Journal of Natural Gas Science and Engineering 48 (2017) 36e45

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering

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

Pore-scale investigation of microscopic remaining oil variation

characteristics in water-wet sandstone using CT scanning
Junjian Li a, Hanqiao Jiang a, Chuan Wang a, Yuyun Zhao a, Yajun Gao a, Yanli Pei a,
Chenchen Wang b, c, *, Hu Dong c
a
State Key Laboratory of Petroleum Resources and Engineering, China University of Petroleum(Beijing), Beijing 102249, China
b
Unconventional Oil and Gas Collaborative Innovation Center, Yangtze University, Hubei, 430100, China
c
iRock Technologies Co., LTD, Beijing, 100094, China

a r t i c l e i n f o a b s t r a c t

Article history: With the aid of CT scanning and image processing techniques, water displacing oil experiments were
Received 6 October 2016 performed to analyze six natural water-wet sandstone cores of different permeability and porosity
Received in revised form visually and quantitatively. Microscopic remaining oil was categorized on the basis of quantitative
4 April 2017
characterization parameters, such as shape factor, contact ratio, Euler number etc. The remaining oil
Accepted 5 April 2017
Available online 7 April 2017
characteristics with different flow types were presented in different water displacement stages. Exper-
imental results show that the remaining oil can be divided into five categories. They are, respectively
membranous flow, droplet flow, columnar flow, multi-porous flow and clustered flow from hard-to-
Keywords:
Water displacement
produce to easy-to-produce. The relative permeability of oil phase is represented by the macro-
CT average relative permeability of all the above five flow regimes. Among them, clustered flow pos-
Shape factor sesses strong producing capacity and high relative permeability, and the other four flow regimes have
Remaining oil weaker producing capacity and lower relative permeability. Variation of number and average volume of
Relative permeability five types remaining oil and several measurements of the ganglion size distribution have been performed
in the literature. The nonlinear knee point on the relative permeability ratio curve is intrinsically caused
by the decreasing of volume and quantity fraction of clustered flow when water saturation increases. This
paper has studied the flowing law of microscopic remaining oil, explained the intrinsic mechanism for
the appearance of an inflection point on the relative permeability ratio curve and microscopic sweep
phenomena, and also presented a new effective way of upscaling to a certain extent.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction
Accurate description of the remaining oil distribution under-
ground has remained a challenging task in the oilfield develop-
ment. Understanding the pore structure morphology, granular
property, fluid distribution and two-phase fluid dynamic migration
is the basis for remaining oil study also the foundation of enhancing
oil recovery. A large number of testing methods have been pro-
posed for the study of multiphase flow movement characteristics,
e.g., numerical simulation method (Dalla et al., 2002; Bryant and
Anna, 2004; Ramstad et al., 2012; Raeini et al., 2015), two-
dimensional micromodels (Sahloul et al., 2002; Chen et al., 2007;
* Corresponding author. Unconventional Oil and Gas Collaborative Innovation
Center, Yangtze University, Hubei, 430100, China.
E-mail address: wcc1220@163.com (C. Wang).
Zhou et al., 2014) and network model (Blunt and King, 1991;
Valvatne and Blunt, 2004; Løvoll et al., 2005). They can be widely
used to predict various flow properties and greatly promoted the
research progress of two-phase flow. However, the limitation of
these methods is that the accurate representations of pore space in
the porous medium, rigorous validation methods and relevant
experimental data may not be fully obtained. The application of
computed tomography (CT) technique in petroleum field eventu-
ally makes it possible to have an accurate recognition of 3D pore
structure and fluid distribution. Through the extraction of pore
structure and phase separation of flow fluid, characteristic pa-
rameters (e.g. porosity and permeability), flow volume and surface
area can be calculated to promote the study of the remaining oil
dynamic characteristics.
CT scanning was firstly applied to analyze the relative perme-
ability of the oil phase and water phase when gel was injected
(Seright et al., 2006). The residual phase saturation, capillary force,

