Characterizing and modeling aggregate wind plant power output in large systems

A fundamental challenge in integrating wind plants into a power system is the inherent stochastic nature of the power they output. In order to identify appropriate operational and technological solutions to integrating wind plants, it is important to characterize the uncertainty, variability and temporal patterns of the power they output. This paper analyzes historical wind power data from the Bonneville Power Administration, the Electric Reliability Council of Texas and the Midwest ISO to qualitatively and quantitatively characterize the wind power in these systems. From the analysis, probabilistic models of the power output, variations of power output and diurnal patterns are developed. Common probability density functions are fit to the data and the strength and timing of diurnal patterns are identified. The resulting parameters of the distribution can be used to model aggregate wind power output in large systems, which has applications in wind integration analysis and for benchmarking purposes. The results of the analysis quantify the challenges of wind plant integration faced by the system operators in each of the studied systems.