The term 'cut-off' refers to a joint effort by geoscientists and engineers to define a value which discriminates non-reservoir rock (shale) from reservoir rock (sand), and has been used in the petroleum industry for several decades. In many instances a single cut-off is not good enough to define sand requiring the use of additional variable cut-offs. The determination of cut-off(s) is highly dependent on a geoscientist's or engineer's experience. There exists no well-defined method that is built on a sound scientific basis.
A systematic method is presented in this paper. The new method separates reservoir rock from non-reservoir rock based on the statistical properties and probability distribution functions derived for the various properties of a reservoir interval. It is common that the histogram of a variable (i.e. porosity) has a mixed interval between reservoir and non-reservoir rocks.
No matter how a cut-off is defined on the histogram, there will always be a number of values that are incorrectly classified. As a result, we have seen errors in hydrocarbon volume calculation which have been in the range of 5 to 30 %. Unlike the typical application of cut-offs, the proposed method identifies a boundary (no or little mixing) between the two rock populations and then separates them based on this boundary. Correct boundary identification, by analyzing all available variables collected from well logs, is the key to success. When there is no clear boundary, integration of engineering data (i.e. relative permeability and capillary pressure) with the well logs helps to identify possible boundaries or cut-offs.
A field study demonstrates the principle of the new method and the improved results includes more than 80 wells. The method is easy to understand and to apply in practice.
Frequently, geoscientists and engineers in the petroleum industry refer to non-reservoir rock and reservoir rock as shale and sand, respectively, despite the lack of pure shale and pure sand compositions for sandstone reservoirs. We will adopt this terminology in this paper, further noting that this method applies for carbonates and siliciclastics alike.
Separating shale from sand is critical in the process of building a geological and reservoir simulation model. Sand and shale have very different characteristics of lithology, porosity, permeability, and water saturation. In contrast with shale, sand typically possesses much larger values of porosity and permeability, and contains much smaller values of water saturation. Sand stores oil and gas, and provides flow channels for reservoir fluids, while shale acts as flow barriers. These flow barriers greatly impact sweep efficiencies of waterflood and EOR processes.
Though defining petrophysical cut-off values is routine work for geoscientists and reservoir engineers, methods based on well log petrophysical data for separating shale from sand are ill-defined. Geoscientists and reservoir engineers still rely heavily on empirical cut-off values (e.g. porosity, permeability, and water saturation etc.) to predict sand/shale distributions and net pay. Different companies and different interpreters have their own criteria for determining cut-off values.
Through application of advanced reservoir characterization techniques, such as geostatistics, capabilities exist to model sand and shale in three dimensions and in great detail. The industry requires a well-defined and consistent method to correctly distinguish shale from sand and to properly model sand/shale distributions. The following sections describe and discuss a statistically-based technique via both theory and an actual field application.