Basic HTML Version
OH – ET – VA - LL: Analysis of Dynamic Data in Shale Gas Reservoirs – Part 1 – Version 2 (December 2010)
p 6/24
In a typical analysis it is usually an overkill to define each fracture individually (although such
a level of detail may be interesting in test designs) – not to mention that it is rather
cumbersome. For this reason, we have also released another external model with a simplified
interface, where all fractures carry identical properties.
Those solutions are available as external models in both Saphir (PTA) and Topaze (PA); they
will be integrated as internal models in Ecrin v4.20 (Q4-2010). Those integrated models will
benefit from new additions: fractures will be able to intersect the horizontal drain at any angle
and it will be possible to combine the model with variable skin and external boundaries, among
other things.
These models are semi-analytical, hence reasonably fast. For the very same reason they suffer
from not being able to solve the real gas flow equation – we use the usual pseudo-pressures
approach that can take
some
nonlinearities (
.z product) into account, but not all (
.c
g
) – not
to mention “exotic” effects such as desorption that cannot be handled analytically in a proper
manner.
6 – KAPPA’s numerical models for fractured horizontal wells
When we first heard about the issues on shale gas we immediately filed it in the much-to-
complicated tray, and waited. Then we realized we could not get away with it and we asked Dr
Raghavan to make a survey and report what we would have to implement in order to properly
address what the industry thinks is happening in shale gas reservoirs. One issue was the
mechanism of gas desorption from the shale, complemented with some unconsolidation. The
other issue was the geometry of the wells which will produce such formations. The very low
permeability of these formations required extensive series of fractures from an initial
horizontal drain.
In complement of the external analytical models in Saphir / Topaze, we needed a numerical
model that would combine the flow geometries with the nonlinearities related to the extremely
poor permeability and strong pressure gradients, and additional nonlinearities related to
desorption and unconsolidation. This was at least required to check if simpler solutions based
on analytical models or straight line analysis worked.
We had already implemented unconsolidation (relation between permeability/porosity and
pressure) for some time in Saphir NL and Topaze NL. Desorption was a relatively mild thing to
model. The real challenge was with the gridding of a fractured horizontal well, and in Ecrin
v4.12 (2009) we ended up with two alternate solutions: a 2D model and a 3D model.
6.1 - The 2D-Model
If the fractures are fully penetrating in the formation, and if we neglect the direct production of
the formation gas into the horizontal drain, the problem becomes purely 2-D. The first series of
pictures show this model in the 2D-Map option of Saphir / Topaze. Interactively the user can
move the well, control the length and direction of the horizontal drain, control the length of the
fractures (they are all the same length and are regularly spaced along the horizontal drain). In
the dedicated well dialog the user will also control the number of fractures. At the end of any
interaction or when a change in the dialog is validated, the grid is automatically calculated and
displayed.