KG-BGCN: A Secondary Power Equipment Classification Model Based on Knowledge Graphs

In modern power systems, secondary electrical equipment is critical in ensuring grid stability. However, the complexity of network environments and evolving security threats pose significant challenges to accurately identifying and managing these assets. This paper proposes a classification model that combines knowledge graph (KG) and deep learning to identify device manufacturers in power systems. By integrating multi-source data into a unified KG, the proposed model, KG-BGCN, captures both attribute and structural information of devices. Experimental validation on real-world data from substations demonstrates that KG-BGCN outperforms traditional models in device type classification, achieving higher top-k accuracy and F1 scores and offering a robust power system device management solution.