Accuracy of WRF for prediction of operational wind farm data and assessment of influence of upwind farms on power production. (1st September 2022)

• Record Type: Journal Article
• Authors: Cuevas-Figueroa, Gabriel; Stansby, Peter K.; Stallard, Timothy
• Is Part Of: Energy. Volume 254:Part B(2022)
• Publication Date: 2022-09-01
• Subjects: Wind farms -- WRF -- Wake -- Wind power -- Renewable energy; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2022.124362
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: Numerical Weather Prediction models such as Weather Research and Forecasting (WRF) are increasingly used to inform planning of both on-shore and off-shore wind-farms. This study aims to establish the accuracy with which WRF can be employed to predict the operational performance of a wind farm located in complex terrain and to assess the impact of nearby wind farms on farm energy production. Analysis is presented for a wind farm operating in complex terrain located in Baja California, Mexico. The WRF model initialized with the North American Regional Reanalysis (NARR) data provides time-variation and annual occurrence of wind speed with root mean square error (RMSE) 12% and 1.4%, using either the Rapid Update Cycle (RUC) or Pleim-Xu (PX) land-surface models. Energy supply from a 10 MW wind farm is predicted to within 5.25% (RUC) and 2% (PX) of measured data. The wind resource at several sites in the region offers the potential for wind farms with capacity factors in the range 32–34% without wake interaction. This reduces to 29% when wind turbine momentum extraction within the farm is modelled. The wakes of these farms are found to extend over 12 km downwind, with greater extent for larger-diameter 5 MW turbines than 2 MW turbines. These farm wakes reduce power output from a downwind farm by over 20% for wind speeds below rated speed resulting in 1.3–1.7% reduction in annual energy. This analysis of WRF accuracy establishes confidence in use of numerical weather prediction tools for wind farm operational data and informing use for wind farm siting in relation to nearby farms. Highlights: Accuracy of WRF is established for wind speed and energy yield of an operational wind farm located in complex terrain. Highest accuracy is obtained using WRF with the PX Land Surface Model and NARR initialisation data. This configuration predicts annual wind occurrence and farm energy yield to within 1.4% and 2%. Wakes from wind farms over 10 km upwind, reduce mean power output by 5% and overall energy yield by 1.7%.