Beta分布在风电预测误差模型中的适用性

Applicability of Beta distribution on wind power forecast error modeling

Beta分布凭借自身的特征优势,已经成为建立风电功率预测条件误差的主要模型。当前,大多数有关Beta分布建模的研究都集中在储能规划和经济调度方面,而针对复杂多变的风电功率预测条件误差,研究Beta分布适用性影响因素的成果较少。为了填补这方面的不足,论文做了如下工作:研究了Beta分布的参数和形态特征;根据这些特征给出了条件误差的非标准定义形式、异常数据的四分位区间处理方法,以及评估Beta分布适用性指标和步骤;针对不同容量的实际功率数据,完成了Beta分布在分组数量、异常点检测、参数估计方法、功率特征等关键因素影响下的适用性仿真;并根据仿真结果总结了Beta适用性的变化规律和一些经验准则。

Beta distribution has become the main model to establish wind power forecast conditional error(WPFCE)by its own characteristics.At present,most of the research results on Beta distribution modeling are focused on energy storage planning and economic dispatch.In view of the complexity and variability of WPFCE,there are few studies on the factors affecting the applicability of Beta distribution.In order to fill the shortcomings,the following work has been done in this paper.Firstly,the parameters and morphological characteristics of Beta distribution have been studied.Then,this paper gives the non-standard definition of conditional errors,the interquartile range method of outliers,and the indicators and steps for assessing the applicability of Beta distribution based on these characteristics.Finally,according to the actual power data of different capacities,the applicability simulation of Beta distribution under the influence of key factors such as partition numbers,outliers detection,estimation methods and power characteristics is completed,and the changing rules of Beta applicability and some empirical criteria are summarized according to the simulation results.