Summary
The steam-assisted gravity-drainage (SAGD) process is used widely to recover
heavy oil and bitumen from formations in which no other recovery method has
proved to be economical. It is an energy-intensive process, and because of
economic and environmental reasons, solvents as additives to the injected steam
are being explored currently to reduce the energy and emissions intensity of
SAGD. The solvent-aided process (SAP), tested in the field and described in the
literature, is one such attempt.
In the SAP, a small amount of hydrocarbon solvent is introduced as an
additive to the injected steam. Thus, the viscosity of the oil is also reduced
because of solvent dilution in addition to heating. The SAP can improve the
energy efficiency of SAGD significantly, thus reducing the heat requirement, as
shown in field trials discussed elsewhere. However, on the use of the right
amount of solvent that can result in best overall performance, there is very
little discussion in the literature. Because of the high cost of such solvents,
there is incentive to optimize their use in SAGD. Recently, various authors
have attempted to address the subject with, for example, arbitrary
time-dependent schemes of solvent injections, assessing their impact on results
or by treating the internal reservoir dynamics as a black box and using
optimization methods, such as genetic algorithms (GAs), to estimate the optimal
amount of solvent. While these approaches orient us to the problem in a
context-specific manner, it is believed a generalized treatment to estimate
optimal use of solvent requires a mechanism-based understanding.
The approach presented in this paper is aimed at estimating the optimal
solvent in the context of SAGD. It combines the existing Butler's oil-drainage
analytical models (Butler 1985, 1988, 1994) for SAGD and vapor extraction
(VAPEX), which deal with heating effect and solvent-dilution effect one at a
time, into one. Then, it calculates the time-dependent steam rates to maintain
the predicted oil rates in conjunction with solvent rates and, thus, estimates
the solvent/steam ratio (SSR) and the steam/oil ratio (SOR). The results are
discussed for a few light-alkane solvents. In the process of this exercise, it
is discovered that to obtain reasonable SSR and SOR, a significant amount of
oil has to drain from a diffuse layer, which has a varying temperature, solvent
concentration, and gas saturation (from maximum gas saturation at the injection
end to zero at the vapor/liquid interface).
© 2012. Society of Petroleum Engineers
View full textPDF
(
525 KB
)
History
- Original manuscript received:
9 November 2011
- Meeting paper published:
31 October 2011
- Revised manuscript received:
20 March 2012
- Manuscript approved:
1 May 2012
- Published online:
18 September 2012
- Version of record:
6 December 2012
View Cart
