You are on page 1of 16

SPE-169235-MS

Experimental Study of CO2 Huff-n-Puff Process for Low-Pressure


Reservoirs
J. Ma, University of Regina; X. Wang, and R. Gao, Yanchang Petroleum; F. Zeng, University of Regina;
C. Huang, Yanchang Petroleum; P. Tontiwachwuthikul, University of Regina

Copyright 2014, Society of Petroleum Engineers


This paper was prepared for presentation at the SPE Latin American and Caribbean Petroleum Engineering Conference held in Maracaibo, Venezuela, 2123 May
2014.
This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
CO2-based EOR technology has been proven as an effective method to enhance light-oil recovery, while
to reduce CO2 emission. Miscible and near-miscible CO2 flooding have been extensively studied and
commercially applied in the past several decades; while CO2 huff-n-puff process did not gain too much
attention. CO2 huff-n-puff process may have better performance than CO2 flooding for a low-pressure
reservoir, in which reservoir pressure is far below the minimum miscible pressure (MMP), since the CO2
injection is able to build up the reservoir pressure, and the increased reservoir pressure will enhance the
mass transfer between the CO2 phase and the oil phase.
In this paper, 8 coreflood tests, totally 35 runs, are conducted to investigate the major factors affecting the
performance of CO2 huffnpuff in a low-pressure reservoir. The coreflood tests are conducted in a 984
mm-long composite core with an average porosity of 19.6% and an average permeability of 117 mD. The
reservoir pressure is 6.58 MPa, far below the measured MMP value of around 23 MPa. The effects of primary
operational parameters, such as CO2 injection rates, slug size, soaking time, injection pressure, pressure
depletion rates, and chasing gas (N2) on the huffnpuff performance have been extensively studied.
The experimental results indicate that the recovery factor for each cycle is reduced to 40% to 60%,
compared with that of the previous cycle. The recovery factor in each cycle is mainly affected by the total gas
slug size injected and the maximum pressure built up by gas injection. Chasing CO2 with N2 can effectively
reduce the CO2 usage, while increase the reservoir pressure to an ideal level, so that the maximum value of the
amount of the oil as recovered by per unit of CO2 injected can be obtained. It is also found that a long soaking
period is necessary to achieve a favorable performance while a long soaking time may greatly decrease the
average oil production rate of a cycle. In this study, it was observed that the first cycle rather than two cycles
following it was more sensitive to a long soaking time. The results of this study indicate that CO2 huff-n-puff
process has potential to rapidly improve the single-well performance in low-pressure reservoirs.

1. Introduction
Greenhouse gases (GHG) emissions are commonly identified as a major contributor to global warming.
CO2-based enhance oil recovery (EOR) techniques have shown great potential to recover remaining oil

SPE-169235-MS

in secondary or tertiary production while offset the GHG emissions by means of sequestrating CO2
underground. CO2 flooding and cyclic CO2 injection (CO2 huff n puff), are two widely applied
CO2-based EOR techniques. The applicability of either EOR techniques mainly depends on reservoir
conditions, reservoir fluids, formation properties, and the availability of CO2 sources.
This study targets at a reservoir located in northwestern China with a very low original pressure of 12.9
MPa, far below its measured MMP value of 23 MPa. In addition, the reservoir formation is very tight,
permeability averaged 2.3 mD, in some regions even as low as 0.978 mD. The original reservoir pressure
is so low that there is no sufficient energy for the reservoir to do primary production. Furthermore,
waterflooding, a popular EOR method across the world, is not an option for this reservoir because of its
quite low permeabilities. The injectivity required by a successful waterflooding is not achievable.
However, for CO2 injection, injectivity problems do not exist. Additionally, there are abundant CO2
sources from large-scale coal chemical plants in the neighborhood of the reservoir and available for CO2
EOR use. Considering that CO2 flooding is not applicable in low-pressure reservoirs because of its poor
performance when an operation pressure is far below the corresponding MMP, CO2 huff-n-puff process
is a rational option.
CO2 huff-n-puff process is a typical single well operation, which was presented as an alternative of
cyclic steam stimulation for enhanced oil recovery (Stright et al., 1977; Patton et al., 1982; MongerMcClure et al., 1991). This process involves the injection of a slug of CO2, followed by a soaking time
allowing gas phase to mix with oil phase in place. Following the soaking time, the well is put into
production. The efforts to investigate the applicability of this process to enhance oil recovery have been
made for several decades with encouraging results. The results of laboratory tests and field treatments
demonstrated that CO2 huff-n-puff process is economically viable in diverse reservoir conditions.
Khatib et al. (1981) reviewed results of previous cyclic coreflood tests and field applications on
miscible CO2 injection and indicated that the use of CO2 is applicable in both heavy and light crude
to enhance oil recovery. Sayegh and Maini (1984) evaluated a huff-n-puff process in a Lioyminster
heavy oil reservoir by performing corefloods using CO2 and recycled produced gases. They
conducted an evaluation of relative permeabilities to gas and water at reservoir conditions as well as
an assessment of the longitudinal distribution of CO2 and the effect of soak period. On the basis of
field-treatment evaluations, Patton et al. (1982) and Haskin and Alston (1989) developed two
correlations to predict the process performance and some criteria to evaluate the extent to which how
a cyclic CO2 injection process is successful. One important economic indicator presented of
successful implementations is CO2 utilization, defined as the volume of CO2 used for per unit volume
of incremental oil produced, in unit of Mscf/STB. Its favorable rang is from 0.5 to 0.8 Mscf/STB.
Monger-McClure and her colleagues developed extensive research works on the feasibility of CO2
huff n puff process on light-oil recovery (Monger-McClure and Coma, 1988; Thomas et al., 1990;
Monger-McClure et al., 1991; Thomas and Monger-McClure, 1991). They investigated the influence
of various critical parameters, including CO2 slug size, the number of cycles, operational pressures,
impurity of CO2, reservoir gas, and gravity segregation and remaining oil saturation, by conducting
laboratory coreflood tests on watered-out cores in conjunction with comprehensive reviews of
hundreds of field applications. It was suggested that light-oil recovery by CO2 huff-n-puff either in
pressure-depleted reservoirs or for waterflood residual oil is promising. In addition, they also
compared the recovery mechanisms between CO2 injections on light-oil with heavy-oil. Torabi and
his team member investigated the performance of CO2 huff-n-puff process in naturally fractured
reservoirs by conducting experimental and simulation studies (Torabi and Asghari 2010; Torabi et al.
2012). Even though the volume ratio between fracture and matrix used in their laboratory model were
much larger than that in real reservoir scenarios, their research work filled the lack of information
relevant to the application of CO2 huff-n-puff process in naturally fractured reservoirs.

