Journal of Geophysics and Engineering

PAPER Related content

- Transient pressure behavior for a
Analytical analysis of fracture conductivity for horizontal well with multiple finite-
conductivity fractures in tight reservoirs
sparse distribution of proppant packs Jingjing Guo, Haitao Wang and Liehui
Zhang

- Experimental and numerical investigations

To cite this article: Jianchun Guo et al 2017 J. Geophys. Eng. 14 599 on the vertical propagation of hydraulic
fractures in laminated shales
Qinglin Shan, Yan Jin, Peng Tan et al.

- Transient pressure analysis of a volume

fracturing well in fractured tight oil
View the article online for updates and enhancements. reservoirs
Cheng Lu, Jiahang Wang, Cong Zhang et
al.

Recent citations
- Crushing characteristics of four different
proppants and implications for fracture
conductivity
Wenbo Zheng et al

- Influence of gravel on the propagation

pattern of hydraulic fracture in the glutenite
reservoir
Zhenhua Rui et al

This content was downloaded from IP address 125.70.254.70 on 25/05/2018 at 04:14

 Journal of Geophysics and Engineering

J. Geophys. Eng. 14 (2017) 599–610 (12pp) https://doi.org/10.1088/1742-2140/aa6215

Analytical analysis of fracture conductivity

for sparse distribution of proppant packs
Jianchun Guo1, Jiandong Wang1,2, Yuxuan Liu1, Zhangxin Chen3 and
Haiyan Zhu1
1
State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum
University, Chengdu 610500, People’s Republic of China
2
Sinopec Shengli Oilfield Company, Shengli 257001, People’s Republic of China
3
Department of Chemical and Petroleum Engineering, Schulich School of Engineering, University of
Calgary, Calgary T2N 1N4, Canada

E-mail: guojianchun@vip.163.com

Received 28 March 2016, revised 13 December 2016

Accepted for publication 22 February 2017
Published 4 April 2017

Abstract
Conductivity optimization is important for hydraulic fracturing due to its key roles in
determining fractured well productivity. Proppant embedment is an important mechanism that
could cause a remarkable reduction in fracture width and, thus, damage the fracture
conductivity. In this work a new analytical model, based on contact mechanics and the
Carman–Kozeny model, is developed to calculate the embedment and conductivity for the
sparse distribution of proppant packs. Features and controlling factors of embedment, residual
width and conductivity are analyzed. The results indicate an optimum distance between
proppant packs that has the potential to maintain the maximum conductivity after proppant
embedment under a sparse distribution condition. A change in the optimum distance is
primarily controlled by closure pressure, the rock elastic modulus and the proppant elastic
modulus. The proppant concentrations and the poroelastic effect do not influence this
optimum distance.
Keywords: proppant, embedment, conductivity, hydraulic fracturing

(Some figures may appear in colour only in the online journal)

Nomenclature
Fl Force located at Lhf<x<Lhf, N m−1
2
Af Area of fracture face, m Fc Fracture conductivity, m2.m
a Radius of contact area, m Fi Force located at part i, N m-1
2
A Flow area, m f Fitting parameters, dimensionless
Af Area of rock face, m2 H Length in the direction of rows, m
As Contact area of single particle, m2 Hp Length of proppant pack in the direction of
C Kozeny–Carman constant, dimensionless rows, m
D2 Thickness of the upper and lower rocks, m h Embedment depth, m
ΔD1 Deformation of proppant packs, m J Layers of proppants, dimensionless
ΔD2 Deformation of rock, m K Stress intensity factor, Pa-m0.5
E1 Elastic modulus of proppant, Pa Ki Stress intensity factor at fracture tip induced
E2 Elastic modulus of rock, Pa by part i, Pa-m0.5
F Applied force on proppant, N k Permeability, m2

1742-2132/17/030599+12$33.00 599 © 2017 Sinopec Geophysical Research Institute Printed in the UK

J. Geophys. Eng. 14 (2017) 599 J Guo et al

L Length in the direction of columns, m

Lp Length of proppant pack in the direction of
columns, m
Lhf Fracture half length, m
M Columns of proppants, dimensionless
N Rows of proppants, dimensionless
NHF Number of parts divided in hydraulic fracture,
dimensionless
pmax Maximum contact pressure Pa
pave Average contact pressure, Pa
pc Closure pressure, Pa
po Overburden pressure, Pa Figure 1. Illustration of a hydraulic fracture.

