摘要
大型油气管网具有源汇节点多、空间跨度大、热力水力过程耦合程度高等特点,导致建模难度大。国家《中长期油气管网规划》指出,管网智能化是未来方向,融合机理知识与数据驱动建模方法,构建物理意义明确、外推泛化能力强的混合模型是实现智慧管网的关键环节。分析机理建模与数据驱动建模方法的特点,通过融合机理模型与数据驱动模型协同描述研究对象的物理特性,充分挖掘现场数据内在关联,探索过程变量演化规律,最终建成了机理-数据双驱动的高保真混合模型。梳理混合建模不同结构及其在油气管道行业应用的可行性,阐明了不同应用场景下混合建模的策略,探讨了机理与数据驱动协同建模技术的研究方向,以期为智慧管网建设提供参考。(图2,参60)
Large oil and gas pipeline networks are characterized by multiple source and sink nodes,large spatial span and high degree of coupling between the thermal and hydraulic processes,which lead to great difficulty in modeling.It is pointed out in the Medium and long-term oil and gas pipeline network planning that the intellectualization of pipeline networks is the development direction in the future,and building a hybrid model with clear physical meaning and strong generalization capabilities by combining the mechanism knowledge and the data-driven modeling method is critical to realize the intelligent pipeline networks.Herein,the characteristics of mechanism modeling and data-driven modeling were analyzed,the physical properties of the study object were described collaboratively by integrating the mechanism model and the datadriven model,the internal connection among the site data was fully mined,the evolution laws of the process variations were explored,and finally a high-fidelity hybrid model was established.In addition,various structures of hybrid models and the feasibility of their application in oil and gas pipeline industry was summarized,the strategies of hybrid modeling in various application scenarios were clarified and the direction of research on the mechanism and data based collaborative modeling technology in future was discussed.Further,the research results can provide reference to the construction of intelligent pipeline networks.(2 Figures,60 References)
作者
刘刚
袁子云
陈雷
左志恒
LIU Gang;YUAN Ziyun;CHEN Lei;ZUO Zhiheng(College of Pipeline and Civil Engineering,China University of Petroleum(East China)//Shandong Key Laboratory of Oil&Gas Storage and Transportation Safety;PipeChina South China Company)
出处
《油气储运》
CAS
北大核心
2021年第9期980-990,共11页
Oil & Gas Storage and Transportation
基金
国家自然科学基金面上项目“轻组分在原油非稳态压力流中的作用机理研究”,51774315
中央高校基本科研业务费专项“输油管道内部流动安全量化评价研究”,20CX02403A
广东省重点领域研发计划项目“油气储运重大基础设施灾害防御关键技术及装备研发与示范”,2019B111102001。
关键词
智慧管网
机理模型
数据驱动模型
混合模型
建模方法
intelligent pipeline network
mechanism model
data-driven model
hybrid model
modeling method