ZEPHYR: Zero-emissions Electricity system Planning with HourlY operational Resolution

Setup instructions

• To set up your python environment, navigate to this directory and run:
  • conda env create -f environment.yml
  • conda activate zephyr
• To use the renewable-energy supply-curve functionality (not required for running the base capacity-expansion model), perform the following steps:
  • Request an NREL API key from https://developer.nrel.gov/signup/ and paste it into apikeys['nrel'] in settings.py.
  • Follow the instructions at https://github.com/NREL/hsds-examples to set up h5pyd for the NREL HSDS endpoint. Namely:
    • Run hsconfigure
    • At the prompt, enter the text after the =:
      • hs_endpoint = https://developer.nrel.gov/api/hsds
      • hs_username = None
      • hs_password = None
      • hs_api_key = the same NREL API key you pasted into settings.py
    • These values will be stored in ~/.hscfg
• Existing transmission capacity is taken from the NREL ReEDS model. ReEDS is open-source, but requires registration to access. Request access at https://www.nrel.gov/analysis/reeds/request-access.html; then you can access the ReEDS repository at https://github.com/NREL/ReEDS_OpenAccess and run the vresc_0-prepare-existing-trans.py script to generate the necessary inputs for the core model.

Usage notes

Capacity-planning model

• zephyr-run-base.py
  • After setting up the environment, the core model can be run for an example full-US scenario using the command: python zephyr-run-base.py 23 -o test -s clp.
    • 23 indicates the case to use from the cases file (in this case, running for the full US at PA resolution over 2007–2013 with new transmission allowed).
    • -o test indicates both the folder in which to store the outputs (out/test) and the cases file from which to read the input settings (cases-test.xlsx).
    • -s clp indicates that the COIN-LP solver should be used. -s gurobi also works if you have gurobi installed with a working license.
  • Alternative cases-{}.xlsx files are also included to reproduce the results from reference [1] below.
• zephyr-get-intraregion-transcost.py
  • Use this script after running all regions in a 'BApAC' scenario to calculate the inter-state intra-region transmission cost adders for a full US run
    • An example output file is io/transmission/transmission-adder-ba-base_case303.csv

Renewable-energy supply curves

• vresc_0-prepare-shapefiles.py
  • Downloads some input data (primarily shapefile maps) from public sources and creates the US states zone map.
• vresc_0-prepare-existing-trans.py
  • Downloads existing inter-BA transmission capacity from the NREL ReEDS model and reshapes into the inter-state and inter-PA formats used in this model.
• vresc_1-icomesh.py
  • Generates the icosahedral mesh used to define NSRDB (solar) points.
• vresc_2-nsrdb-download-icosahedralmesh.py
  • Submits download requests for NSRDB (solar) data. Before use, update nsrdbparams in settings.py with your email address and other info. Once the downloads are prepared you will receive emails at the address you provided with links to download the data. By default, requests are submitted in 1000-site chunks for 41990 sites over the continental US in 1-year batches from 20007–2013, yielding 42 emails and zip files per year.
  • After downloading the zip files from the links in the emails, unzip them into {datapath}/in/NSRDB/ico9/{year}/v3/ (if using a local datapath) or {extdatapath}/in/NSRDB/ico9/{year}/v3/ (if using an external drive).
• vresc_2-download-wtkhsds-geoslice.py
  • Downloads WIND Toolkit (WTK) data. To increase download speed, we download the data in single-timestamp slices over the full array of points, downsampled by a factor of 2x2 (i.e. at 4km resolution instead of 2km resolution).
  • To reproduce the full model at 100m height you'll need to download data for the following fields: windspeed_100m, temperature_100m, relativehumidity_2m, pressure_100m. Each takes roughly two days to download (depending on network speed).
• vresc_3-wtk-hsds-timeslice-to-timeseries.py
  • Reshapes the single-timestamp all-points WTK data into single-point all-timestamps to match the format of the NSRDB data.
• vresc_4-polyavailable-states.py
  • Creates and saves a polygon of available area for wind and solar development within each zone accounting for different land exclusions.
    • Outputs: io/geo/developable-area/{zonesource}/polyavailable-water,parks,native,mountains,urban-{zonesource}_{zone}.poly.p
• vresc_5-sitearea-zoneoverlap.py
  • Calculates the overlap area between the available area for PV and wind calculated in vresc_4-polyavailable-states.py and Vornoi polygons around the NSRDB and WTK points.
    • Outputs: io/geo/developable-area/{zonesource}/{resource}/sitearea-water,parks,native,moutains,urban-{zonesource}_{zone}.csv
• vresc_6-transmission-site-distance.py
  • Determines the interconnection (spur-line + trunk-line reinforcement) cost for each solar and wind site.
    • Outputs: io/cf-2007_2013/{transcostmultiplier}x/{level}/{resource}/distance-station_urbanedge-mincost/{resource}-urbanU-trans230-{level}_{zone}.csv
• vresc_7-cfsim-lcoebins.py, vresc_8-cfsim-binsum.py
  • Simulates the levelized cost of electricity (LCOE) and hourly capacity factor (CF) for each solar/wind site within a given zone. Bins sites by LCOE, then calculates the weighted-average hourly CF (weighted by developable area associated with each site) for each bin. Saves the weighted-average hourly CF and total developable area within each bin.
  • If running for PV on a machine with sufficient memory, use the -f flag when runnning vresc_7-cfsim-lcoebins.py to perform the entire procedure in one shot.
  • If running for wind, run vresc_7-cfsim-lcoebins.py without the -f flag to perform the binning procedure, then run vresc_8-cfsim-binsum.py to generate the hourly CF profiles.

