Statistics of wind farm power output: methods and applications

Statistical and spectral analysis techniques commonly used for characterising wind power output are reviewed. Techniques including correlation and linear regression, power spectra and the coherence function are explained using short example time series. The techniques are then applied to hourly wind speed data from three UK sites. Frequency diagrams of power output, of hourly and 4-hourly power output changes and generation duration curves are created. Statistics as a tool for studying the impact of wind power integration is discussed with focus on the measure of standard deviation.