The Analysis of Permeability Influence on CO2 Plume based on Reservoir Simulation

Abstract

Abstract Permeability is critical in site characterization and geoscience studies for carbon capture and storage projects. There is significant uncertainty of CO2 movement in the subsurface caused by permeability. This study aims to analyze the influence of permeability on the movement of CO2 plumes using a synthetic single-porosity model based on reservoir simulation. By simplifying geomechanics, employing tornado experimental design, as well as utilizing Pearson and Spearman correlations from 30,000 neural network proxy models based on Latin Hypercube and Monte Carlo experimental design, this research successfully explains the impact of permeability on CO2 movement. The reservoir can withstand pressures up to 6500 psi without rock failure. The total CO2 injection with a 90% probability is ∼1.3 TSCF equivalent, with an injection rate of 30 MMSCFD, optimized for a duration of 120 years. Initially, the CO2 movement tends to be upward during injection due to gravitational effects, but over time, it tends to move laterally. Vertical permeability exhibits a negative correlation with total CO2 injection, lateral distribution of CO2 plumes, and CO2 solubility in water. On the other hand, horizontal permeability shows a positive correlation with total CO2 injection. Factors such as perforation location and layer thickness also influence the extent of permeability’s role. However, the lateral and vertical movement of CO2 in this study has not been fully identified, and more complex quantification formulas are needed in commercial simulators. Despite the limitations in data and computational resources, this research successfully explains the movement of CO2 plumes and highlights the importance of other factors such as porosity, perforation location, layer thickness, among others. The recommendations of this study include the use of more complex quantification formulas to model the movement of CO2 plumes, along with increasing the complexity of the model.