pf Fluid pressure in the fracture, Pa

ps Pressure of single proppant acting on rock, Pa
Δp Pressure difference, Pa
Δptt Pressure drop from fracture tip to fracture
toe, Pa
q Flow rate, m3 s−1
R Radius of proppant, m
Rsi(i=1, 2) Proppant distance ratio, dimensionless
u Velocity, m s−1
wf Fracture width, m
wf0 Initial width of fracture, m
Δwf Change of fracture width, m
x Position along fracture length, m
v1 Poisson’s ratio of proppant, dimensionless
v2 Poisson’s ratio of rock, dimensionless
δ Indentation depth, m
μ Viscosity of fluid, Pa.s
f Porosity of the sample, dimensionless
1. Introduction
Hydraulic fracturing is a reservoir stimulation method often
applied to low permeability formations that would otherwise
be uneconomical to produce. The process starts with the high-
pressure injection of a fracturing fluid into a wellbore to
create cracks. Fluids containing sand or other proppants
suspended with the aid of thickening agents are then injected.
When the hydraulic pressure is removed from a well, the sides
of the fractures compress onto the proppants, creating a high-
permeability pathway through which natural gas and petro-
leum flow more freely in deep-rock formations (figure 1).
The response of fractured wells is highly dependent on
fracture conductivity. In conventional reservoirs, a high
viscosity crosslinked fluid is used to create long and sym-
metric fractures. Proppants are believed to be placed uni-
formly in the fracture due to high viscosity and good proppant
carrying capacity (El-M. Shokir and Al-Quraishi 2009).
Factors affecting the proppant placement have been widely
Figure 2. Width reduction due to proppant embedment.
studied (Clark and Guler 1983, Mobbs and Hammond 2001,
Dontsov and Peirce 2014). In unconventional reservoirs, slick
water is widely used to generate the fracture network. The
fluid has low viscosity and poor proppant carrying capacity,
resulting in some nonuniform distributions of proppants in the
fracture network (Awoleke et al 2016). Experimental and
field tests both show that sparse distribution—such as a partial
monolayer in secondary fractures (Fredd et al 2001, Zou
et al 2015) or sparse distribution in channel fracturing
(D’Huteau et al 2011)—can improve the fracture con-
ductivity. The variation of fracture conductivity under sparse
distribution, therefore, is very important in optimizing the
fracturing design in unconventional reservoirs.
Proppant embedment may significantly impair fracture
conductivity due to a decrease in the fracture width (figure 2).
Proppant embedment and deformation play significant roles
in the stimulation response of hydraulic fracturing, especially
in clay-rich, soft formations. Additionally, a stress shadow
effect, caused by multiple fracture interactions, can increase
the effective closure stress (Jin et al 2010), resulting in
enhanced proppant embedment. This is generally regarded as
a geomechanical effect in an unconventional reservoir simu-
lation and should be considered when assessing the produc-
tion forecast in formations with low Young’s modulus (Yu
and Sepehrnoori 2014).

600
J. Geophys. Eng. 14 (2017) 599 J Guo et al

Table 1. Factors considered in different models.

Model Rock deformation Proppant deformation Sparse distribution of conductivity Conductivity model type
Guo and Liu (2012) √ Analytical
Khanna et al (2012) √ √ Numerical
Neto et al (2015) √ √ Semi-analytical
Li et al (2014) √ √ Analytical
Model in this paper √ √ √ Analytical

Currently, it is still impossible to conduct in situ mea- 2. Model development

surements of residual width and conductivity. Several
experimental apparatus were developed to study the features In hydraulic fracture, proppants are packed in multilayers, as
of residual width and conductivity (Volk et al 1981, Hartley shown in figure 3. J layers of proppants support a fracture
and Bosma 1985, Huitt and McGlothlin 1986, Lacy width of wf. In the top view, the M columns and N rows of
et al 1997, Lacy et al 1998, Wen et al 2007, Guo et al 2008, proppants result in an area of the fracture region Af=L×H.
Zhang et al 2015). Based on the results of these experiments,
empirical formulas to calculate residual width and con- 2.1. Theoretical background
ductivity were proposed. Heinrich Hertz has derived the solution of the contact between
In recent years, several analytical models have also been an elastic sphere and an elastic half-space (Johnson 1987),
developed. Guo and Liu (2012) proposed an analytical model as illustrated in figure 4. An elastic sphere of radius R indents
to describe the embedment features in soft rock. Their model, an elastic half-space to depth δ, and thus creates a contact area
however, only considered closely packed cases. Khanna et al of radius
(2012) proposed a simplified approach for calculating the
conductivity of narrow fractures filled with a sparse mono- a= Rd . (1 )
layer of proppant particles. The particle is considered to be a
The applied force F is related to the displacement depth δ by
rigid body. Neto et al (2015) developed a semi-analytical
model for residual width and conductivity based on a dis- 4
F= E *R 2 d 2
1 3
(2 )
tributed dislocation technique with consideration of proppant 3
pack deformation, which is described by Terzaghi’s classical
consolidation model. The sparse distribution of proppant was where
not considered in this model. Li et al (2014) developed an 1 1 - v12 1 - v22
analytical model to calculate proppant embedment based on = + (3 )
E* E1 E2
contact mechanics for sparse placement of proppant. The
Carman–Kozeny model was modified to predict the con- where v1 is Poisson’s ratio of proppants (dimensionless); v2 is
ductivity for proppant packs, although sparse distribution was Poisson’s ratio of rock (dimensionless); E1 is the elastic
not considered. In summary, theoretical studies on proppant modulus of proppants (Pa); E2 is the elastic modulus of
embedment and conductivity have been widely conducted rock (Pa).
(table 1); however, conductivity under sparse distribution has The radius of the contact circle a, depth δ and the max-
received limited attention. The characteristics and controlling imum contact pressure pmax are given by
factors of conductivity under sparse distribution is not clear, 1
especially for multi-layer proppant packs. ⎛ 3FR ⎞ 3
a= ⎜ ⎟ (4 )
For this paper, a new analytical model, based on contact ⎝ 4E * ⎠
mechanics and the Carman–Kozeny model, was developed to
1
address this problem. A number of simplifications were ⎛ 9F 2 ⎞ 3
d=⎜ ⎟ (5 )
adopted, including disregarding the surface roughness of ⎝ 16RE *2 ⎠
fractures, possible proppant crush and secondary cracking of
fracture walls due to very high contact stresses. The proppant ⎛ 6FE *2 ⎞ 3
1

