A Rapid Construction Method for a Knowledge Graph of Relay Protection Operation Management Based on Prompt Engineering

Relay protection plays a critical role in ensuring the safe and stable operation of modern power systems, especially as the scale and complexity of grids continue to grow. However, existing knowledge graph construction methods often rely heavily on manually defined rules or expert systems, resulting in inefficiency, poor scalability, and long update cycles. To overcome these challenges, this paper proposes a rapid construction method for a relay protection operation management knowledge graph based on prompt engineering and large language models (LLMs). By integrating natural language processing techniques with domain-specific prompt templates, the proposed method automatically extracts key information such as equipment components, fault types, and processing time limits from unstructured text. The resulting knowledge graph exhibits strong adaptability and real-time performance, supporting intelligent fault diagnosis and operational decision-making within power grids. Experimental results demonstrate that the method achieves a precision of 0.8367, recall of 0.9762, and $F 1$-score of 0.9011, with a total construction time of only 20.5 seconds. This approach significantly improves the efficiency and accuracy of knowledge extraction and offers a scalable and automated solution for intelligent operation and maintenance in smart grid systems.