Python Utilities

The PROCESS Python utilities are located in the repository folder

process

A number of utilities for PROCESS are available, for instance to modify the input file IN.DAT, or to extract and plot data from the PROCESS output.

The majority of utilities operate on MFILE.DAT files which are created by running PROCESS on an IN.DAT file.

All executables use Python library functions either from the publicly available numpy, scipy and matplotlib libraries or the PROCESS Python libraries. To use the PROCESS Python libraries, make sure their directory is in your Python path.

Python > 3

All Python code has been written for Python 3.

Compare MFILEs

process/io/mfile_comparison.py

Tool for comparing two MFILEs and outputting significant differences in numerical values.

Usage

python process/io/mfile_comparison.py [-f path/to/first_MFILE.DAT path/to/second_MFILE.DAT] [-s] [--acc] [--verbose]

Options

Argument	Description
`-h, --help`	show help message and exit
`-f`	Files to compare
`-s`	Save output to file called comp.txt
`--acc`	Percentage difference threshold for reporting
`--verbose`	Additional output

Output

Outputs variables and their values which differ significantly between the two MFILEs.

CSV Exporter

process/io/mfile_to_csv.py

This script reads from a PROCESS MFILE and writes values into a CSV file. The variable list is given in a .json file which is defined by the user; a pre-made one can be found in process/io/mfile_to_csv_vars.json.

Usage

python process/io/mfile_to_csv.py [-h] [-f path/to/MFILE] [-v path/to/variable_list.json]

Options

Argument	Description
`-h, --help`	show help message and exit
`-f, [filename]`	specify MFILE file path
`-v, VARFILE`	specify variable .json list file path

Output

A .csv file will be saved to the directory of the input file.

PROCESS 4-Page PDF Summary

process/io/plot_proc.py

A utility to produce a four-page PDF summary of the output from PROCESS, including the major parameters, poloidal and toroidal cross-sections, temperature and density profiles and TF coil layout and turn structure.

Usage

python process/io/plot_proc.py [-h] [-f path/to/MFILE.DAT] [-s] [-n N] [-d] [-c COLOUR]

If no -f argument is provided it assumes a file named MFILE.DAT is in the current directory.

Options

Argument	Description
`-h --help`	Show help message and exit
`-f FILENAME`	Specify input/output file path
`-s, --show`	Show plot
`-n, N`	Which scan number to plot
`-d, --DEMO_ranges`	Uses the DEMO dimensions as ranges for all graphics
`-c, COLOUR`	Which colour scheme to use for cross-section plots; 1: Original PROCESS (default), 2: BLUEMIRA

Output

Produces a three-page PDF file in the same directory as the input MFILE. The PDF file name has the same prefix as the input MFILE but ending in SUMMARY.pdf

Parameters Displayed

runtitle - Variable describing the purpose of the run.

PROCESS version - Tagged version of the PROCESS used for the run.

Date - Date of the PROCESS run.

Time - Time of the PROCESS run.

User - Name of the user who ran PROCESS.

Optimisation - Figure of merit (minmax) for constrained optimisation.

Plasma Composition - Number densities of several ion species relative to the electron density.

Coil Currents etc - Peak coil currents of the PF coils in $\text{MA}$ , flux swing of the central solenoid used for startup and total available in $\text{Wb}$ . Total burn time tburn in hrs.

Cost of electricity - This is the cost of electricity in $ $/ \text{MWh}$ . Check the respective cost model for the reference year of the inflation used.

Geometry
Major radius, $R_0$
Minor radius, $a$
Aspect ratio, $A$
Elongation at the 95% flux surface, $\kappa_{95}$
Plasma triangularity at the 95% flux surface, $\delta_{95}$
Plasma surface area
Plasma volume
Number of TF coils
Inboard blanket + shield
Outboard blanket + shield
Total fusion power
Plasma gain factor, $Q_{\text{p}}$

