ERA5-Driven Simulation for Synthetic Wind Power Time Series Data in Germany

Description

This dataset provides synthetic wind power time series for 200 locations across Germany, produced with a measurement- and ERA5-driven simulation framework that accounts for turbine aging. The framework derives hub-height air density from an extended barometric formula, extrapolates wind speed with an ERA5-derived dynamic power law, and applies an age-resolved Weibull degradation model calibrated against real wind farm data. It targets locations without on-site wind measurements and is validated against operational data from 13 German wind farms. Each of the 200 locations coincides with a DWD weather station. The dataset covers 24 July 2023 to 30 April 2026 at hourly resolution. Three file types are included: a turbine parameter table for six wind turbine models (rotor diameter, hub height, cut-in/cut-out/rated wind speed, rated power), a location table with coordinates, altitude, and commissioning date per site, and 200 time-resolved tables, one per location. Each location table has three column groups. ERA5-derived meteorology: wind components at 10 m and 100 m, wind speed, friction velocity, sensible heat flux, boundary-layer height, temperature, dew point, pressure, relative humidity, and air density. Per-turbine synthetic output, one block per turbine model: hub-height wind speed, hub-height air density, a surface-layer validity flag, and power output. Co-located DWD measurements: wind speed and direction, temperature, relative humidity, and pressure, kept with their native measurement gaps and without imputation. Turbine age is sampled per unit from the empirical age distribution of the German onshore turbine fleet, derived from the Marktstammdatenregister (MaStR). The six turbine models were selected from the most common models installed in Germany, chosen to maximize heterogeneity in capacity, hub height, and power curve. The dataset additionally includes numerical weather prediction (NWP) data from the ICON-D2 model for all 200 locations, covering a 48-hour forecast horizon from runs at 9 UTC, with multi-level wind, temperature, and pressure fields.

Steps to reproduce

1. Clone the simulation framework code from the GitHub repository accompanying this dataset. 2. Download ERA5 reanalysis data for the 200 station locations and the period 24 July 2023 to 30 April 2026. Required fields: u/v wind at 10 m and 100 m, 2 m temperature, 2 m dew point temperature, surface pressure, friction velocity, sensible heat flux, and boundary-layer height. 3. Download 10-minute DWD station measurements for the same 200 stations and period from the DWD Climate Data Center (wind speed, wind direction, temperature, relative humidity, pressure). 4. Download Copernicus DEM GLO-30 elevation data and the CORINE Land Cover 2018 map to derive topographic metrics (elevation, slope, aspect, topographic position index, terrain diversity index) and roughness length per location. 5. Download turbine technical specifications (rotor diameter, hub height, cut-in/cut-out/rated wind speed, power curve) for the six selected turbine models from the wind turbine models database. 6. Download turbine commissioning dates from the Marktstammdatenregister (MaStR) to derive the empirical age distribution of the German onshore turbine fleet. 7. Run the preprocessing step to compute relative humidity and air density at 2 m from the ERA5 temperature, dew point, and pressure fields. 8. Run the wind speed extrapolation step. This computes the ERA5-derived dynamic exponent from the 10 m and 100 m wind levels and applies the power law to obtain hub-height wind speed for each of the six turbine configurations. 9. Run the air density extrapolation step to obtain hub-height air density from the extended barometric formula. 10. Sample one age value per turbine unit from the empirical age distribution (step 6) and apply the age-resolved Weibull degradation model to the turbine power curve. 11. Compute power output per hour and turbine configuration from the extrapolated hub-height wind speed, air density, and the aged power curve. 12. Compute the surface-layer validity flag per hour and turbine configuration by comparing hub height to 0.1 times the ERA5 boundary-layer height. 13. Assemble the turbine parameter table, the location parameter table, and the 200 per-location time series tables with the ERA5-derived, per-turbine, and DWD-measurement columns. 14. Optionally, download ICON-D2 forecast data at runs 9, UTC to add the 48-hour NWP columns. ICON-D2 data is only available for the most recent 24 hours and must be collected in real time; it is not archived publicly. 15. Validate the reproduced dataset against the published version by comparing summary statistics per location and turbine configuration.

Institutions

• Hochschule Karlsruhe Technik und Wirtschaft

  Baden-Württemberg, Karlsruhe

Categories

Funders

• Bundesministerium für Forschung, Technologie und Raumfahrt

  North Rhine-Westphalia, Germany

  Grant ID: 13FH587KX1

Related Links

Licence