基于贝叶斯网络的油码头装卸设备溢油失效可能性研究  

Oil-spilling Failure Assessment of Handing Facilities in Oil Wharf Based on Bayesian Network

作　　者：江文[1] 蒋宏业[1] 蒋安荔 罗珊[1] 

机构地区：[1]西南石油大学石油与天然气工程学院,四川成都610500 [2]中国石化天然气川气东送管道分公司,湖北武汉430000

出　　处：《石油工业技术监督》2015年第6期8-12,共5页Technology Supervision in Petroleum Industry

基　　金：国家科技支撑计划课题"基于风险的油气管道事故预防关键技术研究"(编号:2011BAK06B01-11-01)

摘　　要：油码头装卸设备失效是导致溢油并引起火灾爆炸、海洋污染等事故的因素之一,因此,采用科学的方法对设备进行失效可能性研究具有重要意义。首先,运用故障树的贝叶斯网络化方法,构建油码头装卸设备溢油故障的贝叶斯网络拓扑结构;其次,基于模糊集理论,利用Ge NIe软件进行概率推理,定量评估设备失效可能性,可得引起装卸设备故障的主要因素是输油臂密封圈漏油和输油泵快速接头故障;最后,结合后验概率进行事故致因分析,得出最大致因链,发现装卸设备的薄弱环节。实例分析表明,利用贝叶斯网络推理分析,较传统的故障树分析更具合理性,概率分布更为准确。The failure of handing facilities in oil wharf is one of the factors resulting in oil spilling, fire explosion, and marine pollution incidents, and therefore it is very important to assess the failure possibility of the handing facilities using scientific methods. Firstly, Bayesian network topology for the oil spill failure of the handing facilities in oil wharf is established by using Bayesian net-work method; secondly, the possibility of the oil spill failure of the handing facilities is quantitatively assessed based on fuzzy set theory and using GeNie software, and the assessment result shows that the main factors causing the oil spill failure of the handing facilities are the oil spilling at the seal ring of oil loading arm and the failure of quick coupling of oil loading pump; finally, find out the weaknesses of the handling facilities aecording to the analysis of accident causes. The analysis of a ease shows that, Bayesian network inference analysis is more reasonable than traditional fault tree analysis, and the obtained probability distribution is more accurate.

关 键 词：油码头装卸设备 贝叶斯网络 模糊集理论 失效可能性 事故致因 溢油失效 

分 类 号：TE58[石油与天然气工程—油气田开发工程]