http://dx.doi.org/10.1016/j.jngse.2017.04.003
1875-5100/© 2017 Elsevier B.V. All rights reserved.
 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45
and snap-off have been widely studied in recent years (Dann et al.,
2011; Kumar et al., 2011; Iglauer et al., 2012,2013,Iglauer and
Wülling, 2016a; Andrew et al., 2014a,2015). Wettability was
regarded as an essential study factor which could influence the
residual phase distribution after two-phase displacement in porous
media. The experimental data indicated that the droplets quantity
in oil-wet medium was three times more than that of water-wet
medium at similar saturations (Landry et al., 2011), but the resid-
ual oil saturation is lower for the oil-wet core than for an analogue
water-wet core (Iglauer et al., 2012). Meanwhile, the sweep effi-
ciency of dielectric materials with strong wettability is higher than
that of dielectric materials with weak wettability, and the droplet
size distribution under two wetting conditions is different even
under the same oil saturation (Celauro et al., 2013). Wettability has
also been studied in the three-phase flow (Iglauer et al., 2016b).
These results indicate that the residual oil saturation and droplet
size distribution are closely related to wettability.
A lot of core experiments were conducted to study the size and
distribution of trapped clusters (ganglia) under reservoir condi-
tions with the aid of CT technique. In view of medium wettability,
gravity differentiation, fluid viscosity and capillary force, the spatial
distribution of residual oil droplet size, signal frequency, cumula-
tive frequency and residual oil saturation along core longitudinal
profile in water flooding process were studied in a glass bead pack
(Karpyn et al., 2010). After being fully displaced by supercritical
carbon dioxide, the size distribution of residual oil ganglia obeys
power law distributions with exponent in sandstone (Iglauer et al.,
2010, 2012). The result show that the power-law exponent for the
oil-wet core is higher than water-wet system, which indicating
fewer large clusters in the oil-wet case. Similarly, the distribution of
residual gas clusters present similar characteristics (Iglauer et al.,
2011; Andrew et al., 2013,2014b). It indicated that the rock chem-
ical properties have little influence on the size of the residual gas
clusters, and most residual phase could be driven again when the
capillary number was about 10
5 (Andrew et al., 2014a). On the
other hand, the surface area of each cluster against its volume and
curvature distribution were measured in three-phase flow (Iglauer
et al., 2013,2016b; Carroll et al., 2015), which fully demonstrated
the well-known complexity of multiphase flow in rock.
At present, CT technology is focused on two-phase flow research
including saturation change of the residual phase, measurement of
two-phase interface curvature and surface area, and size distribu-
tion characteristics of residual phase clusters. As discussed above
we can easily find that residual phase clusters size and occurrence
state in the core keeps changing during displacement process, but
only a few researchers have studied the change law and the type
classification of the residual phase clusters (Iglauer et al., 2012).
Therefore, in this paper several water flooding experiments on
water-wet sandstone were conducted with the aid of the micro
displacement experiment system, scan images of the core model
were obtained based on CT scanning technology for different water
flooding time, and three-dimensional distribution information of
the remaining oil were extracted by utilizing image processing and
three-dimensional reconstruction technology. Through the estab-
lishment of remaining oil micro occurrence state and classification
method at pore scale, the micro distribution and change law of
different types of remaining oil in cores at different displacement
time have been quantitatively characterized, which can shed light
on further improvement of oil recovery during water flooding
process.
2. Experiment and methods
Three-dimensional reconstruction technology based on CT
scanning is an important means for laboratory study of microscopic
37
remaining oil. In order to study the flow regime and change rule of
microscopic remaining oil, natural cores with different porosity and
permeability were drilled for micro water flooding experiments.
2.1. Experimental apparatus
The experimental apparatus is shown in Fig. 1. The rock and
fluids are imaged using a Versa nanoVoxel-3502E X-ray Micro-
scope. High resolution microscopy detector is used with pixel of
2048  2048 and resolution of 2 mm. Fig. 2 shows the structure of
the core holder. The high temperature and high pressure formation
condition can be simulated by quantitative adjustment of the
temperature sensor and pressure gauge. Due to the small core size,
produced liquid is too little thus capillary metering device with
precision of 0.002 mL is used to measure the produced oil and
water at the outlet. Six pieces of natural cores from Xinjiang oilfield
(located in the Tarim Basin.) are selected for the experiment. The
cores are well sorted and relatively homogeneous, which are the
Carboniferous marine detrital fine sandstone. The porosity and
permeability are measured in the lab by using conventional cores
(one inch in diameter); the wettability is measured by spontaneous
imbibition under room temperature and pressure. The independent
experiments were conducted on separate larger cores
(diameter ¼ 25.4 mm, length ¼ 80 mm). Before the experiment, the
core was vacuumized and saturated with brine (10 wt% sodium
iodide), pressurized to 20 MPa to age 7 days. Firstly, saturate the
core with oil and put it into the water-absorbing device for 10 days
to measure the amount of water adsorption, and then saturate the
core with water and put it into the oil-absorbing device or 10 days
to measure the amount of oil adsorption. The cores used in the CT
scanning were drilled from the larger cores. The petrophysical and
wettability properties of cores are displayed in Table 1.
2.2. Flow strategy
In order to eliminate the influence of fluid properties on
remaining oil classification and make the experimental results
more universal and practical, experiments are designed with
respect to same injection rate and fluid viscosity on core samples in
Table 1. The properties of the fluids used in this study are listed in
Table 2. (1) Experimental oil: white mineral oil and kerosene are
mixed by certain proportion with viscosity of 6 mPa S, and oil-brine
interfacial tension is 0.032N/m; (2) Experimental brine: simulated
formation water, 10% of NaI is added to distinguish oil and water;
(3) Experimental Apparatus: ISCO 100D Hastalloy constant pres-
sure and constant speed pump and micro displacement device, as is
indicated in Fig. 1.
The displacement experiment process is as follows:
(1) Preparation: wash oil and dry the core sample under the
condition of 70  C for 12 h; weigh dry weight; record the core
diameter;
(2) Saturate water: vacuumize the core for 5 h; saturate the core
with brine for more than 24 h until no brine imbibition
happens; weigh the core weight after fully water saturated
and calculate water saturation; then put the cores into the
brine to age for 7 days.
(3) Put the sample into the core holder, connect the whole set of
experimental apparatus, set the core temperature and pres-
sure at 55  C and 15 MPa respectively, let the system stand
for 12 h in order to reach the chemical equilibrium between
rock particles and pore fluid;
(4) Saturate oil: saturate the core with oil at constant injection
rate (0.01 ml/min, this corresponds to a capillary number
Ncap ¼ vm=s ¼ 1:1  10
7 ) and confining pressure of 15 MPa
38 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45

