基于模糊重叠动态社区发现的电网局部自组织临界结构辨识  

Identification of Local Self-organized Critical Structure in Power Grids Based on Detection of Fuzzy Overlapping Dynamic Community

作　　者：魏震波[1] 罗紫航 关翔友 梁政 张伟林 何永祥 WEI Zhenbo;LUO Zihang;GUAN Xiangyou;LIANG Zheng;ZHANG Weilin;HE Yongxiang(College of Electrical Engineering,Sichuan University,Chengdu 610065,Sichuan Province,China;School of Electrical Engineering,Southwest Jiaotong University,Chengdu 610031,Sichuan Province,China;Yingxiuwan Hydro Power Plant,State Grid Sichuan Electric Power Company,Chengdu 611830,Sichuan Province,China;Suining Power Supply Company,State Grid Sichuan Electric Power Company,Suining 629000,Sichuan Province,China)

机构地区：[1]四川大学电气工程学院,四川省成都市610065 [2]西南交通大学电气工程学院,四川省成都市610031 [3]国网四川省电力公司映秀湾水力发电总厂,四川省成都市611830 [4]国网四川省电力公司遂宁供电公司,四川省遂宁市629000

出　　处：《电网技术》2023年第7期2916-2926,共11页Power System Technology

摘　　要：针对以元件级为对象的传统连锁故障关键环节辨识方法存在准确性差、计算量大且时效性低,难以支撑连锁故障预防和控制的问题,该文提出一种基于电网局部自组织临界(self-organized criticality,SOC)结构的区域级关键环节辨识方法。首先,在基于交流最优潮流的ORNL-PSerc-Alaska(OPA)模型的自组织过程中,考虑负荷波动与系统惯性,同时改进连锁故障与调度控制;其次,建立了考虑虚拟节点的动态有向加权电网模型,并计及电网特性对Newman快速算法进行改进;然后,提出了重叠社区发现的中心-重叠-外壳(center-overlap-shell,COS)模型,利用独立的动态社区发现对扰动下的电力社区进行辨识,并采用极大似然估计和p值估计识别局部SOC结构;最后,算例仿真表明:一般电网结构中,大都存在若干接近于或即将处于局部SOC的区域;该类区域在扰动或故障条件触发下能快速产生级联效应,演化为连锁故障,是扰动或故障发生、发展的关键环节。以上工作验证了所提思路和方法的合理性与有效性,为下一步制定连锁故障区域内事前预防和区域间事中控制奠定了理论依据。For the problem that the poor accuracy,heavy calculation and low timeliness in the traditional component-levelled critical link identification of the cascading faults,which is hard to support the prevention and control of the cascading faults,an area-levelled critical link identification based on the local self-organized critical(SOC)structure in power grids is proposed in this paper.Firstly,in the self-organized process of the ORNL-PSerc-Alaska model based on the AC optimal power flow,the load fluctuation and the system inertia are considered,and the cascading fault and the dispatch control are improved.Secondly,a dynamic directed weighted power grid model considering the virtual nodes is established,and the fast Newman algorithm is improved considering the power grid characteristics.Then,the center-overlap-shell model for overlapping community detection is proposed.The power community under disturbances is identified by the independent dynamic community detection,and the local SOC structure is identified by the maximum likelihood estimation and the p-value estimation.Finally,the simulations show that there exist several areas close to or about to be in the local self-organized criticality in the general power grid structure;When triggered by the disturbances or faults,such areas may produce chain effects and result in cascading faults quickly,which becomes the critical link of the disturbance or fault occurrence and development.The work above verifies the rationality and validity of the proposed ideas and methods,providing theoretical basis for the following prevention before occurrence within areas and control during the development between areas of cascading faults.

关 键 词：关键区域辨识 局部自组织临界 模糊重叠动态社区 连锁故障 

分 类 号：TM721[电气工程—电力系统及自动化]