During the development phase of a wind farm, a significant investment is made in the met campaign. Met towers are strategically installed and remote sensing devices are deployed across the proposed project area. Many questions arise during this process including: When can we deem the met campaign complete? When can we confidently say that enough sites have been installed and that the addition of more met sites is no longer reducing the uncertainty?
Wind Flow Model Error versus Number of Met Sites
An effective way to answer these questions is to perform a wind flow model Round Robin analysis. This lets you track how the model error changes as the number of met sites used in the model increases. As soon as the model error plateaus and no longer improves with the addition of another met site, the met campaign can be considered complete.
Conducting Round Robin Analysis in Continuum
In the Continuum (patent pending) wind flow modeling software, there is a built-in Round Robin analysis tool that can be used for this type of exercise. In the Uncertainty Analysis module (shown in the screenshot below), the user selects the met subset size, clicks the button “Do Round Robin Analysis” and then Continuum finds every possible combination of met sites. It generates a site-calibrated model with each combination of met sites, which are used to predict the wind speed at the met sites excluded from the model.
If a Round Robin analysis is conducted using multiple met subset sizes, the plot and table in the upper left-hand corner will display the Round Robin Root-Mean-Square (RMS) error for each met subset size. The results can be exported to a .CSV file to be analyzed in another program.
Is the Met Campaign Complete?
At a project site with eleven met sites, a Round Robin analysis was conducted for every met subset size in Continuum. The results are shown below – the Round Robin RMS error is plotted against the met subset size. The model error plateaus at approximately 1.3% at a met subset size of 8. Since the addition of met sites is no longer reducing the model uncertainty, the met campaign can be deemed complete.
At a different project site, a total of eight met sites have been installed and a similar Round Robin analysis was conducted in Continuum. The Round Robin RMS error for each met subset size is shown below in the table and plot. As more met sites are added to the model creation, the RMS error steadily declines and continues on a downward trend at a subset size of seven met sites. The met campaign at this site is not complete since the model error has not yet reached a plateau.
Conducting a full Round Robin analysis, which uses all met subset sizes is an objective and effective method to both track the impact of new met sites on the wind flow model uncertainty and to determine when enough met sites have been installed.
To perform this type of analysis at one of your project sites, visit our website at www.cancalia.com and request a free trial license of Continuum or send us an email at info@cancalia.com and let us run the analysis for you.