A Supervised Machine Learning Approach to Discriminate Reservoir Fluid Presence and Saturation in the Gulf of Mexico Using Frequency and Spectral Shape Attributes - Thesis

Aug 5, 2026

Upload by Patrick

A supervised machine learning workflow is developed to discriminate reservoir fluid presence and saturation in the Gulf of Mexico using frequency and spectral shape seismic attributes. The approach builds an attribute dataset, applies controlled steps for seismic cropping, attribute creation, and confirmation using well logs and reports, then labels and extracts reservoir voxels for modeling. Interpretation and mapping connect the inferred reservoir properties to geologic context, aiming to improve the reliability of seismic-based fluid characterization and support reservoir decision-making.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages107

Reading Time4 h 30 min

AccessFree to download

Acknowledgements

Table of Contents

Thesis Abstract

Chapter 1: Analysis of Spectral Shape and Discrete Frequency Magnitude Attributes for the Discrimination of Reservoir Fluid Presence and Saturation

Introduction

Previous work

Geologic Setting

Data

Methods

Step 1: Seismic cropping

Step 2: Attribute creation

Step 3: Confirmation of reservoir fluids with well logs and reports

Step 4: Interpretation and mapping

Step 5: Labeling and extraction of reservoir voxels

UNIVERSITY OF OKLAHOMA GRADUATE COLLEGE A SUPERVISED MACHINE LEARNING APPROACH TO DISCRIMINATE RESERVOIR FLUID PRESENCE AND SATURATION IN THE GULF OF MEXICO USING FREQUENCY AND SPECTRAL SHAPE ATTRIBUTES A THESIS SUBMITTED TO THE GRADUATE FACULTY in partial fulfillment of the requirements for the Degree of MASTER OF SCIENCE By Mario Ricardo Ballinas Rondón Norman, Oklahoma A SUPERVISED MACHINE LEARNING APPROACH TO DISCRIMINATE RESERVOIR FLUID PRESENCE AND SATURATION IN THE GULF OF MEXICO USING FREQUENCY AND SPECTRAL SHAPE ATTRIBUTES A THESIS APPROVED FOR THE SCHOOL OF GEOSCIENCES BY THE COMMITTEE CONSISTING OF Dr. Heather Bedle, Chair Dr. Matthew J. Pranter Dr. Deepak Devegowda © Copyright by MARIO RICARDO BALLINAS 2023 All Rights Reserved. ToMom & Dad, My Sister, and My Love For helping me get this far, and more ACKNOWLEDGEMENTS I would like to start by expressing my biggest thanks to my mom Osglayt, dad Mario, and sister Mariana for their unwavering love and support throughout my academic journey. Their encouragement and motivation have been instrumental in my success, and I couldn't have done it without them. I would also like to thank Audrey Schmitt for her love, patience, and hours she spent proofreading my thesis. Her support and understanding have been invaluable during this time. I am also deeply grateful to my advisor Dr. Heather Bedle for believing in me and giving me the opportunity to be a part of her program. Her guidance and encouragement have helped me stay on track and remain focused, even during the most challenging times. I would like to thank Dr. Deepak Devegowda for helping me shape my thesis project and taking my Python coding and machine learning skills to the next level. Additionally, I want to thank Dr. Matthew J. Pranter for sharing his wealth of geological knowledge and teaching me new tricks in Petrel that I use everyday. I would also like to express my appreciation to the friends I have made at this university, for their camaraderie and support during this stressful time. Special thanks to Marcus Maas, Karelia La Marca, Emily Jackson, David Lubo-Robles, Laura Ortiz, Bobby Buist, Diana Salazar, Jacob Clements, Pamela Blanco, Raymond Ng, Grace Barber, Vanesa Rios Perez, John Mcknight, Patience Emuobosa, Alex Vera and Britney Stroud. Their assistance and insights have helped me to make my research more impactful and meaningful. I am grateful for their friendship and the joy they have brought to my life. Furthermore, I want to extend a special thank you to my mentor Mr. Allen Bertagne for taking me on after my bachelor's graduation and igniting my passion for seismic interpretation and geophysics. Without his teachings, I would not be where I am today. A special thanks also goes tomy second mentor Carlos Eduardo Moreno Elcure, who provided me unvaluable knowledge during my internship with Lumina Geophysical. Both of these professionals believed in me before I knew that seismic amplitudes had polarities and frequencies. Jesus Nevarez, although my peer, deserve a big thanks for reaffirming geophysical principles that I did not have the courage to ask my mentors about again. Finally, I would like to thank the American Chemical Society Petroleum Research Fund for providing the funding that made my graduate research possible. Their support has been integral to my success, and I am deeply grateful for their generosity. In conclusion, I would like to thank everyone who has played a part in my academic journey. Your support, encouragement, and guidance have been invaluable, and I am deeply grateful for everything you have done for me TABLE OF CONTENTS ACKNOWLEDGEMENTS ............................................................................................................ v TABLE OF CONTENTS .............................................................................................................. vii THESIS ABSTRACT ........................

FAQs

What is the main objective of the thesis?

To develop a supervised machine learning approach that discriminates reservoir fluid presence and saturation in the Gulf of Mexico using frequency and spectral shape attributes.

Which seismic attributes are used for discrimination?

The study uses frequency and spectral shape attributes, including analyses of spectral shape and discrete frequency magnitude attributes.

How are reservoir fluids confirmed during the workflow?

Reservoir fluids are confirmed using well logs and reports as part of the staged methodology before interpretation and mapping.

Aug 5, 2026

Upload by Patrick