Table of Contents
Dedication
Acknowledgments
Foreword
1 Introduction
1.1 Reservoir Models for Reservoir Management
1.2 What Is Top-Down Modeling?
1.2.1 Role of Physics and Geology
1.2.2 Formulation and Computational Footprint
1.2.3 Expected Outcome of a Top-Down Model
1.2.4 Limitations of TDM
1.2.5 Software Tool for the Development of TDM
1.3 Paradigm Shift
1.3.1 Drilling Operation
1.3.2 Mature Fields
1.3.3 Smart Completions, Smart Wells, and Smart Fields
1.3.4 Production From Shale Assets
1.3.5 Reservoir Simulation Models
2 Data-Driven Problem Solving
2.1 Misunderstanding Data-Driven Reservoir Modeling
3 Reservoir Modeling
4 Data-Driven Technologies
4.1 Data Mining
4.2 Artificial Intelligence
4.3 Artificial Neural Networks
4.3.1 Structure of a Neural Network
4.3.2 Mechanics of Neural Network Operation
4.4 Fuzzy Logic
4.4.1 Fuzzy Set Theory
4.4.2 Approximate Reasoning
4.4.3 Fuzzy Inference
5 Pitfalls of Using Machine Learning in Reservoir Modeling
6 Fact-Based Reservoir Management
6.1 Empirical Models in the E&P Industry
6.1.1 Decline Curve Analysis
6.1.2 Capacitance/Resistance Modeling
7 Top-Down Modeling
7.1 Components of a Top-Down Model
7.2 Formulation and Computational Footprint of TDM
7.3 Curse of Dimensionality
7.4 Correlation Is Not the Same as Causation
7.5 Quality Control and Quality Assurance of the Data
7.5.1 Inspecting the Quality of the Data
7.5.2 QC of the Production Data
8 The Spatio-Temporal Database
8.1 Static Data
8.2 Dynamic Data
8.3 Well Trajectory and Completion Data
8.3.1 Two-Dimensional vs. Three-Dimensional Reservoir Modeling
8.4 Resolution in Time and Space
8.4.1 Resolution in Space
8.4.2 Resolution in Time
8.5 Role of Offset Wells
8.6 Structure of the Spatio-Temporal Database
8.7 Required Quantity and Quality of Data
9 History Matching the Top-Down Model
9.1 Practical Considerations During the Training of a Neural Network
9.1.1 Selection of Input Parameters
9.1.2 Partitioning the Data Set
9.1.3 Structure and Topology
9.1.4 The Training Process
9.1.5 Convergence
9.2 History-Matching Schemes in TDM
9.2.1 Sequential History Matching
9.2.2 Random History Matching
9.2.3 Mixed History Matching
9.3 Validation of the Top-Down Model
9.3.1 Material Balance Check
10 Post-Modeling Analysis of the Top-Down Model
10.1 Forecasting Oil Production, GOR, and WC
10.2 Production Optimization
10.2.1 Choke-Setting Optimization
10.2.2 Artificial-Lift Optimization
10.2.3 Water-Injection Optimization
10.3 Reservoir Characterization
10.4 Determination of Infill Locations
10.5 Recovery Optimization
10.6 Type Curves
10.7 Uncertainty Analysis
10.8 Updating the Top-Down Model
11 Examples and Case Studies
11.1 Case Study No. 1: A Mature Onshore Field in Central America
11.2 Case Study No. 2: Mature Offshore Field in the North Sea
11.3 Case Study No. 3: Mature Onshore Field in the Middle East
11.3.1 Data Used During the Top-Down-Model Development
11.3.2 Top-Down-Model Training and History Matching
11.3.3 Post-Modeling Analysis
11.3.4 Performing a “Stress Test” on the Top-Down Model
12 Limitations of Data-Driven Reservoir Modeling
13 The Future of Data-Driven Reservoir Modeling
References
Index