V1 V2 V8 V10 V11 V13

measurement
V9 V12

flow
V3 V5

Flow Cell
brine oil

V7
brine
V4 V6
brine Pump1 Pump2 brine Pump3
X-Ray Enclosure

Fig. 1. Flow chart of experimental apparatus.

Fig. 2. Carbon fibre core holder.

Table 1
Experimental sample parameters.

No. Porosity Permeability Diameter Length Imbibition Wettability Testing

(%) (mD) (mm) (mm)

Experiment Data Experiment Data Water-Imbibition Oil Discharge (mL) Oil-Imbibition Water Discharge (mL) Wettability

1 16.8 47.2 7.50 50 0.028 0.007 Water-Wet

2 25.3 89.4 7.80 50 0.027 0.016 Water-Wet
3 12.7 86.2 7.64 50 0.017 0.013 Water-Wet
4 18.4 685.2 7.66 50 0.021 0.009 Water-Wet
5 20.2 87.0 8.05 50 0.012 0.004 Water-Wet
6 18.2 289.7 7.66 50 0.026 0.016 Water-Wet

Table 2
Fluid composition and properties.

Fluid Composition Weight proportion Density@293K Viscosity@293K Refractive

(kg/m3) (mPa.s) index

Brine Water(10 wt%NaI) e 1025 1.07 1.333

oil White mineral oil 50% 830 15.2 1.486
Kerosene 50% 800 1.87 1.447

until no water produces; record water production and
calculate water saturation; conduct CT scanning on the
sample for the first time;
(5) Water drive: connect the capillary metering device at the
outlet; conduct water drive at same constant injection rate
(Ncap ¼ vm=s ¼ 1:1  10
7 ) and pressure (15 MPa); record
the injection pressure, water and oil production according to
the page moving distance; calculate the displacement time
based on the displacement speed; scan the sample secondly
after injecting 1 PV (pore volumes) of brine; then conduct
water drive at the same conditions by injecting more than
50 PV brine followed by first displacement and conduct CT
scanning thirdly.
(6) Repeat step (1)-step (5) for different core samples. In
particular, sample 1 and sample 2 were only scanned twice
by CT, the scanning is not performed by injecting 1 PV (pore
volumes) of brine.
Before each CT scanning, the pump was turned off and allowed
to stand for 30 min, so that the fluid was no longer moving to
ensure the accuracy of the scanned image. It is important that the
position of each scan is the same for every sample. The X-ray source
is aligned with the midpoint of the core, that is, from the bottom of
the core 25 mm. The time it takes to perform a scan is about 35
minutes.
2.3. Image processing
In recent years, CT three-dimensional imaging technology has
been widely applied in medicine, materials, water, life science, etc.
With the continuous technique development, the resolution of the
CT image has reached the level of micro/nano, image quality has
been gradually improved, and computer post-processing and 3D
reconstruction technology are also improved constantly (Jiang
et al., 2007). The main processing technique used in this article is
shown later and the details will not be explained here. CT images
obtained from experiments are grey value image in 64-bit tiff
 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45 39

Fig. 3. Image processing procedure.

format. During both of the early vacuum saturation and late
displacement processes, 10% of NaI has been added to the simulated
formation water. In the scan images, the grey value of water is
greater, the grey value of oil is lower, and the grey value of rock
particles is in the middle. Thus it is possible to distinguish oil, water
and rock skeleton from CT scanning images, as shown in Fig. 3. The
images were filtered using a non-local means edge preserving filter
(Buades et al., 2005;2008). The phases in the images were
segmented using Otsu's algorithm (Ohtsu, 1979).
Based on the above image processing, three-dimensional
reconstruction is conducted to obtain the 3D data volume and the
3D image of the scanning core. Also the 3D separation data volume
