Variability in large-scale wind power generation

Juha Kiviluoma (Corresponding Author), Hannele Holttinen, David Weir, Richard Scharff, Lennart Soder

Research output: Contribution to journalArticleScientificpeer-review

54 Citations (Scopus)


The paper demonstrates the characteristics of wind power variability and net load variability in multiple power systems based on real data from multiple years. Demonstrated characteristics include probability distribution for different ramp durations, seasonal and diurnal variability and low net load events. The comparison shows regions with low variability (Sweden, Spain and Germany), medium variability (Portugal, Ireland, Finland and Denmark) and regions with higher variability (Quebec, Bonneville Power Administration and Electric Reliability Council of Texas in North America; Gansu, Jilin and Liaoning in China; and Norway and offshore wind power in Denmark). For regions with low variability, the maximum 1?h wind ramps are below 10% of nominal capacity, and for regions with high variability, they may be close to 30%. Wind power variability is mainly explained by the extent of geographical spread, but also higher capacity factor causes higher variability. It was also shown how wind power ramps are autocorrelated and dependent on the operating output level. When wind power was concentrated in smaller area, there were outliers with high changes in wind output, which were not present in large areas with well-dispersed wind power.
Original languageEnglish
Pages (from-to)1649-1665
JournalWind Energy
Issue number9
Publication statusPublished - 2016
MoE publication typeA1 Journal article-refereed


  • wind power
  • variability
  • net load
  • variable generation
  • power systems


Dive into the research topics of 'Variability in large-scale wind power generation'. Together they form a unique fingerprint.

Cite this