基于改进SBR的风电出力典型场景提取与分析  

Extraction and analysis of wind power typical scenarios based on the improved SBR algorithm

作　　者：余姚果 梅亚东[1] 王现勋 朱迪 吴贞晖 张祥 YU Yaoguo;MEI Yadong;WANG Xianxun;ZHU Di;WU Zhenhui;ZHANG Xiang(State Key Laboratory of Water Resources and Hydropower Engineering Science,Wuhan University,Wuhan 430072,China;College of Resources and Environment,Yangtze University,Jingzhou 434023,China;Central China Branch of State Grid Corporation of China,Wuhan 430077,China)

机构地区：[1]武汉大学水资源与水电工程科学国家重点实验室,湖北武汉430072 [2]长江大学资源与环境学院,湖北荆州434023 [3]国家电网公司华中分部,湖北武汉430077

出　　处：《武汉大学学报（工学版）》2021年第4期346-353,共8页Engineering Journal of Wuhan University

基　　金：国家自然科学基金资助项目(编号:91647204);国家电网科技项目(编号:SGHZ0000DKJS1900230)。

摘　　要：为解决传统聚类算法无法兼顾时间序列数量和形状特征、聚类数确定依据不足等问题,提出了一种基于量-形距离的同步回代缩减算法(simultaneous backward reduction algorithm based on quantity-contour distance,SBRQC)。首先,将反映风电出力过程线数量、形状、峰谷差等特性指标引入同步回代缩减(simultaneous backward reduction,SBR)算法,对风电出力过程场景进行缩减,得到不同场景数下的风电出力过程及对应概率。其次,综合考虑类内相似度和类间差异度,提出新的聚类有效性指标场景缩减(scenario reduction,SD)作为确定提取场景数的依据。最后,以人工数据集和某区域电网风电日出力过程为例,验证了SBR-QC算法的合理性和准确性。In order to solve the problems that traditional clustering algorithms cannot take into account both the number and shape characteristics of time series,and insufficient basis for determining the number of clusters,a simultaneous backward reduction algorithm based on quantity-contour distance(SBR-QC)is proposed.First,the simultaneous backward reduction(SBR)algorithm is used to reduce the wind power output process scenarios by introducing characteristic indicators that reflect the number,shape,peak-valley difference of wind power output process,and obtain the wind power output process and corresponding probability under different scenarios.Secondly,a new clustering validity index scenario reduction(SD)is proposed as a basis for determining the number of extracted scenes by comprehensively considering the intra-class similarity and interclass difference.Finally,the artificial data set and the daily wind power generation process of a certain regional power grid are taken as examples to verify the rationality and accuracy of the SBR-QC algorithm.

关 键 词：风电出力 典型场景 有效性指标 聚类数 

分 类 号：TM71[电气工程—电力系统及自动化]