Power Curve Analysis of the Wieringermeer Wind Farm : Comparing ground-based LiDAR, met-mast, and nacelle anemometer wind speeds and evaluating the power curve under different wind conditions

Sammanfattning

In the global transition to renewable energy, wind power has undoubtedly positioned itself as one of the most competitive energy sources to replace fossil fuels. A major advantage of wind power is that the wind is not only a renewable resource, but also free. The wind however, is a complex phenomenon, often challenging to predict but certainly measurable. The success of a wind power project is heavily tied to the ability to make accurate and cheap measurements of the wind, which de-risk investments. Even after an investment has been made, when operation has started, it is valuable to measure the wind and power output at a wind farm. At this stage, developers and stakeholders want to ensure that their investment is performing in line with expectations. More specifically, in line with the warranty given by the turbine supplier. This is done in a power curve verification which has become a popular measurement campaign. In the Netherlands, at the Vattenfall wind farm Wieringermeer such a verification was done by a third party using ground-based LiDAR (Light Detection And Ranging, hereafter referred to as lidar). In this project, the power curve verification was taken further, by first successfully recreating the results by the third party and then by expanding on them. In this expanded analysis, performance outside of the wind conditions in the power curve warranty was analysed, and correlations between wind speeds from the lidar, a meteorological mast (met-mast) anemometer and nacelle anemometers on site were studied. The results show that certain extreme wind conditions substantially affect the power curve and the resulting annual energy production, particularly factors such as wake‑affected wind directions, where measurements become unrepresentative of the wind reaching the rotor, and higher shear exponent (α), which appeared to positively affect the power curve. Filtering with varying turbulence intensity (TI) resulted in minor changes to performance. When comparing different measurement techniques, it was found that the correlations between lidar, met-mast, and nacelle anemometer wind speeds show the highest agreement when all signals are wake‑free. Filtering the data for α between 0 and 0.2 improved the lidar–nacelle and the lidar-mast agreement, primarily in lidar-waked conditions. Filtering the data for TI between 0 and 0.4 did not greatly strengthen the lidar – nacelle or lidar – met-mast correlation. The TI parameter showed a higher relative importance in the relationship between these signals when the lidar is waked. Additionally, the study found improvements to be made in future investigations such as recreating filters for removing curtailed data more accurately, and using other measurement setups, such as a lidar and met-mast more closely located, or using a nacelle-mounted lidar.