Field-Scale Optimization of CO₂ Injection Schedule with Critical Pressure and Plume Boundary

This study aims to develop a field-scale CO₂ injection optimization workflow to refine the injection strategy and enhance storage capacity while managing the extent of the Area of Review (AoR). We utilize the concepts of critical pressure and plume boundary to evaluate the AoR delineation during injection optimization. The optimization workflow is enhanced by an advanced upgridding technology to accelerate simulations. We evaluate the feasibility of CO₂ storage with AoR delineations using varied injection schedules.

We employ an advanced upgridding technique to accelerate the optimization workflow. Our approach involves an ‘optimal’ layer design through sequential coarsening using a distance-based error metric while minimizing the loss of heterogeneity between fine-scale and coarsened models. A recursive analysis is applied to develop the optimal layering while preserving geological heterogeneity and reducing the number of layers for computational efficiency. After optimal layer design, a novel pressure transient upscaling is introduced to upscale transmissibility based on local flow solutions, which preserves reservoir quality and reservoir flow barrier and is particularly well-suited for high-contrast systems.

Our proposed workflow is tested to optimize the injection schedule for field-scale application of CO₂ storage in saline sands along the Texas Gulf Coast. Field-scale evaluation and optimization of CO₂ storage using a high-resolution compositional model can be computationally demanding. This is compounded by the fact that optimization of the injection schedule requires a large number of flow simulations to examine a multitude of scenarios. First, the performance of the upgridding scheme was evaluated based on computational speed up and comparison of the well responses and dynamic reservoir properties from the fine-scale model. We examined three different approaches for distance-based errors to quantify the loss of heterogeneity: velocity-based distance, slowness-based distance, and hyper-volume weighted distance. The optimal layer design scheme resulted in approximately an order of magnitude speed up in computation time, which greatly facilitated the field-scale CO₂ injection optimization. The injection schedule was optimized to maximize storage capacity while minimizing the AoR extent to avoid potential contamination of USDW. The injection zone selection involved the evaluations of flow unit geometry, reservoir properties, zone interference, and their influence on the AoR extent. The optimization of the injection schedule was carried out iteratively and heuristically, applying learnings from previous simulation scenarios.

The proposed workflow provides an efficient way to evaluate AoR extent and optimize the injection schedule while providing valuable guidelines for future carbon storage projects.