embedment and deformation are considered to be the primary pmax =⎜ 3 2 ⎟ . (6 )

⎝ pR ⎠
mechanism of fracture width change and fracture conductivity
loss. These assumptions limit the developed approach’s The average contact pressure compressing proppants is rela-
ability to accurately predict the conductivity. Even so, the ted to the maximum contact pressure
results can provide insight into the features and factors that
control the conductivity change induced by proppant 2
pave = pmax . (7 )
embedment and deformation. 3

601
J. Geophys. Eng. 14 (2017) 599 J Guo et al

Figure 3. Problem geometry.

As = pa2 . (11)

Combining equations (8)–(11), ps can be derived as

4pc R 2
ps = (12)
pa2

and

ps = pave . (13)

Combining equations (7), (12) and (13), the maximum

contact pressure pmax can be expressed as
Figure 4. Contact of an elastic sphere with an elastic half-space.
3 4pc R 2
pmax = (14)
2 pa2

Combining equations (4)–(6) and (14), the depth δ and

maximum contact pressure pmax can be derived as
2
⎛ 3p ⎞ 3
d = ⎜ c⎟ R (15)
⎝ E* ⎠

1 1
pmax = (24pc E *2) 3 . (16)
p

Figure 5. Stress analysis of proppants.

The thickness of the upper and lower rocks is the same
and assumed as D2. The deformation of rock is reflected in the
2.2. Even distribution of proppants form of proppant embedment (Li et al 2014). Therefore, the
2.2.1. Residual width. Figure 5 shows a stress analysis of a
embedment depth h can be expressed as
proppant pack. The relationships between the load and stress 2
satisfy the following equations: ⎛ 3p ⎞ 3 p
h = d + DD2 = ⎜ c ⎟ R + c D2 (17)
⎝ E* ⎠ E2
pc Af = ps MNAs (8 )
and where h is the embedment depth, (m); ΔD2 is the deformation
of rock (m).
pc = po - pf (9 )
The deformation of proppant packs satisfies the follow-
where pc is the closure pressure (Pa); po is the overburden ing equation:
pressure (Pa); pf is the fluid pressure in a fracture (Pa); ps is
pave
the pressure with which a single proppant acts on rock (Pa); DD1 = wf0 (18)
Af is the area of the propped fracture surface (m2); As is the E1
contact area of a single particle (m2). Af and As can be
where ΔD1 is the deformation of proppant packs (m); wf0 is
expressed as follows:
the initial width of a fracture (m). Combining equations (7),
Af = 4MNR 2 (10) (17) and (18), the change in the fracture width Δwf can be

602
J. Geophys. Eng. 14 (2017) 599 J Guo et al

Figure 6. Illustration of sparse distribution of proppant packs.

