Next Generation Geological Modeling for Hydrocarbon Reservoir Characterization - Advances in Data, Methods, Models and Their Applications in Geoscience - page 221

presence of high density, closely spaced wells, highly computationally intensive and reliant
on the (3D) training image models (Caers, 2005). Multi-Point GeoStatistics (MPS) (Strebelle
and Journel, 2001) is rapidly growing in popularity offering the modeler the ability to create
geological models with complex geometries, while conditioning to large amounts of well
and seismic data. However, as pointed out by Daly and Mariethoz (2011), it is still a
relatively new topic, which has had a long academic history and is now just finding its way
into commercial software. They also pointed several deficiencies in current implementations
related to 1) performance, 2) training image generation, and 3) non-stationarity.
The DecisionSpace  Desktop Earth Modeling approaches facies modeling very differently
from most current software offerings. The facies simulation workflow utilizes a powerful
combination of describing the geological trends with LPMs created from multiple VPCs, as
described previously, integrated with Plurigaussian simulation (PGS), a robust and well-
tested algorithm with a long industrial history. The PGS is an expansion of TGS method and
uses two Gaussian variograms simultaneously. There are a numerous advantages of PGS
over other methods:
 as trends for each facies within each layer and every reservoir interval in the model are
calculated, based on the LPMs that account for spatial non-stationarity, the PGS
methodology, captures most inter- and intra-facies relationships including post
depositional overprinting, such as diagenesis.
 as a pixel-based method, PGS works easily with closely spaced or sparse well control.
 PGS has been available in two commercial software offerings (HERESIM™ and
ISATIS™) since the early 1990s, however its use is not necessarily intuitive.
 While it is possible to overcome some of the challenges presented by traditional
algorithms through the intervention by experts, the implementation of a LPM with PGS
workflow can be presented simply and intuitively.
The next-generation earth model approach to facies modeling introduces a set of facies
templates based on the understanding of realistic depositional environments (Walker and
James, 1992). The DecisionSpace  Desktop Earth Modeling implements a library of more
than forty standard depositional systems, presented with maps and cross-sectional views
(see Fig. 5).
The PGS facies modeling requires a set of rules to establish lithotype relationships, where a
lithotype is a group of facies sharing common depositional and petrophysical properties.
Knowledge of the proportions is not sufficient for accurate modeling of the lithotypes and
the depositional system templates help modelers to visualize relationships between the
facies and to provide an associated lithotype “rule box” (see: upper left schematics, Fig. 5 -
expanded view) that specifies their mathematical relationship. Rules are based on simple or
complex lithotype transitions; a) simple transitions corresponds to a strict transition from
one lithotype to another and are modeled with one variogram, b) complex transitions
require the definition of two lithotype sets, each controlled by its own variogram that can
have different anisotropy directions. A lithotype set is a set of lithotypes that share a
common spatial model, such as a channel and its associated levy. An example of modeling
complex transition of lithotypes, represented by two lithosets and associated variogram
models is given in Fig. 6. In this figure the vertical and horizontal sets are schematic
representations depicting which variogram controls a lithotype set, when in reality the two
variogram models interact to create the final facies model. Because PGS uses lithology
proportions, rather than indicators as in SIS, a wider variety variogram model types are