A Workflow to Identify Target Block and Optimize Cyclic CO₂ Injection for Enhancing Oil Recovery in Shale Reservoirs

Cyclic CO₂ injection has been demonstrated to be an effective method for enhancing oil recovery in shale reservoirs. However, its applications in oil fields are sensitive to uncertainties related to reservoir properties and operational strategies. We propose a workflow based on our species transport model, which can capture key transport mechanisms in shale reservoirs to select the appropriate target block and optimize cyclic CO₂ injection operating parameters to maximize total oil production. A single-well cyclic CO₂ injection compositional simulation is developed based on a multiphase, multicomponent species transport model. This model accounts for key transport mechanisms such as viscous flow, molecular diffusion, and Knudsen diffusion in shale reservoirs. We employ a least-squares support vector regression (LS-SVR) as a proxy for the compositional reservoir simulation model to reduce computational costs in subsequent robust optimization processes. Training and validation samples generated from the compositional reservoir simulation include reservoir properties and operational strategies within the set of optimal variables, with oil recovery as the objective function. Finally, this LS-SVR proxy model serves as a forward model to conduct robust optimization through a genetic algorithm. The results and discussion section presents three optimization scenarios, progressively incorporating more variables. The LS-SVM proxy model accurately predicts oil recovery, demonstrating its high accuracy even with a small training set. In the first scenario, we confirm that neither short-term HnP with more cycles nor long-term HnP with fewer cycles enhances the well performance of CO₂ HnP. A thorough optimization process is crucial to achieve higher oil recovery, potentially increasing CO₂ HnP recovery from 12.23% to 15.58% through the design of operational parameters in the first scenario. In the second scenario, we conduct an optimization based on parameters from the Eagle Ford shale oil reservoir. A workflow that includes compositional simulation, proxy modeling, and optimization algorithms is broadly applicable and effectively reduces the risks associated with conducting CO₂ HnP. A larger volume of CO₂ injection ensures higher enhanced oil recovery, as it allows CO₂ to penetrate deeper into the formation and mix with crude oil. In the final scenario, we seek insights to identify target blocks for CO₂ HnP. Deep reservoirs containing low gas oil ratio (GOR) black oil are particularly suitable for cyclic CO₂ HnP, as the injected CO₂ significantly enhances oil swelling.