Construction and Secondary Optimization of a Voltage Quality Control Knowledge Graph Based on Large Language Models

To address the challenges in voltage quality control within emerging power systems, this paper proposes a knowledge graph construction method based on large language models (LLMs). By leveraging the structured representation capabilities of knowledge graphs, the method effectively integrates multi-source control rules and reveals implicit logical relationships, thus overcoming the limitations of traditional experience-driven approaches. Experimental results show that, after optimization, the number of nodes and relations in the graph increased by 25.4% and 47.3%, respectively, while the F1 score improved from 0.520 to 0.729, which demonstrates significant enhancements in semantic consistency and inference capability.