Study on Probability Distribution of Wind Power Fluctuation Based on NACEMD and Improved Nonparametric Kernel Density Estimation

Abstract

For operation control process of large-scale wind farms, it is of great significance to accurately build probability distribution model of wind power fluctuation characteristics. According to the noise-assisted signal decomposition method based on complex empirical mode decomposition and adaptive nonparametric kernel density estimation method, a method of wind power fluctuation modeling is proposed. Firstly, the wind power is decomposed with the noise-assisted signal decomposition method based on complex empirical mode decomposition, and the fluctuation is extracted. Secondly, the probability characteristics are modeled based on upgraded nonparametric kernel density estimation and adaptively promoted based on the model. Finally, constrained ordinal optimization algorithm is utilized to solve the model. Accuracy and applicability of the modeling, and effectiveness of the model improvement are verified with simulation.