Abstract
Spatial distribution of permeability is an important factor in the prediction of performance of Steam Assisted Gravity Drainage (SAGD) well pairs. Presence of short-scale variability in sand/shale sequences, preferential sampling of core data, and uncertainty in upscaling parameters are complications that make the inference of a reliable porosity – permeability relationship impossible. A simple yet effective way of overcoming these complications is micro-modeling. The central idea in micro-modeling is to use an additional source of information, namely digitized core images, to quantify the uncertainty in power-law averaging parameters and construct the porosity-permeability bivariate relationship by Monte Carlo Simulations (MCS). The work-flow in micro-modeling is comprised of a few steps from digitizing the selected core images to building 3D geo-blocks of binary sand/shale mixture, populating them with porosity/permeability values, upscaling the populated binary mixture by flow simulations, determining the uncertainty in power-law parameters and implementing MCS. The porosity-permeability relationships are constructed on a by-facies basis. Results of this research suggest that effective properties of clean sand are changing with the volume fraction of shale; and it has ultimately resulted in the development of an extended version of power-law formalism.
