Skip to main content
This guide will walk you through using TACT to adjust turbulence intensity measurements using the Site-Specific Simple + Filter (SS-SF) method.

​
Installation

​
From GitHub

  1. Clone the repository:
git clone https://github.com/CFARS/TACT.git
cd TACT
  1. Install requirements:
pip install -r requirements.txt

​
Required Files

​
1. Input Data

Your input CSV file should contain the following columns:
  • Reference wind speed measurements
  • Reference turbulence intensity
  • RSD (Remote Sensing Device) wind speed measurements
  • RSD turbulence intensity
Example data format (data.csv): 
timestamp,ref_ws,ref_ti,rsd_ws,rsd_ti
2024-01-01 00:00:00,5.2,0.12,5.1,0.15
2024-01-01 00:10:00,6.1,0.14,6.0,0.16
...

​
2. Configuration File

Create a config.json file that maps your data columns: 
{
    "input_data_column_mapping": {
        "reference": {
            "wind_speed": "ref_ws",
            "turbulence_intensity": "ref_ti",
            "standard_deviation": "ref_sd"
        },
        "rsd": {
            "primary": {
                "wind_speed": "rsd_ws",
                "turbulence_intensity": "rsd_ti",
                "standard_deviation": "rsd_sd"
            }
        }
    }
}

​
Running the Adjustment

Here’s a complete example of how to use TACT with the SS-SF method: 
from tact import TACT
from tact.utils.load_data import load_data
from tact.utils.setup_processors import setup_processors
from tact.utils.process_statistics import process_statistics
from tact.utils.save_results import save_results
from tact.validation import validate_dnv_rp0661
from tact.visualization import plot_dnv_validation

def main():
    # Configuration
    config = {
        "data_path": "data.csv",              # Your input data file
        "config_path": "config.json",         # Your column mapping file
        "output_dir": "output",              # Where to save results
        "method": "ss-sf",                   # Using SS-SF method
        "parameters": {"split": True}         # Split data into training/testing sets
    }

    # Initialize TACT
    tact = TACT()

    # Load and process data
    data = load_data(config["data_path"])

    # Setup data processors
    binning_processor, ti_data_processor, stats_processor = setup_processors(config["config_path"])

    # Process data through pipeline
    data = binning_processor.process(data)    # Bin wind speed data
    data = ti_data_processor.process(data)    # Calculate TI metrics

    # Perform SS-SF adjustment
    results = tact.adjust(
        data=data,
        method=config["method"],
        parameters={**config["parameters"], "config_path": config["config_path"]}
    )

    # 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"
    )

    # Print validation results
    print(f"MRBE: {validation['overall']['MRBE_%'].iloc[0]:.2f}%")
    print(f"RRMSE: {validation['overall']['RRMSE_%'].iloc[0]:.2f}%")
    print(f"Passed: {validation['overall']['pass'].iloc[0]}")

    # Generate validation 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="SS-SF",
        output_dir=f"{config['output_dir']}/plots"
    )

    # Calculate statistics
    means = process_statistics(results["adjusted_data"], stats_processor)

    # Save all results
    save_results(
        means=means,
        adjusted_data=results["adjusted_data"],
        reg_results=results["reg_results"],
        method=config["method"],
        output_dir=config["output_dir"]
    )

    print(f"\nResults saved to {config['output_dir']}/")

if __name__ == "__main__":
    main()

​
Understanding the Process

The SS-SF adjustment method:
  1. Data Preparation:
    • Bins wind speed data into categories
    • Calculates turbulence intensity metrics
    • Splits data into training (80%) and testing (20%) sets
  2. Adjustment:
    • Filters out TI measurements > 0.3 from training data
    • Performs linear regression between RSD and reference TI
    • Applies the correction to the test data
  3. Validation:
    • Compares adjusted TI against reference measurements
    • Calculates MRBE (Mean Relative Bias Error) and RRMSE (Relative Root Mean Square Error)
    • Checks against DNV RP-0661 criteria (MRBE ≤ 5%, RRMSE ≤ 15%)
  4. Results:
    • adjusted_data: DataFrame containing:
      • Original RSD and reference TI values
      • Adjusted TI values
      • Binning information
    • reg_results: Regression statistics including:
      • Slope and intercept
      • R² value
      • RMSE (Root Mean Square Error)
    • Validation plots showing before/after comparison

​
Output Files

After running the adjustment, you’ll find these files in your output directory: 
output/
├── ss-sf_adjusted_data.csv       # Full dataset with adjusted values
├── ss-sf_reg_results.csv         # Regression model statistics
├── ss-sf_all_stats.csv           # Summary statistics
└── plots/
    ├── ss-sf_mrbe_by_bin.png    # MRBE per wind speed bin
    ├── ss-sf_rrmse_by_bin.png   # RRMSE per wind speed bin
    ├── ss-sf_ti_scatter.png      # Adjusted vs reference scatter plot
    └── ss-sf_ti_comparison.png   # Before/after comparison by bin
The validation plots show:
  • Scatter plots: Adjusted TI vs reference TI with 1:1 line
  • Error metrics: MRBE and RRMSE per wind speed bin with DNV acceptance thresholds
  • Comparison charts: Reference vs unadjusted vs adjusted TI by wind speed bin
Even if your data doesn’t meet DNV criteria initially, the plots are useful for understanding the adjustment performance and identifying where improvements might be needed.

​
Next Steps