Analysis tools and helper scripts

The ReEDS model includes several tools for setting up new cases, managing active cases, and analyzing case results. Most of these scripts have command-line interfaces, the details of which can be printed by running python path/to/scriptname.py -h.

Setting up new cases

Create a new cases_{}.csv file: preprocessing/casemaker.py

This script facilitates the creation of a multidimensional matrix of scenarios, with “dimensions” and “casegroups” (permutations of dimensions) provided via a .yaml file.

An example .yaml file is provided in preprocessing/casematrix_example.yaml. In this example, three dimensions are explored: “policy”, “demand”, and “sensitivity”. Each dimension specifies a dictionary of choices, and each choice specifies the switch values associated with that choice. Default switch values are taken from cases.csv; if a “shared” field is provided in the casematrix .yaml file, the switch/value pairs in the “shared” field overwrite the associated defaults from cases.csv. Cases in the resulting cases_{batchname}.csv file are labeled as {policy choice}_{demand choice}_{sensitivity choice}. For example, a case using the “IRA” policy setting, “DemMd” demand setting, and “core” sensitivity setting would be labeled as IRA_DemMd_core.

The “casegroups” indicate which permutations of dimension values to include in the resulting cases_{batchname}.csv file. The provided example includes 3 policy scenarios, 7 demand scenarios, and 9 sensitivity cases, which would produce 3 × 7 × 9 = 189 cases if all permutations were included. casegroups is a list of n-element lists, where n is the number of dimensions specified above. If an element is an empty list, all options for that dimension are used; if an element specifies a single option or a list of options, only those options are used. The order of elements matches the order of dimensions. For example, since the dimensions are “policy”, “demand”, and “sensitivity”:

  • [[], [], core] includes all policy scenarios, all demand scenarios, and the single “core” sensitivity, for 3 × 7 × 1 = 21 cases.

  • [OBBBA, DemMd, []] includes a single policy scenario, a single demand scenario, and all sensitivity cases, for 1 × 1 × 9 = 9 cases.

  • [[OBBBA, OBBBAcon], DemMd, [GasPriceLo, GasPriceHi]] includes two policy scenarios, a single demand scenario, and two sensitivity cases, for 2 × 1 × 2 = 4 cases.

  • [[], [], []] would include all 3 × 7 × 9 = 189 cases.

Fix representative/stress periods: preprocessing/get_case_periods.py

This script takes as its required argument a filepath to a completed ReEDS case.

  • If the optional -r/--rep flag is added (as in python preprocessing/get_case_periods.py path/to/casename -r), the representative periods for the provided run are written to inputs/variability/period_szn_user_{name}.csv, where name is either provided by the -n/--name argument or (if no name is provided) given by the case name.

    • In a subsequent case, these representative periods can be used by setting GSw_HourlyClusterAlgorithm to user_{name}.

  • If the optional -s/--stress flag is added, the stress periods for the provided run are written to inputs/variability/stressperiods_user_{name}.csv.

    • In a subsequent case, these stress periods can be used by setting GSw_PRM_StressModel to user_{name}.

Managing currently-running cases

Manage HPC runs: restart_runs.py

This script restarts failed runs on the HPC whose case name starts with the provided prefix: python restart_runs.py case_name_prefix

Analyzing finished cases

Compare results: postprocessing/compare_cases.py

This script creates a powerpoint file comparing the results of the cases provided via the caselist argument. The list of cases to compare can be provided in one of two ways: as a space-delimited list of filepaths, or as a single filepath to a .csv file in the format of postprocessing/example.csv. The first case in the list is treated as the base case, and other cases are all compared to that same case.

Example for two cases:

python postprocessing/compare_cases.py /Users/username/github/ReEDS-2.0/runs/v20250310_main_USA /Users/username/github/ReEDS-2.0/runs/v20250310_newthing_USA

Example for three cases:

python postprocessing/compare_cases.py /Users/username/github/ReEDS-2.0/runs/v20250310_main_USA /Users/username/github/ReEDS-2.0/runs/v20250310_newthing1_USA /Users/username/github/ReEDS-2.0/runs/v20250310_newthing2_USA

Example for a .csv file of cases:

python postprocessing/compare_cases.py /Users/username/github/ReEDS-2.0/postprocessing/example.csv

Run PRAS: postprocessing/run_reeds2pras.py

The PRAS model is typically run multiple times during each ReEDS case (as long as GSw_PRM_CapCredit = 0) to ensure resource adequacy. This script reruns PRAS on a finished ReEDS case (provided by the single required command-line argument) and allows the settings to be changed. For example, to use a different number of samples than are specified by the default pras_samples switch, use the -s/--samples command-line argument.

Run a dispatch model: run_pcm.py

This script reruns a completed ReEDS case as a dispatch simulation at higher time resolution. The operational constraints in c_supplymodel.gms are used directly, but the investment and capacity variables are fixed to their previously optimized values; only the operational variables are re-optimized. 365 representative 1-day periods at 1-hour resolution are used by default, but these settings can be changed using the -s/--switch_mods switch.

This approach is distinct from the R2X tool, which formats the results of a ReEDS case as inputs to a separate production cost modeling tool such as Sienna or PLEXOS. Those tools provide more advanced and realistic features like unit commitment and rolling forecast horizons; by contrast, run_pcm.py simply reuses the existing ReEDS formulation at higher time resolution, and is subject to all the normal caveats and limitations of ReEDS (linear variables, pipe-and-bubble transmission flow, etc.).

Generate static plots: postprocessing/single_case_plots.py

This script runs automatically at the end of a ReEDS case and writes static figures to the {case}/outputs/figures folder as .png files.

Generate interactive plots: postprocessing/bokehpivot

The bokehpivot module can be used to visualize the outputs of ReEDS runs. For more information on how to use bokehpivot, see the bokehpivot guide.

If you’re new to bokehpivot, the following YouTube video will be a good starting point: Viewing ReEDS Outputs Using the BokehPivot Module

Calculate hourly prices and technology value metrics: postprocessing/reValue

reValue is used for two main things:

  • extracting regional hourly prices from ReEDS scenarios and years

  • (Optional) using extracted prices to calculate value and competitiveness-related metrics for a set of regional generation or load profiles.

More more information on reValue, see the reValue documentation.

Estimate retail rates: postprocessing/retail_rate_module

The retail rate module can be used after finishing a ReEDS run to calculate retail electricity rates by state and year, where each state is served by its own investor-owned utility (IOU).

For more information on this module, see the retail_rate_module documentation.

Generate a Tableau results viewer: postprocessing/tableau

Tableau can be used for the analysis of ReEDS and ReEDS-to-PLEXOS results in Tableau.

For more information on how to use Tableau with ReEDS, see the Tableau documentation.