Defect Prediction for Electric Power Communication Equipment Based on Temporal Knowledge Graph

As the scale of power facilities continues to expand, monitoring and assessing the operational status of power equipment have increasingly become key tasks in power system management. Equipment fault prediction is a crucial component of power system health monitoring. To address the challenges of fault prediction in large-scale power networks, this paper proposes a novel predictive model based on temporal knowledge graphs, integrating artificial intelligence techniques. The model effectively integrates multimodal data through an attention mechanism and captures the dynamic changes in both equipment and its related information over time by combining graph neural networks (GNN) with recurrent neural networks (RNN). By leveraging temporal representations, the model achieves accurate fault prediction. Additionally, the temporal knowledge graph reasoning approach based on GNN effectively integrates multiple information sources, improving fault prediction accuracy. Experimental results show that the proposed model significantly outperforms conventional methods in performance.