Characterization of Mechanical Discontinuities Using Machine Learning and Knowledge-Driven Causal Model - Dissertation

Aug 4, 2026

Upload by Aurora

Mechanical wave transmission in solids couples with the propagation of mechanical discontinuities, and the resultant multipoint waveforms provide measurable signals. This dissertation introduces a data-driven monitoring framework using a single-impulse source and a multipoint sensor system to track embedded discontinuity propagation. It targets shortcomings of conventional supervised learning by emphasizing causal, geophysical signatures rather than mere statistical associations. The knowledge-driven causal model improves modeling for both linear and crack propagation samples and supports fracture monitoring, prediction, and early warning through improved explainability and generalizability.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages164

Reading Time6 h 53 min

AccessFree to download

Abstract

Acknowledgements

Contributors and Funding Sources

Nomenclature

Table of Contents

CHARACTERIZATION OF MECHANICAL DISCONTINUITIES USING MACHINE LEARNING AND KNOWLEDGE DRIVEN CAUSAL MODEL A Dissertation by RUI LIU Submitted to the Graduate and Professional School of Texas A&M University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Chair of Committee, Committee Members, Head of Department, Siddharth Misra Eduardo Gildin Kan Wu Benchun Duan Akhil Datta-Gupta August 2023 Major Subject: Petroleum Engineering Copyright 2023 Rui Liu ABSTRACT Mechanical wave transmission through a material interacts with the propagation of mechanical discontinuity in the material. Machine learning can be used to monitor the propagation of embedded discontinuities by analyzing the resultant waveforms recorded by a multipoint sensor system placed on the surface of the material. Our study accomplishes a first-of-its-kind monitoring of mechanical discontinuity propagation by using data-driven model to process the multipoint waveform measurements resulting from a single impulse source. While conventional data-driven methods, especially supervised learning, rely primarily on statistical correlation/association and lack domain knowledge and causality. The primary objective of this work is to discover new geophysical causal signatures relevant to the multipoint waveform measures caused by mechanical discontinuity propagation inside a solid material. The use of causal signatures also led to the development of a novel knowledge-driven model that creates a versatile and resilient data system for both linear and physical crack propagation samples. The newly discovered causal signatures confirm that the statistical correlations/associations and conventional feature rankings are not statistically significant indicators of causality. The new developments presented in this work, especially the causal-based knowledge-driven model, have both theoretical and practical implications that can improve fracture monitoring, prediction, and early warning. In near future, similar knowledge-driven approaches will gain traction in mainstream applications of data analytics to overcome the drawbacks of current machine learning approaches such as lack of generalizability and explainability. ACKNOWLEDGEMENTS First and foremost, I would like to express my heartfelt gratitude to my Ph.D. supervisor and mentor Dr. Siddharth Misra for his support, guidance, and mentorship, which have been invaluable in shaping my academic journey. He has inspired me to become an independent researcher and helped me realize the power of critical thinking. I will forever be thankful for his patience and caring approach in guiding me through the challenges of the research process. His guidance, constructive suggestion, and willingness to listen to my ideas and concerns have been a constant source of motivation. Besides my advisor, I am also grateful to the members of my thesis committee, including Dr. Kan Wu, Dr. Eduardo Gildin and Dr. Benchun Duan for their time on reviewing my work and providing feedback. Their support has been invaluable in helping me to refine and improve my research. Then, I am extending my thanks to my friends and colleagues at Texas A&M University, who made my time here a truly great experience. In particular, I am grateful to Dr. Yuteng Jin, Dr. Aditya Chakravarty and Dr. Hao Li, just to name a few, for their helpful discussions, guidance, and support. Finally, I would like to thank my family and my boyfriend for their unwavering support and understanding at any time, any place. Their love and encouragement have been the driving force behind my success, and I could not have achieved this milestone without them. CONTRIBUTORS AND FUNDING SOURCES Contributors This work was supervised by a dissertation Professor Siddharth Misra of the Department of Petroleum. The work in Chapter 1 was supported by Dr. Hao Li during the initial phase of this work. The HOSS simulation work depicted in C

FAQs

How does the dissertation use machine learning to monitor discontinuity propagation?

It analyzes multipoint waveform measurements recorded from a surface sensor array after a single impulse source to infer propagation behavior of embedded mechanical discontinuities.

What key limitation of conventional supervised learning does the work address?

Conventional methods often rely on statistical correlation or association and lack explicit domain knowledge and causality.

What is the purpose of introducing knowledge-driven causal signatures?

Causal signatures are used to develop a knowledge-driven causal model that yields a more versatile and resilient data system and demonstrates that traditional correlations and feature rankings are not statistically significant indicators of causality.

Aug 4, 2026

Upload by Aurora