Journal of Petroleum Science and Engineering 162 (2018) 325–332

Contents lists available at ScienceDirect

Journal of Petroleum Science and Engineering

journal homepage: www.elsevier.com/locate/petrol

Experimental study and numerical simulation of nitrogen-assisted SAGD in

developing heavy oil reservoirs
Zhe Yuan a, Pengcheng Liu a, *, Shengfei Zhang b, Xiuluan Li b, Lanxiang Shi a, Ruifeng Jin b
a
School of Energy Resources, China University of Geosciences, Beijing, 100083, PR China
b
Research Institute of Petroleum Exploration and Development, PetroChina, Beijing, 100083, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: The presence of significant amounts of injected steam and heat loss are unavoidable issues occurred that are
Heavy oil observed during the steam-assisted gravity drainage (SAGD) process in developing heavy oil reservoirs. Because of
Non-condensable gas these limitations and concerns, non-condensable gas co-injection has become a potential technique to enhancing
Nitrogen oil recovery and reducing energy consumption. However, the oil production mechanism is complicated and has
Temperature profile not been totally understood, thereby requiring further discussion. In this work, a two-dimensional physical model
Enhance oil recovery
was designed to investigate the production mechanisms of nitrogen on SAGD. Two comparative injection
schemes, specifically the conventional SAGD and nitrogen assisted SAGD (NA-SAGD), were conducted to inves-
tigate the effect of nitrogen co-injection with steam on the SAGD performance. Dynamic temperature profile
changes during the two experiment processes were monitored and recorded. The observation results indicate that
nitrogen accumulation at the upper part of the model resulted in the effective expansion of the NA-SAGD process
steam chamber along the horizontal direction. The swept area of steam increased significantly and the residual oil
saturation was reduced. The cumulative oil production of NA-SAGD was higher than that of the conventional
SAGD. Moreover, the apparent heat loss reduction, which defines the heat utilization efficiency, exhibited an
increase following nitrogen injection. In addition, a numerical simulation was generated to compare and verify
the results obtained in the two experiments. Nitrogen co-injection with steam effectively enhanced the oil re-
covery. Studies indicate that NA-SAGD is a feasible method for improving oil production in developing heavy oil
reservoirs.

1. Introduction
Thermal methods are the most effective way for the development of
bitumen and heavy oil reservoirs given the presence of a high initial
viscosity at reservoir conditions (Hu et al., 2017). To enhance the
mobility of oil using the strong temperature sensitivity of viscosity, the
steam assisted gravity drainage (SAGD) technique is one of the major
thermal recovery methods in developing heavy oil reservoirs (Butler and
Stephens, 1981; Butler, 1994). However, during the SAGD process, the
low utilization rate of resources is observed as one of the most important
problems, specifically high energy consumption and heat loss through
overburden, which affect the production efficiency (Ji et al., 2015).
Meanwhile, the enormous volume of CO2 emissions also generates
environmental problems. As a result, the cost of the production process is
expensive and applications are economically limited (Kazeem, 2014).
Improving the economy and efficiency of the oil production process
has recently become the focus of many research projects given the
requirement of a more economic method to improve the production ef-
ficiency (Jha et al., 2013; Yuan et al., 2017). To solve these problems,
many attempts have been performed to study the efficiency of chemical
solvent injections to enhance the oil recovery in both physical experi-
ments and oilfield pilot projects (Leyva-Gomez and Babadagli, 2013; Liu
et al., 2013, 2015). To take the advantage of the mass transfer of the light
hydrocarbon component, Nasr et al. (2003) presented an expanding
solvent-SAGD (ES-SAGD), wherein hydrocarbons (alkanes, aromatic hy-
drocarbons) mixed with steam were injected into a heavy oil reservoir, to
reduce the amount of injected steam and enhance oil recovery. As
compared to the SAGD process, the results of these attempts exhibited an
apparent increase in oil production (Souraki et al., 2013; Al-Murayri
et al., 2016; Zheng et al., 2016). However, the selection of the most
correct and appropriate solvent is crucial for the success of solvent-aided
injection processes (Hascakir, 2016). However, some solvents are

* Corresponding author.
E-mail address: lpc@cugb.edu.cn (P. Liu).

https://doi.org/10.1016/j.petrol.2017.12.064
Received 5 July 2017; Received in revised form 23 November 2017; Accepted 20 December 2017
Available online 25 December 2017
0920-4105/© 2017 Elsevier B.V. All rights reserved.
Z. Yuan et al. Journal of Petroleum Science and Engineering 162 (2018) 325–332

