Scenario Management Settings

PowerGenome supports running multiple scenarios in a single execution, systematically varying parameters across cases. This enables sensitivity analyses, policy comparisons, and uncertainty exploration.

Simpler alternative for time-varying parameters

If you only need to change a parameter's value across planning years (not across different cases), you can use year-keyed dictionaries directly in your settings file instead of setting up full scenario management. See Year-Keyed Values for details.

Scenario Workflow

1. Define scenarios in a CSV file with parameter variations
2. Configure parameter swaps in settings_management
3. Run batch execution with run_powergenome
4. Output to separate folders per case

Scenario Definition File

scenario_definitions_fn

Type: String (path) Required: For multi-scenario runs Example: "scenario_definitions.csv"

CSV file defining all scenario variations.

scenario_definitions_fn: scenario_definitions.csv

File structure:

Mandatory columns:

• case_id: Unique identifier for each case
• year: Model year (must match model_year values)

User-defined columns: Parameter dimensions to vary across scenarios

Example file:

case_id,year,solar_cost,wind_cost,gas_price,carbon_policy
baseline_2030,2030,mid,mid,reference,none
high_re_2030,2030,low,low,reference,none
low_re_2030,2030,high,high,reference,none
carbon_2030,2030,mid,mid,reference,aggressive
baseline_2040,2040,mid,mid,reference,none
high_re_2040,2040,low,low,reference,none

Requirements:

• Every case_id × year combination must be unique
• Number of rows = (# unique case_ids) × (# model years)
• User-defined columns reference keys in settings_management

Settings Management

settings_management

Type: Nested dictionary (year → parameter → value → settings) Required: For multi-scenario runs Example: See below

Defines how settings parameters change based on scenario dimensions.

Structure:

settings_management:
  <year>:
    <parameter_name>:
      <parameter_value>:
        <settings_to_modify>

Year specification:

• Use specific years (e.g., 2030, 2040) for year-specific settings
• Use all_years to apply settings across all model years

settings_management:
  all_years:  # Applied to all years in model_year
    solar_cost:
      low:
        resource_financial_case: R&D
      high:
        resource_financial_case: Market

  2030:  # Year-specific overrides
    carbon_tax: 50

  2040:
    carbon_tax: 100

Detailed example:

settings_management:
  2030:
    solar_cost:
      low:
        atb_cost_case: Advanced
      mid:
        atb_cost_case: Moderate
      high:
        atb_cost_case: Conservative

    wind_cost:
      low:
        resource_modifiers:
          landbased_wind:
            technology: LandbasedWind
            tech_detail: Class3
            capex_mw: [mul, 0.85]
      mid: {}  # No changes from baseline
      high:
        resource_modifiers:
          landbased_wind:
            technology: LandbasedWind
            tech_detail: Class3
            capex_mw: [mul, 1.15]

    gas_price:
      reference:
        fuel_scenarios:
          naturalgas: reference
      high:
        fuel_scenarios:
          naturalgas: high_price
      low:
        fuel_scenarios:
          naturalgas: low_price

    carbon_policy:
      none:
        carbon_tax: 0
      moderate:
        carbon_tax: 50
      aggressive:
        carbon_tax: 100

How it works:

1. For each row in scenario_definitions.csv, PowerGenome reads parameter values
2. Looks up corresponding settings in settings_management
3. Applies all settings modifications for that scenario
4. Runs model with modified settings
5. Writes outputs to results/{case_id}/

Parameter Swap Examples

Technology Costs

settings_management:
  2030:
    battery_cost:
      low:
        resource_modifiers:
          batteries:
            technology: Utility-Scale Battery Storage
            tech_detail: Lithium Ion
            capex_mw: [mul, 0.7]
            capex_mwh: [mul, 0.7]
      high:
        resource_modifiers:
          batteries:
            technology: Utility-Scale Battery Storage
            tech_detail: Lithium Ion
            capex_mw: [mul, 1.3]
            capex_mwh: [mul, 1.3]

Technology Availability

settings_management:
  2030:
    nuclear:
      allowed:
        new_resources:
          - [NaturalGas, 2-on-1 Combined Cycle (F-Frame), Moderate, 500]
          - [UtilityPV, Class1, Moderate, 100]
          - [LandbasedWind, Class3, Moderate, 100]
          - [Utility-Scale Battery Storage, Lithium Ion, Moderate, 100]
          - [Nuclear, Nuclear - Large, Moderate, 1000]
      prohibited:
        new_resources:
          - [NaturalGas, 2-on-1 Combined Cycle (F-Frame), Moderate, 500]
          - [UtilityPV, Class1, Moderate, 100]
          - [LandbasedWind, Class3, Moderate, 100]
          - [Utility-Scale Battery Storage, Lithium Ion, Moderate, 100]
        new_gen_not_available:
          ALL_REGIONS:
            - Nuclear

