超短期风电功率预测误差数值特性分层分析方法  

Stratification Analysis Approach of Numerical Characteristics for Ultra-short-term Wind Power Forecasting Error

作　　者：叶林[1] 任成[1] 赵永宁[1] 饶日晟 滕景竹 

机构地区：[1]中国农业大学信息与电气工程学院,北京市海淀区100083

出　　处：《中国电机工程学报》2016年第3期692-700,共9页Proceedings of the CSEE

基　　金：国家自然科学基金项目(51477174;51077126)~~

摘　　要：风电功率预测误差特性分析可以为电力系统优化调度与稳定运行提供更加准确的参考。该文提出一种根据超短期风电功率预测误差概率密度特性对误差进行分层,再依据误差波动性和不同层误差幅值特性进行分类处理的预测误差数值特性分析方法。在概率密度特性提取部分,采用改进后的广义误差分布模型对预测误差概率密度分布进行拟合。该误差分析方法结合了误差模型预测和误差概率密度拟合两种方法的优点,可以更为准确地对超短期风电功率预测误差进行分析和补偿。算例分析结果表明,改进广义误差分布模型的拟合效果优于正态分布、柯西分布和拉普拉斯分布这些常用模型,尤其在尾部特性拟合方面效果更为明显,所提出的误差分层分析方法可以有效减小风电功率预测误差。Characteristics analysis of wind power forecasting error can provide more accurate reference for optimal dispatch and stable operation of power system. This paper proposed a numerical error characteristics analyzing approach which stratified the errors into different layers according to the probability density of the ultra-short-term wind power forecasting errors. Then, the errors were processed separately by their fluctuation and amplitude characteristics. An improved generalized error distribution（GED） model was used to fit the error probability density distribution. The proposed analyzing approach combined the advantages of error forecasting model and error probability density fitting methods. It is more accurate to analyze and compensate the ultra-short-term wind power forecasting errors. Results of case studies indicate that the improved generalized error distribution performs much better than Normal distribution, Cauchy distribution and Laplace distribution, especially in fitting the tails section. The stratification analysis approach is effective to reduce ultra-short-term wind power forecasting errors.

关 键 词：超短期风电功率预测 广义误差分布 分层分析 误差补偿 

分 类 号：TM85[电气工程—高电压与绝缘技术]