Investigation of Cyclic Solvent Injection Process for Heavy Oil Recovery

Summary This paper summarizes numerical and experimental simulation results of a cyclic solvent injection process study, which was part of a continuing investigation into the use of solvents as a follow-up process in Cold Lake and Lloydminster reservoirs that have been pressure-depleted by cold heavy oil production with sand (CHOPS). Typically only 5% - 10% of the original oil in place (OOIP) is recovered during cold production; therefore, an effective follow-up process is required. The cyclic solvent injection (CSI) experiment consisted of primary production followed by six solvent (28% C3H8 - 72% CO2) injection cycles. Oil recovery after primary production and six solvent cycles was 50%, which indicates the potential viability of the CSI process. Concurrently with the laboratory physical simulation, a numerical simulation model was developed to represent the physical behaviour of the experimental results. A history match of the primary production portion of the experiment was obtained using an Alberta Innovates - Technology Futures (AITF) foamy oil model. This resulted in the characterization (fluid saturations and pressures) of the oil sandpack at the start of the solvent injection process. The history match of the subsequent six solvent injection cycles was used to validate the numerical model of the CSI process developed at AITF. This model includes nonequilibrium rate equations that simulated the delay in solvent reaching its equilibrium concentration as it dissolves or exsolves in the oil in response to changes in the pressure and/or gas-phase composition. Dissolution of CH4, C3H8 and CO2 in oil and CO2 in water were considered, as was exsolution of CH4, C3H8 and CO2 from oil and CO2 from water. Reduced gas-phase permeabilities resulting from gas exsolution were also included. The history match simulations indicated that: The important mechanisms were represented in the simulations. Significant oil swelling by solvent dissolution occurs during solvent injection periods. This can reduce solvent injectivity and penetration into a heavy oil reservoir during solvent injection periods. Low oil and gas-phase relative permeabilities are required during production periods to match the experimental oil and gas production during solvent cycles. A parametric simulation study showed that the quantity of gas injected in an injection period was relatively insensitive to the oil-phase diffusion coefficients, but was sensitive to solvent solubility in oil, dissolution rates, gas-phase diffusion coefficients, molar densities in the oil phase, gas-phase relative permeability and capillary pressure. It was shown that oil production is highly dependent on how quickly solvent can dissolve in the oil during injection and exsolve from the oil during production.