Data-Driven vs. Physics-Driven Models: Engineering in a virtual Sub-Surface

It is true to say the oil & gas industry is transforming – in practice, in operations, and economics. But the physics have not changed – the elemental challenge flowing hydrocarbons from a rock matrix to the surface remains the same. Despite advanced numerical techniques including now various forms of advanced Machine Learning and Artificial Intelligence we struggle to accurately simulate and understand physical processes in the reservoir and surrounding subsurface: fluid flow, transient and static fluid pressures, hydraulic fracturing, wellbore integrity, reservoir compaction just to name a few. Conventionally, we have relied on computationally intensive physics-based models. More recently, particularly driven by the large publicly available data sets in US unconventionals, there is an increasing trend to use purely data-driven models and extract the necessary intelligence. In this webinar we explore the challenges, disparities and limitations of physics-driven and data-driven models; and the opportunities of combining both approaches, as well as borrowing from other engineering and science disciplines that share the same or analogous physical laws.

Speaker Bios

Ravinath Kausik

Dr. Ravinath Kausik K.V is a Senior Research scientist in the Applied Math & Data Analytics Department at Schlumberger-Doll Research Center in Cambridge, USA where he focuses on Machine Learning and AI applications for subsurface characterization. He has previously led the development of novel NMR and petrophysical techniques, especially of unconventional shale gas and tight oil formations leading to methods such as TGIP-NMR and RPI for unconventional shale gas and tight oil plays, respectively. He has also worked on the development of next generation NMR diffusion and relaxation measurements for both laboratory and downhole applications. His work has been recognized with the Conrad Schlumberger prize for technical depth and the Henry-Doll prize for innovation within Schlumberger, and he served as the distinguished speaker for SPWLA in 2015, 2016 and 2018. He was elected to the international advisory committee of the Magnetic Resonance in Porous Media (MRPM). He obtained a M.Sc. from IIT Madras, India and Ph.D. degree in physics from the Universität Ulm, Germany. He has worked as a postdoctoral fellow at the University of California, Santa Barbara before joining Schlumberger-Doll Research in 2009. He has co-authored more than 30 peer-reviewed publications and several patent applications and is a scientific reviewer for more than 10 international journals.

Satish Sankaran

Satish Sankaran has over 20 years of diversified industry experience in technology development, consulting, project execution and management working on several international, deepwater and US onshore projects. His areas of specialization include digital oilfield technologies, reservoir management, field development optimization, uncertainty analysis, production operations and advanced process automation.

At Xecta Digital Labs, he leads an engineering team in the development of digital solutions for energy industry by fusing physics and data analytics methods for applications in reservoir, production, facilities and downstream processes.

Prior to this, Sathish had worked at Halliburton and Anadarko in technical consulting, product management and advisor roles in reservoir and production engineering and advanced process automation. In his most recent role at Anadarko, he led an engineering team, focused on technology development for digital operations, advanced simulation modeling and application of data analytics for drilling, completions, production and facility engineering.
Sathish is a member of Society of Petroleum Engineers (SPE) and served in several roles including advisory positions to the Technical Directors, chairperson and committee member in several industry initiatives. He has co-authored several publications and more recently, the SPE Technical Report on Data Analytics in Reservoir Engineering.

Sathish has a B.Eng. (Honors) degree in Chemical Engineering from Birla Institute of Technology and Science (BITS – Pilani, India), M.S degree in Chemical Engineering from University of Cincinnati and Ph.D. degree in Chemical Engineering from University of Houston.

Dr. Sebastien Matringe

Sebastien Matringe is a Principal Advisor for Subsurface Technology at Hess. In his role, he is responsible for identifying and piloting emerging subsurface technologies that could add value to Hess. He has a broad reservoir engineering background, which include work on a number of fields throughout the world (USA, Middle-East, South-America, Africa) both onshore and offshore, conventional reservoirs and shale plays. Prior to joining Hess, Sebastien has worked for Newfield Exploration, Quantum Reservoir Impact and Chevron. His work experience include positions in reservoir simulation, reservoir engineering, research and technology development as well as in team, project and executive management. Sebastien holds a Diplôme d’Ingénieur from ENSEEIHT (France) in Fluid Mechanics as well as a MS and PhD in Petroleum Engineering from Stanford University.

Weichang Li

Weichang Li leads the Machine Learning Group at Aramco’s Houston Research Center. His research involves developing machine learning and signal processing algorithms for geoscience/petroleum engineering applications. Prior to Aramco, he had been with ExxonMobil’s Corporate Strategic Research lab. since 2008 where he led the machine learning team from 2011-2014. Weichang has co-organized 2018 and 2019 SEG machine learning workshop, 2018 SIAM Data Mining workshop in Geoscience Applications, and is the associate editor for Geophysics special section on Machine Learning. Weichang obtained his M.S. (dual) in Electrical Engineering and Computer Sciences, and Ocean Engineering (2002), and Ph.D. in Electrical and Oceanographic Engineering (2006), all from MIT. He also received an Office of Naval Research (ONR) postdoctoral fellowship at Woods Hole Oceanographic Institution from 2006-2007.

Dr. Siddharth Misra

Prof. Siddharth Misra is an Associate Professor in Harold Vance Department of Petroleum Engineering at Texas A&M University. Misra holds a Ph.D. in Petroleum Engineering from The University of Texas at Austin. Prior to that, from 2007 to 2010, he worked as a Wireline Field Engineer in Saudi Arabia, Egypt, and USA with Halliburton. He received his undergraduate degree in Electrical Engineering from Indian Institute of Technology Bombay, India, in 2007. Recently, he was awarded the prestigious Department of Energy Early Career Award, American Chemical Society New Investigator Award, and SPE Mid-Continent Formation Evaluation Award. His research interest include subsurface characterization, machine learning, sensing and sensors, and inverse problems.

Kriti Singh

Kriti Singh is working as a Data Scientist in the Research and Development team at Corva. She has been leading the company’s efforts on using machine learning models for Rate of Penetration Optimization in Real Time Drilling Operations in collaboration with shale operators. She is also actively involved in other Data Analytics projects including automating well-control kick detection for onshore and offshore drilling rigs. Kriti has a B-Tech degree in Civil Engineering from National Institute of Technology, Karnataka and a Master’s and Doctorate in Petroleum Engineering both from University of Tulsa with specialization in Drilling Engineering. She has been a recipient of several science fellowships and has held volunteering roles in SPE as a student and YP, the most recent being a member of the Data Analytics Group from the Gulf Coast Section.