A Research on Gas Safety Knowledge Graph and Retrieval-Augmented Generation Mechanism Based on Large Language Model

Along with the advancement of modernization, gas safety has become more and more an area of concern. However, there is a lack of research on the mechanism of applying large language models(LLM) in the emergency handling of various types of gas safety events and the popularization of safety knowledge. In this paper, we construct a scheme for generating gas safety knowledge graph and retrieval-augmented generation based on big language model. Specifically, firstly, a knowledge graph is constructed by extracting entities from the dataset using LLM and prompt word engineering. Then, evaluation metrics and measurement methods are designed for this scheme. Finally, the feasibility of the proposed method is verified by experiments.