40检测与完整性|Inspection&Integrity2023年1月第42卷第1期网络出版时间:2022-11-16T16:09:03网络出版地址:http://kns.cnki.net/kcms/detail/13.1093.TE.20221115.1706.002.html元坝高酸性气田地面管道内腐蚀预测孙天礼1韩雪2黄仕林1赵春兰2何欢1曾德智21.中国石化西南油气分公司采气二厂;2.西南石油大学·油气藏地质及开发工程国家重点实验室摘要:针对元坝高酸性气田地面管道的内腐蚀问题,提出了一种考虑多因素的BP神经网络腐蚀预测模型。该模型以温度、CO2分压、H2S分压、pH值、Cl-含量、总矿化度、液气比、缓蚀剂残余量8种影响因素的数据作为输入量,再将腐蚀速率作为输出量,通过现场实测获得大量历史样本数据,对BP神经网络进行训练,实现了地面管道的腐蚀速率预测,并利用该模型对元坝高酸性气田地面管道各腐蚀因素的重要程度进行评判。结果表明:随机抽取工况参数,模型预测值与实测值的平均绝对误差在10%以内,预测模型具有较高的准确性和可靠性;影响元坝高酸性气田地面管道腐蚀速率的主控因素为H2S分压,CO2分压、缓蚀剂残余量次之。研究成果可为类似气田地面管道的内腐蚀评估提供技术借鉴。(图2,表3,参22)关键词:高含硫气田;站内管道;腐蚀速率预测;BP神经网络;主控因素中图分类号:TE88文献标识码:A文章编号:1000-8241(2023)01-0040-06DOI:10.6047/j.issn.1000-8241.2023.01.006InternalcorrosionpredictionofgroundpipelineofYuanbahigh-sourgasfieldSUNTianli1,HANXue2,HUANGShilin1,ZHAOChunlan2,HEHuan1,ZENGDezhi21.No.2GasProductionPlant,SINOPECSouthwestOil&GasCompany;2.SouthwestPetroleumUniversity//StateKeyLaboratoryofOilandGasReservoirGeologyandExploitationAbstract:TostudytheinternalcorrosionofgroundpipelinesinYuanbahigh-sourgasfield,acorrosionpredictionmodelbasedonBPneuralnetworkwasproposedwithconsiderationtomultiplefactors.Inthemodel,theinputwas8typesofworkingconditiondata,i.e.,thetemperature,CO2partialpressure,H2Spartialpressure,pHvalue,Cl-concentration,totalsalinity,liquid-to-gasratioandtheresidueofcorrosioninhibitor,whiletheoutputwasthecorrosionrate.Specifically,theBPneuralnetworkwastrainedwithalargeamountofhistoricalsampledatafromfieldmeasurementtorealizethepredictionofthecorrosionrateofgroundpipelines.Besides,themodelwasusedtoeval...