Applying New Rock Typing Methods, and Modelling for Conventional & Unconventional Reservoirs

Abstract

The main objective of this research is to present two new approaches for unconventional gas reservoirs rock typing methods. In our rock typing we automatically adjust rock type classes by optimization routines using specific surface area per unit grain volume and volume of kerogen values. Further, our results are compared to other conventional rock typing methods. The presented methos can enhance unconventional reservoir characterization by developing and/or establishing new correlations. This is exemplified through a real case study of an unconventional shale gas reservoir called Upper Safa formation that is located in the western desert of Egypt. Addition we describe the fluid properties more consistently through a full integration of unconventional rock parameters such as surface roughness factor, gas adsorption, type of kerogen, volume of kerogen, level of maturity and total organic carbon content. The Upper Safa formation is a shale gas unconventional resource play. Interpretation analysis has confirmed the hydrocarbon potential in the Upper Safa formation. Geochemical pyrolysis analysis has been used to confirm the presents of Kerogen type III. Total organic carbon content results are obtained within the ranges of very good petroleum potential according to Rock Eval pyrolysis from 2% to 4% TOC. The Upper Safa formation is a highly heterogeneous formation, with the Dykstra Parson permeability variation giving a heterogeneity value close to 100%. Large variation in the permeability, rangeing from milli-Darcy to nano-Darcy, is common for unconventional shale reservoirs. This large variation in permeability complicates rock typing of the reservoar. Several conventional rock typing methods have been applied to for the zonation problem, including as Amaefule, Discrete rock typing, Flow Zone Indicator, permeability predictive model, Winland, and a modified Winland. For comparison of the different methods, they have all been forced to produce the same number of rock classes. We have established routines for optimal selection of the boundary values distinguishing these rock classes. We also present two new rock typing techniques utilizing specific surface area per unit grain volume and TOC values.