poisonous and cause irreparable damage to the formation. Furthermore, dynamic changes of the steam chamber and its oil recovery following the
due to the high content of asphaltene in super-heavy oil, asphaltene application of different schemes were compared. The experimental per-
precipitation was observed following the mixture of some solvents and formance and produced results were discussed, of which the results
heavy oil (Stachowiak et al., 2005; Gonzale et al., 2006). Asphaltene indicated that nitrogen as non-condensable gas that is co-injection with
precipitation blocks the pore space and prevents the liquid from flowing, steam is an effective technique to enhance the oil recovery. In addition, a
which is bad for the mobility and the production of oil (Mukhametshina numerical simulation was built to verify the results obtained in the two
et al., 2016). experiments. The temperature distribution characteristics and oil re-
Therefore, an improved technique named steam assisted and gas push covery of the simulation results were in good agreement with the
(SAGP) is presented to improve oil recovery in developing heavy oil experimental results. This work can serve as a reference for the selection
reservoirs and obtain the economic benefits (Butler et al., 2000; Jiang of nitrogen injection during SAGD in developing heavy oil reservoirs.
et al., 2000). During the SAGP process, non-condensable gases such as
carbon dioxide, nitrogen, and methane are added by co-injection of 2. Experiments
steam (Ito et al., 2001; Rahnema et al., 2011; Zheng et al., 2013). These
gases can exist in the steam chamber and have insignificant condensa- 2.1. Experimental apparatus
tions to the liquid phase (Yang and Xi, 2012). Several researchers have
examined the effect of methane gas on SAGP through physical experi- The experimental apparatus of the two-dimensional physical simu-
ments and numerical simulations. Bagci and Gumrah (2004) demon- lation experiments for SAGD and NA-SAGD are shown in Fig. 1. The
strated how the injection of gases such as CO2 and methane increased the apparatus mainly consists of four parts, specifically the injection system,
cumulative oil production in physical experiments. Al-Murayri et al. model body, collection system, and data acquisition system.
(2011a, 2011b) developed a simulation model, wherein the amount of An ISCO pump displaced water and oil into the steam generator and
injected steam and the cumulative steam oil ratio (cSOR) were less than model, respectively, or controlled the flow rate of the inlet during the
that of SAGD as a non-condensable gas methane injection. However, the experiments. Model body is composed of a stainless shell with a glass
production rate and the oil recovery were reduced under the same con- cover, which is wrapped in insulation cotton to reduce the heat loss
ditions. Liu et al. (2012) also reported similar results. Gittins et al. (2011) during the experiments. To prevent the glass cover from crushing, the
studied the effect of non-condensate gas on the development of an SAGD maximum working pressure during the experiment was controlled under
steam chamber. Others reports also generated the simulation models to 10.0 MPa. The injection of nitrogen was controlled and monitored by an
investigate the effects of methane as a solution gas or injected gas for air compressor and gas mass flow controller. High temperature steam was
cumulative oil recovery (Ardali et al., 2012; Sharma et al., 2012; Nour- generated by a steam generator, which can produce steam at a maximum
ozieh et al., 2015). However, methane is still relative expensive and is temperature of 300
C.
easily flammable at high temperatures, which limit the widespread A piston container stored the experimental oil sample, which had a
application during the SAGP process. volume of 2.0 L, and the maximum working pressure was set at 20.0 MPa.
During the SAGP process, non-condensate gas such as nitrogen and Back-pressure valves maintained the pressure of the production well and
carbon dioxide was injected with high temperature steam into the for- ensured the security of the experiment. A data acquisition system with 24
mation when the edge of the steam chamber was closed to cap rock. The pressure sensors and 24 temperature sensors recorded the transient
mechanisms of this process include low heat conductivity, thermal pressures and temperatures inside the model. The system also recorded
expansion effect, and interfacial tension reduction (Yuan et al., 2011; the transient flow rate and the cumulative water injection volume of the
Yang and Xi, 2012). Nitrogen is easily and inexpensively prepared, and is pump. A scale was employed to measure the weight of the production
clean and non-poisonous. It also has a large compressibility coefficient emulsions to calculate the oil recovery.
and a low heat conductivity coefficient. This improved technique has
exhibited good production performance in heavy oil reservoirs in recent
2.2. Experimental materials
years (Rios et al., 2010). A nitrogen gas assisted SAGD pilot test was
performed, of which the oil steam ratio exhibited an 80% increase. In
The experimental materials used in this study are as follows:
contrast, the injection of nitrogen in the Liaohe oil field, China exhibited
Model body: The model has a length and width of 100.0 and 4.0 cm,
a 20% oil production increase (Guo et al., 2015). Lu et al. (2014) indi-
respectively. The height is 25.0 cm and the measured volume is
cated that the addition of nitrogen during steam injection can signifi-
10,000 cm2. The injection well and production well are located in the
cantly improve the oil displacement efficiency according to laboratory
research. However, these experiments were one-dimensional sand pack
experiments, which exhibited limited steam chamber development and
could not describe the dynamic pressure and temperature changes in the
steam chamber. In addition, an increase in the injection volume of ni-
trogen does not generate better results given that nitrogen is concen-
trated at the upper part of the steam chamber where it is hardly
produced, thereby increase the mole fraction of nitrogen (Wang et al.,
2015). The pressure of the steam chamber is too high to allow continuous
injection of steam. Therefore, it is better to adopt a plug injection
method.
Nitrogen has been previously applied as a non-condensable gas in
oilfield pilot projects to enhance the oil production performance (Gao
and Liu, 2008; Du et al., 2013). However, these production mechanisms
are complicated and have not been totally understood, thereby requiring
additional discussions. In this study, two-dimensional physical experi-
ments with different injection schemes were performed to investigate the
production mechanisms of nitrogen assisted SAGD, which we named NA-
SAGD, to distinguish it from the conventional SAGD. The present study
aims to examine and evaluate the effects of nitrogen on the development Fig. 1. Schematic of the apparatus employed for the NA-SAGD and conventional SAGD
of the steam chamber and the residual oil saturation distribution. The experiments.

