Wind farm capacity factor forecasting: An Australian case study

Abstract

Wind energy ranks second among all renewable energy sources in generating power. The F(v) is a common probability distribution for determining the frequency of wind speeds. When evaluating wind data, the scale and shape characteristics of the Weibull distribution should be considered. The average wind speed corresponds to the scale factor, whilst the shape parameter represents the standard deviation of the data. Several attempts have been made, using models or formula, to establish a direct link between the Weibull parameters and the capacity factor of the wind farm. Nevertheless, closer inspection of these methods reveals their error rate is much higher than real wind farm capacity factor estimates. This work aims to provide a more workable and practical approach that provides more realistic numbers by connecting the Weibull shape and scale parameters to the capacity factor. Predicting the capacity factor values for wind farms using these parameters is possible. This research examined data from a wind farm on the northwest coast of Tasmania, Australia, from July 1, 2007, to August 31, 2008. Throughout this time, the farm produced power and recorded wind speeds every ten minutes. These observations were used to compute the wind farm's actual capacity factor and average wind speed. The Weibull shape and scale parameters were then determined. Additional computations were performed to determine Weibull parameters and the capacity factor for various periods after dividing the data into half-yearly, quarterly, monthly, 10-day, 5-day, and daily intervals. These interval results show that every combination of shape and scale parameter values is associated with a certain value of the capacity factor, irrespective of the period over which the computations are performed—monthly, yearly, weekly, etc. Consequently, wind speed data available for brief periods may predict the capacity factor value for wind farms over longer time intervals.