Skip to content

Output File Format

PowerGenome writes input files for the GenX capacity expansion model and, optionally, for MacroEnergy.jl (Macro) in its simpleCSVinputs format. This page describes the folder structure, file names, and the key content of each output.

For the full GenX documentation on input file schemas, see the GenX documentation. For Macro, see the MacroEnergy Examples repository.

Folder structure

Each (case_id, planning period) pair produces a folder:

results/
└── <case_id>/
    └── Inputs/
        └── Inputs_p<N>/           ← one per planning period
            ├── system/
            │   ├── Demand_data.csv
            │   ├── Fuels_data.csv
            │   ├── Generators_variability.csv
            │   ├── Network.csv
            │   ├── Period_map.csv
            │   └── Representative_Period.csv
            ├── resources/          ← one CSV per resource type
            │   ├── Thermal.csv
            │   ├── Vre.csv
            │   ├── Storage.csv
            │   ├── Hydro.csv
            │   ├── ...
            │   └── policy_assignments/
            │       ├── Resource_energy_share_requirement.csv
            │       ├── Resource_capacity_reserve_margin.csv
            │       └── ...
            ├── policies/
            │   ├── Capacity_reserve_margin.csv
            │   ├── CO2_cap.csv
            │   ├── Energy_share_requirement.csv
            │   ├── Maximum_capacity_requirement.csv
            │   └── Minimum_capacity_requirement.csv
            └── powergenome_case_settings.yml

When running without scenario definitions, case_id = Inputs and the folder is simply results/Inputs/Inputs_p1/.

System files

Demand_data.csv

Hourly electricity demand for each model zone. One row per hour (or representative period timestep), with columns:

Column Description
Rep_Periods Number of representative periods
Timesteps_per_Rep_Period Hours per representative period
Sub_Weights Weight (hours) each period represents in the full year
Time_Index Sequential hour index
Demand_MW_z<N> Load in MW for zone N

When time domain reduction is off, there is one period covering all 8760 hours with Sub_Weights = 8760.

Fuels_data.csv

Fuel prices by time step. One row per hour with one column per fuel type, plus a Time_Index column.

Generators_variability.csv

Normalized capacity factor (0–1) for each variable renewable resource, by hour. One column per VRE resource, one row per timestep.

Network.csv

Transmission network topology. One row per network line (inter-regional connection):

Column Description
Network_Lines Line identifier
z<N> Zone membership (+1 sending, −1 receiving, 0 not in line)
Line_Max_Flow_MW Maximum power flow in MW
Line_Loss_Percentage Fraction of power lost in transmission
Line_Max_Reinforcement_MW Maximum expansion allowed (if tx_expansion_per_mw is set)
Line_Reinforcement_Cost_per_MWyr Annualized cost to expand line by 1 MW

Period_map.csv and Representative_Period.csv

Present only when time domain reduction is active. Map representative periods back to the original 8760-hour year for chronological analysis.

Resource files

Resources are split into typed CSVs based on their primary model tag:

File Resources included
Thermal.csv THERM = 1
Vre.csv VRE = 1
Storage.csv STOR > 0
Hydro.csv HYDRO = 1
Flex_demand.csv FLEX = 1
Must_run.csv MUST_RUN = 1
Long_duration_storage.csv LDS = 1
Electrolyzer.csv ELECTROLYZER = 1

All resource files share a common set of columns, with additional columns present depending on the resource type.

Key shared columns

Column Description
Resource Unique resource name
Zone Zone number (integer)
Existing_Cap_MW Installed capacity at start of planning period
New_Build 1 if the resource is a candidate for new investment
Cap_Size Unit size in MW (for integer investment decisions)
Inv_Cost_per_MWyr Annualized investment cost ($/MW/yr)
Fixed_OM_Cost_per_MWyr Fixed O&M cost ($/MW/yr)
Var_OM_Cost_per_MWh Variable O&M cost ($/MWh)
Heat_Rate_MMBTU_per_MWh Thermal efficiency (thermal resources only)
Fuel Fuel name (must match a column in Fuels_data.csv)
Min_Power Minimum stable generation as fraction of capacity

Storage columns (additional)

Column Description
Existing_Cap_MWh Installed energy capacity
Inv_Cost_per_MWhyr Investment cost for energy component
Eff_Up Charging efficiency
Eff_Down Discharging efficiency
Max_Duration Maximum hours of storage at rated power
Min_Duration Minimum hours

Policy files

Energy_share_requirement.csv

Defines Renewable Portfolio Standard (RPS) or Clean Energy Standard (CES) constraints. One row per constraint:

Column Description
ESR_<N> Minimum fraction of load that eligible resources must serve

The companion file resources/policy_assignments/Resource_energy_share_requirement.csv indicates which resources are eligible for each constraint (1 = eligible, 0 = not).

