基于知识图谱的核电厂安全审评辅助决策系统设计与工程小样实现  

Design and Engineering Sample Implementation of Nuclear Power Plant Safety Review Decision-Making Aided System based on Knowledge Graph

作　　者：郭超 依岩 骆文 梁侃慧 李春 Guo Chao;Yi Yan;Luo Wen;Liang Kanhui;Li Chun(Nuclear and Radiation Safety Center,MEE,Beijing 100082,China;Shanghai Nustar Nuclear Power Technology Co.,Ltd.,Shanghai 201103,China)

机构地区：[1]生态环境部核与辐射安全中心,北京100082 [2]上海核星核电科技有限公司,上海201103

出　　处：《核安全》2024年第3期55-62,共8页Nuclear Safety

摘　　要：为了响应国家核安全监管体系和监管能力现代化的倡导,进一步提升核电厂安全审评的质量和效率,生态环境部核与辐射安全中心开展了核电厂安全审评优化技术研究。围绕核安全审评业务流程,构建了整体优化业务架构和数据架构,设计了核电厂安全审评辅助决策系统;通过对核电厂严重事故审评专业特性研究,建立严重事故审评知识图谱并嵌入系统实现了工程小样开发。该系统可利用自然语言处理模型对审评相关报告进行解读,并基于所嵌入知识图谱对相应领域审评相关报告进行辅助分析。研究成果可为核电厂安全审评工作的开展提供技术支持。In response to the advocacy of nuclear safety regulatory authorities on the modernization of nuclear safety regulatory system and regulatory capacity,and to further the quality and efficiency of nuclear safety review,research on optimization technology for nuclear power plant safety review has been carried out by Nuclear and Radiation Safety Center.The overall optimized business architecture and data architecture have been constructed around the nuclear safety review business process,and the nuclear power plant safety review auxiliary decisionmaking system has been designed;By studying the professional characteristics of severe accident review in nuclear power plants,a knowledge graph of severe accident review was established and embedded in the system to achieve engineering samples.The system can use natural language processing models to interpret review related reports and assist in analyzing corresponding field review related reports based on the embedded knowledge graph.The research results can provide technical support for the safety review of nuclear power plants.

关 键 词：核安全审评 严重事故 知识图谱 辅助决策系统 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]