326
Z. Yuan et al.
center of model along the vertical direction at a distance of 5.0 cm.
Sand: The model was filled with clean quartz sand with a grain size
80–120 meshes. For the best results and to prevent bubbles, sands was
alternately filled with water, stirred and compacted during filling. The
measured porosity of the model was 35.2%.
Oil: The heavy oil sample for these experiments was collected from
the Z18 Block, which is a typical extra-heavy oil reservoir in Xinjiang,
China. The super heavy oil sample has a viscosity of 1,167,243.1 mPa s
and a density of 0.997 g/cm3 at 20
C. The key reservoir properties used
in the simulation are presented in Table 1. Heavy-oil viscosity versus
temperature is presented in Table 2.
Non-condensable gas: The non-condensable gas used in this study is
industrial nitrogen gas with a purity of 99%. Under normal conditions, it
is a colorless and inactive gas.
2.3. Experimental procedures
The experimental procedures are briefly described as follows:
1) Before the experiment, the density and viscosity of heavy oil sample
was tested. The accuracy of the pressure sensors and temperature
sensors were also checked, and replaced if necessary.
2) After the model was filled with quartz sand, high pressure nitrogen
was injected to test the sealing performance of the model. The model
pressure was controlled at 10.0 MPa for 5.0 h to test the leakage
performance.
3) The model was first saturated with water, after which a heated heavy
oil sample was pressed into the model by an ISCO pump, thereby
displacing the residual water through its pores and to saturate oil. The
experiment for the two-dimensional SAGD was conducted when the
oil temperature cooled down to 20
C.
4) Before steam was injected, the confining pressure of the model was
controlled at 4.5 MPa to simulate the formation pressure by injecting
a certain amount of nitrogen. A small amount of naphtha was then
injected into the injection and production wells to enhance the
communication between the well pairs. The injection valves were
opened and the steam generator was set to the designated tempera-
ture of 250
C. The production values were opened about 30 min
later.
5) Distilled water was continuously injected at a rate of 10 mL/min until
the edge of the steam chamber was closed to the left and right sides of
the model. The dynamic changes of the model temperature and
pressure were continuously monitored and recorded. The cumulative
amount of injected steam was also continuously recorded.
6) The production emulsions were collected and weighted, and subse-
quently heated in a dryer to evaporate water content at 100
C. The
produced oil was then weight to calculate the cumulative oil pro-
duction and oil recovery.
7) After the experiment, the model was cleaned with naphtha, and
subsequently filled with sand and saturated with the oil sample. The
above steps were repeated to conduct the physical experiment for
two-dimensional NA-SAGD.
8) During the NA-SAGD experiment, nitrogen was first injected when
the steam chamber was closed to the cover of the model, and then
plug injected when the pressure of steam chamber was lower than the
saturation pressure of saturated steam at 250
C. The other steps were
consistent with the SAGD experiment.
Table 1
Key reservoir parameters used in numerical simulations.
Top depth (m) 400 Pay zone thickness (m) 23
Reservoir pressure (MPa) 4.5 Reservoir temperature (
C) 20
Porosity (frac) 0.30 Oil saturation (frac) 0.70
Permeability I, J (mD) 2000 Model dimension (m) 500  100  23
Permeability K (mD) 1600 Well length (m) 400
327
Journal of Petroleum Science and Engineering 162 (2018) 325–332
3. Results and discussion
Fig. 2 shows the comparative pictures of the temperature profiles of
the steam chamber during the two experiments for the SAGD and NA-
SAGD processes, respectively. Seven different temperature profile time
points were selected for discussion during each experiment.
From Fig. 2A, the steam chamber exhibited obvious efficient expan-
sion along the vertical direction following the injection of high temper-
ature steam. The profile of the chamber was elongated and thin in the
center part of the model at the 4-h mark (Fig. 2A (a)). After 2 h, the upper
part of the chamber exhibited contact with the cover of the model and
began to expand laterally (Fig. 2A (b)). However, due to the presence of a
significant amount of heat loss at the cover, the upper part expanded
slowly along the horizontal direction, while the other part remained
unchanged at the 8-h mark (Fig. 2A (c)). About 10 h following the in-
jection of steam, the whole chamber further developed laterally, and the
heating area quickly increased (Fig. 2A (d)). In the next 2 h, the upper
part of the chamber exhibited apparent expansion, while the other part of
the steam chamber expanded minimally. The temperature of the sur-
rounding cold oil also increased (Fig. 2A (e)). The entire chamber
exhibited apparent development after 16 h. The temperature distribution
of other part of the model was greater than 80.0
C (Fig. 2A (f)). After
20 h, the profile of the steam chamber formed a non-standard “V” shape
(Fig. 2A (g)). The temperature in the red part was measured to be nearly
250
C, and the temperature in the yellow part was between 150
C and
170
C. However, the temperature distribution of other the blue part was
still below 120
C.
From Fig. 2B, nitrogen was first injected when the steam chamber was
closed to the cover of the model, such that the profile of the steam
chamber was similar to that of SAGD which was also elongated and thin
(Fig. 2B (a)). As nitrogen accumulated in the upper part of the model,
heat loss to the outside reduced and the steam chamber expanded
effectively along the horizontal direction, which resulting in a larger
steam chamber after 6 h (Fig. 2B (b)). The heating area was almost twice
that of SAGD at the same time. However, the temperature of the steam
chamber decreased following every cold nitrogen injection. The tem-
perature then returned after continuous steam injection. The nitrogen
was injected a second time at about 7 h as pressure of the steam chamber
was lower than the saturation pressure of saturated steam. The steam
chamber remained unchanged at 8 h as compared to its conditions 2 h
ago (Fig. 2B (c)). Two hours following steam injection, specifically at
10 h, the whole steam chamber expanded apparently along the hori-
zontal direction (Fig. 2B (d)). The temperature of other part of the model
also increased at this time. After 2 h, the steam chamber exhibited min-
imal developments, whereas the temperature of other part increased
significantly at 12 h (Fig. 2B (e)), thereby indicating the development of
the steam chamber, which was followed by the warming of the sur-
rounding cold oil zone. The steam chamber area was significantly larger
at 16 h (Fig. 2B (f)). The temperature distribution of other part was nearly
120.0
C, which indicated that all the saturated oil was heated. After 20 h
(Fig. 2B (g)), the area of the steam chamber increased significantly. The
temperature in the red and yellow parts was similar with those of the
SAGD. However, the major difference was that the temperature distri-
bution of the other part was all greater than 140
C which was much
higher than that of the SAGD.
During the SAGD process, a steam accumulation zone was formed in
the reservoir following the continuous injection of high-temperature
steam, which is a section called the steam chamber. The inside of the
steam chamber was filled with steam, and the outside was relative cold
heavy oil. As the steam flowed upwards due to steam overlaying, the
chamber expanded vertically until it exhibited contact with the cover of
the model, and subsequently expanded along the horizontal direction.
The steam-heated oil flowed down to the production well due to the
effect of gravity with the condensed water. After the crude oil flowed
away, the area that was originally filled with original crude oil was filled
with steam. The steam chamber expanded up and outwards, thereby
Z. Yuan et al. Journal of Petroleum Science and Engineering 162 (2018) 325–332

