Distribution Network Feeder-Transformer Relationship Identification Method Based On Voltage Data And GIS Fusion

Distribution network feeder-transformer relationship is crucial for efficient management, optimised operation and improved service quality of power grids. The traditional manual inspection or power carrier wave method of investigating abnormal feeder-transformer relationships has the disadvantages of large workload, low efficiency and high cost, which is difficult to adapt to the increasingly complex and frequent changes of the modern distribution network. In order to solve this problem, this paper proposes a feeder-transformer relationship identification method based on the fusion of geographic information system (GIS) and voltage data in distribution networks. The method firstly addresses the three-phase load imbalance, and imputes the distribution transformer outlet voltage to solve the voltage asymmetry problem. Subsequently, in order to better characterise the anomalous feeder-transformer relationship and overcome the limitations of the traditional threshold judgement method, the method combines the feature extraction of voltage sequence data (e.g., similarity metrics, etc.) with the self-encoder algorithm to construct an anomalous feeder-transformer relationship identification model. At the same time, the method also cleverly integrates geographically distributed GIS data to provide auxiliary verification and correction means for the identification results, thus ensuring the reliability and consistency of the results. Through the example verification, the feeder-transformer relationship identification method based on GIS and voltage data fusion proposed in this paper shows good effectiveness and feasibility, and provides a new technical path to improve the level of distribution network operation and maintenance management. The innovation of this method is the organic integration of data-driven and knowledge-driven, which realises the unity of theory and practice, and has important theoretical significance and application value.