A Method for Low-voltage Distribution Network Topology Verification Based on APAA and Convolution Kernel Matching

Aiming at the topology error problem, which is common in the current low-voltage distribution network, a low-voltage distribution topology error detection method based on adaptive piecewise aggregate approximation (APAA) and convolutional kernel matching is proposed in this paper. First, the original voltage time series data of the users are standardized using Z-Score normalization. After that, the dimensionality of the data is reduced using APAA. Then, the Sobel operator is applied to the processed data to generate a convolutional kernel matching matrix. Based on the principle of voltage similarity among nodes on the feeders in the low-voltage distribution network, an improved K-means clustering algorithm is employed to cluster the convolutional kernel matching matrix. This approach enables the verification of the topology and identification of transformer-household relationships in the low-voltage substation. The feasibility and effectiveness of the proposed method are validated through case study analysis.