SPE-169235-MS

Numerical simulations by history-matching field


performance revealed that reduction of oil viscosity,
oil swelling, and gas relative permeability hysteresis
are the principal mechanisms contributing to CO2
huff-n-puff response (Hsu and Brugman 1986; Denoyelle and Lemonnier, 1987).
However, nearly all the available experimental studies mentioned above simulated the displacement
process for remaining oil after waterflooding in a reservoir with a constant pressure boundary or connected
to an aquifer. There were two commonalities existing among the aforementioned experimental operations.
One was of core preparation: the core used in corefloods was firstly flooded with brine completely to
measure its pore volume and then was flushed with oil sample used in tests till no water was displaced
out to establish the oil saturation and connate water saturation. Afterwards, the core was flooded again till
no oil production was observed to ensure the subsequent oil recovery was really tertiary. The other feature
in common for those tests existed in production phase. Those production phases started with opening
injection end of a core to produce oil while pumping brine back into the core from the other end to
simulate a water flux. So, what they simulated is a physical process of CO2 huff-n-puff occurring in a
reservoir with a constant pressure boundary or connecting to an aquifer. During those processes, reservoir
pressure can be maintained. However, it is common in reality, just as the targeted reservoir in this study
that a reservoir has neither a constant pressure boundary nor an aquifer but a closed boundary and also
lacks of original energy to do primary recovery. For a closed boundary reservoir, it is difficult to maintain
its pressure during production phase, actually a pressure deletion process. The pressure depletion
production process should have some differences in performance form that pressure maintained process
mentioned above. However, few literatures are available on the cyclic CO2 injection in a low-permeability
reservoir with closed outer boundary with a pressure depletion process during production. This study
mainly focuses on the investigation of the performance of CO2 huff-n-puff process in a low-pressure and
low-permeability reservoir with a closed boundary and without an aquifer.
The objective of this study is to investigate the influence of primary operating strategies to maximize
the response of CO2 huff-n-puff in a target reservoir through laboratory coreflooding tests. This study
extends the insight into cyclic CO2 treatment at the primary production stage in a closed boundary tight
oil reservoir. The results suggest that the recovery factor of a three-cycle huff-n-puff gas injection can be
as high as 34.65%. 0.1 PV CO2 slug seems to be the optimum injection slug size for the first cycle
operation with a favorable CO2 utilization as low as 0.324 Mscf/STB. N2 is a good chasing gas that can
significantly improves the economy of the operation.
Table 1Properties of dead oil

Desity @ 20C
Average Molecular Weight
C11 Molecular Weight

857.9 kg/m3 (33.51API)


294.92 kg/kmol
354.07 kg/kmol

2. Pressure-Volume-Temperature (PVT) Studies


2.1. Crude oil sample
In this study, cyclic laboratory corefloods were conducted to investigate the performance of CO2 injection
processes in a low-pressure, low-permeability, and light-oil reservoir. A light crude oil from an oilfield
in northwestern China used in the corefloods has an API gravity of 40.34. The oil phase sample was
flashed down to atmospheric conditions to obtain the dead oil and reservoir gas samples. The properties
of the obtained dead oil are listed in Table 1. The average molecular weight of produced gases is 32.125
kg/kmol and its molar fraction in reservoir fluids is 46.95%.
2.2. Recombined live oil properties
The dead oil was combined with produced gases at reservoir conditions to a gas-oil ratio (GOR) of 57.65
m3/m3 and is used in the CO2 coreflooding tests. Other primary pressure-volume-temperature (PVT)
properties of the recombined oil, such as bubble point pressure, density, viscosity, and swelling factor (SF)

SPE-169235-MS

were also measured at reservoir temperature of


44C and summarized in Table 2. The schematic
experimental setups used in this study are shown in
Figure 1.

Table 2Recombined oil properties @ 44C


Bubble point pressure Pb (MPa)
Density (kg/m3)
Bo (m3/m3)
GOR (m3/m3)
Viscosity @ Pb (cP)

7.45
823.6
1.1458
67.2
2.60

Figure 1Schematic diagram of oil phase behavior measurement apparatus


Table 3Equilibrium liquid properties of CO2 and live oil mixture @ 44C
CO2 molar fraction
(mol%)

Pb
(MPa)

Density @ Pb
(kg/m3)

Viscosity @ Pb
(cP)

Swelling factor

23.18
37.24
41.18
59.96
64.65
67.13

10.75
11.95
12.18
15.88
21.30
25.25

814.9
813.4
805.2
803.0
792.0
785.1

1.20
0.99
0.82
0.63
0.50
0.42

1.0616
1.1402
1.1792
1.3427
1.4932
1.5624

2.3. PVT properties for recombined oil-CO2 mixture


To simulate the effects of CO2 injection on the properties of reservoir oil during CO2 huff-n-puff process,
a PVT study was carried out for the equilibrium system of the mixture of recombined oil and CO2 over
the range of CO2 molar fraction from 23.18 to 67.13 mol%. The equilibrium properties of mixture are
listed in Table 3, including bubble point pressure, density, viscosity, and SF. Here, SF is defined as the
volume of CO2-saturated reservoir fluids that is divided by the volume of reservoir fluids. Table 3 shows
that the bubble point pressure and the swelling factor of the mixture significantly increase with the molar
fraction of CO2, whereas the viscosity of oil dramatically decreases with it. Previous literatures suggested
that oil-swelling and viscosity reduction are two principle recovery mechanisms in CO2 injection process.

