考虑时序性的风电序列场景集生成方法  

Wind Power Sequence Scene Set Generation Method Considering Timing

作　　者：汪柳兵 毕锐[1] 丁明[1] 张震[1] 王杰[1] WANG Liubing;BI Rui;DING Ming;ZHANG Zhen;WANG Jie(Anhui Key Lab of New Energy Utilization and Energy Conservation(Hefei University of Technology),Hefei 230009,China)

机构地区：[1]安徽省新能源利用与节能重点实验室(合肥工业大学),合肥230009

出　　处：《电力系统及其自动化学报》2021年第10期81-88,共8页Proceedings of the CSU-EPSA

基　　金：国家自然科学基金区域创新发展联合基金资助项目(U19A20106)。

摘　　要：针对目前单时段/多时段场景集生成方法在保留风电序列时序性方面的不足,提出一种自适应预测箱和状态转移矩阵相结合的风电序列场景集生成方法:在单时段分析阶段,结合自适应预测箱技术生成静态场景集;在多时段分析阶段,以状态转移矩阵滤除风电序列多时段转移过程的不合理波动,经过迭代过程生成既满足风电随机特性,又满足时序性要求的风电序列场景集。以德国TenneT电力公司的风电数据为例,在风电序列时序特征改进、与实际场景的差异性、算法计算效率和面向优化调度的应用效果方面与其他方法进行对比分析,验证所提方法的新颖性、合理性及实用价值。Aimed at the shortcomings of the current single-/multi-period scene set generation methods in retaining the timing of wind power sequences,a wind power sequence scene set generation method that combines the adaptive prediction box and state transition matrix is proposed.At the single-period analysis stage,a static scene set is generated by combining the adaptive prediction box technology.At the multi-period analysis stage,the state transition matrix is used to filter out unreasonable fluctuations in the multi-period transfer process of wind power sequences,and a wind power sequence scene set is generated during the iteration process,which not only satisfies the random characteristics of wind power,but also meets the timing requirements.The wind power data of TenneT(a power company in Germany)is taken as an example,and the proposed method is compared and analyzed with other methods in terms of the improvement in wind power sequence timing characteristics,difference with actual scenes,the calculation efficiency of algorithm,and application effects for optimal scheduling,thereby verifying the novelty,rationality and practical value of the proposed method.

关 键 词：自适应预测箱 状态转移矩阵 场景分析 序列场景集 时序性 

分 类 号：TM721[电气工程—电力系统及自动化]