Skip to main content

Enhance Oil and Gas Exploration with Data-Driven Geophysical and Petrophysical Models

Enhance Oil and Gas Exploration with Data-Driven Geophysical and Petrophysical Models

Keith R. Holdaway, Duncan H. B. Irving

ISBN: 978-1-119-21510-3

Oct 2017

368 pages

In Stock

$95.00

Description

Leverage Big Data analytics methodologies to add value to geophysical and petrophysical exploration data

Enhance Oil & Gas Exploration with Data-Driven Geophysical and Petrophysical Models demonstrates a new approach to geophysics and petrophysics data analysis using the latest methods drawn from Big Data. Written by two geophysicists with a combined 30 years in the industry, this book shows you how to leverage continually maturing computational intelligence to gain deeper insight from specific exploration data. Case studies illustrate the value propositions of this alternative analytical workflow, and in-depth discussion addresses the many Big Data issues in geophysics and petrophysics. From data collection and context through real-world everyday applications, this book provides an essential resource for anyone involved in oil and gas exploration.

Recent and continual advances in machine learning are driving a rapid increase in empirical modeling capabilities. This book shows you how these new tools and methodologies can enhance geophysical and petrophysical data analysis, increasing the value of your exploration data.

  • Apply data-driven modeling concepts in a geophysical and petrophysical context
  • Learn how to get more information out of models and simulations
  • Add value to everyday tasks with the appropriate Big Data application
  • Adjust methodology to suit diverse geophysical and petrophysical contexts

Data-driven modeling focuses on analyzing the total data within a system, with the goal of uncovering connections between input and output without definitive knowledge of the system's physical behavior. This multi-faceted approach pushes the boundaries of conventional modeling, and brings diverse fields of study together to apply new information and technology in new and more valuable ways. Enhance Oil & Gas Exploration with Data-Driven Geophysical and Petrophysical Models takes you beyond traditional deterministic interpretation to the future of exploration data analysis.