of the rock, oil and water will be provided. Scanning results of
sample No.3 of oil saturated state, in the middle of the displace-
ment and at the end of the displacement are shown in Fig. 4. By
comparing the visual images of the above three states, we can get
some qualitative results: the oil phase continues to decline and is
gradually broken up while the water phase continues to increase
and gradually becomes a whole.
The pore network structure of the rock sample can be obtained
by dealing with the 3D data volume shown in Fig.4. To characterize
remaining oil quantitatively, we calculate the contact relationships

Fig. 4. Separation of rock, oil and water (Left: oil(red), Right: water(blue)). (For interpretation of the references to colour in this figure legend, the reader is referred to the web
version of this article.)
40 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45

Fig. 5. Extraction of pore structure ((a) 3D Core Reconstruction; oil is black, brine light grey and rock dark grey, (b) Pore Structure Extraction; pore is yellow and rock dark grey, (c)
Pore Axis Extraction; large pore throat is red, small pore throat is purple). (For interpretation of the references to colour in this figure legend, the reader is referred to the web
version of this article.)

between single pieces of remaining oil, rock skeleton, and the
number of pore throats occupied by a single piece of remaining oil.
In this paper, a representative network of pores and throats was
extracted by a maximum ball algorithm (Dong and Blunt, 2009).
The extraction results of pore structure and pore axis of sample
No.5 are presented in Fig. 5.
3. Results and discussion
3.1. Quantitative characterization and classification of microscopic
remaining oil
Through the connectivity analysis of remaining oil, a discon-
nected oil ganglion can be extracted and characterized
quantitatively according to the shape factor, Euler number and the
contact ratio. By comparing the experimental results of six core
samples, we find that five types of microscopic remaining oil
existed during water flooding process. The category results are
shown in Table 3.
The calculation formula and physical meaning of the shape
factor, Euler number and the contact ratio are presented in Table 4.
Remaining oil can be classified from hard-to-produce to easy-to-
produce according to the five flow regimes in Table 3. It can be
presented as membranous flow, droplet flow, columnar flow, multi-
porous flow and clustered flow. In general, the exploitation method
for membranous flow, droplet flow and columnar flow increases
the microscopic displacement efficiency, such as surfactant flood-
ing; the exploitation method for multi-porous flow and clustered

Table 3
Quantitative characterization of remaining oil.
 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45 41

Table 4
Calculation formula and physical meaning of the characterization parameters.

Characterization Calculation Formula Physical Meaning

Parameter

Shape Factor Shape Factor ¼ Surface Area3 The proximity degree of the shape of a single piece of remaining oil to spheres. The value of the
36pVolume
2
sphere is 1, and the smaller the closer to a sphere, the larger the more irregular of the shape.
Euler Number Euler Number ¼ 1
conectivity þ enclosed cavities Reflect the pore number of single piece of remaining oil number. The smaller the value, the more
the pore holes.
Contact Ratio Contact Ratio ¼ Contact Area Between Oil and Rock Provide the contact relationship between single piece of remaining oil and the hole wall. The
Surface Area of Oil
smaller the contact ratio, the less the remaining oil adhering to pore surface.

