Compositional Reservoir Simulation Sensitivity Studies: Grid, Permeability, and Well Configuration Analysis Using OPM Flow

This technical report presents a comprehensive parametric sensitivity analysis of compositional reservoir simulation using the open-source OPM Flow simulator, extending previous gas injection feasibility studies (Padder, 2026) through three systematic investigations: (1) grid refinement analysis comparing coarse (7×7×3, 147 cells) and refined (14×14×6, 1,176 cells) spatial discretization under injection rates of 12,000–15,000 rb/day, (2) permeability sensitivity analysis examining 0.5× and 2× permeability multipliers under variable injection rates, and (3) multi-well configuration analysis evaluating 1-injector-1-producer, 2-injector-2-producer, and 1-injector-4-producer well patterns. Results establish that the numerical stability boundary at 15,000 rb/day injection is grid-independent, with both coarse and refined grids exhibiting convergence failure, demonstrating that operational limits are governed by physical displacement instabilities rather than spatial discretization artifacts. Grid refinement increases computational cost by 3.7× (4.19s to 15.57s) and linearizations by 15% (1,289 to 1,492) but does not extend the stable injection envelope. High-permeability systems (2× multiplier) achieve 79% reduction in gas-oil ratio (1,280–1,560 vs 6,153–6,382 Mscf/stb) and tolerate aggressive injection rates, while low-permeability systems exhibit rate sensitivity with optimal performance at 14,000 rb/day. Multi-well configurations reduce computational timesteps by 90% (263 to 24–25 timesteps), demonstrating enhanced pressure equilibration and numerical tractability for complex well architectures. These findings provide quantitative design guidance for CO₂-EOR conversion projects, establishing reservoir permeability as the dominant feasibility parameter while confirming robustness of coarse-grid screening simulations for operational envelope assessment.