3. Huff-n-Puff Coreflood Study


In this study, the coreflooding tests were conducted in a 984 mm-long, 25.38 mm-diameter composite core
drilled from the same area as the sampled oil. This core has a reservoir pore volume (PV) of 95.8 mm3,
an average porosity of 19.6%, and an average permeability of 117 mD. The reservoir pressure is 6.58 MPa

SPE-169235-MS

Figure 2Diagram of huff-n-puff coreflood apparatus.


Table 4 Reservoir properties

while reservoir temperature is 44C. The impacts of


primary operational parameters, such as slug size,
the number of cycles, maximum pressure and chas117
19.6
44
6.58
ing gas (N2) on the performance have been investigated. The purity of CO2 used for this study was
99.99% and formation water used was formation water obtained from the same formation as oil sample,
containing 2 wt% CaCl2 and 1.5 wt%MgCl2, pH value ranging from 3.0 to 4.0. The use of formation water
can minimize the clay swelling during the process.
Permeability
(mD)

Prosodity
(%)

Temperature
(C)

Pressure
(MPa)

3.1. Experimental setup


Figure 2 shows the schematic diagram of the huff-n-puff coreflooding experimental setup. The core was
placed horizontally in the middle of the core holder which was coated with an isothermal case to maintain
the temperature at a constant value, 44C in this study. The horizontal core eliminated the gravity effect
on production. The injection gas, reservoir water, recombined oil, and washing liquid were contained
separately in cylinders connecting to transform vessels. Those vessels were used together with the
displacement pumps to inject the fluids into the core. A back pressure regulator was used to maintain the
core pressure. The pressures applying to the both ends of the core and the annulus were monitored using
pressure gauges, respectively. The annulus pressure was maintained at a pressure around 2.1 MPa higher
than the core pressure to protect the core against rupture. The produced fluids were introduced into a
three-phase separator at ambient conditions and the production gas was measured by a gas flow meter.
3.2. Experimental procedure
The huff-n-puff coreflooding tests were conducted with the recombined oil at a temperature of 44C. A
composite core was used to simulate the typical reservoir properties. It was prepared by organizing 22
reservoir core plugs drilled from the same formation with different length, porosity, and permeability. The
reservoir properties are listed in Table 4.
In the traditional coreflooding process, fluid is injected from one end of the core and the oil is produced
from the other end. However, the cyclic CO2 displacement uses one end of the core for both gas injection
and oil production to mimic a single-well operation. In this study, the other end of the core was closed

SPE-169235-MS

Table 5Conditions of CO2 huff-n-puff corefloods @ 44C (with a pressure drop of 0.5 MPa across the core)

to model a reservoirs closed outer boundary frequently encountered in field cases. The reservoir boundary
condition modeled in this study was different from those by previous literatures (Monger-McClure and
Coma, 1988; Thomas et al., 1990; Monger-McClure et al., 1991; Bardon et al., 1994; Shayegi et al., 1996;
Zhang et al., 2006), where a constant reservoir boundary with an aquifer was involved.

SPE-169235-MS

Table 6 Summary of CO2 huff-n-puff coreflooding test results


Runs
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

Maximum pressure (Mpa)

Recovery factor (%)

11.37
5.35
4.85
4.05
13.96
9.25
8.64
8.4
7.69
13.48
8.89
8.21
7.89
7.39
13.33
9.04
8.42
8.61
7.51
11.37
11.37
11.37
11.37
11.37
11.37
11.37
11.37
11.37
11.37
8.15
6.88
12.34
11.74
10.86
12.58

14.28
5.7
3.48
0.7
15.91
7.49
4.63
2.42
1.08
13.78
7.06
4.59
2.46
0.89
16.17
8.48
4.87
2.21
1.17
9.47
7.06
4.76
11.37
1.13
10.2
7.43
4.78
2.91
1.34
7.7
4.62
5.01
8.32
6.33
4.15

Ultimate recovery factor (%)


Test 1

24.16

Test 2

31.53

Test 3

28.78

Test 4

32.9

Test 5

24.8

Test 6

26.66

Test 7

17.33

Test 8

18.80

First, each core was completely saturated with formation water to measure its pore volume. Then, the
recombined oil was pumped into the core till no additional water produced to establish the original oil
saturation and connate water saturation. The core saturated with oil underwent subsequently a process of
aging to re-establish its wettability. All those operations occurred at reservoir conditions (5.68 MPa and
44C). Compared with the huff-n-puff corefloods by other researchers (Monger-McClure and Coma,
1988; Thomas et al., 1990; Monger-McClure et al., 1991; Bardon et al., 1994; Shayegi et al., 1996; Zhang
et al., 2006), the saturated core with oil in this study was not flooded again with formation water since
here CO2 huff-n-puff process was applied at the primary stage of production.
Once oil saturation and connate water saturation were established, the cyclic CO2 displacement process
was initiated by injecting gases from a compressed gas cylinder connected to the core injection end at a
constant injection rate measured at reservoir temperature of 44Cand ambient pressures. When the
designed amount of gases was completely pumped into the core, the gas injection (huff portion) ceased.
The injected gas slug was designed so that the core pressure was built up to a desired level. When chase
gas involved in the corefloods, it was injected after CO2 injection. Upon the cessation of gas injection, the
core was shut in for a period of time (soak period). Then, the core was returned to production (puff

