Page 183 - Machine Learning for Subsurface Characterization
P. 183

158   Machine learning for subsurface characterization


                    Volume fraction of miscible oil
            V o
                    Volume fraction of movable water
            V w
                    Volume fraction of bound fluid
            V b
                    Volume fraction of kerogen
            V k
            Subscripts
            i   Formation (i)
            j   Formation parameter (j)
            o   Oil, optimum
            w   Worst

            1 Introduction

            1.1 Geology of the shale formation
            The shale formation under investigation is a 200-ft-thick formation divided into
            three distinct sections: upper, middle, and lower shale formations. The upper
            and lower sections are black shales, and the middle section is sandy
            siltstone. The upper and lower shale formations are hydrocarbon source
            rocks with total organic carbon (TOC) ranging from 12 to 36 wt%. The clay
            mineral content in these two sections is dominated by illite and quartz. The
            middle shale formation is the hydrocarbon-bearing reservoir and has a low
            TOC content ranging from 0.1 to 0.3 wt%. Below these three shale
            formations, there exists dolostone interbedded with clay-rich conglomeratic
            dolomudstone. The shale formations are characterized by ultralow matrix
            permeability that constrains oil mobilization [1]. Current recovery factor in
            such shale formations is around 3%–6% of the oil in place [2]. High oil-in-
            place estimates with low primary recovery mandates enhanced oil recovery
            (EOR) projects based on light, miscible hydrocarbon injection.



            1.2 Literature survey
            Due to the ultralow permeability and nanoscale pore sizes, the mechanisms of
            EOR in tight reservoir are different from those in conventional reservoir. In
            tight oil formation, EOR efficiency is controlled by the combined effects of
            miscibility, diffusion, sorption, dissolution, and capillary condensation, to
            name a few, out of which diffusion is the dominant mechanism [3]. Large
            permeability contrast between the organic-rich porous matrix and fracture
            can lead to fracture-dominated flow in fractured shales. In tight shale
            formations, the net displacement due to the injection of light hydrocarbon
            generally includes four steps: (1) injected light hydrocarbon flows through
            fractures, (2) injected hydrocarbon goes into the porous organic-rich shale
            matrix by diffusion and imbibition; following that, the injected hydrocarbon
            may interact with connate oil due to miscibility, (3) connate oil migrates into
            fractures via swelling and reduced viscosity, and (4) injected light
            hydrocarbon achieves equalization inside the organic-rich matrix [2]. Our
   178   179   180   181   182   183   184   185   186   187   188