--> Integrating Diagenesis into Reservoir Models for Carbonate Platforms: Developing a Workflow with Dolomite as a Case Study, by Megan Murphy-Bishop, Jean Hsieh, Paul M. (Mitch) Harris, Jeroen A. M. Kenter, Sebastien Strebelle, Marjorie Levy, and Jeff Carvalho, #50133 (2008).
[First Hit]

Datapages, Inc.Print this page

Click to view posters in PDF format (16.5 mb).
        Or right click to download.

 

PSIntegrating Diagenesis into Reservoir Models for Carbonate Platforms: Developing a Workflow with Dolomite as a Case Study*

Megan Murphy-Previous HitBishopNext Hit1, Jean Hsieh1, Paul (Mitch) Harris1, Jeroen A. M. Kenter1, Sebastien Strebelle1, Marjorie Levy1, and Jeff Carvalho1

 

Search and Discovery Article #50133 (2008)

Posted October 31, 2008

 

*Adapted from poster presentation at AAPG International Conference and Exhibition, Cape Town, South Africa, October 26-29, 2008.
Click to view list of articles adapted from presentations by P.M. (Mitch) Harris or by his co-workers and him at AAPG meetings from 2000 to 2008.

 

1 Chevron ETC San Ramon, CA ([email protected]; [email protected])

 

Abstract

Reservoir models are, in part, based on combinations of petrophysical measurements from logs and rock description from core studies, including interpretation of depositional environment and stratigraphic context. However, reservoir quality in carbonates is commonly a function of both depositional rock type and diagenetic overprint. Diagenesis can significantly modify porosity and permeability in carbonate reservoirs, and the overprint can cross rock type and stratigraphic boundaries. Thus diagenetic trends are often difficult to interpret between wells and to include in models.

To address this issue, we developed a novel workflow to use multiple point statistics (MPS) to Previous HitmodelTop diagenetic “facies” distributions in a carbonate reservoir. In this case study dolomite was the diagenetic parameter. Four models for the formation of dolomite were chosen based on literature review: seawater, reflux, transgressive, and fracture. Generally, the shape of dolomite bodies in a reservoir is unknown. However, these different dolomite models produce unique dolomite geobody shapes. Four training images that represent potential dolomite geobodies were created: a massive body that cross cuts time surfaces, a body with a flat top and irregular base that cross cuts time surfaces, a thin body with flat top and base parallel to time surfaces, and a thin, vertically-oriented body that is perpendicular to time surfaces.

Six hypothetical paragenetic sequences for dolomitization were then simulated with MPS using various combinations of the training images. Even if the dolomite in the reservoir represents dolomite produced from different models, a comparison of the various MPS models shows the potential range and character of dolomite distribution. Similarly, other diagenetic parameters such as karstification, calcite cementation or bitumen emplacement, can be integrated into reservoir models to more accurately distribute porosity and permeability resulting from diagenetic overprint throughout carbonate platform reservoirs.

 

uAbstract

uFigures

uIssues

uSolution

uBackground

uExample 1

uExample 2

uMPS models

uConclusions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigures

uIssues

uSolution

uBackground

uExample 1

uExample 2

uMPS models

uConclusions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigures

uIssues

uSolution

uBackground

uExample 1

uExample 2

uMPS models

uConclusions

 

Selected Figures

Example 1—Vertical Dolomite Proportion Curves: based on well data and estimated facies proportions throughout area of interest

Example 1—Dolomite Proportion Maps: based on well data and interpolated away from control points. In each case replace "Facies" with "Dolomite"

General guidelines for various dolomite geobodies that result from reflux, seawater, transgressive, and fracture dolomitization processes.

MPS models—Example 1: Two scenarios for the distribution of dolomitization. Different modes of dolomitization during different time periods occur in the two scenarios to highlight the range of outcomes possible.

MPS models—Example 2: Low, mid, and high case scenarios for the distribution of dolomitized packestone and dolomitized grainstone. The proportion of facies is varied in each case, but the size and shape of the training image remains the same. These models use well data as hard controls and facies proportion curves and depocenter maps as soft controls.

Issues

  1. Diagenesis affects carbonate platforms.
  2. Diagenetic geobodies difficult to interpret between wells.
  3. Diagenetic geobodies difficult to include in 3D reservoir models.

A Solution

  1. Determine distribution of diagenetic parameter
  2. Identify potential geobody shapes and create training images
    1. Determine mechanism for diagenesis
    2. QC mechanism using geologic knowledge
  3. Run multiple point statistic simulations
    1. To distribute diagenetic geobodies away from well control
    2. To capture ranges of uncertainty

Background

MPS = multiple point statistics

- Allows modeler to use conceptual geological cartoons as 3D training images to generate geologically realistic reservoir models conditioned to well data.

Training Images

- Cartoons, drawings, conceptual images of feature to be modeled.

- Non-conditional object-based algorithms.

- Requires description of relative dimensions and shapes of each facies and associations among different facies; e.g., sinuous river channel, facies belts, reflux dolomite geobody.

Example 1

Facies proportion curves and maps, prepared for the distribution of dolomite in the Bashkirian through the Famennian in this carbonate platform, are based on multimin data from46 wells and reflect the presence or absence of dolomite. Dolomite in the Bashkirian is greatest at the base of the time interval and is concentrated in the SE. Dolomite in the Serpukhovian is greatest near the top of the interval and is concentrated in the SW. Dolomite in the Late Visean is greatest near the base of the interval and is concentrated in the E and W. Dolomite in the Early Visean and Tournaisian is equally spread throughout the interval and is concentrated in the SE. Dolomite is equally spread throughout the Famennian and is concentrated in the W.

Example 2

Dolomite is dominantly in two of the three wells. These two wells are adjacent to faults. The non-dolomitized well is not associated with a fault. Location of dolomite and petrographic descriptions were used to hypothesize that the dolomite is of hydrothermal origin and is related to the faults.

MPS Models

Two scenarios for the distribution of dolomitization in Example 1: Different modes of dolomitization during different time periods occur in the two scenarios to highlight the range of outcomes possible.

Low, mid, and high case scenarios for the distribution of dolomitized packestone and dolomitized grainstone in Example 2. The proportion of facies is varied in each case, but the size and shape of the training image remains the same. These models use well data as hard controls and facies proportion curves and depocenter maps as soft controls.

Conclusions

1. It is critical to understand the mode of dolomitization to create appropriate training images for a field.

2. MPS is a useful tool to capture the ranges of uncertainties associated with the size, shape, and proportions of dolomite geobodies.

 

Return to top.