obtained by
2
2pmax ⎛ 3p ⎞ 3 Figure 7. A representative unit in figure 6.
Dwf = DD1 + 2h = wf0 + 2 ⎜ c ⎟ R
3E1 ⎝ E* ⎠
2pc
+ D2 (19) Using the same derivation process it is found that in even
E2
distribution, the depth δ and maximum contact pressure can
and be derived as
wf = wf0 - Dwf (20) 2
⎛ 3pc ⎞ 3
where Δwf is the change in fracture width (m); wf is the final d=⎜ ⎟ R (25)
⎝ E *Rs1Rs2 ⎠
fracture width (m).
1
2.2.2. Fracture conductivity. Results from discrete element 1 ⎛ 24pc E *2 ⎞ 3
pmax = ⎜ ⎟ . (26)
modeling indicate that the Carman–Kozeny model can be p ⎝ Rs1Rs2 ⎠
used to calculate the permeability of proppant packs
The embedment depth, h, and the change in fracture width,
(Mollanouri Shamsi et al 2015, Sanematsu et al 2015). The
Δwf, can be obtained as
permeability is described as follows:
f3 ⎛ 3pc ⎞ 3
2
kf = (2 R )2 (21) h = d + DD2 = ⎜
p
⎟ R + c D2 (27)
36C (1 - f)2 ⎝ E *Rs1Rs2 ⎠ E2
where f is the porosity of a sample; C is the Carman–Kozeny
2
constant. For beds packed with spherical particles, the constant 2p ⎛ 3pc ⎞ 3
C is equal to 5. The fracture conductivity can be obtained by Dwf = DD1 + 2h = max wf0 + ⎜ ⎟ R
3E1 ⎝ E *Rs1Rs2 ⎠
Fc = k f ´ wf . (22) p
+ c D2 . (28)
E2
2.3. Uneven distribution of proppants
2.3.2. Fracture conductivity. A representative unit in figure 6
2.3.1. Residual width. The proppant packs may have a sparse
is shown in figure 7. Conductivity of this unit is developed in
distribution in a hydraulic fracture, especially in channel
this section. The other unit can be derived using the same
fracturing (D’Huteau et al 2011). A simple sparse distribution
process. In this figure, q is the flow rate (m3 s−1), u is the
is illustrated in figure 6.
velocity (m s−1), k is the permeability (m2), and Δp is the
A two-proppant distance ratio (Rs1, Rs2) can be defined as
pressure difference (Pa). The subscripts (1, 2, 3) represent
Hp
2NR the different sections in this figure.
Rs1 = = (23)
H
H According to Darcy’s Law
Lp
2MR
Rs2 = = (24) kA dp
L L q = uA = - (29)
m dx
where Lp and Hp are the length and height of a proppant pack
(m). If the two packs are close to each other, Rs1 and Rs2 are where A is the flow area (m2); μ is the viscosity of a fluid
equal to unity; if not, they are smaller than unity. (Pa.s). According to the relationship between the pressure

603
J. Geophys. Eng. 14 (2017) 599 J Guo et al

Table 2. Basic parameters.

Data source v1 v2 E1 (MPa) E2 (MPa) R (mm) D2 (mm)

Lacy et al (1998) 0.13 0.13 21 306 1172 0.3175 20
Lu et al (2008) 0.144 0.144 35 000 28 770 0.3175 15

difference in sections 1 and 2, the equivalent permeability of

sections 1 and 2 is

1 1 1
1,2
= (1 - Rs2) + Rs2 . (30)
k eq k1 k2

According to the relationship between the flow rate in

sections 1 and 3, the equivalent permeability of the three
sections is
1,2,3
k eq = k eq
1,2
Rs1 + k3 (1 - Rs1) . (31)

The permeability of the different sections is

Figure 8. Comparison of theoretical values of proppant embedment
with experimental data from Lacy et al (1998).
f3
k2 = d2
36C (1 - f)2
wf2
k1 = k3 = . (32)
12

Subsequently, fracture conductivity can be obtained by

Fc = k eq
1,2,3
´ wf . (33)

3. Discussions

3.1. Model verification

To verify the new model, it is necessary to compare the new

Figure 9. Comparison of modified models and experimental data
model with the existing model and experimental data. Lacy from Lacy et al (1998).
et al (1998) and Lu et al (2008) measured the proppant
embedment and width change at different closure pressures. Table 3. Fitting results.
The experimental data they reported were used to test the new
Lacy et al (1998) Lu et al (2008)
model in this paper and the model developed by Li et al
2
(2014). Li et al (2014) summarized the known experimental Model f R f R2
conditions as listed in table 2. New model 1.314 0.92 16.49 0.97
The comparison of the theoretical values of proppant Li et al (2014) 1.322 0.92 21.14 0.95
embedment with experimental data from Lacy et al (1998) is
shown in figure 8. The values calculated with the new model
and the model by Li et al (2014) are very close. The two The same fitting parameters were introduced into the
theoretical models’ values are smaller than the experimental model developed by Li et al (2014). The comparisons of the
data. The reason for the smaller values has been discussed by modified models are shown in figure 9. The revised models fit
Li et al (2014). well with the experimental data and the calculated values of
The fitting parameters were introduced in equation (17): these two models almost overlap with each other. The cor-
rection factors and regression coefficients of the revised
⎡ 2 ⎤ models in figure 9 are listed in table 3.
⎛ 3p ⎞ 3 p
h = f ⎢ ⎜ c ⎟ R + c D2 ⎥ . (34) The same fitting parameters were introduced into the data
⎢⎣ ⎝ E * ⎠ E2 ⎥⎦
provided by Lu et al (2014). The comparisons of the modified