SPE-169235-MS

portion) from the injection end of the core while the other end was being closed. The production stage was
terminated when core pressure depleted to a certain level. The produced fluids were routed to the three
phase separator where oil phase and gas phase were permitted to separate. The produced oil was weighed
and its volume was calculated from the density of the dead oil. Once the volume of produced oil was
gained, the recovery factor for each cycle and the ultimate recovery factor for a whole test were calculated
from the volumetric material balance. The produced gas was routed into a gas meter to be measured there.
Obviously, the production process in this study was a pressure-depletion process, which was also different
from other studys operational pressure of a constant value during the production process (MongerMcClure and Coma, 1988; Thomas et al., 1990; Monger-McClure et al., 1991; Bardon et al., 1994;
Shayegi et al., 1996; Zhang et al., 2006). This injection-soaking-production process described above is one
typical huff-n-puff cycle.
A second cycle of gas injection was carried out immediately after the first cycle oil production ceased
and a third cycle conducted after the second one and so on. The same procedures were repeated in the
subsequent cycles in the similar manner to the first one with an exception that, in some cases, different
slug sizes of gases were injected. Upon the completion of a test, the core was cleaned carefully using the
same procedure described by Monger and Coma (1988) but petroleum ether rather than xylene was used
as cleaning liquid according to relevant Chinese regulations.
During the huff phase, pressure was gradually built up from the original reservoir pressure of 6.58 MPa
to the expected level while during the puff phase, the core pressure was depleted at a constant depletion
rate, 3- or 1.5 MPa/h, from the maximum pressure to a designed level with a pressure drop of around 0.5
MPa across the core. During soaking time, the slight pressure change of the core was observed which
might result from the dissolution of CO2 into oil phase. To investigate the effects of soaking time, three
different durations of soaking time, 10, 3, and 1 h, were implemented in the study. Eight series of cyclic
coresfloods, 35 runs in total were conducted, 32 runs carried out used pure CO2 as injection gas (runs 1
through 32) while runs 33 through 35 injecting N2 after CO2 injection as chasing gas. Experimental
conditions for this study are concluded in Table 5 and oil displacement results are presented in Table 6
and Figures 3 through 11.

4. Result and Discussions


4.1. Recovery factor
To investigate the potential recovery of CO2 cyclic operations for low-pressure light oil reservoirs, 8 series
of multi-cycle corefloods were conducted, the corresponding operation conditions tabulated in Table 5.
Tests 2 through 5, and 6 were five-cycle operations, Test 1 was a four-cycle operation, and Tests 7 and
8 were three-cycle injection processes. Tests 1 through 7 used pure CO2, while Test 8 injected certain
amount of N2 upon the completion of the injection of 0.1 PV CO2 slug. From Table 6, it is observed from
the results of Tests 1 through 4, four series of pure CO2 operations with constant slugs injected, that the
process response significantly declined in the subsequent cycles compared with previous cycles. Take
four-cycle operation Test 1 (Runs 1 4) as an example, the injected pure CO2 slug size in each run was
0.2 PV. The recovery factor of run 2 was 5.7% OOIP, around 40% of that of Run 1, 14.28% OOIP, while
the recovery factor of Run 3 was 3.48% OOIP, 60% of that of Run 2. It is also observed that the recovery
factors of the forth/fifth cycle was negligible when compared to the first three cycles and the oil
production was mainly from the first three cycles. For example, the sum of the recovery factors of the last
two cycles of Test 2, a five-cycle operation with a CO2 slug of 0.3 PV for each of its cycle injections, was
less than 13% of the total oil recovery of the first three cycles. This decline trend of oil recovery in the
cyclic gas injection sequence is consistent with the observations reported in the low-permeability
corefloods (Wang et al., 2013). The variation trend of recovery factor in sequential cycle operations
suggests that an economical cyclic CO2 injection should consist of three cycles.

SPE-169235-MS

Figure 3RF comparison between Test 5 with an injection rate of 720cc/h and Test 6 with an injection rate of 480 cc/h.

The total recovery factor of the first three cycles of Tests 1 through 8 achieved in this study ranged
from 17.33% OOIP for Test 7 to as high as 29.52% OOIP for Test 4 and was observed to be sensitive to
the operation conditions, such as CO2 injection rate, gas slug size, maximum operation pressure, and
soaking time. The CO2 huff-n-puff technique is viable to recover the oil in low-pressure light oil reservoirs
with closed outer boundary.
4.2. Effects of the CO2 injection rate
The injection rate is one of the important operation parameters for CO2 huff-n-puff field applications. In
field operations of CO2 huff-n-puff process, CO2 is usually recommended to be injected at the highest rate
that the formations can hold, helping the gas phase penetrate as far into the rock and contact as much oil
as it can (Karim et al., 1992, Liu et al., 2005). In this study, to address the effects of the injection rate on
light-oil CO2 huff-n-puff process, two analogous five-cycle operations, test 5 (runs 20 through 24) and test
6 (runs 25 through 29), were performed under the same operation conditions except using two different
injection rates, 480 cc/h for test 5 and 720cc/h for teat 6. The test results were concluded in Table 6 and
also compared in Figure 3. The bar charts in Figure 3 show that the CO2 injection rate of 480cc/h, when
compared to the injection rate of 720 cc/h, is able to generate more favorable performance either for one
single cycle or one complete test. The improvement of the performance for the first two cycles is more
noticeable than other cycles. It seems that a smaller CO2 injection rate, such as 480 vs. 720cc/h, may lead
to a more successful CO2 cyclic operation. This trend is contrary to the observations seen in another
low-permeability corefloods (Wang et al, 2013) with the same boundary conditions as this study. In the
low-permeability corefloods, two CO2 injection rates of 60 and 140cc/hr were examined and it was found
that the injection rate of 140cc/h produced some performance improvement. On the basis of the
combination of the observations in those two coreflood, the optimal CO2 injection rate for a CO2
huff-n-puff application with a closed outer boundary should lie between 140- to 480cc/hr. Further
corefloods using intermediate injection rates are needed to identify its optimal vale. Neither a large
injection rate, such as 720cc/hr, nor a small one, such as 60cc/h can benefit the cyclic CO2 injection
processes. This observation is similar to that presented by Karim et al. (1992). In their study, the
injection rates of 60, 100, 140, 160, and 200 cc/h in first- and second-cycle operations of cyclic CO2
injection process were conducted using a watered-out conventional core. They observed that the
intermediate injection rate of 140 cc/hr was the optimal injection rate to maximize the overall process
performance. The observation differences between this study and Karim et al. may mainly lie in the
differences in simulated reservoir boundaries.

