基于知识图谱技术的风险作业智能化管理研究  

Research on Risk Operation Intelligent Management Based on Knowledge Graph Technology

作　　者：赵丹 杨辉 罗舒元 郭丽雪 

机构地区：[1]中国石油西南油气田分公司川中油气矿,四川 遂宁

出　　处：《计算机科学与应用》2022年第12期2983-2993,共11页Computer Science and Application

摘　　要：作业风险因素管理和风险预警管控作为遏制重特大事故、实现安全生产超前预防的主要措施途径,对提升企业安全生产保障水平、提高政府安全监管效能具有至关重要的作用。据国务院安全生产委员会历年安全生产大检查情况通报显示,2007年至2015年全国共排查事故隐患5298万项,年均排查隐患589万项;海量事故隐患数据被科学地记录与存储,成为辅助安全生产的重要因素。然而,由于大数据存续时间较短,大数据分析与处理能力相对薄弱,大多数企业与机构仅能利用事故隐患数据指导整改工作,无法分析数据背后的潜在价值,对安全生产管理工作的指导作用有限,我国安全生产形势仍不容乐观。石油企业现有高风险作业类型多、数量大、管理要求高,作业的风险识别、风险等级划分、控制措施制定等风险因素因人员能力和素质差距导致分析的结果差异较大,同类风险的管控措施不一致、标准不统一,容易出现风险管控的漏项缺项,目前无有效手段将公司每年2万余项高风险作业数据信息进行整合利用,高风险作业管控难度大。本文研究的基于知识图谱技术构建的石油天然气行业风险作业知识库,运用人工智能技术,建立知识自动加工处理流程,通过深入研究和应用,实现风险作业的“知识标签”及“关联关系”。通过整合共享全公司风险作业步骤、作业风险因素、风险控制措施等数据,提升作业风险数据的应用价值,减少同类不安全行为、不安全状态发生的频率,实现各基层单位风险防控措施标准化管理,提醒高风险作业全过程管理的风险源、风险控制措施,能够更高效、规范和安全的开展风险作业,建立人工智能及知识图谱技术平台,构建自主学习词库,完善和丰富高风险作业知识体系。综上,通过智能化信息手段,利用知识图谱技术建立风险作业行业知识图谱,采集Operation risk factor management and risk early warning control, as the main measures to curb major accidents and realize advance prevention of production safety, play a crucial role in improving the level of production safety guarantee of enterprises and improving the efficiency of government safety supervision. From 2007 to 2015, 52.98 million hidden dangers were investigated nationwide, with an average of 5.89 million being checked annually. Massive accident data are rec-orded and stored scientifically, which becomes an important factor to assist production safety. However, due to the short existence time of big data, big data analysis and processing ability is rela-tively weak, most enterprises and organizations can only use the data of hidden accidents to guide the rectification work, unable to analyze the potential value behind the data, the guidance role of the safety production management is limited, our safety production situation is still not optimistic. The existing high-risk operations of petroleum enterprises have many types, large quantities, and high management requirements. The risk factors such as risk identification, risk grade classification, and control measures formulation of the operations, due to the gap in personnel ability and quality, lead to large differences in the analysis results. The control measures of similar risks are inconsistent and the standards are not uniform, which is prone to the occurrence of missing items in risk control. At present, there is no effective means to integrate and utilize the data information of more than 20,000 high-risk operations of the company every year, which makes it difficult to control high-risk operations. In this paper, the knowledge base of risk operations in oil and gas industry is built based on knowledge graph technology. Artificial intelligence technology is used to establish automatic knowledge processing process. Through indepth research and application, the “knowledge label” and “association relationship” of risk operations are re

关 键 词：风险作业 专业词库 知识图谱 智能搜索 精准推送 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]