Three-dimensional prediction of diagenesis in reservoirs.

in: Proc. 27th Gocad Meeting, Nancy

Abstract

In reservoir studies, a gap often exists between the geostatistical modeling of heterogeneities at deposition time and flow simulation in the present-day reservoir grid: diagenesis transforms rock properties, and could completely modify reservoir quality, especially in carbonate reservoirs. Besides the compaction effects that are often implicitly accounted for in facies models, two phenomena modify rock properties: (1) the dolomitization and dissolution of carbonate formations, and, (2) the cementation which reduces porosity and permeability of both clastic and carbonate reservoirs. These rock modifications mainly originate in chemical reactions involving water circulations through the sediments. These circulations of liquid predominantly concern the permeation of meteoric water in emerged carbonate littorals and the migration of fluids in fracture network or in fault damage zones. Therefore, the chemical diagenesis mainly depends (1) on the capacity of original rocks to transmit fluids, and thus on their petrophysical properties and (2) on the distance to fluid sources, being the emersion areas, the faults and fractures. We propose an innovative approach to better constrain the quantitative prediction of diagenesis. Probabilities for each diagenetic rock type are estimated by integrating well observations, sedimentological and lithological models, and structural and mechanical information. Diagenetic rock types are then predicted by stochastic simulations using the generated probability fields. This solution is applied to a faulted carbonate reservoir, and shows its ability to overlay the effects of diagenesis onto facies models. The suggested approach improves the predictive power of reservoir models by integrating important geological knowledge in reservoir descriptions.

Download / Links

    BibTeX Reference

    @inproceedings{P311_Kedzierski,
     abstract = { In reservoir studies, a gap often exists between the geostatistical modeling of heterogeneities at deposition
    time and flow simulation in the present-day reservoir grid: diagenesis transforms rock properties, and
    could completely modify reservoir quality, especially in carbonate reservoirs. Besides the compaction effects
    that are often implicitly accounted for in facies models, two phenomena modify rock properties: (1)
    the dolomitization and dissolution of carbonate formations, and, (2) the cementation which reduces porosity
    and permeability of both clastic and carbonate reservoirs. These rock modifications mainly originate
    in chemical reactions involving water circulations through the sediments. These circulations of liquid predominantly
    concern the permeation of meteoric water in emerged carbonate littorals and the migration of
    fluids in fracture network or in fault damage zones. Therefore, the chemical diagenesis mainly depends (1)
    on the capacity of original rocks to transmit fluids, and thus on their petrophysical properties and (2) on
    the distance to fluid sources, being the emersion areas, the faults and fractures. We propose an innovative
    approach to better constrain the quantitative prediction of diagenesis. Probabilities for each diagenetic rock
    type are estimated by integrating well observations, sedimentological and lithological models, and structural
    and mechanical information. Diagenetic rock types are then predicted by stochastic simulations using the
    generated probability fields. This solution is applied to a faulted carbonate reservoir, and shows its ability to
    overlay the effects of diagenesis onto facies models. The suggested approach improves the predictive power
    of reservoir models by integrating important geological knowledge in reservoir descriptions. },
     author = { Kedzierski, Pierre AND Durand-Riard, Pauline AND Caumon, Guillaume },
     booktitle = { Proc. 27th Gocad Meeting, Nancy },
     title = { Three-dimensional prediction of diagenesis in reservoirs. },
     year = { 2007 }
    }