Rated Power Inference and Data Format Documentation

Overview

The DatabaseGenerator class needs to know the rated power of turbines in the wind farm to normalize bias calculations. This document explains how rated power is determined and how to structure your input files.

Rated Power Inference

The _infer_rated_power() method attempts to find the rated power using three strategies, in order of preference:

Strategy 1: Explicit `rated_power` Key (Recommended)

The cleanest approach is to specify rated_power in your turbine's performance section:

# wind_farm.yaml
turbines:
  name: "IEA 15MW Offshore Reference"
  hub_height: 150.0
  rotor_diameter: 240.0
  performance:
    rated_power: 15000000  # 15 MW in Watts
    Ct_curve:
      Ct_values: [0.0, 0.8, 0.8, 0.4, 0.0]
      Ct_wind_speeds: [0.0, 4.0, 10.0, 15.0, 25.0]

This follows the windIO turbine schema exactly.

Strategy 2: Power Curve Maximum

If rated_power is not specified but a power curve is provided, the maximum value from power_curve.power_values is used:

turbines:
  name: "Generic 10MW Turbine"
  hub_height: 119.0
  rotor_diameter: 178.0
  performance:
    power_curve:
      power_values: [0.0, 1.0e6, 5.0e6, 10.0e6, 10.0e6, 0.0]
      power_wind_speeds: [0.0, 4.0, 8.0, 12.0, 20.0, 25.0]
    Ct_curve:
      Ct_values: [0.0, 0.8, 0.75, 0.4, 0.0]
      Ct_wind_speeds: [0.0, 4.0, 10.0, 15.0, 25.0]

In this case, rated power would be inferred as 10.0e6 (10 MW).

Strategy 3: Turbine Name Parsing (Fallback)

As a last resort, the code parses the turbine name field looking for patterns like "15MW", "22 MW", "10.5mw", etc.:

turbines:
  name: "IEA 22MW Offshore Reference"  # Will parse "22" and multiply by 1e6
  hub_height: 170.0
  rotor_diameter: 282.0
  # Note: No performance data needed if name contains MW rating

Warning: This is fragile and should only be used for quick prototyping.

Where Turbine Data is Searched

The code looks for turbine data in two locations:

Wind Farm File (wind_farm.yaml): turbines at the top level
System File (system.yaml): wind_farm.turbines nested inside

# system.yaml structure
name: MySystem
wind_farm:
  turbines:  # <-- Also checked here
    name: "..."
    performance:
      rated_power: ...

Optional: Override with `rated_power` Parameter

For cases where inferring rated power is problematic, you can add rated_power as an optional parameter to DatabaseGenerator:

# Future enhancement
gen = DatabaseGenerator(
    nsamples=100,
    param_config={...},
    system_yaml_path=...,
    rated_power=15e6,  # Direct override, skip inference
    ...
)

Note: This feature is not yet implemented but is the recommended approach for complex setups.

Legacy `meta.yaml` Files

Some older examples use a meta.yaml file containing metadata like:

# meta.yaml (legacy format)
rated_power: 15000000
system: wind_energy_system/system.yaml
ref_power: observed_output/power.nc
ref_resource: plant_energy_resource/resource.nc

This approach is deprecated in favor of windIO-compliant files. The current workflow expects all necessary information to be derivable from the windIO files themselves.

Troubleshooting

If you see this error:

ValueError: Could not find or infer 'rated_power'.
Tried:
  1. 'rated_power' key in '...turbines.performance'.
  2. 'max(power_curve.power_values)' in '...turbines.performance'.
  3. Parsing 'XMW' from '...turbines.name' field.

Solutions:

Add performance.rated_power to your turbine definition (recommended)
Add a power curve with realistic values
Include "XMW" in your turbine name (e.g., "My 10MW Turbine")
Check that your wind_farm.yaml file is properly formatted

Example: Complete windIO-compliant Setup

# wind_farm.yaml
name: OffshoreWindFarm
layouts:
  - coordinates:
      x: [0, 1000, 2000, 3000]
      y: [0, 0, 0, 0]
turbines:
  name: "IEA 15MW Offshore Reference Turbine"
  hub_height: 150.0
  rotor_diameter: 240.0
  performance:
    rated_power: 15000000  # <-- Best practice: always include this
    rated_wind_speed: 10.59
    cutin_wind_speed: 3.0
    cutout_wind_speed: 25.0
    Ct_curve:
      Ct_values: [0.0, 0.792, 0.792, 0.405, 0.0]
      Ct_wind_speeds: [0.0, 3.0, 10.59, 15.0, 25.0]

This format ensures: - Clear, unambiguous rated power specification - Full windIO schema compliance - Compatibility with all WIFA-UQ features

Rated Power Inference and Data Format Documentation

Overview