Distribution Network Electrical Topology Identification Based on Edge Computing and Improved KNN

The rapid development of smart grids puts forward high requirements on the fine management of the distribution network side, which based on distribution network electrical topology identification. Facing the big data during the operation of the distribution network, edge computing has obvious advantages. Therefore, this paper studies the electrical topology identification algorithm of the distribution network based on edge computing. In this method, the phase of user in each station area is identified by the edge gateway. Based on the principle that the similarity of the voltage curves of the same phase is higher than that of different phases in a station area, this paper proposes an improved KNN (K-Nearest Neighbor) algorithm. In this algorithm, the traditional KNN algorithm is improved by using correlation coefficient instead of Euclidean distance as the distance metric. In addition, this paper proposes a training set update mechanism, which adds the tested voltage data into the training set. The analysis results of the examples show that this algorithm has high accuracy, and the verification of multiple stations proves its effectiveness.