Capacity_reserve_margin.csv

Required reserve margin by zone and constraint:

Column Description
CapRes_<N> Reserve requirement as a fraction of peak demand for each zone

CO2_cap.csv

Emission cap constraints. Each row is one constraint; columns reference zones.

Minimum_capacity_requirement.csv / Maximum_capacity_requirement.csv

Technology-specific build floors and ceilings. One row per constraint, matched by resource name patterns.

Legacy format (old GenX)

If your GenX version uses the older single-file format (Generators_data.csv, Load_data.csv), set:

old_genx_format: true

PowerGenome then writes Generators_data.csv combining all resource types, and Load_data.csv rather than Demand_data.csv.

Macro simpleCSVinputs format

When enabled (see Macro Output Settings), PowerGenome additionally writes a Macro case in the simpleCSVinputs format — the same structure used by Macro's multisector_3zone_simpleCSVinputs example. Only the simpleCSV format is produced (not Macro's JSON asset format). The semantic mapping of generators, transmission, demand, fuel supply, and CO2 caps follows the GenX-to-Macro converter (lb/multistage branch for multi-period runs).

Both formats are written in the same run_powergenome call by default. Because all intermediate data processing (storage clustering, time reduction, market data) is shared, writing GenX and Macro inputs together in one run is faster than running PowerGenome twice (once per model). Use --no-genx / genx_output: false to produce Macro inputs only.

Folder structure

The Macro case is written to a macro_out/-style folder (per planning period when multi-period), containing system_data.json plus three input subfolders:

<macro case folder>/
├── system_data.json          # Top-level manifest: commodities, locations, settings, assets, time_data, nodes
├── settings/
│   ├── case_settings.json    # PeriodLengths, DiscountRate, SolutionAlgorithm, ...
│   └── macro_settings.json   # ConstraintScaling, WriteSubcommodities, AutoCreateNodes, ...
├── system/
│   ├── commodities.json      # Electricity, NaturalGas, CO2, Uranium, Coal, ...
│   ├── locations.json        # Model region names
│   ├── nodes.json            # Electricity demand nodes, fuel supply nodes, CO2 sinks, hydro source
│   ├── time_data.json        # HoursPerSubperiod, HoursPerTimeStep, NumberOfSubperiods, ...
│   ├── Period_map.csv        # Sub-period mapping (only when time reduction is active)
│   ├── demand.csv            # Time_Index, Demand_MW_z<zone>...
│   ├── availability.csv      # Time_Index, <resource>...  (non-constant VRE/hydro/must-run profiles)
│   └── fuel_prices.csv       # Time_Index, <fuel>_<region>... price timeseries per fuel node
└── assets/
    ├── <base fuel>_power.csv # Type=ThermalPower (one file per base fuel)
    ├── vre.csv               # Type=VRE
    ├── electricity_stor.csv  # Type=Battery
    ├── hydropower.csv        # Type=HydroRes
    ├── mustrun.csv           # Type=MustRun
    └── powerlines.csv        # Type=TransmissionLink

Resource mapping highlights

GenX concept Macro output
Existing_Cap_MW existing_capacity / discharge_existing_capacity (thermal, VRE, storage, hydro)
Heat_Rate_MMBTU_per_MWh fuel_consumption = heat rate × 0.29307107 MWh/MMBtu
CO2 emissions emission_rate = fuel CO2 content / 0.29307107, plus a co2_sink node
Eff_Up / Eff_Down charge_efficiency / discharge_efficiency (storage, hydro)
New_Build / Can_Retire can_expand / can_retire
Line_Max_Flow_MW + reinforcement existing_capacity + max_capacity on a TransmissionLink
VRE variability availability column in system/availability.csv
Demand data Demand_MW_z<zone> columns in system/demand.csv + electricity nodes in nodes.json
Fuel prices per-(fuel, region) supply node + system/fuel_prices.csv
CO2 caps CO2 sink node with CO2CapConstraint

Cross-sector assets (hydrogen, liquid fuels, CCS detail, demand-side) are not yet emitted; PowerGenome is electricity-centric and the Macro writer is scoped to the electricity system in v1.