Table 2
Heavy oil viscosity versus temperature.

T (
C) 20 40 80 100 120 150 180 200 250 290

μo (mPa⋅s) 1,167,243.1 57,695.2 8547.8 1759.3 886.7 269.5 65.6 22.6 6.6 3.5

Fig. 2. Dynamic changes of the temperature profile during the two experiments. (A) Temperature profile during SAGD experiment. (B) Temperature profile during NA-SAGD experiment.

increasing the boundary of the steam chamber. However, steam heat
transfer towards the outside of the model resulted in significant heat loss.
Moreover, the injected steam continued to flow upwards and hardly
expanded laterally due to pressure shortcomings in the steam chamber.
The steam chamber expanded slowly along the horizontal direction only
at the upper part of the model. Based on these observations, the resulting
heat utilization efficiency of steam was low.
During the NA-SAGD process, the temperature profile on either side
further increased in both the upper and middle parts of the model in
comparison with that of SAGD. Due to the low specific gravity of the
injected nitrogen, nitrogen accumulated at the top of the steam chamber.
The accumulated nitrogen in the steam chamber took partial pressure
from the steam, resulting in a corresponding decrease in the pressure and
temperature of steam. A temperature transition region was then formed,

328
Z. Yuan et al. Journal of Petroleum Science and Engineering 162 (2018) 325–332

which reduced the heat transfer to the outside of the model. Moreover,
the upper part of the steam chamber was occupied by nitrogen. There-
fore, the pressure at the upper part of the steam chamber maintained and
upwards steam movement slowed down. Under this condition, steam was
forced to move laterally to areas with lower temperature, thereby
increasing the swept area of steam. The heat utilization efficiency of
steam increased significantly. This point is demonstrated by the residual
oil saturation distribution and higher cumulative oil production in the
NA-SAGD process. As presented in Fig. 2B (g), the left and right sides of
the steam chamber are asymmetrical. The major reason may be that the
accumulation of more nitrogen on the left side, which hindered the
extension of steam.
Fig. 3 shows the comparison of the residual oil distribution after the
two experiments for the SAGD and NA-SAGD processes. From Fig. 3A,
after the SAGD experiment, the oil production was mainly concentrated
in the center and upper parts of the model where the residual oil satu-
ration was less than 30%. However, the other part of the model did not
produce oil at all, and exhibited a residual oil saturation of about 100%. Fig. 4. Relationship between the oil recovery and the injected pore volume of steam in the
From Fig. 3B, on the contrary, the NA-SAGD experiment exhibited two experiments.
different recovery degrees of oil that were produced in other areas of the
model in addition to the bottom left and bottom right regions, where the
residual oil saturation was nearly 100% as it was not swept by steam. The
upper part of the model exhibited the highest recovery degree of heavy
oil, while its residual oil saturation was less than 20%. The residual oil
saturation of other region was between 30% and 70%.
The residual oil saturation distribution is consistent with the steam
chamber at the end of the experiments, especially in the bottom left and
bottom right parts of the model, wherein an obvious difference is
observed in the residual oil saturation, which indicates that the steam
chamber effectively expanded along the horizontal direction following
nitrogen injection. During NA-SAGD process, more saturated oil was
heated by steam, which reduced the residual oil saturation and further
enhanced the recovery factor of heavy oil.
Fig. 4 shows the contrastive relationship between oil recovery and
pore volumes of injected stream in the two experiments. Fig. 5 shows the
contrastive relationship between the cumulative steam oil ratio (cSOR)
and the pore volumes of the injected stream. From Figs. 4–5, the oil re-
covery of SAGD was similar to that of NA-SAGD at the beginning of the
Fig. 5. Relationship between the cumulative SOR and the injected pore volume of steam
two experiments when the nitrogen was first injected at about 0.5 PV. in the two experiments.