604
J. Geophys. Eng. 14 (2017) 599
Figure 10. Comparisons of modified models and experimental data
from Lu et al (2008).
Figure 11. Comparisons of the modified models and experimental
data from Lacy et al (1998).
Figure 12. Comparisons of the modified models and experimental
data from Lu et al (2008).
models are shown in figure 10. Similarly, the two models fit
well with the experimental data; the fitting parameters are also
listed in table 3. The results indicate that the new model can
match the experimental data in all the cases studied, with the
same or slightly better fitting results as compared with the
model by Li et al (2014).
A fracture width change is required in the calculation of
fracture conductivity. Consequently, the width change is also
compared with the modified model and experimental data.
605
J Guo et al
Figure 13. Embedment under different proppant distance ratios.
The comparisons of the modified models and experimental
data from Lacy et al (1998) are shown in figure 11. The two
modified models’ values are both smaller than the exper-
imental data; the deviation of the model values from the
experimental data, however, is not large and follows the trend
of the experimental data.
The comparisons of the modified models and experimental
data from Lu et al (2008) are shown in figure 12. The new
modified model in this paper can match most of the data until
the closure pressure increases to 40 MPa. At the same time,
the modified model results of Li et al (2014) are smaller than
the experimental data. The results demonstrate that the new
modified model can be used to calculate the width change.
3.2. Application to the sparse distribution of proppant packs
3.2.1. Changes in embedment, width and conductivity. In this
section, the new models modified by two sets of experimental
data are used to predict the embedment, fracture width change
and conductivity in the sparse distribution of proppant packs.
The results of the new model modified by the data from Lacy
et al (1998) is referred to as case A; the other, with the data
from Lu et al (2008), is referred to as case B.
Embedment can be obtained with the modified new
model at closure pressure Pc=30 MPa, as shown in
figure 13. With an increase in Rs1, embedment decreases, as
expected. The change in embedment is greater when Rs1 is
small. When Rs1 reaches a specific value, the change in
embedment is very small. For values in case A, when Rs1
increases to 0.4, the embedment depth reduces by more than
30%; when the value increases from 0.4 to 1, the embedment
depth changes only 8%. This shows that with an increase in
the proppant distance ratio, the embedment depth, as well as
the reduction amplitude, decreases. For values in case B the
embedment depthalmost reduces by half when Rs1 increases
to 0.2; when Rs1 increases from 0.2 to 1, the embedment
depth changes by 42%.
Figure 14 shows the relationship between a fracture
width change and Rs1 for the two cases. The trend is similar to
that of proppant embedment. With an increase in Rs1, the
fracture width decreases. Its change is greater when Rs1 is
small because, with an increase in Rs1, the pressure applied on
J. Geophys. Eng. 14 (2017) 599
Figure 14. Fracture width change under different proppant distance
ratios.
the rock and proppant packs reduces. From figures 13 and 14,
the trends indicate that Rs1 has a significant influence on
embedment and fracture width.
Furthermore, the stress intensity factor at the crack tip
would yield useful information on fracture closure especially
when the proppants are unevenly and sparsely distributed
within the fracture. The stress intensity factor (K ) is used in
fracture mechanics to predict the stress state near the fracture
tip. When the fracture is open, a positive value of K exists at
the fracture tip. Otherwise, a negative or zero value of K
indicates the fracture closure.
The stress intensity factor K at the fracture tip by a point
force F1 can be obtained from the following expression
(Wells 1965):
Fl L Hf LHf + x
ò-L
K= dx . (35)
2 pLHf Hf LHf - x
An approximate formula can be expressed as
Fi LHf + x
Ki = (36)
2 pLHf LHf - x
NHF
K= å Ki (37)
i=1
where K is the stress intensity factor (Pa-m0.5); Ki is the stress
intensity factor at the fracture tip induced by part i (Pa-m0.5);
Fl is the force located at Lhf<x<Lhf (N m−1); Fi is the
force located at part i (N m−1); Lhf is the fracture half length
(m); x is the position along the fracture length (m); NHF is the
number of parts divided in hydraulic fracture (dimensionless).
Under a sparse distribution, the stress intensity factor for
a narrow fracture filled with proppants has two limits. On one
side, in the proppant free part of the fracture, the stress
intensity factor is zero as the fracture faces have no cohesion
(Neto and Kotousov 2013). This case will happen when there
is no proppant or highly compressible proppant. On the other
side, when the whole fracture is continuously propped by
incompressible particles, the stress intensity factors at the tip
of the fracture have a certain value, K0. The hydraulic fracture
is divided in to NHF parts. Then K0 of each part can be
606
J Guo et al
Figure 15. Relationship between stress intensity factor and Rs1. The
stress intensity factor increases with the increased proppant distance
ratio.
Figure 16. Fracture conductivity under different proppant distance
ratios. There is an optimal distance for maximum conductivity after
proppant embedment and deformation.
obtained by equation (36) and then the integrated value can be
achieved by equation (37).
Figure 15 shows the relationship between the stress
intensity factor and Rs1 at different fracture lengths. The Kn is
a normalized stress intensity factor, which is normalized
against K0. The relationships presented in figure 15 clearly
show the expected tendency of K approaching K0, when the
fracture is fully filled with proppants (Rs1=1). This case also
corresponds to the maximum residual width of the fracture.
Fracture conductivity after proppant embedment and
deformation can also be obtained, as shown in figure 16. The
fracture conductivity is normalized against the conductivity of
a channel with an opening equal to the initial fracture width,
which is wf0 2 12. The two cases have the same trend. With an
increase in Rs1, the fracture conductivity first increases to a
certain value and then decreases. In other words, there is an
optimal distance which has maximum conductivity after
proppant embedment and deformation. This is because
increasing the proppant distance ratio has two different
effects on conductivity: (1) a reduction in proppant deforma-
tion and embedment due to more proppants supporting the
closure pressure, which decreases the loss of conductivity,
and (2) an increase in the loss of conductivity due to a
J. Geophys. Eng. 14 (2017) 599 J Guo et al

