SSWS - Site-Specific Wind Speed Adjustment

Location: tact/adjustments/SSWS.py

Overview

The SSWS (Site-Specific Wind Speed) method adjusts turbulence intensity measurements by first correcting wind speed measurements, then recalculating turbulence intensity using the adjusted wind speed.

Methodology

Algorithm Steps

Train Wind Speed Regression (on training data):
```
Reference_WS = m * RSD_WS + c
```
Apply Wind Speed Adjustment (to test data):
```
Adjusted_WS = m * RSD_WS + c
```
Recalculate Turbulence Intensity:
```
Adjusted_TI = RSD_SD / Adjusted_WS
```

Calculate Representative TI:

Adjusted_RepTI = Adjusted_TI + 1.28 * (RSD_SD / Adjusted_WS)

Key Characteristics

Two-stage adjustment: Wind speed first, then TI calculation
Error propagation: Errors in WS adjustment affect TI calculation
Linear regression: Uses simple linear model for WS correction
Train/test split: Builds model on training data, applies to test data

Usage

Basic Example

from tact import TACT

# Initialize
tact = TACT()

# Run SSWS adjustment
results = tact.adjust(
    data=your_data,
    method="ssws",
    parameters={
        "split": True,              # Enable train/test split
        "config_path": "config.json"  # Path to configuration
    }
)

With DNV Validation

from tact import TACT
from tact.validation import validate_dnv_rp0661
from tact.visualization import plot_dnv_validation

# Run adjustment
tact = TACT()
results = tact.adjust(
    data=data,
    method="ssws",
    parameters={"split": True, "config_path": "config.json"}
)

# Validate against DNV RP-0661
validation = validate_dnv_rp0661(
    adjusted_data=results["adjusted_data"],
    reference_col="ref_ti",
    adjusted_col="adjTI_RSD_TI",
    wind_speed_col="ref_ws",
    bin_col="bins",
    criteria_type="LV"
)

# Generate plots
plot_dnv_validation(
    validation_results=validation,
    adjusted_data=results["adjusted_data"],
    reference_col="ref_ti",
    adjusted_col="adjTI_RSD_TI",
    unadjusted_col="rsd_ti",
    wind_speed_col="ref_ws",
    bin_col="bins",
    method_name="SSWS",
    output_dir="output/plots"
)

Parameters

Required Parameters

Parameter	Type	Description
`config_path`	`str`	Path to configuration JSON file
`split`	`bool`	Whether to split data into train/test sets

Configuration Requirements

The config file must specify column mappings:

{
    "input_data_column_mapping": {
        "reference": {
            "wind_speed": "ref_ws",
            "wind_speed_std": "ref_sd"
        },
        "rsd": {
            "primary": {
                "wind_speed": "rsd_ws",
                "wind_speed_std": "rsd_sd"
            }
        }
    }
}

Required Data Columns

The input data must contain:

Reference wind speed (e.g., ref_ws)
Reference wind speed standard deviation (e.g., ref_sd)
RSD wind speed (e.g., rsd_ws)
RSD wind speed standard deviation (e.g., rsd_sd)
Train/test split indicator (e.g., split)
Wind speed bins (e.g., bins)

Output Format

Returned Dictionary

{
    "adjusted_data": pd.DataFrame,  # Data with adjusted columns
    "reg_results": pd.DataFrame,    # Regression statistics
    "all_stats": pd.DataFrame       # Post-adjustment statistics
}

Adjusted Data Columns

The method adds these columns to your data:

Column	Description
`RSD_adjWS`	Adjusted RSD wind speed
`adjTI_RSD_TI`	Adjusted turbulence intensity
`adjRepTI_RSD_RepTI`	Adjusted representative TI

Regression Results

Column	Description
`sensor`	Sensor identifier
`height`	Measurement height
`adjustment`	Adjustment method name
`m`	Regression slope
`c`	Regression intercept
`rsquared`	R² value

Performance Characteristics

When SSWS Works Well

Strong linear relationship between RSD and reference wind speed
Low noise in wind speed measurements
Consistent wind speed bias across all speeds

When SSWS May Struggle

Error propagation: WS errors amplify in TI calculation
Low wind speeds: Division by small values creates large relative errors
Non-linear relationships: Linear regression can’t capture complex patterns
High scatter: R² < 0.8 indicates poor model fit

Performance on Example Dataset

Based on DNV RP-0661 LV criteria validation:

Metric	Value	Target	Pass
MRBE	+94.77%	≤5%	❌
RRMSE	188.03%	≤15%	❌
N observations	3,054	-	-

Note: SSWS performs worse than baseline (90.39% MRBE) on the example dataset due to error propagation. See Method Comparison for details.

