No abstract

Log derived permeability averages in homogeneous clastic reservoirs most often matches the reservoir scale permeability determined from well testing. However, when it comes to heterolithic, anisotropic reservoirs such as observed in successions interpreted as turbidites, there can be significant differences between reservoir scale estimates and traditional techniques such as arithmetic and geometric averaging. This mismatch is often overlooked, although it can be critical when it comes to history matching production data. The fundamental flaw of the log-only approach is that it assumes permeability to be isotropic, and therefore ignores the three-dimensional nature of permeability. In heterogeneous clastic reservoirs, the anisotropies originating from various small scale sedimentary structures may have an influence on fluid flow at larger scales, impacting the observed reservoir-scale permeability.This paper demonstrates how data acquired from wireline logs and cores from very heterogeneous successions like turbidites may be used to consistently predict the reservoir scale permeability. The workflow consists in processing and integrating the data into a lamina-scale near-wellbore model containing structural as well as petrophysical properties; a flowbased upscaling method is then applied to provide a geologically consistent input into a single-well model, from which a typical well test scenario will be simulated; the resulting pressure transient is then analyzed to determine a simulated permeability-thickness product.Presented in this paper is an application to two wells intersecting a turbidite reservoir. This shows that the permeability obtained from this workflow is closer to the well test permeability when compared to log derived estimates from traditional averaging techniques.This method can be applied to obtain valid reservoir-scale permeability values that can then be compared to the actual well test result. However, it can also be applied in cases where the value of information processes prove well testing to be uneconomical; then, a global reservoir permeability value can still be obtained using the workflow described.