Improved Topology Identification Method for Low-voltage Distribution Networks by Combining Quadratic Programming and Clustering

This study proposes an improved topology identification method combining quadratic programming and clustering for low-voltage distribution networks. Adding regular terms to the traditional quadratic programming objective function effectively solves the problem of cross-layer solution and instability when solving multi-layer topology. According to the relationship between voltage correlation and line distance, the voltage correlation is firstly used to cluster the energy meters. Then, the quadratic programming solution is performed on the sub-clusters after clustering, which realizes the identification of the belonging relationship between meter boxes and energy meters and the topology identification of the low-voltage distribution network. The validation shows that the algorithm is highly accurate, stable, and universal to different types of equipment.