Figure 18. Effect of rock elastic modulus on conductivity. (a) the

Figure 17. Effect of closure pressure on conductivity. (a) the
conductivity increases with the increased rock elastic modulus; (b)
conductivity decreases with the increased closure pressure; (b) the
the optimal distance ratio shows a slight decrease with an increase in
optimal distance ratio shows a slight increase with an increase in
the rock elastic modulus.
closure pressure.

at different Rs1. As shown in figure 17(a), the conductivity

reduced void volume which is occupied by more proppants.
The two mechanisms control the loss of conductivity,
resulting in the complex behavior shown in figure 16. As
seen from this simulation, when the proppant distance ratio is
less than 0.2 and 0.4 for the two cases, respectively, the
reduction in proppant deformation and embedment mainly
controls the change in conductivity. When the proppant
distance ratio increases and exceeds a certain value, the
reduction of a void volume primarily controls the change in
conductivity. The curve trend is the same as the calculations
done by Khanna et al (2012), where the conductivity is
computed by an Ansys CFX with a monolayer of proppants.
A reasonable agreement exists between the model developed
in this paper and their model.
The two cases have the same trend and, therefore, in the
following sections, the basic parameters from Lacy et al
(1998), shown in table 2, are used to study the effect of
different influencing factors. If not specified otherwise, the
closure pressure is 30 MPa and Rs2=1.
3.2.2. Effect of closure pressure. Figure 17 shows a
relationship between conductivity and closure pressure (Pc)
decreases with the closure pressure for all the cases. As
shown in figure 17(b), the curve trends under four closure
pressures are slightly different, especially at the inflection
point. With an increase in the closure pressure, the inflection
point moves to the right, meaning that Rs1 with maximum
conductivity increases. In other words, with an increase in the
closure pressure, the proppant packs should be designed to be
closer to each other in order to achieve the maximum
conductivity. Under a high closure pressure, the conductivity
of a smaller Rs1 reduces to almost zero. This type of proppant
placement will be closed due to proppant embedment and
deformation.
3.2.3. Effect of rock elastic modulus. Figure 18 shows a
relationship between conductivity and rock elastic modulus
(E2) at different Rs1. As shown in figure 18(a), with an
increase in E2, the conductivity increases, and the incremental
amplitude decreases. When E2 is large, a further increase in
E2 has little effect on conductivity. As shown in figure 18(b),
with an increase in E2, the inflection point moves to the left,
which means that Rs1 with maximum conductivity decreases.
Essentially, with an increase in E2, the proppant packs can be

607
J. Geophys. Eng. 14 (2017) 599
Figure 19. Effect of proppant elastic modulus on conductivity. (a)
the conductivity increases with the increased proppant elastic
modulus; (b) the optimal distance ratio shows a slight decrease with
an increase in proppant elastic modulus.
designed more diffusively to each other; with an increase in
E2, the ability to resist embedding increases, while the
deformation of rock under compression also decreases, so the
fracture width can be better maintained. Thus, the maximum
fracture conductivity can be achieved with a larger distance
between proppant packs.
3.2.4. Effect of proppant elastic modulus. Figure 19 shows a
relationship between conductivity and proppant elastic
modulus (E1) at different Rs1. As shown in figure 19(a),
with an increase in E1, the conductivity increases and the
incremental amplitude decreases. The trend is the same as for
E2 in figure 18(a). When E1 is large, a further increase in E1
has little effect on conductivity. As shown in figure 19(b),
with an increase in E1, the inflection point moves to the left,
which means that Rs1 with maximum conductivity decreases.
In other words, with an increase in E1, the proppant packs can
be designed more diffusively to each other, because with an
increase in E1, the deformation of proppants under
608
J Guo et al
Figure 20. Effect of proppant concentration on conductivity. (a) the
conductivity increases with the increased proppant concentration; (b)
the optimal distance ratio does not show an obvious change with the
variation of proppant concentration.
compression decreases, so the fracture width can be better
maintained.
3.2.5. Effect of proppant concentration. Parameter J reflects
the proppant concentration in a fracture. Figure 20 shows a
relationship between conductivity and proppant concentration
(J) at different Rs1. As shown in figure 20(a), with an increase
in J, the conductivity, as well as the incremental amplitude,
increases. The results demonstrate that the relationship is a
power law. As shown in figure 20(b), with an increase in J,
the inflection point does not move obviously, which means
that the Rs1 with maximum conductivity remains the same for
the four cases. The proppant concentration does not show a
discernable influence on the optimal distance.
3.2.6. Effect of poroelastic effect. During hydraulic
fracturing, the fluid pressure near the wellbore is high,
while it is small near the tip. The two faces of the hydraulic
fracture are not parallel to each other. The fracture width
reduces from wellbore to the fracture tip along the
longitudinal direction. The initial fracture width was
J. Geophys. Eng. 14 (2017) 599 J Guo et al

