A Dual-scale Low-voltage Distribution Network Topology Identification Method Based on AMI Data Driving

To improve the operational stability of distribution grids and solve the increasingly serious topology error problem in low-voltage distribution, this paper proposes a dual-scale low-voltage distribution network topology identification method based on AMI data driving. Firstly, an improved K-means algorithm is proposed, combining the first-order derivatives and the original data, and a new idea of distance calculation is adopted to cluster and analyze the user appliance data collected by AMI. Then, according to the principle of correlation of electrical data, the results of the analysis are compared with the topology information recorded in the station archives, and the topological connections with possible errors are searched out. Finally, the validity of the method proposed in this paper is verified after relevant case analysis. Unlike traditional topology identification methods, this method not only reduces labor and hardware costs, but also has higher applicability and can better cope with the dynamic changes and complexity of power system topology.