Wind resource assessment using weather research and forecasting model. A case study of the wind resources at Havøygavlen wind farm

• Main Author: Fossem, Anders Aarhuus
• Format: Master Thesis
• Language: English
• Published: UiT The Arctic University of Norway 2019
• Subjects: VDP::Mathematics and natural science: 400::Information and communication science: 420::Mathematical modeling and numerical methods: 427; VDP::Matematikk og Naturvitenskap: 400::Informasjons- og kommunikasjonsvitenskap: 420::Matematisk modellering og numeriske metoder: 427; Wind resource mapping; Numerical weather models; WRF; Wind in complex terrain; VDP::Technology: 500::Environmental engineering: 610; VDP::Teknologi: 500::Miljøteknologi: 610; Renewable energy; Wind power; Wind power production; EOM-3901; North Cape
• Online Access: https://hdl.handle.net/10037/17252

The need for energy increases globally due to rapid expansion of population and prosperity. To meet this demand while decreasing carbon emission and eventually transition out fossil fuel, efficient utilization of wind power is prominent. This study evaluates the performance of the Weather Research and Forecast model (WRF) with respect to wind speed and wind direction. The area of the study is the northernmost wind farm site in the world, Havøygavlen. It is located just about 50 kilometers southwest of the North cape, consisting of a complex and coastal terrain. The model simulation period was the entire year of 2017, and the resulting estimates where compared to on-site data measured at hub height at each of the 16 turbines located at the site. In terms of forecasting capability, the Model was evaluated using correlation, Root Mean Square Error and Bias. The assessment showed little agreement and implementing finer resolution displayed no apparent improvements. The estimate was particularly vulnerable to sudden changes in wind speed, and performed more accurately in periods of low to moderate wind speeds. Annual weather resource assessment of the site was performed using box plots, annual average wind maps and wind speed histograms. The model is unsuccessful at capturing the high complexity of the terrain, ultimately leading to an underestimation of the wind resources. However, enhanced domain resolution improved the predictive performance, which agreed adequately with the on-site measurements. Furthermore, the annual average wind maps provided valuable knowledge about the local wind patterns surrounding the site. Annual wind roses and wind fields at specific times of high wind speed occurrence was used to evaluate the model’s estimated wind direction. Enhanced domain resolution showed improved directional stability and ability to capture the terrain’s effect on the wind before arriving at the site. As a preliminary wind resource tool, the model performs sufficiently, despite the complex terrain of the studied area.