Posters

Abstract

The development of wind farms often starts in early stages with some pre-screening processes. An important aspect to de-risk the viability of a given project relies on the energy content directly tied to the future financial wealth of the wind farm. The wind resource at the site is to be assessed and on-site measurement is not always a possible option at such stage.  The developers need to rely on other ways to get an idea of the energy content at the site, including indicative assessment without on-site measurement, production data or nearby known wind speed estimates. However, most of those means of assessments heavily rely on wind speed rescaling and a lot of value will be put on the estimated site long term wind speed. However, the energy content of the site also depends on the wind speed distribution and direction which are inherent to the data used for those assessment: reanalysis datasets or atlas wind roses. In this work, DNV has used its extensive internal mast database to provide a wide regional benchmark of the quality of the most used reanalysis datasets and wind atlas. This study focused on the quality and reliability of the frequency distribution and wind roses when rescaled to an expected mean wind speed and the corresponding impact on gross and net energy. Regional bias per region and reanalysis datasets or sources are discussed and the rescaling of datasets and Weibull fitting coherence challenged. The main goal is to show the wind industry the do’s and don’ts of using reanalysis datasets as the base data and to point out potential risks tied to each reference.