Geomechanics Coupling
One- and two-way (explicit as well as iterative) coupling with CMG’s
Geomechanics module has been implemented into GEM and is being
released as a Beta feature.
The geomechanics module handles different constitutive models like,
linear and non-linear elastic, generalized plasticity, elasto-plastic, single
surface, and creep models. For the elasto-plastic model the yield criteria
can be specified via Mohr-Columb or Drucker-Prager that are suitable
for the description of geologic material. There is also an isotropic strain
hardening and softening option. The behavior of cyclic loading and
unloading that occurs during cyclical injection and production can be
modeled. A cap model is also available for yield criteria of elasto-plastic
material. For Non-linear elasticity, hypoelastic and hyperelastic
constitutive models are available. Associated or non-associated flow
rules can be used for Mohr-Columb, Drucker-Prager, single surface, or
generalized plasticity models.
The above constitutive models can be used in 2D or 3D problems in the
Cartesian, corner-point, or axissymmetric grid. Finite element method is
used in the geomechanics module to compute the geomechanical
responses. The user can specify a variety of boundary conditions in 2D
and 3D on nodal points of the finite element.
The geomechanics module solves for the force equilibrium of the
formation and calculates volumetric dilation/compression as a result of
both elastic and plastic straining.
The coupling options allow porosity to be optionally dependent on
deformation. It can be a function of pressure temperature or volumetric
strain; or a function of pressure temperature and total mean stress.
The change in matrix and fracture permeability due to geomechanical
effects can be computed by a number of algorithms including the ones by
Li and Chalaturnyk; and Barton-Bandis.
A large number of possible geomechanics responses are available for
viewing in CMG’s graphical post processor “Results”, and also available

- 6 -

to the text “.out” and “.geo” files. The user can view grids in “Results”
that deform with time, but there are some restrictions for displaying 2D
radial grids.
A new data section GEOMECHANICS has been added which precedes
WELLS AND RECURRENT DATA section in the GEM input data file
to facilitate entry of geomechanics related keywords. Please see the
chapter on geomechanical features for specific keywords.
The coupled geomechanic modeling may be useful for a variety of
situations, like determination of leakage through points of weakness in
the cap rock during GHG sequestration, effect of rock deformation and
porosity change during CBM/ECBM simulation, solid deposition in
asphaltene modeling, and mineral dissolution and precipitation during
geochemical processes, and others.

Improved Parallel Performance
A new method has been implemented for parallel equation and residual
building using a method of domain decomposition based on Parasol
classes. This method is expected to give better scalability on computers
with larger number of processors.
The method is activated by using *JACPAR *ON in the Numerical
Methods section, or by specifying –jacpar on the command line. In order
to use *JACPAR, the parallel computing licensing feature must be

Oil-Wet Rock-Fluid Model
The oil-wet rock fluid model modifies both the initialization and flow
calculations in GEM. See the OILWET subkeyword of the RPT keyword
for details.
The oil wet option initialization assumes that the reservoir was initially
water wet and has undergone a wettability alteration to become an oil
wet reservoir. This assumption has a significant impact on how fluids are
placed during reservoir initialization.
In the oil wet model, three phase flow is calculated assuming water is the
intermediate phase. Therefore Kro is a function of oil saturation, Krg is a
function of gas saturation, while Krw is a function of both oil and gas
saturation (through Krwo and Krwg). The normal three phase flow
models (e.g. Stone 1, Stone 2, segregated, etc) are used to determine
water relative permeability, not oil permeability.

No comments: