基于Copula-POME的负荷与气象因素相关性度量研究

负荷与气象因素的关系是非线性且模糊的，针对传统的线性相关系数不能准确刻画负荷与其气象成因的相依结构。在分析负荷对气象因素响应的基础上，提出了结合Copula函数与最大熵原理（POME）的负荷与气象因素相关性度量方法，该方法基于POME建立了负荷与气象因素的边缘分布，利用Copula函数拟合了负荷与气象多变量系统中的非线性相依结构，并推导了度量相关性的Kendall秩相关系数、Spearman秩相关系数和Copula熵。在实际的负荷和气象系统中的应用表明，Copula-POME方法在分析负荷与其气象成因关系时无先验分布假定，具有灵活的函数形式，能准确表达多变量系统的相依结构；秩相关系数和Copula熵弥补了线性相关系数在度量尾部相关中的不足，能准确度量负荷与气象因素的相关性。

Research on the Measurement of Correlation between Load and Meteorological Factors Based on Copula-POME

The relationship between load and meteorological factors is non-linear and fuzzy. Traditional linear correlation coefficients cannot accurately describe the dependent structure of load and its meteorological causes. Based on the analysis of load response to meteorological factors, a method for measuring the correlation between load and meteorological factors is proposed by combining the Copula function and the Principle of Maximum Entropy (POME). The marginal distribution of load and meteorological factors was established based on POME, and the non-linear dependent structure of load and meteorological multivariable system was fitted using the Copula function. The Kendall rank correlation coefficient, Spearman rank correlation coefficient, and Copula-entropy were derived to measure the correlation. The application in actual load and meteorological system shows that the Copula-POME method has no prior distribution assumption when analyzing the relationship between load and its meteorological genesis, and has a flexible function form, can accurately express the dependent structure of multivariable systems; The rank correlation coefficient and Copula-entropy can make up for the lack of linear correlation coefficients in measuring tail correlation, and can accurately measure the correlation between load and meteorological factors.