Transmission Expansion

settings_management:
  2030:
    tx_expansion:
      limited:
        max_network_reinforcement_mw: 500
        tx_expansion_per_mw: 1500
      baseline:
        max_network_reinforcement_mw: 1000
        tx_expansion_per_mw: 1000
      unlimited:
        enforce_constraints: false  # Copper plate

Multi-Period Scenarios

For multi-period models, each period can have different parameter values. You can use all_years to apply common settings across all years, then override specific years as needed.

scenario_definitions.csv (multi-period, technology cost & carbon only):

case_id,year,tech_cost,carbon
baseline,2030,moderate,none
baseline,2040,moderate,none
baseline,2050,moderate,none
high_tech,2030,advanced,aggressive
high_tech,2040,advanced,aggressive
high_tech,2050,advanced,aggressive

settings_management.yml (using all_years):

settings_management:
  all_years:  # Common settings across all years
    tech_cost:
      moderate:
        resource_financial_case: Market
      advanced:
        resource_financial_case: R&D
    carbon:
      none:
        carbon_tax: 0

  # Year-specific overrides for aggressive carbon case
  2030:
    carbon:
      aggressive:
        carbon_tax: 100

  2040:
    carbon:
      aggressive:
        carbon_tax: 150

  2050:
    carbon:
      aggressive:
        carbon_tax: 200

Alternative (without all_years, explicit per year):

settings_management:
  2030:
    tech_cost:
      moderate:
        resource_financial_case: Market
      advanced:
        resource_financial_case: R&D
    carbon:
      none:
        carbon_tax: 0
      aggressive:
        carbon_tax: 100
  2040:
    tech_cost:
      moderate:
        resource_financial_case: Market
      advanced:
        resource_financial_case: R&D
    carbon:
      none:
        carbon_tax: 0
      aggressive:
        carbon_tax: 150
  2050:
    tech_cost:
      moderate:
        resource_financial_case: Market
      advanced:
        resource_financial_case: R&D
    carbon:
      none:
        carbon_tax: 0
      aggressive:
        carbon_tax: 200

Input/Output Folders

input_folder

Type: String (path) Required: No Default: Current directory Example: "extra_inputs"

Folder with supplementary input files:

• emission_policies.csv
• misc_gen_inputs.csv
• demand_response_profiles.csv

input_folder: extra_inputs

Referenced files are relative to this folder.

Results Folder

Not a setting - specified on command line:

run_powergenome \
  --settings_file settings \
  --results_folder results

Creates structure:

results/
├── baseline_2030/
│   ├── Generators_data.csv
│   ├── Load_data.csv
│   └── ...
├── high_re_2030/
│   ├── Generators_data.csv
│   └── ...
└── carbon_2030/
    └── ...

Scenario Copying

copy_case_id

Use in emission_policies.csv to copy policies from another case:

case_id,year,region,copy_case_id,RPS,CES
baseline,2030,all,,0.5,0.8
high_re,2030,all,baseline,,
carbon,2030,all,,0.5,0.9

high_re copies policies from baseline instead of specifying them explicitly.

Validation

PowerGenome validates scenario definitions:

Missing case_id × year combinations:

Error: case_id 'baseline' missing year 2040

Invalid parameter values:

Error: Parameter value 'ultra_low' not found in settings_management['solar_cost']

Mismatched years:

Error: scenario_definitions contains year 2035, not in model_year [2030, 2040, 2050]

Example Complete Configuration

scenario_definitions.csv (single-period example):

case_id,year,cost_case,carbon
base,2030,mid,none
low_cost,2030,low,none
carbon_50,2030,mid,50
carbon_100,2030,mid,100

settings/scenario_management.yml:

scenario_definitions_fn: scenario_definitions.csv

settings_management:
  2030:
    cost_case:
      low:
        resource_financial_case: R&D
        new_resources:
          - [NaturalGas, 2-on-1 Combined Cycle (F-Frame), Advanced, 500]
          - [UtilityPV, Class1, Advanced, 100]
      mid:
        resource_financial_case: Market
        new_resources:
          - [NaturalGas, 2-on-1 Combined Cycle (F-Frame), Moderate, 500]
          - [UtilityPV, Class1, Moderate, 100]
      high:
        resource_financial_case: Market
        new_resources:
          - [NaturalGas, 2-on-1 Combined Cycle (F-Frame), Conservative, 500]
          - [UtilityPV, Class1, Conservative, 100]
    carbon:
      none:
        carbon_tax: 0
      50:
        carbon_tax: 50
      100:
        carbon_tax: 100

Run command:

run_powergenome \
  --settings_file settings \
  --results_folder results

This generates 4 cases (4 rows in scenario_definitions.csv), running sequentially.

Related Settings

• Model Definition: model_year must match scenario years
• New-Build Resources: Technology cost modifications
• Fuels: Fuel price scenario swapping
• Transmission: Network reinforcement parameters