Comparison with Other Methods

Aspect	SSWS	SS-SF	SSWSStd
Complexity	Medium	Low	High
Parameters	2 (m, c for WS)	2 (m, c for TI)	4 (m, c for WS and SD)
Error Propagation	Yes	No	Yes
Performance	Poor	Best	Medium
Use Case	WS correction needed	General purpose	WS+SD correction

Implementation Details

Class Definition

from tact.core.base import AdjustmentMethod
from tact.core.registry import AdjustmentRegistry

@AdjustmentRegistry.register("ssws")
class SSWS(AdjustmentMethod):
    """Site-Specific Wind Speed adjustment method."""

    def required_model_parameters(self) -> Dict[str, type]:
        return {"config_path": str, "split": bool}

    def required_data_columns(self) -> list:
        return [
            "reference.wind_speed",
            "reference.wind_speed_std",
            "rsd.height_1.wind_speed",
            "rsd.height_1.wind_speed_std"
        ]

    def adjust(self, data: pd.DataFrame, parameters: Dict[str, Any]):
        # Implementation...

Key Functions Used

get_regression() - Performs linear regression
post_adjustment_stats() - Calculates statistics
Column mapping from config file

Troubleshooting

Common Issues

Issue: High MRBE/RRMSE despite good R² Cause: Error propagation from WS adjustment to TI calculation Solution: Try SS-SF method instead (adjusts TI directly)

Issue: Negative adjusted wind speeds Cause: Large negative intercept with low wind speeds Solution: Check regression intercept, consider filtering low WS data

Issue: Very different train vs test performance Cause: Overfitting or non-representative split Solution: Verify train/test split is random and representative

References

DNV GL: DNV-RP-0661 - Remote Sensing Measurement Verification
IEC 61400-12-1 - Wind turbine power performance testing

Source Code

Full implementation: tact/adjustments/SSWS.py

API Reference

Ssws

SSWS - Site-Specific Wind Speed Adjustment

Overview

Methodology

Algorithm Steps

Key Characteristics

Usage

Basic Example

With DNV Validation

Parameters

Required Parameters

Configuration Requirements

Required Data Columns

Output Format

Returned Dictionary

Adjusted Data Columns

Regression Results

Performance Characteristics

When SSWS Works Well

When SSWS May Struggle

Performance on Example Dataset

Comparison with Other Methods

Implementation Details

Class Definition

Key Functions Used

Troubleshooting

Common Issues

See Also

References

Source Code

API Reference

​SSWS - Site-Specific Wind Speed Adjustment

​Overview

​Methodology

​Algorithm Steps

​Key Characteristics

​Usage

​Basic Example

​With DNV Validation

​Parameters

​Required Parameters

​Configuration Requirements

​Required Data Columns

​Output Format

​Returned Dictionary

​Adjusted Data Columns

​Regression Results

​Performance Characteristics

​When SSWS Works Well

​When SSWS May Struggle

​Performance on Example Dataset

​Comparison with Other Methods

​Implementation Details

​Class Definition

​Key Functions Used

​Troubleshooting

​Common Issues

​See Also

​References

​Source Code

SSWS - Site-Specific Wind Speed Adjustment

Overview

Methodology

Algorithm Steps

Key Characteristics

Usage

Basic Example

With DNV Validation

Parameters

Required Parameters

Configuration Requirements

Required Data Columns

Output Format

Returned Dictionary

Adjusted Data Columns

Regression Results

Performance Characteristics

When SSWS Works Well

When SSWS May Struggle

Performance on Example Dataset

Comparison with Other Methods

Implementation Details

Class Definition

Key Functions Used

Troubleshooting

Common Issues

See Also

References

Source Code