Applying machine learning to the interpretation of seismic data
Seismic data gathered on the surface can be used to generate numerous seismic attributes that enable better understanding of subsurface geological structures and stratigraphic features. With an ever-increasing volume of seismic data available, machine learning augments faster data processing and interpretation of complex subsurface geology.
Meta-Attributes and Artificial Networking: A New Tool for Seismic Interpretation explores how artificial neural networks can be used for the automatic interpretation of 2D and 3D seismic data.
Volume highlights include:
- Historic evolution of seismic attributes
- Overview of meta-attributes and how to design them
- Workflows for the computation of meta-attributes from seismic data
- Case studies demonstrating the application of meta-attributes
- Sets of exercises with solutions provided
- Sample data sets available for hands-on exercises
The American Geophysical Union promotes discovery in Earth and space science for the benefit of humanity. Its publications disseminate scientific knowledge and provide resources for researchers, students, and professionals.
Table of Contents
Preface
About the Authors
Abbreviations
List of Symbols and Operators
PART I: SEISMIC ATTRIBUTES
1. An Overview of Seismic Attributes
1.1 Introduction
1.2 Historical evolution of seismic attributes
1.3 Characteristics of Seismic Attributes
1.4 A glance at seismic characteristics
1.4.1 Amplitude
1.4.2 Phase
1.4.3 Frequency
1.4.4 Bandwidth
1.4.5 Amplitude Change
1.4.6 Slope Dip and Azimuth
1.4.7 Curvature
1.4.8 Seismic Discontinuity
1.5 Summary
References
2. Complex Trace, Structural and Stratigraphic Attributes
2.1 Introduction
2.2 Complex Trace Attributes: Mathematical Formulations and Derivations
2.3 Other Derived Complex Trace Attributes
2.3.1 Instantaneous Frequency
2.3.2 Sweetness
2.3.3 Relative Amplitude Change and Instantaneous Bandwidth
2.3.4 RMS Frequency
2.3.5 Q-factor
2.4 Structural and Stratigraphic Attributes
2.4.1 Dip and Azimuth Attributes
Slope and Dip Exaggeration
Dip-steering
2.4.2 Coherence Attribute
2.4.3 Similarity Attribute
2.4.4 Curvature Attribute
2.4.5 Advanced structural attributes
Ridge Enhancement Filter (REF) attribute
Thin Fault Likelihood (TFL) attribute
Pseudo Relief attribute
2.4.6 Amplitude Variance
2.4.7 Reflection Spacing
2.4.8 Reflection Divergence
2.4.9 Reflection Parallelism
2.4.10 Spectral Decomposition
2.4.11 Velocity, Reflectivity and Attenuation attributes
2.5 A glance on interpretation pitfalls
2.6 Summary
References
3. Be an Interpreter: Brainstorming Session
3.1 Task 1
3.2 Task 2
3.3 Task 3
3.4 Task 4
3.5 Task 5
3.6 Task 6
3.7 Task 7
3.8 Task 8
3.9 Task 9
3.10 Task 10
PART II: META-ATTRIBUTES
4. An Overview of Meta-attributes
4.1 Introduction
4.2 Meta-attributes
4.3 Types of Meta-attributes
4.3.1 Hydrocarbon Probability meta-attribute
4.3.2 Chimney Cube meta-attribute
4.3.3 Fault Cube meta-attribute
4.3.4 Intrusion Cube meta-attribute
4.3.5 Sill Cube meta-attribute
4.3.6 Mass Transport Deposit Cube meta-attribute
4.3.7 Lithology meta-attribute
4.4 Summary
References
5. An Overview of Artificial Neural Networks
5.1 Introduction
5.2 Historical Evolution
5.3 Biological Neuron Vs Mathematical Neuron
5.3.1 Biological Neuron
5.3.2 Mathematical Neuron
5.4 Activation or Transfer Function
5.5 Types of Learning
5.6 Multi-layer Perceptron (MLP) and the Backpropagation Algorithm
5.7 Different Types of ANNs
