摘要
沉积正演数值模拟是研究沉积过程的重要手段,由大量的数值模型支撑,但如何选择和应用数值模型仍是难点。为此,梳理了沉积数值模型的时间尺度、驱动机制、河道演化、遵循规则、沉积结果和适用对象,着重阐述了Delft3D和DIONISOS模型的运算原理、参数选择、模拟结果和局限性。不同湖平面级次控制下的层序演化和发育过程与沉积过程尺度相对应,沉积物组分特征和模拟目标是优选数值模型的基本原则,井-震数据和地质建模算法是验证和优化训练模型的必要手段,基于大数据和人工智能的深度学习和强化学习模型是沉积正演数值模拟的发展趋势。
Sedimentary forward numerical modeling is a key method to study depositional processes supported by numerous numerical models.However,the model selection and application are difficult.This review summarized model features,in cluding sedimentary time scales,driving mechanism,channel evolution,following rules,sedimentary outcomes,and applica ble objects.The mathematical principles,parameter selection,simulation outcomes and restrictions of Delft3D and DION ISOS are illustrated as well.It shows that the sequence evolution controlled by lake level change hierarchy correspond to the sedimentary scales.Combining the characteristics of sediment components and simulation objectives is the basic princi ple of optimizing the numerical model.The well-seismic data and geological modeling algorithm are the necessary means to verify and optimize the training model.The deep-learning and reinforcement-learning numerical models based on big data and AI are main trends for sedimentary forward numerical modeling.
作者
杜威
纪友亮
李其海
王子涵
席家辉
唐林
高星星
DU Wei;JI Youliang;LI Qihai;WANG Zihan;XI Jiahui;TANG Lin;GAO Xingxing(State Key Laboratory of Petroleum Resource and Prospecting,China University of Petroleum(Beijing),Beijing City,102249,China;Xianhe Oil Production Plant,Shengli Oilfield Company,SINOPEC,Dongying City,Shandong Province,257000,China;Boxi Operation Area,Tianjin Branch,CNOOC(China),Tianjin City,300452,China;No.12 Oil Production Plant,Changqing Oilfield Company,Qingyang City,Shaanxi Province,745400,China)
出处
《油气地质与采收率》
CAS
CSCD
北大核心
2020年第2期62-71,共10页
Petroleum Geology and Recovery Efficiency
基金
国家自然科学基金项目“高频湖平面变化背景下复合三角洲砂体的结构及分布模式”(41672098)。
关键词
数值模型
储层建模
浅水湖盆
沉积过程
层序演化
numerical model
reservoir modeling
shallow water lake basin
sedimentary process
sequence evolution
