Use of stripping ratios to identify dispersion levels and displacement mechanisms in miscible gas floods

Oil recovery estimated from numerical simulation of miscible gas floods is highly dependent on the grid-block sizes used and the level of reservoir dispersion. Stripping ratios, defined as the ratio of certain heavy to light components, have been used recently to estimate the level of in situ dispersion at the field scale. The use of stripping ratios, however, has not been thoroughly investigated. The focus of this paper is to validate the use of stripping ratios and to determine its appropriate definition for certain gas injection processes. The approach is applied to the condensing/vaporizing displacement of twelve-component oil by a five-component miscible injectant (MI). Analytical gas injection theory based on the method of characteristics is used to help identify the correct choice of components to calculate the stripping ratio. Numerical simulations are conducted for both 1-D and 2-D gas floods. The results from 1-D displacements show that the condensing and vaporizing region can be identified from stripping ratios when the stripping ratio is calculated from the heaviest component in the oil and the lightest component in the oil not present in the injected gas. The stripping ratio for 1-D field scale displacements is shown to be useful in identifying (as measured from core data) if a large level of dispersion and mixing is present for such reservoirs. For more realistic 2-D (and 3-D) displacements, however, crossflow or re-saturation occurs, which can complicate the interpretation of the stripping ratio and its use as a practical reservoir engineering tool. We show that the level of dispersion at field scale might be qualitatively confirmed by comparison of stripping ratio and oil saturation data from cores to those determined from simulation. When dispersion is low (near a longitudinal dispersivity of about 1.0 ft) resaturated regions give large oil saturations accompanied by very large stripping ratios. These features are not present at dispersivity levels greater than 5.0 ft.