Existing Generators Settings¶

These parameters control how existing power plants are processed, clustered, and configured for the model.

Clustering Parameters¶

`num_clusters`¶

Type: Integer or nested dictionary Required: Yes Example: See below

Default number of clusters to create for each technology in each region. Can be specified globally or per region/technology.

Global (same for all regions/technologies):

num_clusters:
  Conventional Steam Coal: 3
  Nuclear: 1
  Natural Gas Fired Combined Cycle: ~ # Does not cluster plants

Regional (different per region):

alt_num_clusters:
  CA_N:
    Conventional Steam Coal: 5
    Natural Gas Fired Combined Cycle: 10
    Onshore Wind Turbine: 3
  CA_S:
    Natural Gas Fired Combined Cycle: 8

How it works:

K-means clustering, by default on heat rate and fixed O&M
Each cluster represents a group of similar generators
Reduces thousands of plants to manageable number of resources
"None" (~) skips clustering and returns individual generators

Practical guidance (ReEDS/IPM-scale regions)

Small regions often work with num_clusters: 1 as a starting point
Use alt_num_clusters to split very large or heterogeneous fleets (e.g., >3–4 GW of combustion turbines with a wide heat-rate IQR)
The generated thermal inputs include heat rate inter-quartile range and standard deviation columns. Check those to decide where an extra cluster is warranted.

`alt_num_clusters`¶

Type: Dictionary (region → technology → number) Required: No Example: See below

Overrides num_clusters for specific region/technology combinations.

num_clusters:
  Nuclear: 2  # Default

alt_num_clusters:
  CA_N:
    Nuclear: 1  # Each nuclear plant is unique
    Biomass: 2  # Only 2 biomass clusters needed

This allows fine-grained control without repeating all technologies in the main num_clusters parameter.

`generator_cluster_columns`¶

Type: List of strings Required: No Default: ["heat_rate_mmbtu_mwh", "fixed_o_m_mw"] Example: See below

Columns (features) used for k-means clustering. By default, generators are clustered based on heat rate and fixed O&M costs.

Default behavior (uses heat rate and fixed O&M implicitly):

generator_cluster_columns:
  - heat_rate_mmbtu_mwh
  - fixed_o_m_mw

Custom features:

generator_cluster_columns:
  - heat_rate_mmbtu_mwh
  - capacity_mw
  - operating_year

Use cases:

Add capacity_mw to keep large/small plants separate
Add operating_year to segregate old vs. new plants
Add minimum_load_mw for minimum load constraints
Use only heat_rate_mmbtu_mwh for thermal efficiency clustering

Available columns depend on your generation data schema. Common options:

heat_rate_mmbtu_mwh: Thermal efficiency
capacity_mw: Plant size
fixed_o_m_mw: Fixed O&M costs
variable_o_m_mwh: Variable O&M costs
operating_year: Plant age proxy
minimum_load_mw: Minimum stable operation level

Clustering Quality

Clustering features should be normalized or on similar scales for best results. PowerGenome uses the scikit-learn StandardScaler() method to normalize features.

Technology Grouping¶

`tech_groups`¶

Type: Dictionary (group name → list of technologies) Required: No Example: See below

Combines multiple similar technologies into a single group before clustering. Useful for technologies that are functionally equivalent.

tech_groups:
  Peaker:
    - Natural Gas Fired Combustion Turbine
    - Natural Gas Steam Turbine
  Biomass:
    - Biomass
    - Municipal Solid Waste
    - Landfill Gas

Benefits:

Reduces cluster count for minor technology variations
Groups functionally similar resources
Simplifies output analysis

Matching: Technology names are matched case-insensitively with partial matching.

