Come meet the poster presenters to ask them questions and discuss their work
Check the programme for our poster viewing moments. For more details on each poster, click on the poster titles to read the abstract. On Wednesday, 6 April at 15:30-16:15, join us on Level 3 of the Conference area for the Poster Awards!
PO267: Seasonal effects in the long-term correction of short-term wind measurements: Comparison of linear models and machine learning approaches
Alexander Basse, Research Associate, University of Kassel
To estimate the wind potential at a site, wind measurements are corrected to a long-term wind climate using simple statistical methods. According to today's guidelines, one year of measurement data is necessary. In our work, we investigated how different methods, including linear models as well as machine learning approaches, can be used to shorten the measurement period and, thus, significantly reduce measurement costs. Different reanalysis data sets are used to correct measurement data from 11 sites in Germany. The results show that systematic, seasonal biases occur, hence, a dependence of the results on the measurement period (season). These biases are strongly connected to the applied method and show remarkably different behavior for two different linear models (Variance Ratio and Linear Regression). Using a simple machine learning technique such as random forest, however, decreases the seasonal biases in mean wind speed significantly.