Fig. 6. Remaining oil classification result.

30% 100% the middle of water flooding. The volume proportion statistical
water saturation 37.40%
volume fraction discontinuous stream

histogram of all flow regimes is illustrated in Fig. 7. The volume

25% fraction of clustered flow is higher than that of the other flow
volume fraction of clustered flow

75.37% 80%
19.55% regimes.
20%
60%
15% 3.2. Microscopic remaining oil variation characteristics
40%
10% The experimental results are shown in Table 5.
20%
Based on statistical analysis of the experimental results, volume
5%
2.16% 1.42% 1.50% fractions of the different disconnected oil categories are plotted as
functions of water saturation in Fig. 8.
0% 0%
membranous droplet columnar multi-porous clustered Fig. 9 have shown the total number of five types of remaining oil
water saturation and their average volume as functions of water saturation.
The results show that: (1) with the increase of water saturation,
Fig. 7. Volume ratio of different remaining oil. oil phase is gradually broken up and the degree of discontinuity
increases, thus leading to increased quantity, decreased average;
(2) The main cause of discontinuity aggravation of oil phase is the
remaining oil type transformation from clustered flow to multi-
flow is to increase microscopic sweep efficiency, such as polymer
porous flow; (3) as water saturation increases, the volume frac-
flooding or polymer microspheres displacement.
tion of discontinuous flow (membranous flow, droplet flow,
Fig. 6 shows the visual classification results of sample No.3 in
columnar flow and multi-porous flow) gradually increases while

Table 5
Experimental results.

No. Porosity (%) Oil Saturation (%) Recovery (%)

Experiment Data CT Scanning Relative Error Oil Saturated State 1 PV of Water Drive End of Water Drive
(%)

1 16.8 17.87 6.0 67.13 59.64 39.65 40.94

2 25.3 23.42 8.0 76.55 34.51 54.92
3 12.7 15.52 18.2 77.27 62.60 33.15 57.10
4 18.4 16.03 14.8 66.51 57.87 34.53 48.08
5 20.2 19.02 6.2 63.66 35.79 43.78
6 18.2 18.27 0.3 76.23 63.78 35.61 53.29
42 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45

Fig. 8. Volume ratio variation of different remaining oil types along with the change of water saturation.
 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45 43

600 1000000

N-membranous
500

V \ average volume [voxel]

N-droplet 100000
400 N-columnar
N-multi-porous

300 N-clustred 10000

N / number V-membranous
200 V-droplet
V-columnar 1000
100 V-multi-porous
V-clustered
0 100
20% 30% 40% 50% 60% 70% 80%
water saturation

Fig. 9. Variation of number and average volume of five types of remaining oil clusters (sample NO.3). (The dashed line represents the cluster number of the remaining oil; The
solid line represents average volume of five types of remaining oil clusters).

shown in Fig. 10, the relative permeability ratio curves of Sample

100 NO.3 is obtained by lab experiment, which show a knee point when
the water saturation is about 60%. Simultaneously it can be found in
10
Fig. 9 that both the number and average volume of clustered flow
knee point
1 decrease with the increasing of the number and average volume of
Kro /Krw

the other four types of residual oil. There is an obvious increase in

0.1 the number of droplet flow and porous flow. The authors deduce
that the five kinds of oil phase flow regimes have different
0.01
permeability. As the continuous phase, clustered flow has relatively
0.001 high relative permeability and is easy to produce. For the other four
20% 40% 60% 80% 100% discontinuous phases, they have relatively low relative perme-
water saturation ability and are difficult to produce. The relative permeability of the
oil phase is the macro relative permeability average of the five
Fig. 10. Relative permeability ratio curve (sample NO.3).
kinds of flow regimes. When the water saturation is close to 50%,
the quantity of clustered flow decreased significantly, which lower
down the relative mobility of oil phase and lead to a quick decrease
1 of the oil-water permeability ratio.
In the paper, we also perform several measurements of the
ganglion size distribution in different water drive stages. In some
0.1
early studies, researchers have demonstrated that ganglia of re-
sidual phase can range in size from a single pore to multiple-pore
0.01
structures (Kumar et al., 2011; Karpyn et al., 2010). Fig. 11 shows
S(s); n(s)

