Applications of Machine Learning in Subsurface Reservoir Simulation—A Review—Part II

Abstract

In recent years, Machine Learning (ML) has become a buzzword in the petroleum industry, with numerous applications which guide engineers in better decision making. The most powerful tool that most production development decisions rely on is reservoir simulation with applications in multiple modeling procedures, such as individual simulation runs, history matching and production forecast and optimization. However, all of these applications lead to considerable computational time and computer resource-associated costs, rendering reservoir simulators as not fast and robust enough, and thus introducing the need for more time-efficient and intelligent tools, such as ML models which are able to adapt and provide fast and competent results that mimic the simulator’s performance within an acceptable error margin. In a recent paper, the developed ML applications in a subsurface reservoir simulation were reviewed, focusing on improving the speed and accuracy of individual reservoir simulation runs and history matching. This paper consists of the second part of that study, offering a detailed review of ML-based Production Forecast Optimization (PFO). This review can assist engineers as a complete source for applied ML techniques in reservoir simulation since, with the generation of large-scale data in everyday activities, ML is becoming a necessity for future and more efficient applications.

Enhancement of Machine-Learning-Based Flash Calculations near Criticality Using a Resampling Approach

“…The equality of fugacities dictates that at equilibrium, the fugacities of each component in the two phases must be equal, i.e., f V i = f L i , such that the Gibbs free energy of the final two-phase system is minimized [11]. Reservoir model consisting of millions of grid blocks [6].…”

“…Imposing specific k-values essentially ensures the equality of component fugacities. Reservoir model consisting of millions of grid blocks [6].…”

Optimization of Well Placement in Carbon Capture and Storage (CCS): Bayesian Optimization Framework under Permutation Invariance

“…Petrophysical and fluid properties, such as porosity, permeability, fluid saturation, compressibility, and relative permeability, are crucial inputs for reservoir simulation [22,23]. Accurate characterization of these properties is essential for realistic simulation results.…”

Prospects of an Acid Gas Re-Injection Process into a Mature Reservoir

“…This output is then integrated into simulation models that represent physical components within the hydrocarbon production chain, which encompass systems responsible for producing fluids at the surface (wellbore) and their subsequent processing (surface facilities). This comprehensive modeling approach extends all the way to the final sales point, ensuring a thorough representation of the entire production system and allowing its detailed study and optimization [32,33].…”

“…Contrariwise, a dynamic reservoir model operates as a transient, time-dependent simulation of fluid flow within the reservoir and builds upon the static model by integrating production history, fluid properties, and reservoir management strategies. Dynamic models are essential in predicting reservoir behavior, optimizing production, and evaluating diverse development scenarios, such as EOR strategies and the effects of new wells on production [32,33].…”