Power flows
Nominal neutron wall load²
Normalised radius of the 'core' region $\rho_{core}$ used in the radiation correction of the onfinement scaling³ ⁴
The electron density at the pedestal top, $n_{\text{e,ped}}$
The normalised radius $\rho=r/a$ at the pedestal top
The helium fraction relative to the electron density
The core radiation $P_{\text{rad}} (\rho<\rho_{\text{core}})$ subtracted from $P_{\text{heat}}$ in confinement scaling
The total radiation inside the separatrix (LCFS), $W_{\text{th}}$
Nuclear heating power to blanket $P_{\text{nuc,blkt}}= P_{\text{neutr}} \left(1-e^{-\frac{\Delta x_{\text{blkt}}}{\lambda_{\text{decay}}}}\right)$
Nuclear heating power to the shield $P_{\text{nuc,shld}}=P_{\text{neutr}}-P_{\text{nuc,blkt}}$
TF cryogenic power
Power to the divertor
Divertor lifetime in years
Primary high grade heat for electricity production, $P_{\text{therm}}$
Gross cycle efficiency, $P_{\text{e,gross}}/P_{\text{therm}}$
Net cycle efficiency, $\frac{P_{\text{e,gross}}-P_{\text{heat,pump}}}{P_{\text{therm}}-P_{\text{heat,pump}}}$
Net electric power, $P_{\text{e,net}}=P_{\text{e,gross}}-P_{\text{recirc}}$
Fusion-to-electric efficiency, $P_{\text{e,net}}/P_{\text{fus}}$

Physics
Plasma current, $I_{\text{P}}$
Vaccuum magnetic field at in the plasma centre, $B_{\text{T}}(R_0)$
Safety factor at the 95% flux surface, $q_{95}$
Definitions of $\beta$ as given in ¹
Volume averaged electron temperature $\langle T_e\rangle$ and density $\langle n_e\rangle$
Fraction of the line averaged electron density over the Greenwald density, $\langle n_{\text{e,line}}\rangle / n_{\text{GW}}$
Peaking of the electron temperature $T_{\text{e,0}}/\langle T_{\text{e}}\rangle$ and density $n_{\text{e,0}}/\langle n_{\text{e,vol}}\rangle$
Plasma effective charge, $Z_{\text{eff}}=\sum_i f_iZ_i^2$
Impurity fraction, $f_Z=n_Z/\langle n_e\rangle$
H-factor and confinement time calculated from a radiation corrected confinement scaling³ ⁴.
L-H threshold power, $P_{\text{LH}}$
The confinement time scaling law used

Heating & Current Drive
The steady state auxiliary power used for heating and current drive during the flat top phase (NOT to be confused with the start up or ramp down power requirements)
Part of the auxiliary power that is used for heating only, but not current drive
Current drive fractions for the bootstrap auxiliary and inductive current
The neutral beam current drive efficiency, $\gamma_{NB}$ (If NBI used)
The neutral beam energy (If NBI used)
The plasma heating used in the calculation of the confinement scaling / H-factor, $P_{\text{aux}} + P_\alpha - P_{\text{rad,core}}$
The normalised current drive efficiency
The divertor figures of merit, $\frac{P_{\text{sep}}}{R}$ & $\frac{P_{\text{sep}}}{\langle n_e\rangle R}$
Fraction of the power crossing the separatrix with respect to the LH-threshold power $P_{\text{sep}}/P_{\text{LH}}$
Non-radiation corrected H-factor, $\text{H*}$ (Calculated for info only)

TF and WP structure
Inboard TF nose case thickness
Inboard TF minimum distance between side case and WP
Radial width of inboard TF leg
Thickness of insualtion surrounding WP
Number of turns in WP
WP current density
Radial width of WP
Inter-turn insulation thickness
Turn steel conduit thickness
Turn cable space width and surface area
Turn cooling pipe diameter

Sankey Diagram

process/io/plot_sankey.py

The power flows of the power plant will be extracted from MFILE.DAT and used to populate a Sankey diagram. The diagram will start from the initial fusion power and show the inputs and outputs for the power flows. The Recirculated power will finish by connecting the plasma heating back into the fusion power.

Usage

python process/io/plot_sankey.py [-h] [-e END] [-m path/to/MFILE.DAT]

If no -m argument is provided it assumes a file named MFILE.DAT is in the current directory.

Options

Argument	Description
`-h --help`	show help message and exit
`-e --end`	file format, default = pdf
`-m --mfile`	mfile name, default = MFILE.DAT

Output

A .pdf file is created called 'SankeyPowerFlow.pdf' in the directory the utility was run. N.B. Rounding to whole integer can cause errors of $\pm$ 1 between adjacent arrows.

Example Output

Sankey flow chart of large tokamak scenario — Figure 1: Sankey flow chart of power flows for the large tokamak scenario.

TF Stress distribution plots

process/io/plot_stress_tf.py

Program to plot stress, strain and displacement radial distributions at the inboard mid-plane section of the TF coil. This program uses the SIG_TF.json file created by running PROCESS, that stores stress distributions of the VMCON point and stores the output plots in the SIG_TF_plots/ folder, created if not existing.

Discussion of the stress modelling assumptions

In case of a resisitive coil, the stress is calculated from a generalized plane strain model, hence providing vertical stress radial distribution, alongside the radial and the toroidal ones. This is not the case for superconducting magnets as a plane stress modelling is used for now. The reason is that a transverse orthotropic formulation of the generalized plane strain is needed to correctly take the difference of the casing in the vertical direction properly. This will be done in the near future.

