Page 53 - Machine Learning for Subsurface Characterization
P. 53
Chapter 2
Unsupervised clustering
methods for noninvasive
characterization of
fracture-induced
geomechanical alterations
Siddharth Misra*, Aditya Chakravarty*, Pritesh Bhoumick †,a and
Chandra S. Rai ‡
*
Harold Vance Department of Petroleum Engineering, Texas A&M University, College Station,
†
‡
TX, United States, PricewaterhouseCoopers (PwC), Houston, TX, United States, The University
of Oklahoma, Norman, OK, United States
Chapter outline
1 Introduction 39 6 Results and discussions 53
2 Objective of this study 41 6.1 Effect of feature engineering 53
3 Laboratory setup and 6.2 Effect of clustering method 55
measurements 41 6.3 Effect of dimensionality
4 Clustering methods for the reduction 57
proposed noninvasive 6.4 Effect of using features
visualization of geomechanical derived from both prefracture
alterations 44 and postfracture waveforms 58
4.1 K-means clustering 44 7 Physical basis of the fracture-
4.2 Hierarchical clustering 45 induced geomechanical
4.3 DBSCAN 47 alteration index 59
5 Features/attributes for the 8 Conclusions 61
proposed noninvasive Acknowledgments 62
visualization of geomechanical Declarations 62
alteration 48 References 62
5.1 Feature engineering 49
5.2 Dimensionality reduction 52
a
Present address: Pricewaterhouse Coopers, Houston, TX, United States.
Machine Learning for Subsurface Characterization. https://doi.org/10.1016/B978-0-12-817736-5.00006-5
© 2020 Elsevier Inc. All rights reserved. 39
