Relative permeability estimation in a CO2/heavy oil system
MetadataShow full item record
With the decline in the world reserves of conventional hydrocarbon resources, attention is shifting to the exploitation of unconventional resources like heavy oil. Though the thermal recovery method is quite common with heavy oil, studies and field trials have shown that in certain situations, non-thermal methods like CO2-flooding could be more appropriate in exploiting heavy oil reservoirs. Relative permeability for oil, water, and gas are necessary requirements in carrying out numerical simulation for the purpose of predicting and optimizing recovery from these non-thermal methods. The scarcity of information on heavy oil relative permeability often leads to the erroneous application of the conventional oil relative-permeability assumptions in the simulation of heavy oil processes. Studies have shown that viscous fingering, oil swelling, and drastic reduction of viscosity due to dissolution of gas in oil are common phenomena in the recovery processes of heavy oil, thus it is expected that the properties of the relative permeability of heavy oil should be different from the relative permeability from conventional oil which clearly has a different recovery mechanism. In this work, the results of series of coreflood unsteady-state two-phase displacement experiments were used to investigate the characteristics of two phase relative permeability in heavy oil production processes. The experiments were carried out on two different heavy oils with different viscosity on cores of similar characteristics. Firstly, two phase relative permeability of the heavy oil processes were estimated by the history matching and the analytical approach for oil-water, and oil-gas systems. Also, in order to investigate the suitability of existing three phase models in simulating three phase flow in heavy oil systems, the two phase relative permeability curves from the history matching technique were used to generate three phase relative permeability curves. Also investigated was the appropriateness of the analytical method to generate three phase relative permeability curves. In this work, the relative permeability curves from several heavy oil core flood experiments using 1D and 2D models is estimated. These experiments involve different heavy oil viscosities, and different core orientations. This study shows that when gas or water is injected to improve the recovery of heavy oil, the values of the residual oil saturation vary with oil viscosity. As a result of these effects, the relative permeability curves estimated from these core flood experiments are also found to vary with the oil viscosity. Besides the effect of oil viscosity on the relative permeability curves of heavy oil systems, the relative permeability curves obtained from cores in the vertical direction during flooding are also observed to be different from those obtained from cores placed in the horizontal direction. The performance of the correlation-generated three-phase relative permeability models were assessed by comparing the results of the simulation against the experimental data. From the results, it is seen that there could be large prediction errors if an inappropriate three-phase relative permeability model is used for the simulation. The analytical method was also found deficient in estimating three-phase relative permeability curves for heavy oil systems. An automatic history matching of the three phase experimental results of heavy oil was then used to give very accurate relative permeability results, though this could be laborious and time consuming. The main conclusions derived from this work are that (1) while the relative permeability from conventional oil are only dependent on saturation, the relative permeability from heavy oil (in addition to saturation), is dependent on oil viscosity and direction of flow; (2) once there is a proven case of instability in the heavy oil system, a 2D grid system should be used to estimate relative permeability curves in order that the viscous fingering in the system can be modelled; (3) The analytical corrected Nitrogen-heavy oil relative permeability can be used as a substitute relative permeability to model CO2/heavy oil system; (4) The three phase relative permeability models in existing reservoir simulation packages are not adequate in modelling three phase flow in heavy oil; the three phase relative permeability for modelling heavy oil processes should be instead estimated from the automatic history matching of the three phase experimental results of the heavy oil.