Usage

python process/io/plot_stress_tf.py [-h] [-f path/to/SIG_TF.json] [-p [PLOT_SELEC]] [-sf [SAVE_FORMAT]] [-as [AXIS_FONT_SIZE]]

Option

Argument	Description
`-h, --help`	show help message and exit
`-f, --input-file`	`SIG_TF.json` input file
`-p, --plot_selec [PLOT_SELEC]`	Plot selection string :
-	- if the string contains `sig`, plot the stress distributions
-	- if the string contains `strain`, plot the strain distributions
-	- if the string contains `disp`, plot the radial displacement distribution
-	- if the string contains `all`, plot stress and displacement distributions
`-sf, --save_format [SAVE_FORMAT]`	output format (default='pdf')
`-as, --axis_font_size [AXIS_FONT_SIZE]`	Axis label font size selection (default=18)

Turn output into input

process/io/write_new_in_dat.py

This program creates a new IN.DAT file with the initial values of all the iteration variables replaced by their results in OUT.DAT, if that output is a feasible solution.

When a scan has been run, by default this program uses the last feasible point in that scan to write the new starting values. There is also an option to select the first feasible solution from a scan.

Input: IN.DAT, MFILE.DAT

Output: new_IN.DAT

Usage

python process/io/write_new_in_dat.py [-h] [-f path/to/MFILE.DAT] [-i path/to/IN.DAT] [-o path/to/new_IN.DAT]

Options

Argument	Description
`-h, --help`	show help message and exit
`-f`	file to read as MFILE.DAT
`-i`	file to read as IN.DAT
`-o`	file to write as new IN.DAT
`-lfp`	use the last feasible point from a scan (default)
`-ffp`	use the first feasible point from a scan

Plot scan results

process/io/plot_scans.py

This utility plots the output of a PROCESS scan. PROCESS must be run on a scan-enabled input file to create an MFILE on which plot_scans.py can be run. More than one input file can be used and the different files will be plotted on the same graph.

Input: MFILE.DAT

Output scan_var1_vs_var2.pdf (var1 by default is bmaxtf, var2 specified by user)

Usage

python process/io/plot_scans.py [-h] [-f path/to/MFILE(s)] [-yv output vars] [-yv2 2nd axis output variable] [-o [path/to/directory]] [-out] [-sf [SAVE_FORMAT]] [-as [AXIS_FONT_SIZE]] [-ln LABEL_NAME] [-2DC] [-stc]

Options

Argument	Description
`-h, --help`	show help message and exit
`-f, --input_files`	Specify input file(s) path(s) (default = MFILE.DAT).More than one input file can be used eg: -f 'A_MFILE.DAT B_MFILE.DAT'. You can only specify the folder containing the MFILE. The different files scan will be plotted on the same graph. The scans must use the same scan variation.
`-yv, --y_vars`	Select the output variables. More than one output can be plotted eg: -yv 'var1 var2'. A separate plot will be created for each inputs variables
`-yv2, --y_vars2`	Select the 2^nd axis output variable eg: -yv2 'var'. 2^nd variable will be plotted on shared figure inputsvariable.
`-o, --outputdir`	Output directory for plots, defaults to current working directory.
`-out, --term_output`	Option to show scans values on terminal directory.
`-sf, --save_format`	Output format (default='pdf')
`-as, --axis_font_size`	Axis label font size selection (default=18)
`-ln, --label_name`	Label names for plot legend. If multiple input files used then list the same number of label names eg: -nl 'leg1 leg2', (default = MFile file name)
`-2DC, --two_dimensional_contour`	Option to plot 2D scans as a coloured contour plot instead of a line plot. Note: Non convergent points will show up with a value of zero Note: The scan paramters must both be in increasing orderl
`-stc, --stack_plots`	Option to plot multiple 1D plots in a column of subplots. Variables will be plotted in order of input

Plot a pie chart of the cost breakdown

process/io/costs_pie.py

This utility plots the cost breakdown as a pie chart giving each component as a percentage. This allows for the most expensive areas to be easily identified. For the 1990 cost model, an additional plot showing how direct, indirect and contingency costs contribute to the overall budget is shown.

Input: MFILE.DAT

Output: Displays plot of the cost breakdown to screen. For the 1990 cost model, the breakdown for direct, indirect and contingency are also shown. These can be saved with -s argument (cost_pie.pdf and direct_cost_pie.pdf).