n(s)-1 residual oil cluster size distributions of Sample NO.3 under three
S(s)-1 residual oil saturation(Sor). As we can see, the size distribution of
0.001 n(s)-2 the three scanning all appears to peak at some typical ganglion size
S(s)-2 and then decay rapidly. The three distribution curves show that: In
n(s)-3
the saturated oil state (Sor ¼ 77.27%) and first water drive(1 PV)
S(s)-3
0.0001 state (Sor ¼ 62.60%), the size distribution frequency of the small
1 10 100 1000 10000 100000 clusters is lower, so that the contribution of the oil saturation
S mainly comes from the ganglions with the size larger than 100
Fig. 11. Residual oil cluster size distributions of Sample NO.3 for the water flood
voxel; With the injection of water (Sor ¼ 33.15%), the frequency of
sequence;(S is the cluster volume [voxel]; S(s) is the cumulative normalised distri- small oil droplets increases, the contribution of oil saturation
bution and n(s)is the normalised distribution. S(1) is the residual phase mainly comes from the capture of larger clusters. The small oil
saturation(Sor ¼ 77.27%). n(s)-1 and S(s)-1 are oil cluster size distributions of saturated droplets frequency is increased due to the big clusters are scattered.
oil state; n(s)-2 and S(s)-2 are oil cluster size distributions after first water
The continuous oil clusters, which cannot be scattered by water
drive(Sor ¼ 62.60%); n(s)-3 and S(s)-3 are oil cluster size distributions of second fully
water drive (Sor ¼ 33.15%)). flooding, are captured by the pore throat, resulting in residual oil
with the voxel range of 104 to 105.
In summary, large continuous oil clusters are dispersed and
the volume fraction of continuous flow (clustered flow) gradually portions captured by the pores is the main reason for the decrease
decreases (Fig. 7); (4) the volume ratio variation of clustered flow of oil relative permeability. The intrinsic reason for the nonlinear
presents nonlinear downward bend trend along with the increase inflection point on the relative permeability ratio curve is the vol-
of water saturation (Fig. 8). ume fraction decrease of clustered flow with the increase of water
It is well-known that the relative permeability ratio curves for saturation. That is, in the late high water cut period, oil phase will
most of the rocks are under nonlinear downward bend trend. As transform from the continuous phase to the discontinuous phase,
44 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45
Fig. 12. Oil and water dynamic distribution during water flooding.
which results in a decline in the relative permeability and thereby
the nonlinear bending phenomenon appears on the curve. In this
paper, clustered flow will change to other discontinuous phases at
the late period of water displacement and the downward bending
clustered flow volume fraction is consistent with the nonlinear
bending phenomenon on the relative permeability ratio curve.
3.3. Microscopic sweep phenomenon
In the process of conventional water flooding, there exists an
obvious water flood front on the water saturation profile. In the
view of macro scale, we believe a water drive front is formed and
the macroscopic sweep efficiency of the water flooding transition
zone is one. However, in the view of microscopic pore scale, it can
be obviously found that not all pores can be swept in the process of
water flooding. Combined with 3D reconstruction images of oil and
water distribution, it shows that there is no water drive front at the
pore scale and the “sweep efficiency” concept under macroscopic
condition does not apply for microscopic pore scale.
As shown in Fig. 12, there is no water sweep for part of the pores
in all three stages thus the oil phase has not been driven out. For
two-phase flow in other pores, the flow position will redistribute
during the displacement process. After first time water drive, oil
phase will enter pores initially saturated with water phase.
Continue water drive again until no oil is produced, and the results
reveal that the oil phase in pores is blocked by water phase so that
the sweep efficiency for a single pore will decrease rather than
