Simultaneous three-phase flow of gas, oil and water is a common phenomenon in enhanced oil recovery techniques such as water-alternating-gas (WAG) injection. Reliable reservoir simulations are required to predict the performance of these injections before field application. However, heavy oil displacement by gas or water can lead to viscous fingering due to the unfavorable mobility ratio between heavy oil and the displacing fluid. In addition, the injection of partially dissolvable gases such as CO 2 can result in compositional effects, which can bring about a significant reduction of oil viscosity and hence can cause variations of the mobility ratio. Estimations of three-phase relative permeability under such conditions are extremely complex, and using conventional techniques for the estimation can lead to erroneous results. We used the results of four coreflood experiments, carried out on a core, to estimate two-phase and three-phase relative permeability. A new history matching methodology for laboratory experiments was used that takes into account the instability and the compositional effects in the estimation processes. The results demonstrate that a simultaneous CO 2 and water injection (CO 2 -simultaneous water and gas (SWAG)) can be adequately matched using the relative permeabilities of a secondary gas/liquid and a tertiary oil/water. In heavy oil WAG injection, the injected water follows the CO 2 path due its lower resistance as a result of the CO 2 dissolution in the oil and the resultant reduction of the oil viscosity. This is contrary to WAG injection in conventional oils, where gas and water open up separate saturations paths. It is also important to include capillary pressure (Pc), even in high permeable porous media, as we observed that the inclusion of capillary pressure dampened the propagation of the viscous fingers and hence helped the front to become stabilized, leading to a more realistic simulated sweep efficiency.
