考虑不完全信息的电力系统黑启动方案评估方法  

Evaluation method of power system black start scheme considering incomplete information

作　　者：冷亚军 刘鹏飞 时浩 LENG Yajun;LIU Pengfei;SHI Hao(College of Economics and Management,Shanghai University of Electric Power,Shanghai 200090,China)

机构地区：[1]上海电力大学经济与管理学院,上海200090

出　　处：《电力科学与技术学报》2020年第6期99-109,共11页Journal of Electric Power Science And Technology

基　　金：国家自然科学基金(71601109)。

摘　　要：合理筛选黑启动方案是电力系统安全防御的重要措施。区别于以往完全信息下黑启动方案评估问题的研究,该文主要解决信息不完全情况黑启动方案的优选问题。提出一种基于AP聚类和Slope One填补的黑启动方案评价方法。首先,采用AP算法对黑启动方案进行聚类,生成相似的方案集合;然后,根据同一聚类中黑启动方案的评价值,采用Slope One算法对存在的评价空值进行预测填补;最后,基于填补后完备的评价信息,采用标准差权重法和TOPSIS法完成对所有黑启动方案的完全排序。在广东电网黑启动数据上对该文方法进行验证,比较该文方法与4种现有完备信息黑启动评估方法的优劣。实验结果表明,该文方法可以有效处理不完全黑启动评价信息,得到较为满意的黑启动方案优选结果。The reasonable selection of black-start scheme is significant to the fast restoration of power systems however most previous studies are based on complete evaluation information without consideration for cases with incomplete evaluation information.Under this background,the Affinity Propagation(AP)algorithm and Slope One method are adopted to assess black-start schemes with incomplete evaluation information in this paper.Firstly,the Affinity Propagation algorithm is utilized to cluster the black-start schemes to generate similar sets of schemes.Then,the Slope One algorithm is employed to predict the absent values according to the evaluation value of the black-start schemes in the same cluster.Finally,the standard deviation and the TOPSIS methods are applied to rank the schemes.The relative data of Guangdong power system are selected to evaluate the performance of the proposed method.Four existing black-start decision making methods with complete information are compared with the proposed method.The result shows that the proposed method solves the problem of incomplete evaluation information well.

关 键 词：不完全信息 黑启动 近邻传播聚类 Slope One算法 

分 类 号：TM72[电气工程—电力系统及自动化]