An efficient multiscale simulation framework integrating dynamic heterogeneity for accurate waterflooding prediction.

Area of Science:

• Petroleum Engineering
• Reservoir Simulation
• Computational Geoscience

Background:

• Waterflooding is vital for oil and gas production in China.
• Commercial simulators struggle to accurately model dynamic reservoir heterogeneity induced by waterflooding.
• This limitation hinders precise prediction of remaining oil, especially in high water-cut stages.

Purpose of the Study:

• To develop and validate an efficient and accurate multi-scale simulation method for modeling dynamic reservoir heterogeneity.
• To incorporate time-varying absolute permeability (k) and relative permeability (k_r) driven by surface flux.
• To improve the prediction of oil recovery in high water-cut reservoirs.

Main Methods:

• Proposed an improved multi-scale finite volume (IMsFV) method to solve pressure equations on multi-scale grids.
• Incorporated time-varying absolute permeability (k) and relative permeability (k_r) driven by surface flux.
• Validated the method using the SPE10 benchmark, comparing simulation time and accuracy against a fully implicit method.

Main Results:

• Achieved significant reductions in simulation time (95.07% total, 98.19% linear solver) with errors < 5% compared to the fully implicit method.
• Demonstrated that dynamic relative permeability (k_r) enhances fluid mobility and reduces residual oil saturation, dominating over dynamic absolute permeability (k) effects.
• Sensitivity analysis showed enhanced recovery by 28.88-32.87% at 99% water cut and amplified gains with increasing injection rates.

Conclusions:

• The proposed multi-scale simulation method accurately captures dynamic heterogeneity, outperforming commercial simulators.
• Dynamic relative permeability plays a crucial role in improving oil recovery during high water-cut stages.
• This approach offers an efficient and accurate tool for predicting remaining oil in mature, high water-cut reservoirs.