However, following nitrogen injection, a remarkable difference was
observed between the two experiments. At the end of the two experi-
11,278.7 mL (3.23 PV), whereas their cSOR were calculated to be 9.45
ments, the ultimate oil recoveries of SAGD and NA-SAGD were 42.7%
and 5.79 respectively.
and 56.1%, respectively. The cumulative injected volumes of steam for
Based on the comparative SAGD and NA-SAGD results, the ultimate
the SAGD and NA-SAGD experiments were 14,291.2 mL (4.06 PV) and

Fig. 3. Comparison of the residual oil distribution after the two experiments. (A) Residual oil distribution after the SAGD experiment. (B) Residual oil distribution after the NA-SAGD
experiment.

329
Z. Yuan et al.
oil recovery of the NA-SAGD process was 13.4% higher than that of the
SAGD process, whereas the injected volumes of steam for the NA-SAGD
process was 0.83 PV less. In addition, the decrease of water consump-
tion was significant. More oil was produced at a lower steam injection
amount, which indicated an increase in the heat utilization efficiency.
The amount of energy consumption and carbon dioxide emissions to
generate high temperature steam was effectively reduced which aids in
the environmental goals of this study. At the end part of the NA-SAGD
experiment, as the area of the steam chamber increased significantly
and all the saturated oil was heated, the oil production rate increased
significantly. The produced oil that remained in the pipeline, which
resulted in the same phenomenon.
From Fig. 5, the cSOR was relatively high at first due to the low heat
utilization efficiency during the communication period. After the steam
chamber was formed, the cSOR decreased following an increase in the oil
production. During the horizontal expansion period of the steam cham-
ber, the cSOR decreased due to significant heat loss. The cSOR curve of
NA-SAGD was much smaller as the effect of the injected nitrogen. Every
time nitrogen was plug-injected, the cumulative steam oil ratio decreased
following an increase in the oil production or a decrease in the steam
injection.
Combined with the analysis of the temperature profiles and the re-
sidual oil saturation distribution, the production mechanism of NA-SAGD
Fig. 6. Dynamic changes of the temperature profile during the two injection strategies. (A) Temperature profile during SAGD process. (B) Temperature profile during NA-SAGD process.
330
Journal of Petroleum Science and Engineering 162 (2018) 325–332
indicates that following nitrogen injection, a transition region is formed
on the upper part of the steam chamber as the nitrogen accumulated. This
region effectively reduces heat loss to the outside of the model and also
maintains the pressure of the steam chamber, which slows down the
injection rate of steam. Steam chamber effectively expands along the
horizontal direction and increases the swept area of steam, thereby
allowing more heavy oil to be heated and produced. In addition, a
cemented area was formed in the center of the model after the first NA-
SAGD experiment. However, this condition was not observed in the
second additional experiment possibly because the area may be an
alkaline substance that was produced by nitrogen at a high temperature.
The changes in the oil composition, the formation of the cement area, and
the effects of the other operation parameters on oil production requires
further investigations.
4. Numerical simulation
To compare and verify the results obtained in the two experiments, a
conceptual homogeneous model of a single well pair, wherein a CMG-
STARS was used as the simulator, was employed to investigate the
SAGD and NA-SAGD processes.
Z. Yuan et al. Journal of Petroleum Science and Engineering 162 (2018) 325–332

4.1. Model description

The single well pair model was homogeneous with regards to its
porosity and permeability. The model had fundamental grid dimensions
of Nx ¼ 20, Ny ¼ 50, and Nz ¼ 46. The grid size is 25 m in the X-direction,
2 m in the Y-direction, and 0.5 m in the Z-direction. Horizontal injection
well was drilled at the bottom of the reservoir and was about 5 m above
the production well. The key reservoir properties used in the simulation
are presented in Table 1. Heavy-oil viscosity versus temperature is given
in Table 2.

