Research of modeling method based on non-parametric kernel density estimation of probability of wind power fluctuations [基于非参数核密度估计的风功率波动性概率密度建模方法]

作者：Yang N.^*; Zhou Z.; Chen D.; Wang X.; Li H.; Li S.

来源：Acta Energiae Solaris Sinica, 2019, 40(7): 2028-2035.

• Bandwidth optimization
• Non-parametric kernel density estimation
• Ordinal optimization
• Probability density
• Wavelet decomposition
• Wind power fluctuation

Research on fluctuation probability distribution model of wind power can positively affect wind power integration and operation. This paper presents a probability characteristic model of wind power fluctuation based on non-parametric kernel density estimation method. Firstly, fluctuation components of wind power are extracted with wavelet decomposition. Then an optimization model applied to bandwidth selection is built. In order to solve bandwidth selection problem, a constrained bandwidth optimization model is constructed with goodness-of-fit test as its constraint. Finally, the bandwidth optimization model is solved with constrained ordinal optimization algorithm. Simulation results showed that the improved probabilistic density model of wind power fluctuation components possesses higher accuracy and applicability than traditional model built with parameter estimation method, and the constrained ordinal optimization algorithm can greatly improve computational efficiency of non-parametric estimation methods.

• 单位
  三峡大学