Wind Power Probability Model Based on Spatiotemporal Correlation

In this paper, a method to establish the output probability distribution model of wind farm is proposed. According to the spatiotemporal characteristics of wind, the time and space correlation of wind power are analyzed respectively, and the output of wind farm is obtained. The kernel density estimation method and Gaussian mixture model are proposed to fit the wind power probability density distribution. The EM algorithm is used to solve the parameters of the Gaussian mixture model, and the error evaluation index is statistically established to analyze the wind power error. By comparing the actual operation data of a wind farm in a certain area with the kernel density estimation method, the superiority of the Gaussian mixture model is proved. The model can describe the probabilistic characteristics of wind farm output well and provide a valuable reference for operators in wind power dispatching operation.