| 引用本文: | 刘丽军,吴桐,翁涵迪,郑文迪,徐启峰.基于标签传播的配电网动态分区方法[J].电力自动化设备,2021,41(12): |
| LIU Lijun,WU Tong,WENG Handi,ZHENG Wendi,XU Qifeng.Dynamic partitioning method of distribution network based on label propagation[J].Electric Power Automation Equipment,2021,41(12): |
|
| 摘要: |
| 为实现接入分布式可再生能源(RDG)的配电网分区在结构与能量层面上的区内强耦合、区间弱耦合以及分区内部功率储备充足的目标，除广泛应用于电力系统分区的模块度指标外，引入复杂网络理论中的标签传播算法和稳定度指标。为降低分区后各分区在能量与结构上的耦合程度，提高分区后各区内部电源的控制能力，定义了适用于配电网分区的标签传播算法；利用潮流追踪算法得出配电网电源供能和负荷吸能网络，基于此定义稳定度指标以量化节点与节点、节点与分区在能量上的紧密程度。并结合以灵敏度矩阵为基础的模块度指标，从结构和功能的角度推演出标签传播的判断依据：增益函数与损失函数。最后，考虑RDG接入配电网对分区的影响，将标签传播算法与RDG出力预测场景构建方法相结合，并以动态分区的形式提高分区应对RDG不确定性的能力，为分层分区调控方法提供理论支撑。 |
| 关键词: 配电网 标签传播算法 模块度 稳定度 潮流追踪 动态分区 |
| DOI：10.16081/j.epae.202111020 |
| 分类号:TM761 |
| 基金项目:福建省自然科学基金资助项目(2017J01480) |
|
| Dynamic partitioning method of distribution network based on label propagation |
|
LIU Lijun, WU Tong, WENG Handi, ZHENG Wendi, XU Qifeng
|
|
College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108, China
|
| Abstract: |
| In order to achieve the goals of strong coupling within partition, low coupling between partitions and sufficient power reserve inside partition in terms of structure and energy for partitioning of distribution network connected with RDG(Renewable Distributed Generation),in addition to the modularity index widely used in partitioning of power system, the label propagation algorithm and permanence index in complex network theory are introduced. In order to reduce the coupling degree of energy and structure between partitions and improve the control ability of the internal power supplies of partitions after partitioning, a label propagation algorithm suitable for distribution network partitioning is defined. The power supply and the load energy absorption networks of distribution network are obtained by using power flow tracking algorithm. Based on this, the permanence index is defined to quantify the tightness degree about energy of bus to bus and bus to partition. Combined with modularity index based on sensitivity matrix, the judgment basis of label propagation such as gain and loss function is deduced from the point of structure and function. Finally, considering the impact of RDG accessed to distribution network on partition, the label propagation algorithm is combined with prediction scenario construction method of RDG output, and the ability of the partition to deal with the uncertainty of RDG is improved in form of dynamic partitioning, which provides theoretical support for hierarchical partition regulation methods. |
| Key words: distribution network label propagation algorithm modularity permanence power flow tracing dynamic partitioning |
