A Sensor-Driven Wind-Farm Parameterization in WRF Using a Skewed Gaussian Wind-Turbine Wake Model

Le résumé fourni par la source

Abstract The coarse resolution of numerical weather prediction models limits their ability to account for turbine-to-turbine wake interactions within wind farms. The widely adopted Fitch’s wind-farm parameterization neglects wake interactions between turbines located within the same grid cell, leading to nonnegligible errors in performance estimation. Existing models that address this limitation rely on using the wake-affected flow to drive engineering wake models, introducing a risk of double-counting wake effects. To overcome these limitations, we introduce a new wind-farm parameterization in which turbine-to-turbine wake interactions are represented using a skewed Gaussian wake model driven by wind conditions sampled from the atmospheric flow. To avoid double-counting wake effects, sensors are positioned around the simulated farm to capture the undisturbed inflow while accounting for the spatial variability in the incoming wind field. The new parameterization is implemented in the Weather Research and Forecasting (WRF) Model, v4.4.3, and is evaluated through both idealized and real-case WRF simulations of the Anholt wind farm. Under idealized conditions, the new parameterization reduced the mean error in power production relative to power recordings from 17%–42% (using Fitch’s parameterization) to 4%–8%, depending on the simulation conditions. For real-case atmospheric conditions, the new parameterization reduced the mean error of the farm’s power production by around 50% compared to Fitch’s parameterization. The new parameterization also exhibited less sensitivity to both the model’s grid spacing and the uncertainty associated with the turbines’ coordinates. Overall, the proposed parameterization enhances the representation of wind-turbine wakes within numerical weather prediction models, enabling improved assessment of wind-farm performance.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.