Fast and Robust Data-driven surface and subsurface simulation for leveraging decarbonization efforts in the Oil & Gas Industry

Sustainable hydrocarbon production in light of a decarbonization paradigm demands complex decision-support strategies involving fast risk assessment and optimal injection-production scheduling. At the core of these decisions is the prediction of reservoir performance and surface networks (e.g., CO2 pipeline), usually done by running computationally demanding complex simulators.  As a substitute, physics-aware machine learning (ML) techniques have been used to endow data-driven proxy models with features closely related to the ones encountered in nature, especially conservation laws.  They can lead to fast, reliable, and interpretable simulations used in many reservoir management workflows. In this talk, I will build upon our recently developed deep-learning-based reduced-order modeling framework for fast and reliable proxy for reservoir simulation and drilling operations. I will show examples in C02 sequestration and storage, optimal drilling operations and in designing  large C02 pipeline networks.