See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/341336299

Geology and geomechanics of hydraulic fracturing in the Marcellus shale gas

play and their potential applications to the Fuling shale gas development

Article  in  Energy Geoscience · May 2020

DOI: 10.1016/j.engeos.2020.05.002

CITATIONS READS

38 1,528

3 authors, including:

Jon Jincai Zhang

Hess Corporation
121 PUBLICATIONS   2,831 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Jon Jincai Zhang on 21 May 2020.

The user has requested enhancement of the downloaded file.

 Energy Geoscience 1 (2020) 36e46

Contents lists available at ScienceDirect

Energy Geoscience
journal homepage: www.keaipublishing.com/en/journals/energy-geoscience

Geology and geomechanics of hydraulic fracturing in the Marcellus

shale gas play and their potential applications to the Fuling shale gas
development
Herong Zheng a, *, Jincai Zhang b, Yuanchang Qi b
a
Petroleum Exploration and Production Research Institute, SINOPEC, Beijing, 100083, China
b
Sinopec Tech Houston, Houston, TX, 77056, USA

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

Article history: Geological characteristics, geomechanical behavior and hydraulic fracture propagation mechanism in the
Received 1 April 2020 Marcellus shale gas play are analyzed and compared with China’s Fuling shale play. Successful experi-
Received in revised form ences in hydraulic fracturing and shale gas development in the Marcellus shale gas play are summarized,
24 April 2020
which might be applicable in other shale plays. The main factors contributing to the successful devel-
Accepted 6 May 2020
opment of the Marcellus shale gas play include adoption of advanced drilling and completion technol-
ogies, increases of hydraulic fracturing stages, proppant concentration and fluid injection volume. The
Keywords:
geological and geomechanical mechanisms related to those technologies are analyzed, particularly the
Marcellus shale
Fuling shale
in-situ stress impacts on hydraulic fracturing. The minimum horizontal stress controls where the frac-
Geology tures are initiated, and the maximum horizontal stress dominates the direction of the hydraulic fracture
Geomechanics propagation. Hydraulic fracturing performed in the shale reservoir normally has no stress barriers in
In-situ stresses most cases because the shale has a high minimum horizontal stress, inducing hydraulic fractures
Hydraulic fracturing propagating beyond the reservoir zone, resulting in inefficient stimulation. This is a common problem in
shale plays, and its mechanism is studied in the paper. It is also found that the on-azimuth well has a
higher productivity than the off-azimuth well, because shear fractures are created in the off-azimuth
well, causing main fractures to kink and increasing fracture tortuosity and friction. The Fuling shale
gas play has a markedly higher minimum horizontal stress and much smaller horizontal stress differ-
ence. The high minimum horizontal stress causes a much higher formation breakdown pressure;
therefore, hydraulic fracturing in the Fuling shale gas play needs a higher treatment pressure, which
implies higher difficulty in fracture propagation. The small difference in the two horizontal stresses in
the Fuling shale gas play generates shorter and more complex hydraulic fractures, because hydraulic
fractures in this case are prone to curve to preexisting fractures. To overcome these difficulties, we
recommend reducing well spacing and increasing proppant concentration to increase gas productivity
for the Fuling shale gas development.
© 2020 Sinopec Petroleum Exploration and Production Research Institute. Production and hosting by
Elsevier B.V. on behalf of KeAi. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction Marcellus shale gas play accounts for about 25.3% of total U.S. dry
natural gas production and 32.2% of total U.S. dry shale gas pro-
The Marcellus shale play is the largest shale gas reservoir and duction in February 2020 (Fig. 1). The Marcellus and Utica/Point
gas producer in the United States. The annual gas production in the Pleasant shale plays provide 91% of U.S. shale gas production
growth since the start of 2012e2018. The gas production in the
Marcellus shale gas play increased from 6.3 Bcf/d (1.78
108 m3/d)
* Corresponding author.
in 2012 to 23.5 Bcf/d (6.7
108 m3/d) in January 2020. Besides the
E-mail address: [email protected] (H. Zheng). geological factors of the shale reservoir, increased adoption of
advanced drilling and completion techniques, increased number of
stages in hydraulic fracturing operations, and applications of zipper
fracturing and simultaneous fracturing are also contributing factors
Production and Hosting by Elsevier on behalf of KeAi to the great increase in natural gas production from the Marcellus

https://doi.org/10.1016/j.engeos.2020.05.002
2666-7592/© 2020 Sinopec Petroleum Exploration and Production Research Institute. Production and hosting by Elsevier B.V. on behalf of KeAi. This is an open access article
under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46 37

Fig. 1. Historical daily production of the U.S. major shale gas plays (plotted from EIA data) (1BCF ¼ 0.0283 billion m3).