10

SPE-169235-MS

4.3. Effects of CO2 slug size


CO2 slug size is a primary factor for an economically successful CO2 huff-n-puff process. To investigate the effects of the CO2 slug size on the operation response and optimize it, three series of
corefloods, Tests 1, 3, and 7, were conducted using
CO2 slugs ranging from 0.1 PV to 0.3 PV. Test 1
was a four-cycle operation, Runs 1 through 4, each
Run using constant CO2 slug size of 0.2 PV. Test 2
was a five-cycle operation using constant CO2 slug
size of 0.3 PV in each cycle while Test 7 used 0.1,
0.2, and 0.3 PV slug sequentially in its three cycles.
Test results were concluded in Table 6. Comparing
the performance of Runs 1, 5, and 30, the first-cycle
operations of Tests 1, 2, and 7, respectively, shows
that Run 5, achieving a recovery factor of 15.91%
Figure 4 Cycle recovery factor for 0.1 PVs CO2 injection for Tests 1,
OOIP, was more favorable than Run 1, 14.28%
2 and 7.
OOIP, and Run 1 was more favorable than Run 30,
7.7% OOIP. It is suggested that injecting a large
CO2 slug in the first cycle benefits the cycle performance a lot. The same trend can be seen for the
subsequent cycles as well.
Comparing the total oil recovery and the corresponding CO2 consumption for the first three cycles
between Tests 1 and 7, 23.46% OOIP and 0.6 PV
vs. 17.33% OOIP and 0.7 PV, shows that Test 1
using a constant CO2 slug of 0.2 PV in each of its
cycles was more favorable and economic than Test
7, using sequentially doubled CO2 slugs for its three
cycles. It is suggested that, compared to injecting
increasing CO2 slugs in the subsequent sequences,
using a constant CO2 slug in each cycle operation is
Figure 5The dependence of the recovery factor on the maximum
an ideal operation scheme.
pressure of the core Run 30 vs. 33.
To analyze the CO2 efficiency of the operations
of Tests 1, 2, and 7, the average cycle performance
with respect to 0.1 PV CO2 slug, is calculated and the comparison of it is done among those three tests,
as shown in Figure 4. Figure 4 displays that Run 30 was most favorable when compared to Runs 1 and
5 and Run 1 is more efficient than Run 5 in terms of the average cycle response corresponding to 0.1 PV
CO2 slug injected. It implies that increasing the CO2 slug in the first cycle operation does not improve the
CO2 efficiency of the operations, a key economic indicator meaning a successful operation.
It can be concluded that injecting a large CO2 slug increases the oil recovery of a cycle operation while
decreases its CO2 efficiency and that injecting a fixed CO2 slug in each cycle of a multi-cycle operation
is an ideal operation strategy to improve its total economy.
4.4. Effects of the maximum pressures
Just as literatures reported, the process response increases with the maximum pressure because a certain
level of pressure support is essential for sustaining production during puff portion (Monger-McClure and
Coma, 1988; Mohammed-Singh et al., 2006, Wang et al., 2013). In addition, High level pressure is the

SPE-169235-MS

11

primary energy to displace the oil in place for the


pressure-depleted production processes studied in
this paper.
To examine the effects of the maximum pressures on process response, the performance comparison between Runs 30 and 33 were conducted. Both
Runs were first-cycle operations with similar operation conditions, listed in Figure 6, except that Run
33 using 0.1 PV N2 slug following the injection of
0.1 PV CO2 slug. The injection of 0.1 PV N2 slug
after 0.1 PV CO2 slug in Run 33 helped to build up
the maximum pressure to 11.71 MPa, much higher
than that of Run 30, 8.15 MPa. Correspondingly, the
response of Run 33, 8.32% OOIP, is more favorable
Figure 6 Effects of the maximum pressure on oil production process than that of Run 30, 7.7% OOIP. The dependence of
for Run 30 vs. Run 33.
the process response on the maximum pressure is
plotted in Figure 5, suggesting that the higher the
maximum pressure of the core is, the more favorable the process is. The same trend can be seen from
the performance of Runs 1 vs. 5 and 31 vs. 34. The
level of the maximum pressure should be as high as
the rock can hold.
Figure 6 displays how the oil production rates
varied with the core pressure for Runs 30 and 33,
respectively, over the whole puff portion the core
pressure depleted from the maximum pressure to the
termination pressure of 3 MPa. It is suggested that
building up the maximum pressure to a high level is
helpful to extend the production duration with reFigure 7Comparison of CO2 efficiency between Tests 7 and 8.
spect to the fixed termination pressure, e.g. 3 MPa
for Runs 30 and 33, and the oil production rates are
generally much larger at the early production stage,
the core pressure depleted from the maximum value
to a certain level, e.g. around 6 MPa in this study, than at the later production stage, pressure depleted from
6 MPa to the termination pressure. Therefore, the performance in the early production stage is more
favorable than in the later production stage. It can be confirmed by analyzing the contribution of different
production stages to the oil production of a Run. Take Run 33 as an example, building up the maximum
pressure of 11.74 MPa by injecting 0.1 PV CO2 slug plus 0.1 PV N2 and terminated at the pressure of 3
MPa. So, the pressure drop of Run 33s puff portion was 8.74 MPa. The first production stage of it
consumed less than two-thirds of the whole pressure drop but produced nearly 85% of its oil production.
It indicates that for the pressure depleted production processes, the early production stage with higher
level pressures is the major contributor to the oil production, more favorable and efficient than the later
stage when the pressure significantly depleted to a certain level.
4.5. Effects of chasing gas
As mentioned above, the high level of maximum pressure caused by gases injection improves the recovery
significantly. The high maximum pressure may be achieved by injecting a large CO2 slug or injecting a small
CO2 slug plus an additional chasing gas slug, such as N2. However, just as Figure 4 displayed, the increasing