increase. This also explains why large oil clusters are not displaced
after water flooding in Fig. 11. In conclusion, macroscopic sweep
efficiency is not applicable in microscopic pore scale and local pore
recovery does not represent the overall core recovery. The injection
of water flow in the micro-channeling along the larger pores result
in local residual oil enrichment.
4. Conclusions
With the aid of CT scanning and image processing techniques,
water displacing oil experiments were performed on six natural
cores of different permeability and porosity for visual and quanti-
tative analysis. The microscopic remaining oil is categorized and
the change law of different flow regimes is clarified.
(1) In the process of water flooding, microscopic remaining oil
can be quantitatively characterized as membranous flow,
droplet flow, columnar flow, multi-porous flow and clustered
flow from hard-to-produce to easy-to-produce based on
shape factor, contact ratio and Euler number.
(2) The relative permeability of oil phase is the macro relative
permeability average of five kinds of flow regimes, among
which clustered flow is the easiest to produce due to rela-
tively high relative permeability while other four regimes are
more difficult to produce due to relatively low relative
permeability.
(3) The intrinsic reason for the nonlinear inflection point on the
relative permeability ratio curve is the quantity and average
volume decrease of continuous clustered flow with the in-
crease of water saturation. The turning point appearance
time on the relative permeability ratio curve happens to be
the start point of discontinuous phase development. There-
fore, if we keep the ratio of the continuous phase higher than
a certain value, it is possible to delay the inflection point
appearance time and improve oil recovery.
(4) Macroscopic sweep efficiency is not applicable under micro
pore-scale condition. There is no conventional water drive
front in micro pores and micro inter-porosity flow causes the
local enrichment of residual oil.
(5) In the aspect of the microscopic flow mechanism of
remaining oil, this paper has explained the intrinsic reason
for the appearance of an inflection point on the relative
permeability ratio curve, which also presented a new effec-
tive way of upscaling to a certain extent.
References
Andrew, M., Bijeljic, B., Blunt, M.J., 2013. Pore-scale imaging of geological carbon
dioxide storage under in situ conditions. Geophys. Res. Lett. 40 (15), 3915e3918.
Andrew, M., Bijeljic, B., Blunt, M.J., 2014a. Pore-by-pore capillary pressure mea-
surements using x-ray microtomography at reservoir conditions: curvature,
snap-off, and remobilization of residual co 2. Water Resour. Res. 50, 8760e8774.
Andrew, M., Bijeljic, B., Blunt, M.J., 2014b. Pore-scale imaging of trapped super-
critical carbon dioxide in sandstones and carbonates. Int. J. Greenh. Gas Control
22, 1e14.
Andrew, M., Menke, H., Blunt, M.J., Bijeljic, B., 2015. The imaging of dynamic
multiphase fluid flow using synchrotron-based x-ray microtomography at
reservoir conditions. Transp. Porous Media 110, 1e24.
Bryant, S., Anna, J., 2004. Bulk and film contributions to fluid/fluid interfacial area in
granular media. Chem. Eng. Commun. 12, 1660e1670.
Blunt, M., King, P., 1991. Relative permeabilities from two- and three-dimensional
pore-scale network modelling. Transp. Porous Media 6 (4), 407e433.
Buades, A., Coll, B., Morel, J.M., 2005. A non-local algorithm for image denoising. In:
IEEE Computer Society Conference on Computer Vision & Pattern Recognition,
Vol. 2. IEEE Computer Society, pp. 60e65 vol. 2.
Buades, A., Coll, B., Morel, J.M., 2008. Nonlocal image and movie denoising. Int. J.
 J. Li et al. / Journal of Natural Gas Science and Engineering 48 (2017) 36e45
Comput. Vis. 76 (2), 123e139.
Carroll, K.C., Mcdonald, K., Marble, J., Russo, A.E., Brusseau, M.L., 2015. The impact of
transitions between two-fluid and three-fluid phases on fluid configuration and
fluid-fluid interfacial area in porous media. Water Resour. Res. 51 (9), 572e581.
