 |
| Titel |
Multi-decadal Variability of the Wind Power Output |
| VerfasserIn |
Nicolas Kirchner Bossi, Ricardo García-Herrera, Luis Prieto, Ricardo M. Trigo |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250097978
|
| Publikation (Nr.) |
 EGU/EGU2014-13612.pdf |
|
|
|
|
|
| Zusammenfassung |
| The knowledge of the long-term wind power variability is essential to provide a realistic
outlook on the power output during the lifetime of a planned wind power project. In this
work, the Power Output (Po) of a market wind turbine is simulated with a daily resolution for
the period 1871-2009 at two different locations in Spain, one at the Central Iberian Plateau
and another at the Gibraltar Strait Area.
This is attained through a statistical downscaling of the daily wind conditions. It
implements a Greedy Algorithm as classificator of a geostrophic-based wind predictor, which
is derived by considering the SLP daily field from the 56 ensemble members of the longest
homogeneous reanalysis available (20CR, 1871-2009). For calibration and validation
purposes we use 10 years of wind observations (the predictand) at both sites. As a result, a
series of 139 annual wind speed Probability Density Functions (PDF) are obtained, with a
good performance in terms of wind speed uncertainty reduction (average daily wind speed
MAE=1.48 m/s).
The obtained centennial series allow to investigate the multi-decadal variability of
wind power from different points of view. Significant periodicities around the 25-yr
frequency band, as well as long-term linear trends are detected at both locations.
In addition, a negative correlation is found between annual Po at both locations,
evidencing the differences in the dynamical mechanisms ruling them (and possible
complementary behavior). Furthermore, the impact that the three leading large-scale
circulation patterns over Iberia (NAO, EA and SCAND) exert over wind power output is
evaluated. Results show distinct (and non-stationary) couplings to these forcings
depending on the geographical position and season or month. Moreover, significant
non-stationary correlations are observed with the slow varying Atlantic Multidecadal
Oscillation (AMO) index for both case studies. Finally, an empirical relationship
is explored between the annual Po and the parameters of the Weibull PDF. This
allowed us to derive a linear model to estimate the annual power output from those
parameters, which results especially useful when no wind power data is available. |
|
|
|
|
|
|