectTorreyPines/GGDUtils.jl.git");
+julia> Pkg.add(; url="https://github.com/ProjectTorreyPines/SOLPS2IMAS.jl.git");
+julia> Pkg.add(; url="https://github.com/JuliaFusion/EFIT.jl.git");
+julia> Pkg.add(; url="https://github.com/ProjectTorreyPines/SD4SOLPS.jl.git");
+julia> Pkg.instantiate()Top file handling functions
SD4SOLPS.find_files_in_allowed_folders — Functionfind_files_in_allowed_folders(
+ input_dirs::String...;
+ eqdsk_file::String,
+ recursive::Bool=true,
+)Searches a list of allowed folders for a set of filenames that will provide information about the SOLPS case. Returns a list of filenames with complete paths.
Example:
SD4SOLPS.find_files_in_allowed_folders(
+ "<your samples folder>/D3D_Ma_184833_03600";
+ eqdsk_file="g184833.03600",
+)SD4SOLPS.geqdsk_to_imas! — Functiongeqdsk_to_imas!(
+ eqdsk_file::String,
+ dd::IMASDD.dd;
+ set_time::Union{Nothing, Float64}=nothing,
+ time_index::Int=1,
+)Transfers the equilibrium reconstruction from an EFIT-style gEQDSK file into the IMAS DD structure.
SD4SOLPS.preparation — Functionpreparation(
+ eqdsk_file::String,
+ dirs::String...;
+ core_method::String="simple",
+ filename::String="sd_input_data",
+ output_format::String="json",
+ eqdsk_set_time::Union{Nothing, Float64}=nothing,
+ eq_time_index::Int64=1,
+)::IMASDD.ddGathers SOLPS and EFIT files and loads them into IMAS structure. Extrapolates profiles as needed to get a complete picture.
Repairing/filling out partial equilibrium files
Tools for repairing/filling out partial equilibrium files.
Some of the added fields may not be totally accurate, so it is recommended to use this tool mainly for test cases, as a utility. For a real equilibrium, problems should be fixed properly.
SD4SOLPS.add_rho_to_equilibrium! — Functionfunction add_rho_to_equilibrium(dd:IMASDD.dd)Adds equilibrium rho profile to the DD
SD4SOLPS.check_rho_1d — Functioncheck_rho_1d(
+ dd::IMASDD.dd;
+ time_slice::Int64=1,
+ throw_on_fail::Bool=false,
+)::BoolChecks to see if rho exists and is valid in the equilibrium 1d profiles
Extrapolations
Utilities for extrapolating profiles
Core profile extrapolations
SD4SOLPS.extrapolate_core — Functionextrapolate_core(
+ edge_rho::Vector{Float64},
+ edge_quantity::Vector{Float64},
+ rho_output::Vector{Float64},
+)::Vector{Float64}Function for assuming a core profile when given edge profile data.
Concept:
- value and derivative should be continuous when joining real data
- derivative at magnetic axis is known to be 0 when making profile vs. rho, by the definition of rho
- derivative probably does something fancier between the pedestal and axis than just linear interpolation, so add an extra point in there
- there's a joint between the steep pedestal and the shallow core that needs an extra knot to manage it properly
- after making up a very simple gradient profile out of a few line segments, integrate it to get the profile of the quantity in question
SD4SOLPS.fill_in_extrapolated_core_profile! — Functionfill_in_extrapolated_core_profile!(
+ dd::IMASDD.dd,
+ quantity_name::String;
+ method::String="simple",
+ eq_time_idx::Int64=1,
+ eq_profiles_2d_idx::Int64=1,
+ grid_ggd_idx::Int64=1,
+ space_idx::Int64=1,
+)This function accepts a DD that should be populated with equilibrium and edge_profiles as well as a request for a quantity to extrapolate into the core. It then maps edge_profiles data to rho, calls the function that performs the extrapolation (which is not a simple linear extrapolation but has some trickery to attempt to make a somewhat convincing profile shape), and writes the result to core_profiles. This involves a bunch of interpolations and stuff.
