Description of the SCADA dataset of the EDP onshore wind farm in Portugal

Source: Energias de Portugal (EDP)
Coverage: 2016–2017
Format: Excel (.xlsx), 8 files, 219 MB total
Turbines monitored: T01, T06, T07, T11
Sampling interval: 10 minutes (UTC timestamps in ISO 8601 format)

The dataset contains four types of data:

SCADA signals ~417,000 rows across two files (one per year), each with 83 columns. Records operational telemetry per turbine per timestamp, including generator speed and temperatures, gearbox and hydraulic oil temperatures, rotor RPM, wind speed and direction, blade pitch angles, active and reactive power output, grid voltage/current/frequency, and temperatures across numerous mechanical and electrical components.

Meteorological mast data ~87,500 rows, 41 columns. Independent onsite environmental measurements including two anemometer channels (wind speed, direction), ambient temperature, pressure, humidity, precipitation, and rain detection, along with anemometer calibration parameters.

Event logs ~256,000 rows, 5 columns. Timestamped operational events and alarms per turbine with free-text remarks. The reset time column is almost empty, indicating most entries are one-sided event detections rather than resolved incidents.

Failure logs - 28 rows total, 4 columns. Manual records of confirmed component failures including gearbox, generator, transformer, and hydraulic group, with timestamps and short descriptions.

The SCADA signal files are virtually complete (0 NaN values in 2016; 8 NaN values in 2017). The met-mast files have minor gaps of fewer than 15 values each. The event logs have structurally empty columns by design.