Synthetic Simulation Data of Wind Power Time Series in Germany

This dataset contains synthetic wind power generation time series for 160 locations in Germany. The data is provided at an hourly resolution covering a two-year period. It was generated using an open-source, reproducible simulation framework described in the publication "Synthetic Data Simulation Framework for Wind Power Time Series". To address different modeling needs, the dataset is structured into three main folders:

era5_wind_hourly_age: Synthetic data based exclusively on state-of-the-art ERA5 numerical reanalysis data.

dwd_era5_wind_hourly_age: Synthetic data based on local DWD weather station measurements. Wind speed extrapolation utilizes dynamic vertical wind shear calculated from ERA5 multi-level data between 10 and 100 m.

icon_d2_nwp_time_series: Accompanying numerical weather prediction (NWP) data from the ICON-D2 model. This features a 48-hour forecast horizon from 6, 9, 12, and 15 UTC runs.

Key features of the simulation framework include: Hub Height Air Density: Air density is dynamically calculated using an extended barometric formula. This avoids the inaccurate assumption of a constant standard air density. Aging Effects: Power coefficients (Cp) are dynamically adjusted based on turbine age. This uses an annual degradation rate (ADR) of 0.63%. It applies a real-world age distribution from the German market data register (MaStR). Physics-Informed Simulation: The final output is calculated using the physical power equation. This applies the extrapolated meteorological values and the adjusted Cp. The time series files contain both the meteorological variables at ground/hub height and the synthetic power generation for six common turbine configurations:

Enercon E-70 E4 (57m),
E-82 E2 (138m),
E-115 (149m),
Vestas V90 (95m),
V112-3.45 (119m),
V80-1.8 (78m).

Metadata tables detailing turbine specifications and site parameters are also included. The framework was comprehensively validated against Renewables.ninja and operational data from 13 real wind parks across nine German states. Results prove that the ERA5-driven approach and the integration of turbine aging significantly improve long-term accuracy, particularly for older fleets. This dataset is highly suitable for benchmarking, developing machine learning forecasting models, grid planning, and energy system analysis.