Consumer-transformer Relationship Identification Based on UMAP and Mini Batch K-Means

The accuracy of the relationship between single-phase users and transformers in low-voltage areas has an important impact on the analysis of line loss management. At the moment, the identification of consumer-transformer relationship mainly relies on manual checking, which is lack of efficiency. Therefore, an efficient and reliable method for identifying the relationship between consumer and transformer is urgently needed. In this paper, a consumer-transformer relationship identification method based on voltage data recorded by a smart meter is proposed, which provides a reference for the line loss analysis in low power distribution network. Firstly, the UMAP technique is chosen to reduce the dimensionality of the voltage data to deal with the redundancy caused by the high dimensionality of the original load characteristics of the users in the station area. After that, the mini batch K-means algorithm is applied to realize the identification of the consumer-transformer relationship in the low-voltage area. Finally, the proposed method is validated in simulation models. The results indicate that the method works.