5.7.1 Radial Basis Function (RBF) Network
5.7.2 Probabilistic Neural Network (PNN)
5.7.3 Generalized Regression Neural Network (GRNN)
5.7.4 Modular Neural Network (MNN)
5.7.5 Self Organizing Maps (SOM)
5.8 Summary
References
6. How to Design Meta-attributes
6.1 Introduction
6.2 Meta-attribute design
6.2.1 Seismic Data conditioning
Mean Filter (or Running-Average filter)
Median Filter
Alpha-Trimmed Mean Filter
6.2.2 Selection and Extraction of Seismic Attributes
6.2.3 Example Location
6.2.4 NN operation
Evaluation of intelligent neural model
6.2.5 Validation
6.3 RGB Blending and Geo-body Extraction
6.4 Summary
References
PART III: CASE STUDIES OF META-ATTRIBUTES
7. Chimney interpretation using meta-attribute
7.1 Gas Chimneys: a clue for hydrocarbon exploration
7.2 Research Methodology
7.3 Chimney Validation
7.3.1 Geological Validation
7.3.2 Petrophysical Validation
7.3.3 Soft sediment deformation anomalies
7.4 Interpretation using Chimney Cube
7.5 Summary
References
8. Fault Interpretation Using Meta-attribute
8.1 Fault meta-attribute: a motivation
8.2 Research Methodology
8.3 Results and Interpretation
8.4 Efficiency of the optimized TFC
8.5 Summary
References
9. Fault and Fluid Migration Interpretation Using Meta-attribute
9.1 Introduction
9.2 Geophysical Data
9.3 Results and Interpretation
9.3.1 Thinned Fault Cube (TFC) and Fluid Cube (FlC)
9.3.2 Neural Design for the TFC and FlC
9.3.3 Interpretation using TFC and FlC
9.4 Summary
References
10. Magmatic Sill Interpretation Using Meta-attribute (Part 1: Taranaki Basin example)
10.1 Magmatic Sills: Interpretation techniques
10.2 Research Methods
10.2.1 Structural conditioning
10.2.2 Selection of attributes
10.2.3 Example Locations
10.2.4 Neural Network
10.2.5 Validation
10.3 Results and Interpretation
10.4 Discussion
10.4.1 Sill cube an efficient interpretation tool for magmatic sills
10.4.2 Limitations of the Sill Cube automated approach
10.5 Conclusions
References
11. Magmatic Sill Interpretation Using Meta-attribute (Part 2: Vøring Basin example)
11.1 Introduction: The Vøring Basin case
11.2 Description of the Data
11.3 Interpretation based on SC meta-attribute computation
11.4 Summary
References
12. Magmatic Sill and Fluid Plumbing Interpretation Using Meta-attribute (Canterbury Basin example)
12.1 Introduction: The Canterbury Basin case
12.2 Description of the Data
12.3 Results and Interpretation
12.3.1 Data Enhancement, Attribute Analysis and Neural Operation
12.3.2 Interpretation through Sill Cube (SC) and Fluid Cube (FlC) meta-attributes
12.3.3 Limitation of the automated approach
12.4 Summary
References
13. Volcanic System Interpretation Using Meta-attribute
13.1 Introduction
13.2 Research Workflow
13.3 Results and Interpretation
13.3.1 Seismic Data Enhancement
13.3.2 Neural Networks: Analysis and Optimization
13.3.3 Geologic interpretation using IC meta-attribute
13.3.4 Validation of the IC meta-attribute
13.4 Summary
References
14. Interpretation of Mass Transport Deposits Using Meta-attribute
14.1 Introduction
14.2 Data and Research Workflow
14.3 Results and Interpretation
14.4 Summary
References
Appendix A
A.1 Mathematical formulation of some common series and transformation
A.1.1 Fourier Series
A.1.2 Fourier and Inverse Fourier Transforms
A.1.3 Hilbert Transform
A.1.4 Convolution
A.2 Dip-Steering
Appendix B
B.1 Answers to seismic cross-section interpretation (Tasks 1-6)
B.2 Answers to numerical tasks (Tasks 7-10)
Glossary
