DSDE: In Theory

In this study, a deep-neural-network-based workflow with enhanced efficiency and scalability is developed for solving complex history-matching problems.

This study presents a production-optimization method that uses a deep-learning-based proxy model for the prediction of state variables and well outputs to solve nonlinearly constrained optimization with geological uncertainty.

The paper describes a parameter inversion of reservoirs based on featured points, using a semi-iterative well-test-curve-matching approach that addresses problems of imbalanced inversion accuracy and efficiency.

Technology and partnerships play a pivotal role in how the oil industry finds and produces energy from frontier regions and brownfields, both now and in the future.

The authors propose a hybrid virtual flow and pressure metering algorithm that merges physics-based and machine-learning models for enhanced data collection.

The authors introduce a novel framework combining dynamic mode decomposition, a data-driven model-reduction technique, with direct data assimilation to streamline the calibration of carbon-dioxide plume evolution models.

A detailed comparison of two leading software platforms demonstrates the effective application of these platforms in modeling complex reservoir dynamics and biochemical reactions in geological formations for risk assessment in underground hydrogen storage.

Regional pore-pressure variations in the Leonardian- and Wolfcampian-age producing strata in the Midland and Delaware basins are studied using a variety of subsurface data.

The authors of this paper present a novel joint-domain full waveform inversion framework optimizing travel-time accuracy in both data and model domains.

The main goal of this research work was to determine subseismic faults and fracture corridors and their characteristics, including density and orientation, for a Paleocene fractured carbonate reservoir.