Sources

• Additional documentation is provided in the following reference:
  • [1] Brown, P.R.; Botterud, A. "The value of inter-regional coordination and transmission in decarbonizing the US electricity system." Joule 2020, 5, 1–20. https://doi.org/10.1016/j.joule.2020.11.013
• Land exclusions
  • Mountains
    • The original data source for mountains is the USGS Global Mountain Explorer dataset (https://rmgsc.cr.usgs.gov/gme/). We use the "High Mountains" and "Scattered High Mountains" fileds from the K3 datafile at https://rmgsc.cr.usgs.gov/outgoing/ecosystems/Global/GlobalMountainsK3Classes.zip. These data are in raster format, so we use the gdal_polygonize.py script (https://gdal.org/programs/gdal_polygonize.html) from the GDAL library (https://github.com/OSGeo/gdal) to convert to polygons and truncate to US mountain ranges. These steps are not included in this repository; instead we provide the intermediate io/geo/mountains/usgsgmek3_high_-170,-30lon_5,70lat and io/geo/mountains/usgsgmek3_highscattered_-170,-30lon_5,70lat shapefiles.
• PV modeling
  • Additional documentation on PV modeling is provided in the following papers and repositories. Some of the code in this repository is copied from the Zenodo repositories linked below.
    • [2] Brown, P.R.; O'Sullivan, F. "Shaping photovoltaic array output to align with changing wholesale electricity price profiles." Applied Energy 2019, 256, 113734. https://doi.org/10.1016/j.apenergy.2019.113734, https://zenodo.org/record/3368397
    • [3] Brown, P.R.; O'Sullivan, F. "Spatial and temporal variation in the value of solar power across United States Electricity Markets." Renewable and Sustainable Energy Reviews 2020, 121, 109594. https://doi.org/10.1016/j.rser.2019.109594, https://zenodo.org/record/3562896
• In some cases we include intermediate processed data rather than the raw data-processing scripts. These scripts may be added in the future. Intermediate processed data (with references for original data provided in [1]) include:
  • Existing hydropower capacity and availability (io/hydro/*)
  • NREL EFS load (io/load/EFS/*)
  • Existing nuclear capacity (io/nuclear/*)
  • Wind turbine power curves (io/wind/*)
  • Inter-state and -PA urban centroid distances (io/transmission/{}-distance-urbancentroid-km.csv)

Additional notes

• More recent versions of python packages than are specified in environment.yml may lead to faster run times, but are not fully tested. geopandas 0.8.1 vs 0.6.1 appears to be one such example. Feel free to try more recent versions, but note that doing so may lead to inconsistent behavior elsewhere.
• This repository was developed and tested on macintosh and linux machines. While efforts have been made to support platform independence, there may be issues on windows machines.
• The abbreviations 'BA' (balancing area) and 'PA' (planning area) are used interchangeably in many scripts; in all cases they refer to the planning areas in reference [1].