A new experimental method for comparing solvents in steam-solvent coinjection for bitumen recovery under controlled thermodynamic conditions

Abstract

Solvent-aided steam-assisted gravity drainage (SA-SAGD) involves the interplay between phase behavior and fluid flow near the edge of a steam chamber, which affects the mixing of solvent with bitumen. The mixing of solvent with bitumen (i.e., dispersion) results in dilution of the bitumen and can improve the energy efficiency of SAGD. However, it is often difficult to analyze this complex interplay through large-scale steam injection experiments because of the transient chamber-edge thermodynamic conditions.This paper presents an experimental program that compares the bitumen gravity drainage with steam injection (SAGD) and solvent-steam co-injection (SA-SAGD) under controlled thermodynamic conditions. In addition to SAGD as the base case, two sets of SA-SAGD were performed with 20 mol% C4 and 10 mol% C8 in the coinjected vapor at 3500 kPa. The experiments used a sand-pack of 3-inch diameter and 15-inch length, which was placed in a 25-L cylindrical pressure vessel. The sand-pack was surrounded by one-inch annular space, into which the vapor phase was injected under controlled pressure, temperature, and composition. Oil production and temperature profiles inside and outside the sand-pack were recorded for all experiments. Excavated samples from the sand-pack were analyzed after the experiments.The total recovery factors for SAGD, C4-SAGD, and C8-SAGD were 78%, 84%, and 89%, respectively. The recovery factors at 1 h for SAGD, C4-SAGD, and C8-SAGD were 71%, 80%, and 85%, respectively. The peak oil rate was 9.8 cm3/min with SAGD, 14.6 cm3/min with C4-SAGD, and 31.3 cm3/min with C8-SAGD. The SA-SAGD cases resulted in markedly better results than the SAGD case, and C8-SAGD yielded more rapid oil drainage than C4-SAGD.The SAGD experimental data were history-matched using a numerical simulation model. Based on the calibrated numerical model, the SA-SAGD experimental data were history-matched by adjusting the dispersion coefficient to model the mixing between the solvent and bitumen. The apparent dispersion coefficients for C4 and C8 in bitumen were determined to be 0.012 m2/day and 0.093 m2/day, respectively. The experimental program verified in this research offers a way to systematically compare different solvents for SA-SAGD with their quantifiable dispersion coefficients under given chamber conditions.