Wind Turbine SCADA Data From 36 Turbines For Anomaly Detection

89 cumulative years of real-world SCADA data from 36 wind turbines across 3 farms, organized into 95 datasets. This collection, originally published as 'CARE to Compare: Wind Turbine Anomaly Detection Dataset,' is designed for developing early fault detection algorithms. The dataset was uploaded by author 'kevykibbz' to Hugging Face in February 2025.

Use Cases

• Detect operational anomalies using SCADA time-series features like turbine speed, power output, and temperature.
• Develop predictive models for component failure by analyzing patterns in vibration or bearing temperature data.
• Compare anomaly detection algorithm performance across 95 distinct datasets representing different turbines and fault conditions.
• Analyze the impact of environmental conditions on turbine health by correlating SCADA metrics with weather data.

Strengths

• 89 cumulative years of operational data provides a substantial temporal scope for analysis.
• Data originates from 36 real-world turbines across 3 wind farms, offering diversity in operational conditions.
• The dataset is structured into 95 distinct subsets, facilitating cross-validation and robust algorithm testing.

Limitations

• Specific row counts, column names, and data completeness for each SCADA metric are unknown.
• The dataset's geographic origin and specific temporal range for each turbine are not detailed, limiting contextual analysis.
• Potential class imbalance between normal and faulty operation records is not quantified, which could affect model training.

Provenance

Source

Originally published as the 'CARE to Compare: Wind Turbine Anomaly Detection Dataset.'

Collection Method

Collected from real-world Supervisory Control and Data Acquisition (SCADA) systems on operational wind turbines.

Time Range

Covers 89 cumulative years of operation; specific start and end dates per turbine are unknown.

Freshness

Dataset was last updated on the Hugging Face platform in February 2025.

Geography

Data comes from turbines located across 3 wind farms; specific locations are unknown.