Depleted gas reservoirs are viable choices for large-scale CO₂ storage and to displace remaining methane volumes to further increase the storage capacity (EGR). However, deployment of such projects depends on an informed knowledge of the magnitude of mixing of the miscible gases, efficiency in displacing in-situ methane by CO₂, composition of the produced gas, and CO₂ storage capacity. This study focuses on the fundamental analysis of mixing during CO₂-EGR using a numerical approach. We propose to conduct very fine grid compositional simulations to provide insights into the mixing of CO₂ and methane in a gas reservoir at different reservoir and operational conditions. We first analyze a stratified layer model to understand the basic mechanisms of scale-dependency of dispersion and the significance of reservoir heterogeneity on fluid mixing. To consider more realistic reservoir heterogeneity, a two-dimensional stochastic reservoir model is analyzed to estimate dispersivity generated as fluids flow in porous media at different scales. Reservoir heterogeneity is represented by the Dykstra Parsons coefficient (V_DP) and autocorrelation length, and fluid properties are modeled depending on pressure and temperature conditions. Field-scale simulation is also performed to discuss the way dispersion is modeled in reservoir simulation affects simulated gas recovery. Our study shows that the variance of permeability and convective spreading are the primary causes of fluid mixing at any scale. In addition, molecular diffusion is not always negligible in gas mixing even in large-scale heterogeneous reservoirs since gas has much larger diffusivity than liquid. Furthermore, the mechanism of fluid mixing during CO₂-EGR is complex with the interplay between convective spreading, transverse dispersion (including molecular diffusion), and gravity segregation. Although geoscientists often assume numerical dispersion can represent physical dispersion, our study indicates this is an oversimplification and could cause significant errors in calculated gas recovery. Permeability heterogeneity is essential for the dispersion growth process and the final displacement behavior. Reservoir heterogeneity should be modeled with high-resolution grid models to analyze mixing behaviors more accurately.

Depleted gas reservoirs are attractive sites for Carbon Capture and Storage (CCS) due to their huge storage capacities, proven seal integrity, existing infrastructure and subsurface data availability. However, CO₂ injection into depleted formations can potentially lead to hydrate formation near the wellbore due to Joule-Thomson cooling, which might cause injectivity issues. Some challenges encountered when modeling and simulating this process are the computational time caused by Newton's convergence issues and instability. The objective of this work is to propose a novel approach for hydrate risk assessment during CO₂ injection into depleted gas reservoirs using physics-based Machine Learning (ML) approach. First, the selection of input parameters for the ML models is performed based on sensitivity study results using an analytical solution for different operational and petrophysical values. Then the ML models are tuned and tested using datasets from numerical reservoir simulation results based on a wide range of input parameter values. To the best of our knowledge, this is the first time that an ML approach is used for risk assessment of CO₂ hydrate in its storage in depleted gas reservoirs. The ML models developed in this study presented an efficient performance to predict hydrate-forming events. The deep neural network model performed best with a 95% recall value and 84% precision value. These results show that the ML model can be further utilized for risk assessment in the screening stage, and the combination of screening by ML, followed by detailed analysis with numerical simulation in high-risk cases can be an efficient probing workflow for future CCS projects.