Surface-Based Reservoir Modeling: Generating Realistic Sedimentologic Heterogeneity for Reservoir Modeling and Simulation

Abstract

Building spatially realistic representations of sedimentologic heterogeneity in reservoir models is a challenging task that is limited by the inflexibility of predefined pillar or cornerpoint grids. Rock types with diverse petrophysical characteristics are ‘averaged’ within grid cells of arbitrary size and shape, and the continuity and connectivity of baffles, barriers or high-permeability zones is often lost. Because the same areal grid resolution is needed everywhere in the model, large features are over-resolved and small features are under-resolved or omitted. We present a surface-based modelling workflow using grid-free surfaces that allows efficient creation of geologic models without the limitations of predefined grids.

Surface-based reservoir modelling uses a boundary representation approach. All heterogeneity of interest (structural, stratigraphic, sedimentologic, diagenetic) is modelled by its bounding surfaces, independent of any grid. Smooth surfaces are modelled across scales using a parametric NURBS (non-uniform rational basis spline) description. These surfaces are efficient to generate and manipulate, and allow fast creation of multiple realizations of geometrically realistic reservoir models that contain detail at appropriate scales. Surfaces are explicitly constructed by (1) extruding a cross section along a plan-view trajectory, or (2) using geostatistical models. Both approaches seamlessly integrate with commonly available information such as sedimentologic descriptions from cross sections and map views in outcrop or seismic data, or sill, range and anisotropy of variogram models.

We demonstrate this surface-based approach using geometric data from literature for commonly occurring geometries, such as clinoform orientation, dip and length, clinothem facies proportions, cement distribution and channelised sandbody depth, width and sinuosity. Geometric datasets are sampled stochastically to generate surfaces, which are combined into a reservoir model that preserves realistic geometries. This is applied to a coastal-plain and overlying wave-dominated shoreface succession, analogous to an upper Brent Group reservoir, North Sea (e.g. SPE10 benchmark reservoir model). The resulting models can be used directly for flow simulation on adaptive tetrahedral grids that preserve input geometry without the need for upscaling.

AAPG Datapages/Search and Discovery Article #90323 ©2018 AAPG Annual Convention and Exhibition, Salt Lake City, Utah, May 20-23, 2018