12

SPE-169235-MS

CO2 slug would decrease the CO2 efficiency and


therefore lower the economy of operations. It was
proven that injecting N2 following a CO2 slug has
greatly potential to improve the process response of
CO2-based EOR processes (Rowe et al., 1982; Karim
et al., 1992; Shayegi et al., 1996). In this study, the
viability of using N2 as chasing gas for CO2-based
close-boundary cyclic gas injections was investigated
by conducting a tree-cycle operation of Test 8.
The three cycles of Test 8, Runs 33 through 35,
sequentially injected 0.1 PV CO2 slug plus 0.1 PVs N2
slug, 0.1 PV CO2 slug plus 0.2 PV N2 slug, and 0.1 PV
CO2 slug plus 0.3 PV N2 slug. The maximum pressure
Figure 8 Effects of the pressure depletion rates on the cycle recovery achieved in Run 33, the first cycle of Test 8, was 11.74
factor
MPa, larger than 11.37 MPa, the maximum pressure
caused by 0.2 PV pure CO2 slug in the first cycle
operations. The maximum pressures built up in the
second cycle of Test 8, Run 34, was 10.86 MPa, larger
than that caused by 0.3 PV CO2 injected in the secondcycle operations, 9.04 MPa, and the Run 35s maximum pressure reached the same level of that caused by
injection 0.4 PV CO2 slug in Run 32, the third cycle
operation of Test 7. It is suggested that with respect to
a desirable level of maximum pressure, injecting N2
after the injection of a small slug of CO2 dramatically
reduce the usage of CO2 and therefore improve the
CO2 efficiency of operations.
To evaluate the effects of injecting N2 immediately after the injection of CO2 on the CO2 effiFigure 9 Effects of the pressure depletion rate on the cycles average
ciency of processes, the comparison between Tests
oil production rates for Tests 2 and 4
7 and 8, pure CO2 injection vs. CO2 plus N2 injection with the similar conditions, was conducted by considering the average cycle response caused by 0.1
PV CO2 slug and was displayed in Figure 7. It was suggested that using N2 as chasing gas had great
potential to improve the CO2 efficiency of operations and this improvement was more significant in
subsequent cycles when compared to the previous ones.
4.6. Effects of the pressure depletion rate
To address the effects of pressure depletion rate on light oils huff-n-puff processes, two pressure
depletion rates, 3- and 1.5 MPa/h, were examined. 3 MPa/h was used for Tests 1 through 3, 5, and 6, and
1.5 MPa/h for Tests 4, 7, and 8. Among those 8 Tests, Tests 2 and 4 were analogous except using different
pressure depletion rates and were compared from two aspects of the cycle response: cycle recovery factor
and the average oil production rate of the cycle, illustrated in Figures 8 and 9, respectively. From Figure
8, it is seen clearly that the recovery factor of each cycle of Test 4, using the pressure depletion rate of
1.5 MPa/h, is more favorable than that of the same cycle of Test 2, using the pressure depletion rate of
3 MPa, while the improvement resulted from the decrease of pressure depletion rates is negligible.
Inversely, Figure 9 shows that the pressure depletion rate of 3 MPa/h for Test 2 induced some tiny
improvements of the cycles average oil production rate when compared to the pressure depletion rate of
1.5 MPa/h for Test 4.

SPE-169235-MS

13

Based on the test results in this study, it is suggested that using different pressure depletion rates
may affect the process response positively or negatively, but the resulted effects were unmeasurable.
4.7. Effects of soaking time
Soaking time is a major consideration when designing a CO2 huff-n-puff process. To investigate
its effects on the process response and figure out
the optimal soaking period of the cyclic CO2 injections in low-pressure light oil reservoirs with
closed outer boundary, Tests 2 and 3 were conducted under similar conditions with an exception
of allowing different soaking periods. Long soaking periods were used in the successive cycles of
Figure 10 The performance comparison between Run 5 with 10 h
Test 2 (10-h soak time for the first cycle, Run 5, vs.
soaking time and Run 10 with 1 h soaking
3-h soaking time for the remaining four cycles,
Runs 6 through 9.) to have the core pressure stabilized, while a short soaking time of 1 hour was used
in each cycle of Test 3, Runs 10 through 14.
The results of those two Tests (listed in Table 6)
indicated that, for the first cycle, the oil recovery was
significantly increased by 15.5 % when the soaking
time was extended from 1 hour for Run 10 with a
recovery factor of 13.78% OOIP to 10 hours for Run
5 with a recovery factor of 15.91% OOIP. However, if
compared to Run 5 in terms of the average oil production rate, as plotted in Figure 10, the performance
of Run 10 was much favorable than that of Run 5. Run
10 achieved an average oil production rate of
Figure 11The performance comparison between Run 6 with 3 h
14.14108 BOPD, nearly as twice as that of Run 5,
soaking time and Run 11 with 1 h soaking
7.9108BOPD. It suggested that an over-extended
soaking time may greatly reduce the average oil production rate of the cycle operations.
For the second-cycle operations, as plotted in Figure 11, Run 6, allowing 3 hours soaking, exhibited
the slight improvement in oil recovery, 7.49% OOIP vs. 7.06% OOIP for Run 11, using 1 hour soaking,
but the measurable decrease in the average oil production rate, 7.1108 vs. 8.75108 BOPD for Run
11. The same observations can be obtained for the third-cycle operations, Runs 7 vs. 12.
It is indicated that a long soaking period, such as 10 hours in this study, helped a lot to improve the oil
recovery in the first cycle but simultaneously significantly decreased the average oil production rate of the
process. However, for the second- and third-cycle injections, allowing 3 hours vs. 1 hour to shut in did not show
a noticeable improvement in oil recovery while resulted in unfavorable average oil production rate. Different
from the observations of Wang et al. (2013), for low-pressure light oil CO2 huff-n-puff injections, the first cycle
rather than the second and third cycles was more sensitive to a long soaking time.

