Machine Learning in the Oil and Gas Industry: Including Geosciences, Reservoir Engineering, and Production Engineering with Python

ISBN-13: 9781484260937 / Angielski / Miękka / 2020 / 300 str.

Yogendra Narayan Pandey; Luigi Saputelli; Srimoyee Bhattacharya

Spis treści

Machine Learning in the Oil and Gas Industry with Python 

Chapter 1:  Towards Oil and Gas 4.0

Chapter Goal: This chapter provides an overview of the digital transformation state-of-the-art in the Oil & Gas industry. The overview includes a literature review of the publications from the academic and industrial institutions, available in the public domain. It follows a theme of end-to-end Oil & Gas exploration and production project lifecycle.

Chapter 2:  Python Programming Primer

Chapter Goal: This chapter provides a brief primer of the Python programming language. The idea is to make the user familiar with the basic syntax on Python programming language. This chapter also briefly touches on the numpy, pandas, and a selected visualization (to be selected from matplotlib/seaborn/bokeh) library.

Chapter 3: Overview of Machine and Deep Learning Concepts

Chapter Goal: This chapter introduces supervised and unsupervised machine learning concepts with the code examples using simplistic and clean data sets. The aim is to provide readers with understanding of practical concepts of different machine and deep learning algorithms, along with simple coding examples. Scikit-learn and Keras will be used for machine and deep learning code samples respectively.

Chapter 4: Geophysics and Seismic Data Processing

Chapter Goal: This chapter will focus on using seismic data available from open data sources, e.g., Equinor Volve project, to provide two example applications for seismic data interpolation, and fault identification. Further, it will also discuss other problems, such as, horizon identification, and salt dome identification, without going in to too much details, while providing enough pointers and resources to the interested users.

Chapter 5: Geomodeling

Chapter Goal: This chapter focuses on the geological modeling problems, including unsupervised learning for clustering different rock types based upon the petrophysical well logs, and estimation of the petrophysical properties away from the well locations by applying supervised machine learning techniques.

Chapter 6: Reservoir Engineering

Chapter Goal: This chapter focused on the approaches for developing machine learning based proxy models to replace a full-physics reservoir simulator, and the use of these proxy models for generating production forecasts. The chapter will also cover related topics of interest including well placement optimization, and planning future wells based upon the historical production data.

Chapter 7: Production Engineering

Chapter Goal: This chapter will cover the topic of production modeling using machine learning methodologies. The topics will include identification of specific completion design for a well to achieve optimal production rates, and identifying the producing wells, which may benefit from the workover activities. A part of chapter will also provide methodology for equipment failure analytics, and predictive maintenance for production equipment, e.g., electrical submersible pumps (ESPs).

Chapter 8: Opportunities, Challenges, and Expected Future Trends

Chapter Goal: This chapters gleans over the challenges arising in the execution of the machine learning based digital transformation projects, the pitfalls leading to the project failure. Also, the opportunities that inherently lie in addressing these challenges are discussed from both the executive and practitioners’ perspective. Finally, an overview of the expected roadmap for the industry over the next decade will be discussed.

O autorze

Yogendra Pandey is a senior product manager at Oracle Cloud Infrastructure. He has more than 14 years of experience in orchestrating intelligent systems for the oil and gas, utilities, and chemical industries. He has worked in different capacities with oil and gas, and utilities companies, including Halliburton, ExxonMobil, and ADNOC. Yogendra holds a bachelor’s degree in chemical engineering from the Indian Institute of Technology (BHU), and a PhD from the University of Houston, with specialization in high-performance computing applications to complex engineering problems. He served as an executive editor for the Journal of Natural Gas Science and Engineering. Also, he has authored/co-authored more than 25 peer-reviewed journal articles, conference publications, and patent applications. He is a member of the Society of Petroleum Engineers.

 

Ayush Rastogi is a data scientist at BPX Energy, Denver CO. His research interests are based on multi-phase fluid flow modeling and integrating physics-based and data-driven algorithms to develop robust predictive models. He has published his work in the field of machine learning and data-driven predictive modeling in the oil and gas industry. He has previously worked with Liberty Oilfield Services in the technology team in Denver, prior to which he worked as a field engineer in TX, ND, and CO as a part of his internship. He also has experience working as a petroleum engineering consultant in Houston, TX. Ayush holds a PhD in petroleum engineering with a minor in computer science from Colorado School of Mines, and is an active member of the Society of Petroleum Engineers. 

 

Sribharath Kainkaryam leads a team of data scientists and data engineers at TGS. Prior to joining TGS in 2018, he was a research scientist working on imaging and velocity model building challenges at Schlumberger. He graduated with a masters in computational geophysics from Purdue University and has an undergraduate degree from the Indian Institute of Technology, Kharagpur.

 

Srimoyee Bhattacharya is a reservoir engineer in the Permian asset team in the Shell Exploration and Production Company. She has over 11 years of combined academic and professional experience in the oil and gas industry. She has worked in reservoir modeling, enhanced oil recovery, history matching, fracture design, production optimization, proxy modelling, and applications of multivariate analysis methods. She also worked with Halliburton as a research intern on digitalization of oil fields and field-wide data analysis using statistical methods. Srimoyee holds a PhD in chemical engineering from the University of Houston, and a bachelor’s degree from the Indian Institute of Technology, Kharagpur. She has served as a technical reviewer for the SPE Journal, Journal of Natural Gas Science and Engineering, and Journal of Sustainable Energy Engineering. She has authored/co-authored more than 25 peer-reviewed journal articles, conference publications, technical reports, and patent application.

 

Luigi Saputelli is a reservoir management expert advisor to ADNOC and Frontender Corporation with over 28 years of experience. He worked in various operators and services companies around the world including PDVSA, Hess, and Halliburton. He is a founding member of the Real-time Optimization TIG and Petroleum Data-driven Analytics technical section of the Society of Petroleum Engineers, and recipient of the 2015 Society of Petroleum Engineers international production and operations award. He also received the 2007 employee of the year award from Halliburton. He has published more than 90 industry papers on applied technologies related to reservoir management, real-time optimization, and production operations. Saputelli is an electronic engineer with a masters in petroleum engineering, and a PhD in chemical engineering. He also serves as managing partner in Frontender Corporation, a petroleum engineering services firm based in Houston.

Blurb

Apply machine and deep learning to solve some of the challenges in the oil and gas industry. The book begins with a brief discussion of the oil and gas exploration and production life cycle in the context of data flow through the different stages of industry operations. This leads to a survey of some interesting problems, which are good candidates for applying machine and deep learning approaches. The initial chapters provide a primer on the Python programming language used for implementing the algorithms; this is followed by an overview of supervised and unsupervised machine learning concepts. The authors provide industry examples using open source data sets along with practical explanations of the algorithms, without diving too deep into the theoretical aspects of the algorithms employed. Machine Learning in the Oil and Gas Industrycovers problems encompassing diverse industry topics, including geophysics (seismic interpretation), geological modeling, reservoir engineering, and production engineering. 

Throughout the book, the emphasis is on providing a practical approach with step-by-step explanations and code examples for implementing machine and deep learning algorithms for solving real-life problems in the oil and gas industry. 

You will:

• Understanding the end-to-end industry life cycle and flow of data in the industrial operations of the oil and gas industry
• Get the basic concepts of computer programming and machine and deep learning required for implementing the algorithms used
• Study interesting industry problems that are good candidates for being solved by machine and deep learning
• Discover the practical considerations and challenges for executing machine and deep learning projects in the oil and gas industry