Abstract

Wind farm time series power data is widely used in wind power integration studies and wind power forecasting. Historical time series wind power data is frequently corrupted by communication errors, wind turbine outages and curtailment, which introduce outliers. The identification and correction of outliers is important in ensuring the integrity of wind power integration studies, wind power forecasts and other analysis employing such data. An outlier identification methodology based on a probabilistic wind farm power curve and typical outlier characteristics is described. Additionally, an outlier correction methodology that exploits the spatial correlation between adjacent wind farms is proposed. These methods are tested on US modelled wind power data and applied to a case study of real operational wind power data from Northwest China. The cost and curtailment benefits of the identification and correction methods are demonstrated through a day-ahead unit commitment example.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call