卓越的讲座项目 Primary funding is provided by The SPE Foundation through member donations and a contribution from Offshore Europe The Society is grateful to those companies that allow their professionals to serve as lecturers 主要资金来源于 SPE 会员捐款和欧洲近海贡献 协会非常感谢允许他们的专业人士担任讲师的这些公司 AIME 提供额外的支持 Additional support provided by AIME Society of Petroleum Engineers Distinguished Lecturer Program www.spe.org/dl SPE First 1
Methods and Workflows to Process Dynamic Data in Intelligent Fields Methods & Workflows Olivier Houzé 动态数据的处理方法及其工作流在智能油 田的运用 2
Analysis of Dynamic Data 动态数据的分析 create a flow sequence 创建一个流序列 measure rates, pressures, temperature 测量流量 压力 温度 at any scale in time or space 在时间或空间上的任何尺度 analyse and match data with a model 分析和数据与模型的拟合 use the analysis for operational decisions 运用分析结果做出决策 use the model to forecast scenarios 使用模型来预测 Intelligent Fields: all this in real time 智能油田 : 所有这一切都是 实时的 3
Dynamic Data Sources 动态数据源 Well tests 测井数据 : q(t), P(t), T Formation tests 地层测试 : q(t), P(t), T vs. z Production logs 生产测井 : q(z), P(z), T(z) vs. t Production data 生产数据 : q(t), Q(t) PDG data: P(t), T Fiberoptics 纤维光学 : T(z,t), P(z,t) 4
PDG Data (Permanent Downhole Gauges) 永久井下压力计 Est: 75,000 to 100,000 PDG + 15,000 /year 5,000 IFM +800 /year 150 DFM 5
PDG Usage: PTA (Pressure Transient Analysis) 非稳态试井分析 6
PDG Usage: PA (Production Analysis) PDG 使用方法 : 生产分析 7
PDG Usage: History Matching PDG 使用方法 : 历史拟合 8
Issues 问题 Data overflow Information overflow Scarce human resources > 50% of time wasted fighting data Process rather than engineering Large crew change to come 大量的人员变化 9
Tools / Methods 工具 / 方法 1. Wavelet filtering 微波过滤 2. Events identification 事件识别 3. Deconvolution 反褶积 4. Flexible numerical models 灵活的数值模型 5. Proxys and virtual metering 代理模型和虚拟计量 10
1. Wavelet Filtering 微波过滤 11656 points too low too high OK Post-filtering 70 points 11
First generation of wavelets (2000) 第一代微波 (2000) 12
Second generation of wavelets (2010) 第二代微波 (2010) 13
Second generation of wavelets (2010) 第二代微波 (2010) 14
2. Events identification 事件鉴定 15
Identifying events 事件识别 Pressure or temperature vs. time (scale irrelevant) 16
Locating events 事件定位 Pressure or temperature vs. time (scale irrelevant) 17
Setting thresholds 设置临界值 10 hr & 100 psi 30 mn & 50 psi Pressure vs. time 18
3. Deconvolution 反褶积 we have we want 19
We change this Deconvolution: How? 反褶积 : 怎么做? Convolution to match that 20
Deconvolution in action 反褶积行为 12 34 56 A Extraction Initial First and Final couple later iteration deconvolved guess of of iterations of two of the the consistent response deconvolution nonlinear later build-ups regression response 21
Deconvolution explained 反褶积解释 22
4. Flexible Numerical Models 灵活的数值模型 PTA - 2000 Cells PA - 200 Cells Simulation - 5 Cells 23
Adaptative gridding PTA - 2000 Cells 可适应的网格 PA - 200 Cells Simulation - 5 Cells 24
Comparison of grid refinements 网格细化比较 PTA grid: matches the analytical model PA grid: matches after less than 1 hour Simulation grid: matches after 20 hours Dp PA & Simulation well indexes calibrated in situ from PTA grid 0.01 0.1 1 10 Dt (hr) 25
5. Proxys & Virtual Metering 代理模型及虚拟计量 Virtual Metering D D D Dynamic Data Analysis 26
Dynamic data in intelligent fields 智能化油田的动态数据 PDG Data Historians 井下测量历史数据 Intelligent Field 智能化油田 Data Model 数据模型 Production Databases 生产数据库 Engineering Applications 工程应用 Dynamic data processing 动态数据处理 27
The goal: Dynamic Data Workflow 目标 : 动态数据工作流 Raw Data 原始数据 Extracted Data 提取数据 Analyses 分析 (PTA+ PA HM) Filtered & Decimated Data 过滤和数据处理 Synchronized Data 同步化数据 Proxy Model 代理模型 Human Process Reallocated Data 数据再分配 ALARMS! 报警 28
The risk: from Engineering to Process 风险 : 从工程到处理 Raw Data 原始数据 Extracted Data 提取数据 Analyses some 分析 (PTA+ kh & PA skin HM) Filtered & Decimated Data 过滤和数据处理 Synchronized Data 同步化数据 some Proxy proxy Model straight 代理模型 line Human Process Reallocated some Data results 数据再分配 ALARMS! some 报警 report 29
What s next? DDA RTFM 30
Conclusion 总结 Masses of potentially useful data 大量的潜在的有用的数据 Data overflow 数据溢出 Information overflow 信息溢出 More Research & Development needed 更多的研究和发展需求 Leave processes to computers 计算机处理进程 Leave engineering thinking to Engineers 工程师进行工程思考 More education & training (crew change) 更多的教育和培训 ( 人员变化 ) More awareness from stakeholders (SPE) 更多利益相关者的意识 (SPE) 31
A final remark on research 关于研究的最后一个评论 Most technologies described here were initiated by research projects in the Petroleum Eng. Department of various universities: Stanford, Tulsa, Imperial College, Texas A&M, etc.yesterday s investment in these projects turned out to be marginal compared to today s benefits 这里描述的很多研究项目都是在不同大学的石油工程部门研究的 : 如斯坦福大学, 塔尔萨, 帝国理工学院 德州农工大学等 对比现在的收益, 昨天的投资显得微不足道. 32
Thank you! Gracias! Grazie! Спасибо! ありがとう! Merci! 谢谢! Дякую! Danke! Obrigado! Зур рахмəт! 33
SPE Last 卓越的讲座项目 Your Feedback is Important 您的反馈意见很重要请对这次演讲进行评估 : Enter your section in the DL Evaluation Contest by completing the evaluation form for this presentation : Click on: Section Evaluation Society of Petroleum Engineers Distinguished Lecturer Program www.spe.org/dl 34