4.2. Simulation results

Fig. 6 shows the simulation results of the temperature profile at

different stages following the two injection strategies.
From Fig. 6, after 2 years, the heating area of the SAGD process was
slender in the center of the reservoir, and only the upper part of the steam
chamber extended along the horizontal direction. However, the steam
Fig. 7. Cumulative oil recovery after the two injection strategies.
chamber of the NA-SAGD process was obviously larger for the same time
points. In addition, the whole chamber exhibited an apparent lateral
expansion. The temperature distribution characteristics of the simulation
results were in good agreement with the experimental results. After 6
years of steam injection, the SAGD steam chamber exhibited a “V” shape
profile, whereas the NA-SAGD steam chamber exhibited a “U” shape
profile. The swept area of steam was larger, and more oil was heated and
produced.
Fig. 7 displays the cumulative oil recovery after the two injection
strategies. According to Fig. 7, the cumulative oil recovery after two in-
jection strategies was 39.4% and 47.5%, respectively, after 9 years.
However, following the subsequent steam injection, the production gap
gradually decreased as the model was finite. These findings indicate the
beneficial effect of the nitrogen injection on the oil recovery. A reduction
in heat loss was observed in the upper part of the reservoir following
nitrogen accumulation, which effectively developed the steam chamber.
An increase in the heat transfer enhanced the growth rate of the NA-
SAGD steam chamber and accelerated the rate of oil drainage, thereby
indicating that the cumulative oil production was significant higher
following the nitrogen treatment. Fig. 8. Effect of the injected nitrogen mole fraction on the oil rate and cumulative oil
recovery.

4.3. Sensitivity analyses

Sensitivity analyses were performed to evaluate the effects of the

different parameters on the NA-SAGD process performance. The results
are presented as follows.

4.3.1. Effect of nitrogen mole fraction

In order to study the effect of the nitrogen mole fraction on the NA-
SAGD performance, three cases specifically at a nitrogen mole fraction
of 0.1, 0.2 (base case), and 0.3 were considered. Fig. 8 shows the effect of
the injected nitrogen mole fraction on the oil rate and cumulative oil
recovery. Following the three injection strategies, the cumulative oil
recovery was measured to be 34.1%, 47.5%, and 46.4%, respectively. An
increase in the mole fraction of nitrogen from 0.2 to 0.3 resulted in a
decrease in the cumulative oil recovery. The accumulation of too much
nitrogen in the upper part of the steam chamber hindered the subsequent
injection of steam, thereby resulting in a decrease in the oil production
rate at a later stage and ultimately significantly lower oil recovery rate. In
general, the optimized steam injection rate must be between 0.1 and 0.2. Fig. 9. Effect of the steam injection rate on the oil rate and cumulative oil recovery.

4.3.2. Effect of steam injection rate
Three steam injection rate cases, specifically at 100, 400 (base case),
and 1000 m3/d, were selected to examine the effects of the steam in-
jection rate on the NA-SAGD performance. Fig. 9 shows the effect of the
steam injection rate on the oil rate and cumulative oil recovery.
Following the three injection strategies, the cumulative oil recovery was
measured to be 44.1%, 47.5%, and 64.8%, respectively. An increase in
the steam injection rate generated an increase in the oil recovery
following an increase in the oil production rate during the early stages. In
later stages of production, a higher injection rate resulted in steam
rapidly reaching the top of the reservoir. Therefore, the 1000 m3/d oil
production rate case exhibited an apparent decrease, while the two other
production rates were stable. Moreover, although the increase in the
injection rate improved the oil production, the CSOR increased as well.