Celauro, J.G., Torrealba, V.A., Karpyn, Z.T., Klise, K.A., Mckenna, S.A., 2013. Pore-scale
multiphase flow experiments in bead packs of variable wettability. Geofluids
14, 95e105.
Chen, D., Pyrak-Nolte, L.J., Jessica, G., Giordano, N.J., 2007. Measurement of inter-
facial area per volume for drainage and imbibition. Water Resour. Res. 43,
55e60.
Dalla, E., Hilpert, M., Miller, C.T., 2002. Computation of the interfacial area for two-
fluid porous medium systems. J. Contam. Hydrology 56, 25e48.
Dann, R., Turner, M., Close, M., Knackstedt, M., 2011. Multi-scale characterisation of
coastal sand aquifer media for contaminant transport using x-ray computed
tomography. Environ. Earth Sci. 63 (5), 1125e1137.
Dong, H., Blunt, M.J., 2009. Pore-network extraction from micro-computerized-
tomography images. Phys. Rev. E 80 (2), 1957e1974.
Iglauer, S., Favretto, S., Spinelli, G., Schena, G., Blunt, M.J., 2010. X-ray tomography
measurements of power-law cluster size distributions for the nonwetting phase
in sandstones. Phys. Rev. E 82 (5Pt2), 2757e2768.
Iglauer, S., Paluszny, A., Pentland, C.H., Blunt, M.J., 2011. Residual CO2 imaged with
X-ray micro-tomography. Geophys. Res. Lett. 38 (21), 1440e1441.
Iglauer, S., Fernø, M.A., Shearing, P., Blunt, M.J., 2012. Comparison of residual oil
cluster size distribution, morphology and saturation in oil-wet and water-wet
sandstone. J. Colloid Interface Sci. 375, 187e192.
Iglauer, S., Paluszny, A., Blunt, M.J., 2013. Simultaneous oil recovery and residual gas
storage: a pore-level analysis using in situ x-ray micro-tomography. fuel 15,
905e914.
Iglauer, S., Wülling, W., 2016a. The scaling exponent of residual non-wetting phase
cluster size distributions in porous media. Geophys. Res. Lett. 43, 11,253e11,260.
Iglauer, S., Rahman, T., Sarmadivaleh, M., et al., 2016b. Influence of wettability on
residual gas trapping and enhanced oil recovery in three-phase flow: a pore-
scale analysis by use of microcomputed tomography. SPE J. 179727, 1e14.
Jiang, Z., Wu, K., Couples, G., Van Dijke, M.I.J., Sorbie, K.S., Ma, J., 2007. Efficient
45
extraction of networks from three-dimensional porous media. Water Resour.
Res. 43, 2578e2584.
Karpyn, Z.T., Piri, M., Singh, G., 2010. Experimental investigation of trapped oil
clusters in a water-wet bead pack using x-ray microtomography. Water Resour.
Res. 46, 475e478.
Kumar, M., Senden, T.J., Sheppard, A.P., et al., 2011. Probing the archie's exponent
under variable saturation conditions. Petrophysics 52 (2), 124e134.
Landry, C.J., Karpyn, Z.T., Piri, M., 2011. Pore-scale analysis of trapped immiscible
fluid structures and fluid interfacial areas in oil-wet and water-wet bead packs.
Geofluids 11, 209e227.
Løvoll, G., Meheust, Y., Måløy, K.J., Aker, E., Schmittbuhl, J., 2005. Competition of
gravity, capillary and viscous forces during drainage in a two-dimensional
porous medium, a pore scale study. Energy 30, 861e872.
Ohtsu, N., 1979. A threshold selection method from gray-level histograms. IEEE
Trans. Syst. Man Cybern. 9 (1), 62e66.
Raeini, A.Q., Bijeljic, B., Blunt, M.J., 2015. Modelling capillary trapping using finite-
volume simulation of two-phase flow directly on micro-CT images. Adv. wa-
ter Resour. 83, 102e110.
Ramstad, T., Idowu, N., Nardi, C., Øren, P.E., 2012. Relative permeability calculations
from two-phase flow simulations directly on digital images of porous rocks.
Transp Porous Media 94, 487e504.
Sahloul, N.A., Ioannidis, M.A., Chatzis, I., 2002. Dissolution of residual non-aqueous
phase liquids in porous media: pore-scale mechanisms and mass transfer rates.
Adv. water Resour. 25, 33e49.
Seright, R.S., Prodanovic, M., Lindquist, W.B., 2006. X-Ray computed micro-
tomography studies of fluid partitioning in imbibition and drainage before and
after gel placement: disproportionate permeability reduction. SPE J. 11,
159e170.
Valvatne, P.H., Blunt, M.J., 2004. Predictive pore-scale modeling of two-phase flow
in mixed wet media. Water Resour. Res. 40 (7), 187.
Zhou, Y., Helland, J., Hatzignatiou, D.G., 2014. Pore-scale modeling of waterflooding
in mixed-wet-rock images: effects of initial saturation and wettability. Spe J. 19,
88e100.