Attribute Modifiers¶

`resource_attr_modifiers`¶

Type: Dictionary (technology → list of modifier dictionaries) Required: No Example: See below

Apply custom formulas to modify generator attributes based on other data fields. Useful for age-based cost adjustments, regional scaling, or custom business logic.

resource_attr_modifiers:
  conventional steam coal:
    - attribute: fom_per_mwyr
      formula:
        op: add
        rate: 126  # $/kW-yr per year of age
        multiplier: age

Structure:

Each technology can have multiple modifiers. Each modifier requires:

attribute: Column name to modify (e.g., fom_per_mwyr, heat_rate_mmbtu_mwh)
formula: Dictionary with operation details
op: Operation type (add, mul, sub, truediv, or replace)
rate: Numeric rate/coefficient
multiplier: Column name containing values to multiply by rate

Operations:

Operation	Formula	Example
`add`	`new = old + (rate × multiplier)`	Add age-based O&M
`mul`	`new = old × (rate × multiplier)`	Scale by efficiency
`sub`	`new = old - (rate × multiplier)`	Reduce value
`truediv`	`new = old / (rate × multiplier)`	Divide value
`replace`	`new = rate × multiplier`	Ignore old value

Multiple attributes:

resource_attr_modifiers:
  natural gas fired combined cycle:
    - attribute: fom_per_mwyr
      formula:
        op: add
        rate: 75
        multiplier: age
    - attribute: heat_rate_mmbtu_mwh
      formula:
        op: mul
        rate: 1.002  # 0.2% degradation per year
        multiplier: age

Multiple technologies:

resource_attr_modifiers:
  conventional steam coal:
    - attribute: fom_per_mwyr
      formula:
        op: add
        rate: 126
        multiplier: age
  nuclear:
    - attribute: fom_per_mwyr
      formula:
        op: add
        rate: 200
        multiplier: age

Technology matching: Case-insensitive substring matching (e.g., "coal" matches "Conventional Steam Coal").

Common attributes:

fom_per_mwyr: Fixed O&M cost ($/MW-yr)
vom_per_mwh: Variable O&M cost ($/MWh)
heat_rate_mmbtu_mwh: Heat rate (MMBtu/MWh)
capex_mw: Capital cost ($/MW)
minimum_load_mw: Minimum load (MW)
ramp_rate_mw_per_min: Ramp rate (MW/min)

Common multipliers:

age: Plant age in years
capacity_mw: Nameplate capacity
operating_year: Year plant started operation

Column Requirements

Both attribute and multiplier columns must exist in the generator dataframe. Missing columns raise KeyError.

Execution Order

Applied after clustering and cost loading, so you can reference calculated fields like age.

See Modify Generator Attributes How-To Guide for detailed examples and validation.

Data Table Configuration¶

`generation_table`¶

Type: String or dictionary Required: Yes Example: See below

Source of existing generator data. Can be a simple filename or advanced configuration with filters.

Simple:

generation_table: generators_2030.csv

Advanced:

generation_table:
  table_name: generators.parquet
  filters:
    - - [operating_year, '<=', 2030]
      - [retirement_year, '>', 2030]
  columns: [plant_id, technology, capacity_mw, heat_rate, region]

See Data Tables for full filter syntax.

`plant_region_table`¶

Type: String or dictionary Required: No Example: "plant_regions.csv"

Mapping of plants to base regions.

plant_region_table: plant_region_map.csv

Expected columns:

plant_id: Plant identifier
region: Region name (matching model_regions)

Example Configuration¶

Complete existing generator settings for a regional model:

# Clustering
num_clusters:
  Conventional Steam Coal: 3
  Nuclear: 2
  Hydroelectric Pumped Storage: 2
  Peaker: 1

alt_num_clusters:
  CA_N:
    Nuclear: 1
    Hydroelectric Pumped Storage: 1
  CA_S:
    Nuclear: 1

# Technology grouping
tech_groups:
  Peaker:
    - Natural Gas Fired Combustion Turbine
    - Natural Gas Steam Turbine

# Data source
generation_table:
  table_name: generators.parquet
  filters:
    - - [operating_year, '<=', 2030]

Model Definition: Planning years affect retirement calculations
Regions: model_regions used in clustering
Resource Tags: Tag assignment for clustered generators
Data Tables: Table configuration details

Existing Generators Settings¶

Clustering Parameters¶

num_clusters¶

alt_num_clusters¶

generator_cluster_columns¶