--> Geocellular Modeling of Fluvial Meander-Belt Reservoirs: A Rule-Based Method Conditioned on Seismically Imaged Geometries

AAPG ACE 2018

Datapages, Inc.Print this page

Geocellular Modeling of Fluvial Meander-Belt Reservoirs: A Rule-Based Method Conditioned on Seismically Imaged Geometries

Abstract

A novel method is presented for building static models of fluvial reservoirs that are composed of large point-bar architectural elements, based on the application of a specialized forward stratigraphic model – Point-Bar Sedimentary Architecture Numerical Deduction (PB-SAND).

The PB-SAND model uses interpreted horizontal slices from 3D seismic datasets to reconstruct the planform evolution of meander loops, on which basis the geometry of point-bar deposits and associated accretion units can be simulated deterministically. The resulting meander-belt geometry is then populated with different types of facies through a rule-based algorithm capable of generating facies distributions that reflect geologic understanding, conditioned by analogs from a database of case-study examples. This enables establishment of linkages between mode of meander evolution (e.g., growth via expansion, translation, rotation, or combinations thereof) and facies distributions. Input parameters define the proportions, geometries and distributions of types of point-bar deposits, as captured from subsurface data or a database of geologic analogs. Multiple stochastic realizations of facies architecture can be generated.

To demonstrate this modeling approach, the workflow has been applied to a meander-belt reservoir where large point-bar and channel-fill elements are imaged in seismic. A detailed example illustrates workflows that can be used to build high-resolution sector models in pre-drill contexts, suitable for guiding development plans. A second example shows how to achieve well match for densely drilled sectors, by means of a hybrid approach that combines the new algorithm with traditional geostatistical techniques.

Results demonstrate how point-bar evolution styles, as inferred from seismic data, can be used as input to model the predicted distribution and geometry of heterogeneities, and how this facilitates the reproduction of geologic features that control the static connectivity of point-bar reservoirs (e.g., distribution and characteristics of bar-front mud drapes, and mud-prone packages). A comparison with traditional variogram-based methods is undertaken to compare metrics that describe intra-point-bar static connectivity and that represent proxies for the degree of compartmentalization of upper-bar sand by mud drapes. Results indicate that constructing point-bar reservoir models using PB-SAND yields more variable scenarios of static connectivity of net-reservoir volumes.