文献类型：期刊文献

英文题名：Exploring the Application of Large Language Model-Assisted Risk Identification in HAZOP Analysis

作者：Bao, Haotian[1]; Wang, Bing[1]; Du, Wenli[1]

机构：[1] School of Information Science and Engineering, East China University of Science and Technology, Shanghai, China

年份：2026

起止页码：1158

外文期刊名：38th Chinese Control and Decision Conference, CCDC 2026

收录：EI(收录号：20262821063013)

语种：英文

外文关键词：Cost benefit analysis - Risk analysis - Risk assessment - Safety engineering - Wages

摘要：As a qualitative risk assessment method highly dependent on expert experience, Hazard and Operability Analysis (HAZOP) faces challenges such as long analysis cycles, tedious tasks, and high labor costs in practice. However, since HAZOP follows a standardized reasoning process and logical framework, it is an ideal scenario for Large Language Model (LLM)-assisted safety analysis. This paper develops an automated HAZOP analysis framework based on LLM. By constructing a six-tuple knowledge representation, the framework successfully transforms Piping and Instrumentation Diagram (P&ID) and Instrument Logic Diagrams (ILD) into structured data, overcoming the limitations of LLM in physical topology perception and finegrained entity recognition. Four frontier Large Language Models validate the feasibility of this framework in industrial risk identification in terms of coverage, expansion rate, and validity. ? 2026 IEEE.