Probabilistic generation model for grid connected wind DG

Abstract

The uncertainty associated with renewable energy sources, particularly wind, makes them an unpredictable means of power generation. To guarantee continuous and reliable power supply, wind speed variability modeling is considered a vital step for meeting the planning and operational challenges of an electric power grid. In this paper, a novel probabilistic generation model is developed to estimate and generate time-coupled wind speed patterns. A 1 h time step is considered to construct the time-coupled probabilistic wind speed patterns based on a two-parameter Weibull distribution. These parameters of the Weibull distribution are found by considering the variations of reference wind speed patterns at two successive time steps. The probabilistic model is then used to create a number of aggregate wind speed generation scenarios. The validity of our proposed approach is evaluated with the help of goodness-of-fit test indicators such as average mean absolute percentage error and the Kolmogorov-Smirnov test error. The results of goodness of fit tests and comparison of output power determined through the proposed model with the existing model in the literature suggest that the proposed model is appropriate for wind speed uncertainty modeling and can be applied in power system planning studies.