331
Z. Yuan et al.
Therefore, the optimized steam injection rate was selected based on the
oil and steam generation prices. The 1000 m3/d case produced more oil
but had a higher steam generation price that restricted the application of
a higher steam injection rate.
5. Conclusions
Based on the present experimental study and numerical simulation,
we can draw the following conclusions.
1) The accumulation of nitrogen maintains the pressure of the steam
chamber, and consequently slows down the upwards movement of
steam and reduces the heat transfer of steam to the cover of the
reservoir. The steam chamber effectively expands along the hori-
zontal direction. The steam exhibits a significantly larger sweep area
and generates a reduction in the residual oil saturation.
2) According to the experimental results, the NA-SAGD process en-
hances the ultimate oil recovery accompanied by lower injected
steam volume requirements as compared to the SAGD process. In
addition, the process exhibits a reduction in the cost consumption,
effectively improves the economic benefit, and significantly improves
the heat utilization efficiency.
3) A numerical simulation is generated to compare and verify the results
obtained in the two experiments. The temperature distribution
characteristics and oil recovery of the simulation results are well in
agreement with the experimental results. An optimized steam injec-
tion rate between 0.1 and 0.2 is selected. The oil production rate
increases following an increase in the steam injection rate, though
this decreases rapidly in the later stages. In addition, a higher steam
generation price may restrict the application of a higher steam in-
jection rate.
Acknowledgements
This work was financially supported by the National Natural Science
Foundation of China (No. 51774227), Fundamental Research Funds for
the Central Universities (2-9-2017-275) and the National Oil & Gas Key
Special Project of China for 2016ZX05016-006 the contributions of
which and the permissions for publication are gratefully acknowledged.
References
Al-Murayri, M.T., Harding, T.G., Maini, B.B., 2011a. Solubility of methane, nitrogen, and
carbon dioxide in bitumen and water for SAGD modeling. J. Can. Petrol. Technol. 50
(7), 34–45.
Al-Murayri, M.T., Harding, T.G., Maini, B.B., 2011b. Impact of noncondensable gas on
performance of steam-assisted gravity drainage. J. Can. Petrol. Technol. 50 (7),
46–54.
Al-Murayri, M.T., Maini, B.B., Harding, T.g., Oskouei, J., 2016. Multicomponent solvent
co-injection with steam in heavy and extra-heavy oil reservoirs. Energy Fuel. 30 (4),
2604–2616.
Ardali, M., Barrufet, M., Mamora, D.D., 2012. Effect of non-condensable gas on solvent-
aided SAGD processes. In: SPE 146996, Present at SPE Heavy Oil Conference Canada.
Calgary, Alberta, Canada, 12–14 June.
Bagci, A.S., Gumrah, F., 2004. Effect of CO2 and CH4 addition to steam on recovery of
west Kozluca heavy oil. In: SPE 86953, Present at SPE International Thermal
Operations and Heavy Oil Symposium and Western Regional Meeting. Bakersfield,
California, US, 16–18 March.
Butler, R.M., Stephens, D.J., 1981. The gravity drainage of steam heated to parallel
horizontal wells. J. Can. Petrol. Technol. 20 (2), 90–96.
Butler, R.M., 1994. Steam-assisted gravity drainage: concept, development, performance
and future. J. Can. Petrol. Technol. 33 (2), 44–50.
Butler, R.M., Jiang, Q., Yee, C.T., 2000. Steam and gas push (SAGP)-3: recent theoretical
developments and laboratory results. J. Can. Petrol. Technol. 39 (08), 11–20.
Du, Y., Wang, Y., Jiang, P., Ge, J.J., Zhang, G.C., 2013. Mechanism and feasibility study of
nitrogen assisted cyclic steam stimulation for Ultra-heavy oil reservoir. In: SPE
165212, Present at SPE Enhanced Oil Recovery Conference. Kuala Lumpur, Malaysia,
2–4 July.
332
Journal of Petroleum Science and Engineering 162 (2018) 325–332
Gao, Y., Liu, S., 2008. Improving oil recovery by adding N2 in SAGD process for super-
heavy crude reservoir with top-water. In: SPE 114590, Present at SPE Russian Oil and
Gas Technical Conference and Exhibition. Moscow, Russia, 28–30 October.
Gonzale, G., Sousa, M.A., Lucas, E.F., 2006. Asphaltenes precipitation from crude oil and
hydrocarbon media. Energy Fuel. 20 (6), 25544–2551.
Gittins, S., Gupta, S., Zaman, M., 2011. Simulation of noncondensable gases in SAGD
steam chanbers. In: SPE 149503, Present at Canadian Unconventional Resources
Conference. Calgary, Alberta, Canada, 15–17 November.
Guo, E., Jiang, Y., Gao, Y., Wang, H., Yu, P., 2015. Discussion on the first N2–SAGD pilot
test in China. In: SPE 174655, Present at SPE Asia Pacific Enhanced Oil Recovery
Conference. Kuala Lumpur, Malaysia, 11–13 August.
Hascakir, B., 2016. How to select the right solvent for solvent-aided steam injection