5 Conclusions
Based on the results of this study, some conclusions can be summarized as follows:
1. The CO2 huff-n-puff technique is viable to enhace oil recovery in low-pressure tight oil reservoirs
with a closed outer boundary.

14

SPE-169235-MS

2. The economically optimum huff-n-puff gas injection consists of three cycles with a potential
recovery factor of as high as 29.52% OOIP.
3. For the first cycle, a small CO2 slug size showed an attractive result. 0.1 PV CO2 plus a proper
amount of N2 as the chasing gas to build up the reservoir pressure seems to be the best injection
strategy in terms of the recovery factor and CO2 efficiency. Using more CO2 sequential cycles is
more favorable than than using large CO2 slug sizes.
4. The injection of N2 after a CO2 slug can dramatically reduce the usage of CO2. The performance
strongly depends on the maximum pressure built up during the gas injection period. It is
recommended that the injection pressure of CO2 huff-n-puff is better to be built up to as high as
rock permits.
5. High injection rates of CO2 is unfavorable for the oil production.
6. A soaking time is required in cyclic CO2 injection process with closed boundary to achieve
favorable recovery efficiency while a long soaking time may greatly decrease the average oil
production rate of a cycle. In this study, it was observed that the first cycle rather than two cycles
following it was more sensitive to a long soaking time.
7. The influence of pressure depletion rates were marginally to the performance of CO2 huff-n-puff.

Acknowledgements
The authors would like to acknowledge the permission for the pubilication of this paper from Yanchang
Petroleum (Group), China.

References
Bardon, C., Corlay, P., Longeron, D. et al. 1994. CO2 Huff-n-puff Revives Shallow Light-OilDepleted Reservoirs. SPE Res Eng. 9(2):92100. SPE 22650-PA. http://dx.doi.org/10.2118/22650-PA.
Brock, W.R. and Bryan, L.A. 1989. Summary Results of CO2 EOR Field Tests, 19721987. Paper
SPE-18977-MS presented at Low Permeability Reservoirs Symposium, Denver, Colorado, 6 8 March.
http://dx.doi.org/10.2118/18977-MS.
Bryant, D.W. and Monger, T.G. 1988. Multiple-Contact Phase Behavior Measurement and Application with Mixtures of CO2 and Highly Asphaltic Crude. SPE Reservoir Engineering, 3(2):701710.
http://dx.doi.org/10.2118/14438-PA.
Denoyelle, L.C., Lemonnier, P. 1987. Simulation of CO2 Huff n Puff Using Relative Permeability
Hysteresis. Paper SPE-16710-MS presented at SPE Annual Technical Conference and Exhibition, Dallas,
Texas, 2730 September. http://dx.doi.org/10.2118/16710-MS.
Haskin, H.K. and Alston, R.B. 1989. An Evaluation of CO2Huff-n-puff Tests in Texas. Journal of
Petroleum Technology 41(2):177184. http://dx.doi.org/10.2118/15502-PA.
Huang, E. T.S. and Tracht, J. H. 1974. The Displacement of Residual Oil By Carbon Dioxide. Paper
SPE 4735-MS presented at SPE Improved Oil Recovery Symposium, Tulsa, Oklahoma, 2224 April.http://dx.doi.org/10.2118/4735-MS.
Hsu, H.-H., Brugman, R.J. 1986. CO2 Huff-Puff Simulation Using a Compositional Reservoir
Simulator. Paper SPE-15503 presented at SPE Annual Technical Conference and Exhibition, New
Orleans, Louisiana, 5 8 October. http://dx.doi.org/10.2118/15503-MS.
Karim, F., Berzins, T.V., Schenewerk, P.A., et al. 1992. Light Oil Recovery from Cyclic CO2
Injection: Influence of Drive Gas, CO2 Injection Rate, and Reservoir Dip. Paper SPE 24336-MS presented
at the SPE Rocky Mountain Regional Meeting, Casper, Wyoming, 18 21 May. http://dx.doi.org/10.2118/
24336-MS.
Khatib, A. K., Earlougher, R.C., Kantar, K. 1981. CO2 Injection as an Immiscible Application for
Enhanced Recovery in Heavy Oil Reservoirs. Paper SPE-9928-MS presented at SPE California Regional
Meeting, Bakersfield, California, and 2527 March. http://dx.doi.org/10.2118/9928-MS.

