Diagnosing systematic differences in predicted wind turbine array-array interactions

Abstract

Improved quantification of the spatial extent and intensity of wind farm wakes is urgently needed given the rapid pace of expansion of installed capacity both on-and off-shore. We present and analyse long-term, high-resolution simulations of whole wind farm wakes conducted for real-world wind turbine installations performed with the two wind farm parameterizations (EWP and Fitch) designed for use with the Weather Research and Forecasting model. We document differences in the formulation of these two parameterizations and demonstrate their impact on simulated wind farm wakes. Divergence between the schemes in terms of wake spatial extent and magnitude is maximized under low/moderate turbulent kinetic energy (TKE < 0.3 m2s-2) and wind speeds between cut-in and rated (U ≈ 4-12 ms-1). Thus, it is under those conditions that model predictions of the intensity/spatial extent of wind farm wakes are inferred to have highest uncertainty. A framework is introduced based on these simulations that can be used to aid planning for experiments such as AWAKEN. It could be used to identify where and when observational data would be most beneficial in differentiating relative skill of the two parameterizations and identifying areas where modifications to the schemes are necessary to improve fidelity.