processes. J. Petrol. Sci. Eng. 146, 746–751.
Hu, L., Li, H.A., Babadagli, T., Ahmadloo, M., 2017. Experimental investigation of
combined electromagnetic heating and solvent-assisted gravity drainage for heavy oil
recovery. J. Petrol. Sci. Eng. 154, 589–601.
Ito, Y., Ichikawa, M., Hirata, T., 2001. The effect of gas injection on oil recovery during
SAGD projects. J. Can. Petrol. Technol. 40 (1), 38–43.
Jha, R.K., Kumar, M., Benson, I., Hanzlik, E., 2013. New insights into steam/solvent-
coinjection-process mechanism. SPE J. 18 (5), 867–877.
Ji, D.Q., Dong, M.Z., Chen, Z.X., 2015. Analysis of steam-solvent-bitumn phase behavior
and solvent mass transfer for improving the performance of the ES-SAGD process.
J. Petrol. Sci. Eng. 133, 826–837.
Jiang, Q., Bulter, R., Yee, C.T., 2000. The steam and gas push (SAGP)-2: mechanism
analysis and physical model testing. J. Can. Petrol. Technol. 39 (4), 52–61.
Kazeem, A.L., 2014. Economics of steam-assisted gravity drainage for the Nigerian
bitumen deposit. J. Petrol. Sci. Eng. 116, 28–35.
Leyva-Gomez, H., Babadagli, T., 2013. Numerical simulation of heavy-oil/bitumen
recovery by solvent injection at elevated temperatures. J. Petrol. Sci. Eng. 110,
199–209.
Liu, P., Li, W., Shen, D., 2015. Experimental study and pilot test of urea-and urea-and-
foam-assisted steam flooding in heavy oil reservoirs. J. Petrol. Sci. Eng. 135,
291–298.
Liu, P., Wu, Y., Li, X., 2013. Experimental study on the stability of the foamy oil in
developing heavy oil reservoirs. Fuel 111, 12–19.
Liu, Y., Xi, C., Liu, S., Liu, C., 2012. Impact of non-condensable gas on SAGD performance.
In: SPE 150539, Present at SPE Heavy Oil Conference Canada. Calgary, Alberta,
Cananda, 12–14 June.
Lu, C., Liu, H., Liu, Q., Lu, K., Wang, L., 2014. Research on the effect of non-condensable
gas and viscosity reducer for better SAGD performance. In: SPE 170026, Present at
SPE Heavy Oil Conference Canada. Calgary, Alberta, Canada, 10–12 June.
Mukhametshina, A., Kar, T., Hascakir, B., 2016. Asphaltene precipitation during bitumen
extraction with expanding-solvent steam-assisted gravity drainage: effects on pore-
scale displacement. SPE J. 21 (2), 380–392.
Nasr, T.N., Beaulieu, G., Golbeck, H., Heck, G., 2003. Novel expanding solvent-SAGD
process “ES-SAGD”. J. Can. Petrol. Technol. 42 (1), 13–16.
Nourozieh, H., Ranjbar, E., Kumar, A., 2015. Modeling of non-condensable gas injection
in SAGD process-important mechanisms and their impact on field scale simulation
models. In: SPE 174494, Present at SPE Canada Heavy Oil Technical Conference.
Calgary, Alberta, Canada, 9–11 June.
Rahnema, H., Barrufet, M., Mamora, D., 2011. Experimental study of air injection in
SAGD chamber. In: SPE 149195, Present at Canadian Unconventional Resources
Conference. Calgary, Alberta, Canada, 15–17 November.
Rios, V.S., Laboissiere, P., Trevisan, O.V., 2010. Economic evaluation of steam and
nitrogen injection on SAGD process. In: SPE 139319, Present at SPE Latin American
and Caribbean Petroleum Engineering. Lima, Peru, 1–3 December.
Sharma, J., Moore, R.G., Mehta, R., 2012. Effect of methane co-injection in SAGD-
analytical and simulation study. SPE J. 17 (3), 687–704.
Stachowiak, C., Viguie, J.R., Grolier, J.P.E., Rogalski, M., 2005. Effect of n-alkanes on
asphaltene structuring in petroleum oils. Langmuir 21 (11), 4824–4829.
Souraki, S., Ashrafi, M., Torsaete, O., 2013. A comparative field-scale simulation study on
feasibility of SAGD and ES-SAGD processes in naturally fractured bitumen reservoirs.
Energy Environ. Res. 3 (1), 49–63.
Wang, C., Zhong, L.G., Li, J.P., Chen, G., Zang, Y., Wei, F., 2015. Laboratory study on
steam and nitrogen co-injection for mid-deep heavy oil reservoirs. In: SPE 175266,
Present at SPE Kuwait Oil and Gas Show and Conference. Mishref, Kuwait, 11–14
October.
Yang, L., Xi, C.F., 2012. Impact of Non-condensable gas on SAGD performance. In: SPE
150539, Present at SPE Heavy Oil Conference Canada. Calgary, Alberta, Cannda,
12–14 June.
Yuan, J.Y., Chen, J.X., Pierce, G., Wiwchar, B., Golbeck, H., Wang, X., Beaulieu, G.,
Cameron, S., 2011. Noncondensable gas distribution in SAGD chambers. SPE J. 50
(3), 11–20.
Yuan, Z., Liu, P., Zhang, S., Jiao, Y., Li, X., 2017. Experimental study and numerical
simulation of a solvent-assisted start-up for SAGD wells in heavy oil reservoirs.
J. Petrol. Sci. Eng. 154, 521–527.
Zheng, S., Li, H., Yang, D., 2013. Pressure maintenance and improving oil recovery with
immiscible CO2 injection in thin heavy oil reservoirs. J. Petrol. Sci. Eng. 112,
139–152.
Zheng, S., Li, H., Sun, H., Yang, D., 2016. Determination of diffusion coefficient for
Alkane solvent–CO2 mixtures in heavy oil with consideration of swelling effect. Ind.
Eng. Chem. Res. 55 (6), 1533–1549.