Summary
We have been investigating the sequestration of atmospheric pollutants by
injection into coal seams while at the same time enhancing hydrocarbon
productivity by displacement of methane with pollutants. Our effort is one of
several field-based research efforts into CO2 storage in coal seams that are
currently operating in Canada (Mavor et al. 2004), China (Robinson et al.
2004), Japan (Nago and Komaki 2004), and Poland (Van Bergen et al. 2004).
During the course of our field measurements, we have been using single-well
injection, soak, and production tests to collect data required to understand
and predict enhanced coalbed methane (ECBM) recovery potential and
sequestration capacity. We found that changing the composition of the gas
sorbed into the coal changes the porosity and permeability of the coal
natural-fracture system owing to gas-content changes, which cause matrix
swelling or shrinkage due to relative adsorption of different gases.
We collected sufficient information to develop a method for predicting the
permeability and porosity of a coalbed as a function of the secondary porosity
system (SPS) pressure and the gas content and composition of the primary
porosity system (PPS). The method uses data from injection/falloff tests with
water and/or a weaker adsorbing gas (WAG) than CH4 and a stronger adsorbing gas
(SAG) than CH4. Estimates of effective permeability to gas and water obtained
from these tests are used with an iterative computation procedure subject to
constraints to solve for equivalent SPS porosity and absolute permeability at
atmospheric pressure.
Once calibrated, the model can be used to predict a coalbed’s permeability
and porosity as a function of injection pressure and injected-fluid
composition, which in turn are used to predict injection performance. The model
is applicable to production forecasts to account for SPS permeability and
porosity changes as reservoir pressure declines with changes in gas
composition.
This paper describes the new model and discusses well-test procedures to
obtain the data required for model calibration. Also included are coal property
estimates resulting from Alberta Medicine River (Manville) coal core and test
data and an example model calibration.
Background
Commercial production of gas from coal seams is highly dependent upon
natural fractures that control the absolute permeability magnitude and
distribution throughout the reservoir. Natural-fracture absolute permeability
and porosity vary as a function of location, pressure within the
natural-fracture system, and the composition of gas within the coal matrix.
Variations in fracture porosity cause variations in fluid saturations that in
turn cause variations in the relative and effective permeability to gas and
water.
Coal-gas reservoirs are dual-porosity reservoirs consisting of primary and
secondary storage and mass transfer systems. The PPS contains the vast majority
of the gas-in-place volume stored by sorption, while the SPS provides the
conduit for mass transfer to wells (Mavor and Nelson 1997). The SPS generally
consists of two or more natural-fracture sets created at different times. The
more continuous, greater-permeability natural-fracture set is referred to as
face cleats. There is often a more discontinuous, lower-permeability set
referred to as butt cleats oriented at roughly right angles to face cleats.
There also can be additional fracture sets present (Close 1993).
Coal permeability has been known for some time to be dependent upon net
stress and changes in sorbed-gas content. Gray (1987) discussed both of these
phenomena. The general belief was that as reservoirs are depleted, the increase
in net stress decreases the aperture of natural fractures, which decreases the
absolute permeability. However, Gray reported that permeability might increase
as pressure and gas content decreased because an absolute permeability increase
caused by gas desorption could overcome the decrease caused by net stress
increase.
© 2006. Society of Petroleum Engineers
View full textPDF
(
1,990 KB
)
History
- Original manuscript received:
19 May 2004
- Revised manuscript received:
11 December 2005
- Manuscript approved:
3 January 2006
- Version of record:
20 April 2006
View Cart