Input arguments:
dd: an IMAS data dictionaryquantity_name: the name of a quantity inedge_profiles.profiles_2dandcore_profiles.profiles_1d, such as "electrons.density"method: Extrapolation method.eq_time_idx: index of the equilibrium time slice to use. For a typical SOLPS run, the SOLPS mesh will be based on the equilibrium reconstruction at a single time, so the DD associated with the SOLPS run only needs one equilibrium time slice to be loaded. However, one could combine the complete equilibrium time series with the SOLPS run and then have to specify which slice of the equilibrium corresponds to the SOLPS mesh.eq_profiles_2d_idx: index of theprofiles_2Din equilibriumtime_slice.grid_ggd_idx: index of thegrid_ggdto use. For a typical SOLPS run, the SOLPS grid is fixed, so this index defaults to 1. But in future, if a time varying grid is used, then this index will need to be specified.space_idx: index of the space to use. For a typical SOLPS run, there will be only one space so this index will mostly remain at 1.cell_subset_idx: index of the subset of cells to use for the extrapolation. The default is 5, which is the subset of all cells. Ifedge_profilesdata is instead present for a different subset, for instance, -5, which are b2.5 cells only, then this index should be set to -5.
Edge profiles extrapolations
These functions have not been fully tested and/or supported yet.
SD4SOLPS.mesh_psi_spacing — Functionmesh_psi_spacing(
+ dd::IMASDD.dd;
+ eq_time_idx::Int64=1,
+ eq_profiles_2d_idx::Int64=1,
+ grid_ggd_idx::Int64=1,
+ space_idx::Int64=1,
+ avoid_guard_cell::Bool=true,
+ spacing_rule="mean",
+)Inspects the mesh to see how far apart faces are in psi_N. Requires that GGD and equilibrium are populated.
Input Arguments:
dd: a data dictionary instance with required data loaded into iteq_time_idx: index of the equilibrium time slice to use. For a typical SOLPS run, the SOLPS mesh will be based on the equilibrium reconstruction at a single time, so the DD associated with the SOLPS run only needs one equilibrium time slice to be loaded. However, one could combine the complete equilibrium time series with the SOLPS run and then have to specify which slice of the equilibrium corresponds to the SOLPS mesh.eq_profiles_2d_idx: index of theprofiles_2Din equilibriumtime_slice.grid_ggd_idx: index of thegrid_ggdto use. For a typical SOLPS run, the SOLPS grid is fixed, so this index defaults to 1. But in future, if a time varying grid is used, then this index will need to be specified.space_idx: index of the space to use. For a typical SOLPS run, there will be only one space so this index will mostly remain at 1.avoid_guard_cell: assume that the last cell is a guard cell so takeend-2andend-1instead ofendandend-1spacing_rule: "edge" or "mean" to make spacing of new cells (inpsi_N) be the same as the spacing at the edge of the mesh, or the same as the average spacing
SD4SOLPS.cached_mesh_extension! — Functioncached_mesh_extension!(
+ dd::IMASDD.dd,
+ eqdsk_file::String,
+ b2fgmtry::String;
+ eq_time_idx::Int64=1,
+ eq_profiles_2d_idx::Int64=1,
+ grid_ggd_idx::Int64=1,
+ space_idx::Int64=1,
+ clear_cache::Bool=false,
+)::StringAdds an extended mesh to a data dictionary, possibly from a cached result.
Input Arguments:
dd: The data dictionary. It will be modified in place.eqdsk_file: the name of the EQDSK file that was used to get equilibrium data in the dd.b2fgmtry: the name of the SOLPS geometry file that was used to get GGD info inedge_profilesin the dd.eq_time_idx: Index of the time slice in equilibriumeq_profiles_2d_idx: Index of the 2D profile set in equilibrium (there is usually only one)grid_ggd_idx: Index of thegrid_ggdset in edge_profilesspace_idx: Index of the spaceclear_cache: delete any existing cache file (for use in testing)
Unit conversion utilities
SD4SOLPS.gas_unit_converter — Functiongas_unit_converter(
+ value_in::Float64,
+ units_in::String,
+ units_out::String;
+ species::String="H",
+ temperature::Float64=293.15,
+)Converts gas flows between different units. Uses ideal gas law to convert between Pressure * volume type flows / quantities and count / current types of units. There is a version that accepts floats in and outputs floats, and another that deals in Unitful quantities.
gas_unit_converter(
+ value_in::Unitful.Quantity,
+ units_in::String,
+ units_out::String;
+ species::String="H",
+ temperature=293.15 * Unitful.K,
+)Converts gas flows between different units. Uses ideal gas law to convert between Pressure * volume type flows / quantities and count / current types of units. This is the Unitful version.
Output will be unitful, but the units are not simplified automatically. You can perform operations such as
(output |> Unitful.upreferred).valUnitful.uconvert(Unitful.whatever, output).val
to handle simplification or conversion of units.
Although this function pretends torr L s$^{-1}$ and Pa m$^3$ s$^{-1}$ are different, use of Unitful should cause them to behave the same way as long as you simplify or convert units at the end. This means that you can use other pressure*volume type gas units and call them torr L s$^{-1}$ and the script will deal with them up to having messy units in the output.