SPE-169235-MS

15

Liu, H., Wang, M. C., Zhou, X 2005. EOS Simulation for CO2 Huff-n-Puff Process. Paper PETSOC2005-120 presented at Canadian International Petroleum Conference, Calgary, Alberta 79 June. http://
dx.doi.org/10.2118/2005-120.
Mohammed-Singh, L., Singhal, A.K., and Sim, S. 2006. Screening Criteria for CO2 Huff n Puff
Operations. SPE 100044-MS presented at SPE/DOE Symposium on Improved Oil Recovery, 2226 April
2006, Tulsa, Oklahoma. http://dx.doi.org/10.2118/100044-MS.
Monger-McClure, T.G., Coma, J.M. 1988. A Laboratory and Field Evaluation of the CO2 Huff n
Puff Process for Light-Oil Recovery. SPE Res Eng. 3(4):1168 1176. SPE 15501-PA. http://dx.doi.org/
10.2118/15501-PA.
Monger-McClure, T.G., Ramos, J.C., Thomas, Jacob 1991. Light Oil Recovery from Cyclic CO2
Injection: Influence of Low Pressures Impure CO2, and Reservoir Gas. SPE Res Eng. 6(1):2532. SPE
18084-PA. http://dx.doi.org/10.2118/18084-PA.
Miller, B.J. 1990. Design and Results of a Shallow, Light Oilfield-Wide Application of CO2 Huffn
Puff Process. Paper SPE-20268-MS presented at SPE/DOE Enhanced Oil Recovery Symposium, Tulsa,
Oklahoma, 2225 April. http://dx.doi.org/10.2118/20268-MS.
Miller, B.J., Bretagne and Bardon, C.P. 1994. CO2 Huff n Puff Field Case: Five-Year Program
Update. Paper SPE-27677-MS presented at Permian Basin Oil and Gas Recovery Conference, Midland,
Texas, 16 18 March. http://dx.doi.org/10.2118/27677-MS.
Mohammed-Singh, L., Petrotrin, Singhal, A.K. and et al. 2006. Screening Criteria for CO2 Huff n
Puff Operations. Paper SPE-100044-MS presented at SPE/DOE Symposium on Improved Oil Recovery,
Tulsa, Oklahoma, 2226 April. http://dx.doi.org/10.2118/100044-MS.
Palmer, F. S., Landry, R.W., and Bou-Mikael, S. 1986. Design and Implementation of Immiscible
Carbon Dioxide Displacement Projects (CO2 Huff-Puff) in South Louisiana. Paper SPE15497 presented
at the 61st Annual Technical Conference and Exhibition of the Society of Petroleum Engineers held in
New Orleans, LA, October 5 8. http://dx.doi.org/10.2118/36687-MS.
Patton, J.T., Coats, Keith H., Spence, K. 1982A. Carbon Dioxide Well Stimulation: Part 1-A
Parametric Study. Journal of Petroleum Technology, 34(8): 18051810. http://dx.doi.org/10.2118/9228PA.
Patton, J. T., Sigmund, P., Evans, B. 1982B. Carbon Dioxide Well Stimulation: Part 2 -- Design of
Aminoils North Bolsa Strip Project (includes associated papers 11928 and 12234). Journal of Petroleum
Technology, 34(8): 1798 1804. http://dx.doi.org/10.2118/8897-PA.
Shayegi, S., Jin, Z., Schenewerk, P., et al. 1996. Improved Cyclic Stimulation Using Gas Mixtures.
Paper SPE 36687-MS presented at SPE Annual Technical Conference and Exhibition, Denver, Colorado,
6 9 October. http://dx.doi.org/10.2118/36687-MS.
Simpson, M. R. 1988. The CO2 Huff n Puff Process in a Bottomwater-Drive Reservoir. Journal of
Petroleum Technology, 40(7):887893. http://dx.doi.org/10.2118/16720-PA.
Sayegh, S. G. and Maini, B. B. 1984. Laboratory Evaluation of The CO2 Huff-N-Puff Process For
Heavy Oil Reservoirs. Journal of Canadian Petroleum Technology, 23(3): 29 36. http://dx.doi.org/
10.2118/84-03-02-PA.
Stright D.H.Jr.,, Aziz, K., et al. 1977. Carbon Dioxide Injection into Bottom-Water, Undersaturated
Viscous Oil Reservoirs. Journal of Petroleum Technology, 29(10):1248 1258. http://dx.doi.org/10.2118/
6116-PA.
Torabi, F. and Asghari, K. 2010. Effect of operating Pressure, Matrix Permeability and Connate Water
Saturation on Performance of CO2 Huff-n-Puff Process in Matrix-Fracture Experimental Model. Fuel, 89:
29852990.
Torabi, F. et al. 2012. Comparative Evaluation of Immiscible, near miscible, and miscible CO2
Huff-n-Puff to Enhance Oil Recovery from a Single Mtrix-Fracture System (Experimental and Simulation
Studies). Fuel, 93: 443453.

16

SPE-169235-MS

Thomas, J., Berzins, T.V., and Monger-McClure, T.G. et al. 1990. Light Oil Recovery from Cyclic
CO2 Injection: Influence of Gravity Segregation and Remaining Oil. Paper SPE 20531-MS presented at
SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana 2326 September. http://
dx.doi.org/10.2118/20531-MS.
Thomas, G.A. and Monger-McClure, T.G. 1991. The feasibility of Cyclic CO2 injection for Light-Oil
Recovery. SPE Res Eng. 6(2):179 184. SPE-20208-PA. http://dx.doi.org/10.2118/20208-PA.
Wang, Z.X., Ma, J.H., Zeng, F.H. et al. 2013. Optimizing Cyclic CO2 Injection for Low-permeability
Oil Reservoirs through Experimental Study. Paper SPE 167193-MS presented at SPE Unconventional
Resources Conference Canada, Calgary, Alberta 5 6 November. http://dx.doi.org/10.2118/167193-MS.
Zhang, Y.P., Sayegh, S. G., Huang, S. et al. 2006. Laboratory Investigation of Enhanced Light-Oil
Recovery by CO/Flue Gas Huff-n-Puff Process. J. Can. Pet. Tech. 45(2):24 32. http://dx.doi.org/10.2118/
06-02-01.