Identification of Feeder-transformer Connectivity in Distribution Network Based on Electrical Characteristics and Geographic Relevance

The accurate feeder-transformer connectivity of distribution network serves as a crucial data basis for power grid operations, such as loss calculation, operation and maintenance. The manual investigation method, however, incurs high cost and exhibits low efficiency, making it unsuitable for its application in the distribution network characterized by increasingly complex network structure and frequent changes. In view of this, a method to identify the feeder-transformer connectivity is proposed based on the measurement data and the equipment archives information. Firstly, the transformers that are electrically close to each other are merged based on the principle of voltage similarity, and then a virtual transformer is constructed as the subsequent input. After that, the evaluation indexes for feeder-transformer energy difference and voltage similarity are constructed respectively, followed by the establishment of an optimal identification model for feeder-transformer connectivity based on electrical characteristic indexation. Finally, to reduce the impact of measurement data quality on the model, the geographical location information of distribution transformer is added to the original model. Consequently, an improved optimal identification model considering geographical correlation is obtained. It has been verified in the actual distribution system that this method demonstrates significant advantages in improving the identification accuracy of line transformer relationship, and holds certain practical engineering application value as well as a guiding role.