Posters

Abstract

Introduction   As the EU targets 300 GW of offshore wind by 2050, offshore wind farm density will rise significantly, and wakes spanning multiple farms may reduce output at neighboring sites. This work focuses on assessing the performance of wake modeling tools in simulating interfarm wakes, considering the influence of atmospheric conditions.  Innovative contributions   A central question is whether the representation of atmospheric processes, such as temperature-momentum coupling and inversion dynamics, influences the predicted wake extent. The novelty of this work lies in systematically examining the role of realistic atmospheric dynamics in wake extent predictions through a direct comparison of engineering and mesoscale models.  Method    Engineering wake models from PyWake [2] (NOJ, Gaussian, TurboGaussian) and the MesoNH mesoscale model [3] are applied over the full year 2020 using ERA5 inflow conditions. To this end, the Abkar and Porté‑Agel parametrization was implemented in MesoNH, therefore enabling the simulation of large windfarms in the model. This study focuses on direct model intercomparison, analyzing predicted powers and wake lengths as a function of atmospheric characteristics. Results are also qualitatively compared to WINS50 data generated with the LES model ASPIRE [4]. Yet comparing WINS50 data with observation data shows a positive bias in wind speed, as documented in [4], limiting their use as reference.  Case study   The case study focuses on the Gemini offshore wind farm, located 75 km north of the Dutch coast. Gemini consists of two sub-farms, Gemini West and Gemini East, each with 75 identical turbines and separated by approximately 5 km along a west-east axis. This configuration allows evaluation of wake effects over medium distances (interactions between the two sub-farms) and longer distances (combined Gemini wake extending further downstream).  Results   Results are binned by inflow speed and direction, with a focus on the 180°-270° sector as it captures the transition for Gemini East (downstream) from partial to full wake shadowing by Gemini West (upstream).  Medium-distance farm-to-farm interactions are quantified through power production differences between Gemini West and East. PyWake results indicate that the NOJ model underestimates farm-to-farm wake losses, while TurboGaussian overestimates them and the Gaussian models better reproduce the trends observed with WINS50 data for Gemini.   For longer-distance effects, wake lengths are systematically examined as a function of boundary-layer height to better understand atmospheric control on far-reaching wakes.  Conclusion   This study examines how engineering and mesoscale models represent mid- and far-reaching windfarm wakes. These insights are crucial for accurate annual energy production estimates as offshore wind farm density continues to grow. Preliminary results show engineering models vary in accuracy, while mesoscale simulations will clarify the role of atmospheric conditions in the development and modeling of far-reaching wakes.  References [1] Porté-Agel et al., Wind-Turbine and Wind-Farm Flows: A Review, Boundary-Layer Meteorology, 2020.  [2] Pedersen et al., PyWake 2.5.0: An open-source wind farm simulation tool, 2023, DTU Wind.  [3] Lac et al., Overview of the Meso-NH model version 5.4 and its applications, Geoscientific Model Development, 2018.  [4] Baas, Winds of the North Sea in 2050 - Public final report, Whiffle, 2024.