Usage

python process/io/costs_pie.py [-h] [-f path/to/MFILE] [-s]

If no -f argument is provided it assumes a file named MFILE.DAT is in the current directory.

Options

Argument	Description
`-h, --help`	show help message and exit
`-f MFILE`	specify the MFILE
`-s, --save`	save figure

Plot a bar chart of the cost breakdown

process/io/costs_bar.py

This utility plots the cost breakdown as a bar chart giving the cost of each component. This allows for the most expensive areas to be easily identified. For the 1990 cost model, an additional plot showing how the direct, indirect and contingency costs contribute to the overall budget is shown. Multiple MFILEs can be specified allowing for different PROCESS runs to be compared on the same plot. An inflation factor can be specified using the -inf argument, which multiplies all the costs by that value.

Input: MFILE.DAT

Output: Displays plot of the cost breakdown to screen. For the 1990 cost model, the breakdown for direct, indirect and contingency is also shown. These can be saved with -s argument (cost_bar.pdf and direct_cost_bar.pdf).

Usage

python process/io/costs_bar.py [-h] [-f f [f ...]] [-s] [-inf INF]

Options

Argument	Description
`-h, --help`	show help message and exit
`-f MFILE`	specify the MFILE(s) to plot
`-s, --save`	save figure
`-inf INF`	Inflation Factor (multiplies costs)

Uncertainty Tools

In this section, we explain the usage of the PROCESS tools to both evaluate the uncertainties of a design point and display them using a simple plotting facility.

The uncertainty evaluation tool has a significantly longer run time than typical evaluations of PROCESS design points and therefore should only be used once a suitable design point has been found. As only user selected output data is kept, the user is recommended to put careful thought into the list of needed output variables.

`evaluate_uncertainties.py`

This program evaluates the uncertainties of a single PROCESS design point by use of Monte Carlo method as described in⁵ by default, and can also use the Morris method and Sobol techniques. It is recommended to submit this script as a batch job to Freia when 1000s of sample points are required.

Input

This script requires two files to run:

config_evaluate_uncertainties.json: A configuration file which details the uncertain parameters under investigation. These are described by probability distributions such as Gaussian, lower half Gaussian, flat top, etc.
IN.DAT: A PROCESS input file which describes the relevant design point. The path to this file should be specified in the config_evaluate_uncertainties.json file.

The configuration file config_evaluate_uncertainties.json uses the JSON format, and has the following style:

{
    "_description": "Configuration file for uncertainties evaluation in PROCESS",
    "_author": "Process McCoder",
    "config": {
        "runtitle": "testrun for uncertainty tool",
        "IN.DAT_path": "path_to_input_file/IN.DAT",
        "working_directory": "path_to_output_folder/",
        "pseudorandom_seed": 16,
        "no_iter": 1
    },
    "uncertainties": [
        {
            "Varname": "boundu(9)",
            "Errortype": "LowerHalfGaussian",
            "Mean": 1.2,
            "Std": 0.1
        },
        {
            "Varname": "boundu(10)",
            "Errortype": "LowerHalfGaussian",
            "Mean": 1.2,
            "Std": 0.1
        },
        {
            "Varname": "coreradius",
            "Errortype": "Gaussian",
            "Mean": 0.6,
            "Std": 0.15
        }
    ],
    "output_vars": [],
    "no_scans": 1,
    "no_samples": 100,
    "output_mean": 8056.98,
    "figure_of_merit": "rmajor",
    "vary_iteration_variables": false,
    "latin_hypercube_level": 4
    ...

By convention, we have designated metadata about the PROCESS runs as having a preceding underscore to distinguish these values from the other configuration data used directly by the tools or PROCESS itself. Furthermore, all the optional attributes that can be changed when running PROCESS from most Python utilities can be specified in the "config" section. All these values have default values and do not need to be set.