Figure 21. Comparison of parallel faces and elliptical shape. Fracture

conductivity is not a constant in an elliptical shape.

calculated by an analytical model (PKN) (Perkins and

Kern 1961, Nordgren 1972). It is assumed that the dynamic
fracture width calculated using the PKN model is completely
filled with proppants. Values for embedment and conductivity
were obtained by the model proposed in this paper.
Figure 21 shows a relationship between conductivity and
position along the fracture length direction at different
fracture shapes. The conductivity in parallel fractures is a
constant, while it decreases along the fracture length direction
in the elliptical shape. Generally, fracture conductivity is a
constant for a single fracture in a reservoir simulation.
Figure 21, however, shows that the possible actual case is
totally different from a constant. This difference may have an Figure 22. Effect of pressure drop on conductivity. (a) the
impact on simulated well productivity. conductivity decreases with the increased pressure drop; (b) the
In the production process, after the fluid diffusion from optimal distance ratio does not show an obvious change with the
variation of pressure drop.
the porous matrix to the fracture and then to the wellbore, the
poroelastic effect occurs. On the one hand, the pore pressure
decrease and the effective closure pressure increase will
increase the embedment depth. On the other hand, fluid flow with consideration of the sparse distribution of proppant
within the fracture causes a pressure drop (Δptt) along the packs.
fracture length direction. In order to better illustrate the effect (2) With an increase in the proppant distance ratio, the
of poroelastic, an elliptical shape was used. fracture conductivity first increases to a certain value
Because there is pressure drop inside the fracture. The and then decreases. An optimum proppant distance ratio
fluid pressure at the tip is higher than that at the near wellbore maintains the maximum conductivity after proppant
region. Therefore, the effective stress increases from the embedment and deformation.
fracture tip to near the wellbore region. With an increased (3) A change in the optimum proppant distance ratio
pressure drop, the effective stress near the wellbore increases. increases with closure pressure and decreases with the
Therefore, the conductivity near the wellbore is more rock elastic modulus and proppant elastic modulus.
sensitive to the pressure drop, as shown in figure 22(a). There is no obvious change with the variation of
However, the varied amplitude is limited, which does not proppant concentration.
show an effect on the optimal distance (figure 22(b)). (4) The stress intensity factor at the fracture tip increases
with the increased proppant distance ratio.

4. Conclusions
Based on the study, the following conclusions can be drawn.
(1) New mathematical models have been derived to
calculate changes in fracture width and conductivity
Acknowledgments
The authors are grateful for the financial support of National
Natural Science Foundation of China (No. 51374178),
National Natural Science Foundation of China (No.
51604232), the Young Scholars Development Fund of SWPU

