DLP Webinar on A Scalable Deep Learning Surrogate Model for Efficient Reservoir Performance unde Prediction r Geological Uncertainties by Prof. Mustafa Onur

DLP Webinar: A Scalable Deep Learning Surrogate Model for Efficient Reservoir Performance Prediction under Geological Uncertainties 

Webinar details

Instructor:	Prof. Mustafa Onur
Duration:	40 min include Q&A
Main topics:	Closed-Loop Reservoir Management and Life- Cycle Production Optimization  Machine Learning Applications in Production Optimization  Gradient-Based Optimization under Nonlinear Constraints (SQP + StoSAG)  Integration of High-Fidelity Simulators and Deep-Learning Proxy Models  Field-Scale Application and Benchmarking
Language:	English

Next Delivery: 18 August 2026 at 9AM CEST

 Request this Webinar   Overview of all scheduled webinars    Read Paper in EarthDoc

Register Here (Click)

Attending webinars and access to recent EarthDoc material is free of charge for EAGE members, join here.

Description

This lecture presents advanced methods for life-cycle production optimization within closed-loop reservoir management to maximize net present value (NPV). It introduces a scalable, efficient deep learning (Embed-to-Control- based Reduced Order) surrogate coupled with a gradient-based framework using sequential quadratic programming (SQP) with stochastic gradients for nonlinearly constrained problems. Integration with high-fidelity simulators and this machine learning-based proxy is demonstrated on a large-scale two and three-phase reservoir cases, highlighting efficient, practical solutions under realistic operational constraints.

Participant profile

Geoscientists, Reservoir Engineers & Asset Teams,  Production & Petroleum Engineers, Reservoir Simulation & Optimization Specialists, Data Scientists / ML Engineers in Energy, R&D Professionals in Energy Companies & Service Firms, Applied Mathematicians, Graduate Students & Researchers

About Instructor

Mustafa Onur is the McMan Endowed Chair Professor at the McDougall School of Petroleum Engineering at the University of Tulsa and Director of the TU Petroleum Reservoir Exploitation Projects (TUPREP). He is also Emeritus Professor in the Department of Petroleum and Natural Gas Engineering at Istanbul Technical University, Türkiye. He previously served as department head at both the University of Tulsa (2016–2020) and Istanbul Technical University (2005–2011), and held the Schlumberger Professorial Chair at Universiti Teknologi Petronas (2012–2014), Malaysia. He was also a faculty member at King Saud University (1994–1997), Saudi Arabia. His research focuses on inverse problem theory, mathematical optimization, and data science, with applications to reservoir management, history matching, uncertainty quantification, formation testing, PTA, RTA, and TTA methods, and hydrothermal and enhanced geothermal systems. Dr. Onur holds BS, MS, and PhD degrees in petroleum engineering from Middle East Technical University and the University of Tulsa. He serves as Technical Editor for the SPE Journal and Associate Editor for Geoenergy Science and Engineering. He is a recipient of the 2010 SPE Formation Evaluation Award and the 2018 SPE Reservoir Description and Dynamics Award, and has been an SPE Distinguished Member since 2014.