Influence of Permeability Heterogeneity on Miscible CO2 Flooding Efficiency in Sandstone Reservoirs: An Experimental Investigation

Abstract

In this study, we systematically investigate the effect of core-scale heterogeneity on the performance of miscible CO2 flooding under various injection modes (secondary and tertiary). Manufactured heterogeneous core plugs are used to simulate vertical and horizontal heterogeneity that may be present in a reservoir. A sample with vertical heterogeneity (i.e. a layered sample) is constructed using two axially cut half plugs each with a distinctly different permeability value. In these samples, the permeability ratio (PR) defines the ratio between the permeabilities of adjacent half plugs. Horizontal heterogeneity (i.e. a composite sample) is introduced by stacking two or three short cylindrical core segments each with a different permeability value. Our special sample construction techniques have also enabled us to investigate the effect of permeability ratio and crossflow in layered samples and axial arrangement of core segments in composite samples on the ultimate recovery of the floods. Core flooding experiments are conducted with an n-Decane–brine–CO2 system at a pore pressure of 17.2 MPa and a temperature of 343 K. At this temperature, the minimum miscibility pressure of CO2 with n-Decane is 12.6–12.7 MPa so it is expected that at 17.2 MPa CO2 is fully miscible with n-Decane. The results obtained for both the composite and layered samples indicate that CO2 injection would achieve the highest recovery factor (RF) when performed under the secondary mode (e.g. layered: 79.00%, composite: 89.83%) compared with the tertiary mode (e.g. layered: 73.2%, composite: 86.2%). This may be attributed to the effect of water shielding which impedes the access of the injected CO2 to the residual oil under the tertiary injection mode. It is also found that the oil recovery from a layered sample decreases noticeably with an increase in the PR as higher PR makes the displacement more uneven due to CO2 channelling. The RFs of 93.4, 87.89, 77.9 and 69.8% correspond to PRs of 1, 2.5, 5, and 12.5, respectively. In addition, for the layered samples, crossflow was found to have an important role during the recovery process; however, due to excessive channelling, this effect tends to diminish as PR increases. Compared with the layered heterogeneity, the effect of composite heterogeneity on the RF seems to be very subtle as the RF is found to be almost independent from the permeability sequence along the length of a composite sample. This outcome may have been caused by the small diameter of the plugs resulting in invariable 1-D floods.