Effective permeability in the inter-well reservoir volume is notoriously difficult to estimate due to its scale dependence and sensitivity to heterogeneity. Using stochastic models of realistic 3D sedimentary heterogeneity at the 10cm to 10m scales, we established a series of functional relationships between effective permeability, bedding type and volume fraction (sand, mud, etc.). These functional relationships (bedding type curves) can be understood in terms of effective medium and percolation theory for mixed sand/mud systems, and are a significant improvement of conventional averaging procedures. The functions can be integrated with core plug and wireline data to give improved initial property estimates of the reservoir volume. We applied the method in field studies from several heterolithic reservoir formations of the Norwegian Sea. Improved estimates of vertical permeability are found to be particularly valuable for interpreting unexpected production data and dynamic reservoir history.
Philip S. Ringrose, Erik Skjetne, Carsten Elfenbein, Statoil ASA