I’m happy to announce the release of PowerGenome 0.3.0 with just about all of the data needed for power system capacity expansion models in the US. This release includes a set of example notebooks that demonstrate how to use the underlying PowerGenome functions to build the following data for custom model regions based on aggregations of IPM regions:
- Clusters of existing generators within model regions, including the capacity, min load, heat rate, and O&M (a few operating characteristics still have to be supplied by the user). Existing generators are currently clustered using unit heat rate and fixed O&M. All cost data in PowerGenome is adjusted to a single dollar year.
- New resources (renewable/thermal generators and demand response) in each model region. These are based on NREL ATB technologies, which can be modified in-place or copied and modified.
- Fuel costs for a planning period, based on data from EIA AEO 2020 scenarios. Fuel costs and carbon intensity are both adjusted for CCS technologies.
- New utility PV, onshore wind, and offshore wind resources:
- Select the type of resource, maximum capacity, number of clusters, and maximum LCOE (optional) in a model region. For offshore wind, specify fixed/floating and if the sites are part of BOEM lease areas.
- PowerGenome will aggregate pre-grouped clusters (~20-50 per IPM region, which are aggregations of candidate project areas) of each resource to create hourly generation profiles and interconnection costs. Generation profiles are based on data from Vibrant Clean Energy, and will eventually be put into a data repository for appropriate citation.
- Renewable resource profiles for existing wind, solar, and hydro. These are based on a 2012 weather year, and we plan to add more years of data in the future.
- Transmission constraints between regions, network line loss, line reinforcement costs, and the maximum amount of reinforcement allowed in a planning period. Spur line and transmission reinforcement costs must be specified by a user, but you can use this mapping of IPM to ReEDS regions (and costs) as a starting point.
- Hourly demand in each model region.
- The starting data is based on FERC 714 profiles from 2012, and inflated to future planning years using historical demand growth through 2018, then growth in EIA’s AEO 2020.
- Users can supply hourly profiles for distributed generation and specify either a capacity value (MW) or percent of demand that is satisfied by DG in each planning period. These profiles are subtracted from starting demand.
- Users can supply hourly profiles for additional electrification loads and include them as demand response resources, specifying how much of the load is shiftable and how many hours it can be shifted. These profiles are added to the starting demand.
Instructions for accessing data are included in the repository README. At the moment users need to supply the following data on their own:
- Some operating characteristics of generators (e.g. ramp rates, up/down time, etc.)
- DG (rooftop solar) generation profiles and future generation levels (capacity or percent of load).
- New electrification demand profiles, such as EVs, heat pumps, or industrial electrification. NREL’s Electrification Futures Study - which is also based on a 2012 weather year - can be used to build these profiles.
- Electricity policies such as state-level RPS/CES.
All of the data needed to run an example system is included in the repository. To understand how to set up a new system, I recommend looking at the settings YAML file, the “extra_inputs” CSV files, and the notebooks. Ask questions on here or report bugs/feature requests using the repository “issues”.
Finally, I plan to hold a couple sessions later this month to walk through PowerGenome and answer questions. Those will probably be coordinated through the PowerGenome groups.io forum, so sign up there if you’re interested.
Greg