Production Optimization of CO₂ Huff-n-Puff Process in Unconventional Oil Reservoirs with Effects of Gas Relative Permeability Hysteresis, Geomechanics, and Capillary Pressure

In this study, we introduce a physics-based numerical optimization framework that integrates key factors like hysteresis in gas permeability, the impact of geomechanics, and the effects of high capillary pressure, enabling a more realistic optimization of production in CO₂ Huff-and-Puff (HnP) processes. We utilize a commercial multi-porosity compositional simulator, fully coupled with geomechanics, to simulate the CO₂ HnP process for unconventional oil resources. In the numerical model, compositional PVT flash calculations are coupled with the capillary pressures within nano-sized pores, while molecular diffusion remains consistently present throughout the simulations. The Killough's hysteresis model is used for relative permeability calculations. For maximizing net present value (NPV) in production optimization with the CO₂ HnP process, both the Stochastic Simplex Approximate Gradient (StoSAG) method and the iterative least-squares support regression (LSSVR) proxy method are employed and compared. The optimization (or design) variables could be any combination of the variables like CO₂ injection rate, production BHP, durations of the injection, soaking, and production periods overall length of each cycle. Nano-pore confinement results in the suppression of the saturation curve in PVT flash calculations, consequently leading to an increase in oil production. Conversely, the presence of hysteresis, when combined with capillary pressures, tends to have detrimental effects. CO₂ injection around the minimum miscibility pressure point can enhance the recovery of oil, complemented by the positive effect of molecular diffusion. The iterative LSSVR proxy method is highly effective for production optimization in the complex CO₂ HnP process, significantly reducing the computation time at least 5-10 times faster depending on the case over a stochastic gradient method.

Production from unconventional oil systems like shale oil or tight liquid reservoirs through primary methods yields low production rates due to rapid declines caused by extremely low reservoir matrix permeability. Despite the advantages of high reservoir pressure and low oil viscosity, challenges arise due to nano-scale pore sizes and reduced connectivity compared to conventional reservoirs (Ataceri et al., 2023). Typically, the primary oil recovery factor (RF) from these formations falls below 10-12% (Clark 2009; Gherabati et al., 2018; Kuuskraa 2019). Thus, implementing enhanced oil recovery techniques in tight-oil reservoirs following primary production holds the potential to considerably enhance oil recovery.