609
J. Geophys. Eng. 14 (2017) 599
(201599010084) and NSERC/AIEES/Foundation CMG and
AITF Chairs (iCORE), and National Science and Technology
Major Project (No. 2016ZX05002-002).
References
Awoleke O O, Zhu D and Hill A D 2016 New propped-fracture-
conductivity models for tight gas sands SPE J. 21 1508–17
Clark P E and Guler N 1983 Prop transport in vertical fractures:
settling velocity correlations SPE/DOE Low Permeability Gas
Reservoirs Symposium (Denver, CO, 14–16 March) (Tulsa,
OK: Society of Petroleum Engineers) SPE-11636-MS
D’Huteau E et al 2011 Open channel fracturing—a fast track to
production Oilfield Rev. Autumn 2011 23
Dontsov E V and Peirce A P 2014 Slurry flow, gravitational settling
and a proppant transport model for hydraulic fractures J. Fluid
Mech. 760 567–90
El-M.Shokir E M and Al-Quraishi A A 2009 Experimental and
numerical investigation of proppant placement in hydraulic
fractures Pet. Sci. Technol. 27 1690–703
Fredd C N, McConnell S B, Boney C L and England K W 2001
Experimental study of fracture conductivity for water-
fracturing and conventional fracturing applications SPE J. 6
288–98
Guo J C and Liu Y X 2012 Modeling of proppant embedment:
elastic deformation and creep deformation SPE International
Production and Operations Conference & Exhibition (Doha,
Qatar, 14–16 May) (Tulsa, OK: Society of Petroleum
Engineers) SPE-157449-MS
Guo J C, Lu C, Zhao J Z and Wang W Y 2008 Experimental
research on proppant embedment J. China Coal Soc. 33 661–4
(in Chinese)
Hartley R and Bosma M 1985 Fracturing in chalk completions
(includes associated papers 14483 and 14682) SPE J. Pet.
Technol. 37 73–9
Huitt J L and McGlothlin B B Jr 1986 The propping of fractures in
formations susceptible to propping-sand embedment Drilling
and Production Practice (New York, 1 January) (Washington,
DC: American Petroleum Institute) API-58-115
Jin Z H, Johnson S E and Fan Z Q 2010 Subcritical propagation and
coalescence of oil-filled cracks: getting the oil out of low-
permeability source rocks Geophys. Res. Lett. 37 L01305
Johnson K L 1987 Contact Mechanics (Cambridge: Cambridge
University Press) pp 92–3
Khanna A, Kotousov A, Sobey J and Weller P 2012 Conductivity of
narrow fractures filled with a proppant monolayer J. Petrol.
Sci. Eng. 100 9–13
Lacy L L, Rickards A R and Ali S A 1997 Embedment and fracture
conductivity in soft formations associated with HEC, borate
and water-based fracture designs PE Annual Technical
610
J Guo et al
Conference and Exhibition (San Antonio, TX, 5–8 October)
(Tulsa, OK: Society of Petroleum Engineers) SPE-38590-MS
Lacy L L, Rickards A R and Bilden D M 1998 Fracture width and
embedment testing in soft reservoir sandstone SPE Drill
Completion 13 25–9
Li K W, Gao Y P, Lyu Y C and Wang M 2014 New Mathematical
models for calculating proppant embedment and fracture
conductivity SPE J. 20 496–507
Lu C, Guo J C, Wang W Y, Deng Y and Liu D F 2008 Experimental
research on proppant embedment and its damage to fractures
conductivity Nat. Gas Ind. 28 99–101
Mobbs A T and Hammond P S 2001 Computer simulations of
proppant transport in a hydraulic fracture SPE Prod. Facil. 16
112–21
Mollanouri Shamsi M M, Farhadi Nia S and Jessen K 2015
Conductivity of proppant-packs under variable stress
conditions: an integrated 3D discrete element and lattice
boltzman method approach SPE Western Regional Meeting
(Garden Grove, CA, 27–30 April) (Tulsa, OK: Society of
Petroleum Engineers) SPE-174046-MS
Neto L B, Khanna A and Kotousov A 2015 Conductivity and
performance of hydraulic fractures partially filled with
compressible proppant packs Int. J. Rock Mech. Min. 74 1–9
Neto L B and Kotousov A 2013 Residual opening of hydraulic
fractures filled with compressible proppant Int. J. Rock Mech.
Min. 61 223–30
Nordgren R P 1972 Propagation of a vertical hydraulic fracture SPE
J. 12 306–14
Perkins T K and Kern L R 1961 Widths of hydraulic fractures J. Pet.
Technol. 13 937–49
Sanematsu P et al 2015 Image-based Stokes flow modeling in bulk
proppant packs and propped fractures under high loading
stresses J. Petrol. Sci. Eng. 135 391–402
Volk L J, Raible C J, Carroll H B and Spears J S 1981 Embedment
of high strength proppant into low-permeability reservoir rock
SPE/DOE Low Permeability Gas Reservoirs Symposium
(Denver, CO, 27–29 May) (Tulsa, OK: Society of Petroleum
Engineers) SPE-9867-MS
Wells A A 1965 Notched bar tests, fracture mechanics and the brittle
strengths of welded structures Br. Weld. J. 12 2–13
Wen Q Z, Zhang S C, Wang L, Liu Y S and Li X P 2007 The effect
of proppant embedment upon the long-term conductivity of
fractures J. Petrol. Sci. Eng. 55 221–7
Yu W and Sepehrnoori K 2014 Simulation of gas desorption and
geomechanics effects for unconventional gas reservoirs Fuel
116 455–64
Zhang J J, Ouyang L C, Zhu D and Hill A D 2015 Experimental and
numerical studies of reduced fracture conductivity due to proppant
embedment in the shale reservoir J. Petrol. Sci. Eng. 130 37–45
Zou Y, Ma X, Zhang S, Zhou T, Ehlig-Economides C and Li H 2015
The origins of low-fracture conductivity in soft shale
formations: an experimental study Energy Technol. 3 1233–42