Evaluation of probabilistic models of wind plant power output characteristics

Abstract

The power output by weather-driven renewable resources such as wind energy conversion systems can be appropriately described as being stochastic. To manage these resources, probabilistic models of wind power are being increasingly employed by power system stakeholders in applications such as stochastic unit-commitment programs and wind power forecast systems. This paper evaluates probabilistic models-specifically the probability density functions-of aggregate wind plant power output and conditional and unconditional variations of aggregate wind plant power output. The parameters of the models are fit to historical aggregate wind plant power data from three large North American systems. Parametric and non-parametric evaluations of the suitability of the models are performed in the form of χ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> goodness-of-fit tests and through the inspection of probability plots and histograms. It is shown that Beta distributions are appropriate models for the aggregate power output and Laplace distributions are appropriate models for wind power variability. Conditional wind power variation follows a generalized extreme value distribution.