runtitle: is a one line description of the purpose of the run to be saved in README.txt in the working directory as well as the runtitle parameter in the OUT.DAT and MFILE.DAT files. Per default it is empty.
IN.DAT_path: is the name/path of the IN.DAT file describing the design point. If not specified it is assumed to be IN.DAT.
working_directory: directs to the working directory in which PROCESS will be executed. It is recommended to create a directory for each run as this can aide organisation while several runs are executed with slightly different configs.
pseudorandom_seed: is the value of the seed for the random number generator. It can be any integer value. If it is not specified, its default value is taken from the system clock.
no_iter: sets Niter, the maximum number of retries that the tool will attempt if PROCESS fails to find a feasible solution. The default value is 10, but this can be changed depending on the user's preference for speed and solutions.
factor: varies the start values of the iteration variables by a factor of the original values. This does not change the physical meaning of the input file, but can help the solver to find a better starting point for its iteration. The default value is factor=1.5.
uncertainties: any uncertain parameters should be specified in the uncertainties section. Each parameter is specified in its own sub-directory in the config file example above. For each entry, the Varname and Errortype need to be specified and each Errorrtype must be include the appropriate boundaries, listed below:
Errortype :
- Gaussian (Mean and Std)
- LowerHalfGaussian (Mean and Std)
- UpperHalfGaussian (Mean and Std)
- Uniform (Lowerbound and Upperbound)
- Relative (Mean and Percentage)
Please note that all distributions are cut off at the boundaries for the input values for PROCESS! At least one uncertain parameter has to be specified for the program to run and there is no upper limit to how many uncertain parameters can be used. However, for large numbers of uncertain parameters it is recommended to increase the number of sampling points.
no_samples: sets the number of sample points in the Monte Carlo method. It is by default set to its recommended minimum value of 1000, but the user should contemplate higher values especially if a large number of uncertain parameters are involved.
no_scans: can be used to set the number of scan runs in each MC sample point. Only the last scan point is stored in the data ouput. Older versions of the code made more use of this feature and it is recommended to set this to 1.
no_allowed_unfeasible: is the number of allowed unfeasible points in a run which is set as 2 by default.
vary_iteration_variables: This enables a shuffle of the iteration variables in the Monte Carlo method. By default it is set to false and may be set to true to recreate old runs of the MC code.

Output

uncertainties_data.h5: This file contains the output variables of each successfully converged PROCESS run generated by the evaluate_uncertainties.py script. PROCESS output variables can be plotted using using the hdf_to_scatter_plot.py script. This file uses the HDF format and requires software to view its contents in a human legible format.
README.txt, process.log, MFILE.DAT, OUT.DAT, SIG_TF.json: Typical PROCESS output generated by the last run.

Usage

The evaluate_uncertainties.py script is run with with the option -f to specify the path to the config_evaluate_uncertainties.json file:

python process/uncertainties/evaluate_uncertainties.py -f path/to/config_evaluate_uncertainties.json -m method

Options

Argument	Description
`-h, --help`	show help message and exit
`-f CONFIGFILE`	specify the path to the config file
`-m METHOD`	type of uncertainty analysis performed, default = monte_carlo, other options = sobol_method, morris_method

The uncertainty analysis technique used can be specified using '-m monte_carlo/sobol_method/morris_method' but the default is Monte Carlo. Use -h or --help for help.

Sobol Plotting

process/uncertainties/sobol_plotting.py

Program to plot the output of the the Sobols sensitivity analysis at a given PROCESS design point. It creates a bar chart showing both the first order and total Sobol indices for each variable and gives the 95% confidence intervals.

Usage

python process/uncertainties/sobol_plotting.py [-h] [-f path/to/DATAFILE] [-o OUTPUTFILE]

If no -f argument is provided it assumes a file named sobol.txt is in the current directory.

Options

Argument	Description
`-h, --help`	show help message and exit
`-f DATAFILE`	path to datafile for plotting, default = sobol.txt
`-o OUTPUTFILE`	filename of outputed pdf file, default = sobol_output.pdf

Configuration File

The tool reads the data contained sobol.txt produced from running evaluate_uncertainties.py with the sobol_method. The name and location of the data file can be modified using the option DATAFILE.

Output

A .pdf file is created called sobol_output.pdf. The name of the produced pdf file can be specified using the option OUTPUTFILE.

References

M. Kovari, R. Kemp, H. Lux, P. Knight, J. Morris, D. J. Ward "PROCESS: a systems code for fusion power plants - Part 1: Physics", Fusion Engineering and Design 89, 30543069 (2014), http://dx.doi.org/10.1016/j.fusengdes.2014.09.018 ↩
M. Kovari, F. Fox, C. Harrington, R. Kembleton, P. Knight, H. Lux, J. Morris "PROCESS: a systems code for fusion power plants - Part 2: Engineering", Fus. Eng. & Des. 104, 9-20 (2016) ↩
H. Lux, R. Kemp, D.J. Ward, M. Sertoli "Impurity radiation in DEMO systems modelling", Fus. Eng. & Des. 101, 42-51 (2015) ↩↩
H. Lux, R. Kemp, E. Fable, R. Wenninger, "Radiation and confinement in 0D fusion systems codes", PPCF, 58, 7, 075001 (2016) ↩↩
H. Lux, R. Kemp, R. Wenninger, W. Biel, G. Federici, W. Morris, H. Zohm, "Uncertainties in power plant design point evaluations", Fusion Engineering and Design, Vol 123, 63-66, 2017 ↩