基于动态贝叶斯网络的高后果区油气管道风险评价-cfb95baf.pdf

故障与安全

DOI：10.3969/j.issn.1001-2206.2023.05.006

基于动态贝叶斯网络的高后果区油气管道风险

评价

123121

胡奕，荆晶，陈敏，张逸宁，秦虹，梁昌晶

1.中国石油华北油田分公司，河北任丘062252

2.中国石油华北油田分公司第一采油厂，河北任丘062252

3.中国石油华北油田分公司第二采油厂，河北霸州065700

摘要：为实时动态反映油气管道在高后果区的风险水平，结合蝴蝶结（BT）和动态贝叶斯网络（DBN）模

型对高后果区管道泄漏行为进行了分析，利用模糊概率和LeakyNoisy-orGate模型分别确定DBN模型的先验

概率和条件概率，并在人为因素、腐蚀情况等节点引入时间维度进行动态设置，得到管道泄漏概率和事故后果

发生概率的时变特性。结果表明：管道泄漏概率呈先增大后减小再增大的趋势，在投产初期和运行后期的泄漏

概率较大，应引起足够重视；管道占压和违章施工是引发高后果区管道泄漏的主控因素；随着时间的延长，不

同事故的发生概率逐渐增大，事故发生概率从大到小依次为油气聚集、喷射火、闪火、油池火、火球和蒸气云

爆炸。研究结果可为油气管道高后果区的动态系统分析和安全管理提供实际参考。

关键词：贝叶斯网络；动态；油气管道；高后果区；风险评价

Riskassessmentofoilandgaspipelinesinhighconsequenceareasbasedon

dynamicBayesiannetwork

123121

HUYi,JINGJing,CHENMin,ZHANGYining,QINHong,LIANGChangjing

1.PetroChinaHuabeiOilfieldCompany,Renqiu062252,China

2.No.1OilProductionPlantofPetroChinaHuabeiOilfieldCompany,Renqiu062252,China

3.No.2OilProductionPlantofPetroChinaHuabeiOilfieldCompany,Bazhou065700,China

Abstract：Inordertodynamicallyreflecttherisklevelofoilandgaspipelinesinthehighconsequenceareainrealtime,thebowknot

(BT)andthedynamicBayesiannetwork(DBN)modelarecombinedtoanalyzethepipelineleakagebehaviorinthehighconsequence

area.FuzzyprobabilityanalysisandLeakyNoisy-orGatemodelareusedtodeterminethepriorprobabilityandconditionalprobability

oftheDBNmodel,respectively.Inaddition,thetimedimensionisintroducedintothenodesofhumanfactorsandcorrosionconditions

fordynamicsettings,andthetime-varyingcharacteristicsofpipelineleakageprobabilityandaccidentconsequenceprobabilityare

obtained.Theresultsshowthattheprobabilityofpipelineleakageincreasesfirst,thendecreases,andthenincreases.Theprobabilityof

pipelineleakageislargerintheearlystageofproductionandlaterstageofoperation,whichshouldbepaidenoughattentionto.Pipeline

occupationandillegalconstructionarethemaincontrollingfactorsofpipelineleakageinhighconsequenceareas.Withtheextensionof

time,theoc