Page 189 - Machine Learning for Subsurface Characterization
P. 189

164   Machine learning for subsurface characterization


               The proposed R-index calculation (Fig. 6.2) requires the following ranking
            characteristic parameters to be obtained from well logs: (1) Kerogen volume
            obtained from Schlumberger’s Quanti-ELAN inversion, (2) apparent pore
            throat radius obtained from a Winland-type analysis of formation
            permeability and porosity derived from logs, (3) water saturation from the
            interpretation of induction resistivity measurements, (4) porosity from
            neutron-density logs, and (5) permeability from NMR T2 distribution log.
            The porosity and water porosity logs in Fig. 6.2 are derived using the
            Interactive Petrophysics NMR interpretation module. The oil porosity is
            derived from Techlog Quanti-ELAN mineralogy inversion module. As a
            result the water porosity log and oil porosity log do not add up to the total
            porosity log. The r35 log is an approximation of pore aperture radius
            corresponding to the 35th percentile mercury saturation.
               The calculated R-index is shown in track 8 of Fig. 6.2. As per the R-index,
            light-hydrocarbon injection will have better EOR efficiency in the middle
            shale as compared with the upper and lower shales. Most parts of upper and
            lower shales have a low EOR efficiency. There are three disadvantages of
            the R-index: First, TOC has a relatively higher importance when calculating
            R-index that results in relatively lower R-index values in the upper and
            lower shales, which have high TOC. Second, the R-index is a relative
            indicator of EOR efficiency. Finally, the entire middle shale exhibits high
            R-index indicating the low vertical resolution of this index.


            4 Methodology to generate the microscopic displacement
            (MD) index
            4.1 Description of the MD-index

            MD-index is a measure of pore-scale displacement efficiency of the injected light
            hydrocarbon in a hydrocarbon-bearing zone. Theoretically, microscopic
            displacement efficiency can be calculated by dividing the displaced oil volume
            by the initial oil volume [28]. This necessitates calculations of the volume
            fractions of free oil, free water, bound oil, and bound water in pores to
            generate the depth-specific MD-index. To have a good microscopic
            displacement efficiency during the miscible light-hydrocarbon injection,
            injection pressure and reservoir pressure should be above the MMP of the
            injected hydrocarbon and the pore-filling connate hydrocarbon. The MMP is
            determined based on the hydrocarbon composition and reservoir temperature.
            The fluid properties change in nanopores due to pore-confinement effect,
            which is a consequence of the large capillary pressures, electrostatic forces,
            and van der Waals forces giving rise to the changes in structural properties of
            the fluid [29]. MMP will be notably altered due to pore-confinement effect in
            nanoscale pores. Hydrocarbon residing in differently sized pores will have
            different MMP. Consequently, under similar reservoir temperature and
            pressure, oil in unconfined pores and nanopores may have different
            miscibility. Due to the pore-confinement effects, hydrocarbon in nanopores
   184   185   186   187   188   189   190   191   192   193   194