摘要
非常规储层油气资源丰富,压裂是释放非常规储层油气的必要手段,但压裂优化是一个多模态、高维度、大尺度、细时空的复杂大系统问题.为实现非常规储层压裂系统开发,梳理了油气压裂人工智能3个应用场景:透明油气藏数智化构建和压裂双甜点智能优选、机理—数据联合驱动的压裂工艺参数智能优化和压裂风险预警及在线监控智能调控,然后明确了智能压裂3个方面技术特征,并展望了智能压裂未来5个方面的技术攻关方向.研究结果表明:①形成了地质多源数据、建模迭代更新精准化技术,助力了优选地质工程压裂双甜点;②建立了机理数据驱动多目标优化压裂参数方法,促进了裂缝均衡扩展、扩大体积和提高产量;③开发了压裂动态监测及在线调控流程,支撑了压裂高效、可持续开发,提高压裂效率和油气采收率;④提出了"甜点选择数智化—压裂参数智能化—在线监测调控精准化—四维可视化"的地质—工程"动态一体化"智能压裂新理念.结论认为,强化数字孪生智能压裂可视化技术、构建生成式智能压裂大数据生态系统、融合机理数据的工艺参数优化方法、开发压裂动态智能预测监测技术、创新远程智能压裂决策控制系统,可为智能压裂革新发展提供理论指导和技术支撑,为未来压裂人机交互智能决策和闭环调控的实现指明了方向.
Abstract
Unconventional reservoirs are abundant in hydrocarbon resources,and fracturing is a necessary technique to release such resources from the reservoirs.However,fracturing optimization presents complex system issues characterized by multimodal,high-dimensional,large-scale,and fine spatiotemporal challenges.In order to develop the unconventional reservoir fracturing system,three application scenarios of artificial intelligence in reservoir fracturing are outlined:1)the intelligent construction of transparent oil and gas reservoirs and optimal selection of dual sweet spots for fracturing;2)mechanism-data driven intelligent optimization of fracturing parameters;and 3)risk warning and online monitoring for intelligent regulation of fracturing operations.This study clarifies three characteristics of intelligent fracturing and anticipates five research directions for this technology.The following results are obtained.First,a precise technology for geological multi-source data and iterative modeling updates is established,facilitating the optimal selection of geological and engineering sweet spots for fracturing.Second,a mechanism-data driven method for multi-objective optimization of fracturing parameters is developed,promoting balanced fracture propagation,increased volume,and enhanced yield.Third,a dynamic monitoring and online regulating process for fracturing is created,supporting efficient and sustainable development,thereby improving fracturing efficiency and hydrocarbon recovery.Fourth,a"dynamic integration"concept for intelligent fracturing is proposed,which includes data-driven sweet spot selection,intelligent fracturing parameterization,precise online monitoring regulation,and four-dimensional visualization,within geological-engineering contexts.In conclusion,strengthening digital twin visualization technologies for intelligent fracturing,constructing a generative big data ecosystem for intelligent fracturing,integrating mechanism and data for process parameter optimization,developing dynamic intelligent prediction and monitoring technologies for fracturing,and innovating remote intelligent decision-making and control systems can provide theoretical guidance and technical support for the innovative development of intelligent fracturing towards human-machine interaction for decision-making and closed-loop regulation in the future.
基金项目
国家自然科学基金面上项目(52374045)
国家自然科学基金企业创新发展联合基金项目(U23B6004)
四川省自然科学基金面上项目(2023NSFSC0424)
维普
万方数据