Validierung und Verbesserung von Nachlaufmodellen zur standortspezifischen Last- und Leistungsberechnung in Windparks

Abstract

Wind energy plays a leading role in the expansion of renewable energies. Even though renewable energies have experienced immense growth in the past decade, the amount of energy is currently insufficient to fully replace conventional forms of energy, which is why the increase in installed wind energy capacity must continue to be driven forward. Particularly in densely populated countries like Germany, the efficient use of available land holds special importance. In order to use the land optimally, wind farms are built as densely as possible, which results in low distances between turbines. Due to these short distances, wake effects gain influence. These lead to power losses as well as increased loads due to neighboring wind turbines. This dissertation addresses the issue and validates and extends existing wake models for onshore wind farms with dense spacing. Current wake models are validated based on wind field measurements applying lidar systems as well as power, load, and met mast measurements in two onshore wind farms with low distances between the turbines. Special focus is devoted to the usage of nacelle-mounted lidar systems for wake model validation. Moreover, lidar measurements have been used to recalibrate the dynamic wake meandering (DWM) model. As a result, the model proves to be more accurate regarding wind characteristics and fatigue loads under wake conditions at onshore sites with small turbine distances and flat terrain. Additionally, different methods to evaluate multiple wakes in the DWM model are evaluated regarding the wind characteristics in the wake as well as power and fatigue loads. Finally, an extension of the DWM model towards a static model for site-specific load approximations is presented and proves to be a significant improvement to the commonly used Frandsen wake-added turbulence model, especially for short turbine distances. In a common site-specific load calculation process, time-consuming aeroelastic simulations are usually avoided and the loads are estimated based on interpolations of prior-performed load simulations (e.g. response surface method). The current version of the DWM model is not combinable with these load approximations. The static version of the DWM model addresses this issue and provides a damage-equivalent single turbulence intensity value. The extension of the model is built in such a way that the computational costs are very low, thus enabling an implementation into wind farm layout optimization processes.