Foreword xv

Preface xxi

Acknowledgments xxiii

Chapter 1 Introduction to Data-Driven Concepts 1

Introduction 2

Current Approaches 2

Is There a Crisis in Geophysical and Petrophysical Analysis? 3

Applying an Analytical Approach 4

What Are Analytics and Data Science? 5

Meanwhile, Back in the Oil Industry 8

How Do I Do Analytics and Data Science? 10

What Are the Constituent Parts of an Upstream Data Science Team? 13

A Data-Driven Study Timeline 15

What Is Data Engineering? 18

A Workflow for Getting Started 19

Is It Induction or Deduction? 30

References 32

Chapter 2 Data-Driven Analytical Methods Used in E&P 34

Introduction 35

Spatial Datasets 36

Temporal Datasets 37

Soft Computing Techniques 39

Data Mining Nomenclature 40

Decision Trees 43

Rules-Based Methods 44

Regression 45

Classification Tasks 45

Ensemble Methodology 48

Partial Least Squares 50

Traditional Neural Networks: The Details 51

Simple Neural Networks 54

Random Forests 59

Gradient Boosting 60

Gradient Descent 60

Factorized Machine Learning 62

Evolutionary Computing and Genetic Algorithms 62

Artificial Intelligence: Machine and Deep Learning 64

References 65

Chapter 3 Advanced Geophysical and Petrophysical Methodologies 68

Introduction 69

Advanced Geophysical Methodologies 69

How Many Clusters? 70

Case Study: North Sea Mature Reservoir Synopsis 72

Case Study: Working with Passive Seismic Data 74

Advanced Petrophysical Methodologies 78

Well Logging and Petrophysical Data Types 78

Data Collection and Data Quality 82

What Does Well Logging Data Tell Us? 84

Stratigraphic Information 86

Integration with Stratigraphic Data 87

Extracting Useful Information from Well Reports 89

Integration with Other Well Information 90

Integration with Other Technical Domains at the Well Level 90

Fundamental Insights 92

Feature Engineering in Well Logs 95

Toward Machine Learning 98

Use Cases 98

Concluding Remarks 99

References 99

Chapter 4 Continuous Monitoring 102

Introduction 103

Continuous Monitoring in the Reservoir 104

Machine Learning Techniques for Temporal Data 105

Spatiotemporal Perspectives 106

Time Series Analysis 107

Advanced Time Series Prediction 108

Production Gap Analysis 112

Digital Signal Processing Theory 117

Hydraulic Fracture Monitoring and Mapping 117

Completions Evaluation 118

Reservoir Monitoring: Real-Time Data Quality 119

Distributed Acoustic Sensing 122

Distributed Temperature Sensing 123

Case Study: Time Series to Optimize Hydraulic Fracture

Strategy 129

Reservoir Characterization and Tukey Diagrams 131

References 138

Chapter 5 Seismic Reservoir Characterization 140

Introduction 141

Seismic Reservoir Characterization: Key Parameters 141

Principal Component Analysis 146

Self-Organizing Maps 146

Modular Artificial Neural Networks 147

Wavelet Analysis 148

Wavelet Scalograms 157

Spectral Decomposition 159

First Arrivals 160

Noise Suppression 161

References 171

Chapter 6 Seismic Attribute Analysis 174

Introduction 175

Types of Seismic Attributes 176

Seismic Attribute Workflows 180

SEMMA Process 181

Seismic Facies Classification 183

Seismic Facies Dataset 188

Seismic Facies Study: Preprocessing 189

Hierarchical Clustering 190

k-means Clustering 193

Self-Organizing Maps (SOMs) 194

Normal Mixtures 195

Latent Class Analysis 196

Principal Component Analysis (PCA) 198

Statistical Assessment 200

References 204

Chapter 7 Geostatistics: Integrating Seismic and Petrophysical Data 206

Introduction 207

Data Description 208

Interpretation 210

Estimation 210

The Covariance and the Variogram 211

Case Study: Spatially Predicted Model of Anisotropic Permeability 214

What Is Anisotropy? 214

Analysis with Surface Trend Removal 215

Kriging and Co-kriging 224

Geostatistical Inversion 229

Geophysical Attribute: Acoustic Impedance 230

Petrophysical Properties: Density and Lithology 230

Knowledge Synthesis: Bayesian Maximum Entropy (BME) 231

References 237

Chapter 8 Artificial Intelligence: Machine and Deep Learning 240

Introduction 241

Data Management 243

Machine Learning Methodologies 243

Supervised Learning 244

Unsupervised Learning 245

Semi-Supervised Learning 245

Deep Learning Techniques 247

Semi-Supervised Learning 249

Supervised Learning 250

Unsupervised Learning 250

Deep Neural Network Architectures 251

Deep Forward Neural Network 251

Convolutional Deep Neural Network 253

Recurrent Deep Neural Network 260

Stacked Denoising Autoencoder 262

Seismic Feature Identification Workflow 268

Efficient Pattern Recognition Approach 268

Methods and Technologies: Decomposing Images into Patches 270

Representing Patches with a Dictionary 271

Stacked Autoencoder 272

References 274

Chapter 9 Case Studies: Deep Learning in E&P 276

Introduction 277

Reservoir Characterization 277

Case Study: Seismic Profile Analysis 280

Supervised and Unsupervised Experiments 280

Unsupervised Results 282

Case Study: Estimated Ultimate Recovery 288

Deep Learning for Time Series Modeling 289

Scaling Issues with Large Datasets 292

Conclusions 292

Case Study: Deep Learning Applied to Well Data 293

Introduction 293

Restricted Boltzmann Machines 294

Mathematics 297

Case Study: Geophysical Feature Extraction: Deep Neural Networks 298

CDNN Layer Development 299

Case Study: Well Log Data-Driven Evaluation for Petrophysical Insights 302

Case Study: Functional Data Analysis in Reservoir Management 306

References 312

Glossary 314

About the Authors 320

Index 323