基于数据驱动的配电网拓扑识别及线路阻抗估计

Date-Driven Distribution Network Topology Identification and Line ImpedanceEstimation

中低压配电网的拓扑识别及线路阻抗参数估计是未来智能配电网实现各种功能的基础。依托AMI(高级量测体系)提供的电量信息,提出了一种仅依靠配电网节点电压及功率数据驱动的中低压配电网拓扑识别及线路阻抗估计方法。利用核密度估计方法计算各节点电压数据间的互信息并据此分析各节点间相关性;根据图的最小生成树算法生成以邻接矩阵形式表示的配电网拓扑;结合线性回归及Distflow潮流模型对拓扑进行校正,检验拓扑中是否存在AMI系统中没有相关数据的汇流节点并计算线路阻抗;最终得到准确的配电网拓扑及线路阻抗参数。通过IEEE 33节点中压配电网及典型低压配电网算例对所提方法进行了验证,结果表明所提方法能够准确辨识拓扑及线路阻抗参数,即使在感知设备不足的低压配电网中仍有较好的辨识效果。

The topological structure and line impedance parameters of low-and medium-voltage distribution networks are the foundation of realizing various functions of the future intelligent distribution networks.Given the shortcomings of traditional distribution network topology identification methods such as poor timeliness and inability to accurately obtain line impedance parameters,a topology identification and line impedance estimation method for low-and medium-voltage distribution networks driven only by voltage and power data of distribution network nodes is proposed with the backing of power information provided by AMI(advanced metering infrastructure).Firstly,the kernel density estimation method is used to calculate the mutual information(MI)among node voltage data and analyze the correlation among nodes accordingly.Secondly,the distribution network topology is generated in the form of an adjacency matrix according to the minimum spanning tree algorithm.Thirdly,the topology is corrected by combining linear regression and Distflow model to check whether there are outflow nodes without relevant data in the AMI system,and the line impedance is calculated.Finally,the accurate distribution network topology and line impedance parameters are obtained.The proposed method is validated by using examples of IEEE 33-node mediumvoltage distribution networks and typical low-voltage distribution networks.The results show that the proposed method can accurately identify the topology and line impedance parameters even in the absence of sensing equipment for low-voltage distribution networks.