A critical analysis of the uncertainty in the production estimation of wind parks in complex terrains

Assessing the wind resource in complex terrains with the level of accuracy required by industry is not a trivial task. The present study critically analyzes the different elements involved in the process and how they affect the predicted Annual Energy Production (AEP) of a wind farm. Five different sites in Italy, for which the installing company had a posteriori AEP data available, are considered. All sites are modeled with the windPRO software. To separate the effects related to the quality of experimental measurements from those due to the selected wind modeling techniques, two different approaches were followed. In the “horizontal” one, the minimum amount of wind data common to all installation sites were used, i.e., the measurements of one mast at one selected altitude for a single full year, thus highlighting the role of the terrain modelling. In the “vertical” approach, the analysis focused on the site with the most complex terrain, using wind data of different quality and duration, as well as different processing techniques, from the ubiquitous WAsP to WAsP-CFD. The results of the study confirm that, even in presence of high-quality measurements and a standard workflow, the accuracy of the farm's AEP estimation is strongly site-dependent. For mildly-complex terrains, the lack of accuracy in mast data can be compensated with a more advanced spatial scaling method, and vice versa. On the other hand, in presence of steep slopes or sneaky ridges, attention must be paid to both the experimental and modelling components. A scaling approach based on Computational Fluid Dynamics (CFD) is mandatory. In all cases, thermal effects such as atmospheric stability must be carefully evaluated and included in the assessment process.