ABSTRACT: Pore Type Based Relative Permeability Models: Scaling up from Cores to Reservoirs in Heterogeneous Systems

ETRIS, NED, and ROBERT EHRLICH, University of South Carolina, Columbia, SC

Geologists have long recognized that heterogeneity in rock fabric is common across scales ranging from several millimeters to dozens of meters; engineers have long recognized that relative permeability curves used in reservoir simulation do not resemble curves generated on lab samples. The engineering problem is a direct result of the geologic phenomenon. Where wettability changes are small, changes in rock fabric determine the relative permeability curve. Therefore, no vertical sequence can be represented by a single curve derived from a small portion of that sequence. Scaling up procedures based on such limited sampling fail for this reason. It is the geologist's job to translate the geologic information into engineering data for use in reservoir simulation. We have derived numerica models for relative permeability, based on a combination of thin section and capillary pressure data, that adequately recreate values measured on lab scale, sandstone reservoir samples, for both steady state and unsteady state processes. Critical to this method is the objective determination of pore types from thin section. Because pore types vary as a function of grain size and cementation, pore type data can be easily related to macroscopic core description, making the acquisition of pore type data possible on any scale. Thus, the relative permeability curve at any scale can be calculated by determining the relative abundances of pore types at the scale of interest. This ability will allow us to understand better the scaling up problem itself, and the nature of the variability between core and reservoir physical properties.

AAPG Search and Discovery Article #91012©1992 AAPG Annual Meeting, Calgary, Alberta, Canada, June 22-25, 1992 (2009)