高精度油藏数值模拟技术研究进展及应用  

Research progress and application of high‑accuracy reservoir numerical simulation technology

作　　者：赵国忠 兰玉波[1,2,3] 匡铁 何鑫[1,2,3] 王青振 李椋楠[1,2,3] 石亮 ZHAO Guozhong;LAN Yubo;KUANG Tie;HE Xin;WANG Qingzhen;LI Liangnan;SHI Liang(State Key Laboratory of Continental Shale Oil,Daqing 163712,China;Heilongjiang Provincial Key Laboratory of Reservoir Physics&Fluid Mechanics in Porous Medium,Daqing 163712,China;Exploration and Development Research Institute of PetroChina Daqing Oilfield Co.,Ltd.,Daqing 163712,China)

机构地区：[1]多资源协同陆相页岩油绿色开采全国重点实验室,黑龙江大庆163712 [2]黑龙江省油层物理与渗流力学重点实验室,黑龙江大庆163712 [3]中国石油大庆油田有限责任公司勘探开发研究院,黑龙江大庆163712

出　　处：《大庆石油地质与开发》2024年第4期152-160,共9页Petroleum Geology & Oilfield Development in Daqing

基　　金：国家科技重大专项“井震结合油藏精细结构表征技术”(2011ZX05010-001);中国石油股份有限公司重大科技专项“大庆长垣特高含水油田提高采收率示范工程”(2016ZX05054)。

摘　　要：受算力和油藏模拟方法的制约,常规油藏数值模拟技术的模拟规模和效率,难以满足中高渗老油田特高含水后期剩余油精准挖潜的需求,人工智能加速油藏模拟技术的应用仍处于初步应用阶段。“十四五”期间,大庆油田在大规模并行模拟技术、页岩油藏模拟技术研究、人工智能方法和油藏模拟技术应用4个方面开展了攻关。通过对分布式黑油油藏模拟网格求解方法和数据传输方式的改进,大规模并行模拟规模突破1亿节点;考虑页岩油藏微孔限域效应特征,研发了页岩油藏数值模拟软件;基于傅里叶神经算子神经网络,实现了油藏数值模拟的加速;应用数值模拟辅助历史拟合技术,结合物理模拟结果,建立聚合物相对渗透率曲线求取方法。分布式并行模拟技术的改进完善、页岩油藏模拟器的建立和人工智能方法的创新应用,为中高渗老油田剩余油挖潜和非常规油藏开发提供了技术保障。Due to the constraints of computility and reservoir simulation methods,simulation scale and efficiency of conventional numerical simulation techniques cannot meet the need of precise tapping of remaining oil in late stage of ultra-high water cut in medium-high permeability mature oilfields.Artificial intelligence accelerated reservoir simulation technology is still in initial application stage.During the 14th Five-Year Plan,Daqing Oilfield carried out research in 4 techniques including large-scale parallel simulation,shale oil reservoir simulation,artificial intelligence methods and reservoir simulation application.By improving grid solving method and data transmission method for distributed black oil reservoir simulation,large-scale parallel simulation exceeds 100 million nodes.Considering the characteristics of microporous confinement effect,a numerical simulation software for shale oil reservoirs is developed.Based on neural network with Fourier neural operator,reservoir numerical simulation is accelerated.Numerical simulation assisted history matching technique is applied and combined with physical simulation results to establish a method for calculating relative permeability curve of polymer.Improvement of distributed parallel simulation technique,establishment of shale oil reservoir simulator and innovative application of artificial intelligence methods provide technical support for potential tapping of remaining oil in medium-high permeability mature oilfields and development of unconventional reservoirs.

关 键 词：水驱开发 页岩油 油藏模拟 深度学习 辅助拟合 

分 类 号：TE319[石油与天然气工程—油气田开发工程]