基于自适应多变量非参数核密度估计的多风电场出力相关性建模

Modeling of Output Correlation of Multiple Wind Farms Based on Adaptive Multivariable Nonparametric Kernel Density Estimation

研究多个风电场的联合概率密度,对于风电大规模并网及电力系统运行控制具有重要意义。该文提出一种基于自适应多变量非参数核密度估计的多风电场联合概率密度建模方法。首先以可变带宽代替固定带宽建立一种自适应的多变量非参数核密度估计模型,然后针对模型带宽选择问题,构造了一种以欧氏距离和最大距离为拟合性指标的带宽优化模型,最后利用序优化算法对其进行求解。实际算例仿真结果表明,该文方法不仅较传统基于copula函数的参数估计方法具有更高的精度和适用性,而且还较好地解决了传统多变量非参数核密度估计方法的局部适应性问题。

The study of joint probability density of multiple wind farms is of great significance for large-scale wind power grid connection and power system operation control. In this paper, a method of joint probability density modeling for multiple wind farms based on adaptive multivariable nonparametric kernel density estimation was proposed. First, an adaptive multivariate nonparametric kernel density estimation model was built with the variable bandwidth instead of fixed bandwidth. Then, in order to select the optimum model bandwidth, a bandwidth optimization model was constructed based on Euclidean distance and maximum distance. Finally, ordinal optimization algorithm was used to solve the model. Examples illustrate that this method not only has higher accuracy and applicability than the traditional parameter estimation method based on copula function, but also solves the local adaptability problem of traditional multivariate nonparametric kernel density estimation method.