shale gas play. Additionally, the Marcellus Shale Energy and Envi- and production of the Fuling shale play; however, compared with
ronment Laboratory (MSEEL) sponsored by the Department of En- the Marcellus shale play, there may still be room for improvements
ergy of USA has provided a long-term field site to develop and to enhance gas production. Comparisons of geological and geo-
validate new knowledge and promote research and applications of mechanical characteristics in the Marcellus and China’s Fuling shale
new technologies. The Fuling shale gas play is the first and largest gas plays are made. Successful experiences in hydraulic fracturing
commercial shale gas field in China, which produced about and shale gas development in the Marcellus shale play are
1.8
107 m3/d of natural gas in 2019. Many new technologies have analyzed, which may be applicable for the Fuling and other shale
been applied to the geological characterization, drilling and gas plays (Guo et al., 2016; Sun and Bao, 2018; Hu and Liu, 2018; Ma
completion of the Fuling shale play to reduce cost and increase gas et al., 2018).
production. Technical advances have been achieved in exploration

Fig. 2. The area and isopach map of net thickness (in ft.) of the Marcellus shale gas play (Wrightstone, 2008). (1ft ¼ 0.3048m).
38 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46

Fig. 3. The outcrop of the Marcellus shale taken at the Bullpasture river side in West Virginia (left) (photo courtesy of Drs. Herong Zheng and Zhucheng Shang). Cores of the Lower
Marcellus shale and its underlying Onondaga limestone (contact interface at 2286.4 m) (right).

2. Geologic characteristics of the Marcellus shale gas play underlying formation of the Onondaga limestone are seen in the
photo on the left in Fig. 3. And the black shale of the Lower Mar-
2.1. Introduction to the Marcellus shale gas play cellus formation and its base of the greyish Onondaga limestone are
presented in the photo on the right in Fig. 3. Densely distributed
The Marcellus shale gas play located in the Appalachian Basin is natural fractures can also be found in the outcrop and core samples
the Devonian-age marine-deposited shale formations, containing in the Marcellus shale formation. The outcrop of the Marcellus
abundant organic matters. This shale play encompasses 104,000 shale formation contains two sets of nearly orthogonal natural
square miles (269,359 km2) stretching across Pennsylvania and fractures (joints) J1 and J2 (Fig. 4), and one set of natural fractures
West Virginia into eastern Ohio and western New York in the can be found on the outcrop picture of Fig. 3. The regional orien-
eastern United States. It is the largest source of natural gas in the tation of the maximum horizontal stress in the Marcellus play is
United States, and its major producing area covers 50,600 square N45-80 E, consistent to the direction of J1 joint set (Fig. 5). The in-
miles (131,053 km2). In its 2011 Annual Energy Outlook report, EIA situ stresses in most areas of the Marcellus shale play belong to the
(2017) reported that the technically recoverable reserves are strike-slip faulting stress regime. Microseismic measurements of
approximately 400 trillion cubic feet (11.33 trillion m3) of natural hydraulic fracturing operations demonstrate the hydraulic frac-
gas in the Marcellus shale gas play. tures propagate along the directions of the maximum horizontal
The total gross thickness of the Marcellus formations is up to stress and vertical stress. Microseismic observations also indicate
290 m, but the net thickness of the shale gas formations is much that one of the propagating directions of the hydraulic fractures is
smaller. The areas with higher net thickness are located in the the same as the J1 direction (Neuhaus et al., 2013), which might be
northeastern and southwestern core areas. The Marcellus net one of the major reasons for the high productivity in the Marcellus
thickness is the highest and more than 37 m in the northeastern play.
core area located in Pennsylvania, as shown in Fig. 2. The thickness The X-ray diffraction analyses from 195 samples in 18 wells
reduces gradually from the northeastern core area toward the west (Wang and Carr, 2013) indicate that the most abundant clay mineral
and south. The Marcellus net thickness in West Virginia and Ohio is in the Marcellus shale reservoirs is illite, averaging 25% by volume,
small and varies from 0 to 24 m. However, most production areas in and the Marcellus shale reservoirs contain large volumes of brittle
the Marcellus shale gas play are those reservoirs with a net thick- minerals, 40%e60% quartz and 5%e40% calcite. Therefore, the
ness greater than 6 m. Marcellus shale is a reservoir of high brittleness, as demonstrated in
Fig. 6a. One of the objectives in shale reservoir exploration is to find
2.2. Characteristics of the Marcellus shale reservoirs and natural high TOC and brittle shales. The TOC content in the Marcellus shale
fractures increases as the brittle minerals (quartz and feldspar) increase
(Fig. 6b), and this implies that if high TOC shale can be found, then it
The Marcellus shale gas play consists of the Upper and Lower should be a good reservoir high in both TOC and fracability.
Marcellus shale reservoirs which are separated by a thin Cherry
Valley limestone with a thickness of about 6 m. Although the Lower 3. Key aspects of hydraulic fracturing in the Marcellus shale
Marcellus is much thinner, it has a much higher total organic car- gas development
bon (TOC) than the Upper Marcellus. The TOC in the Marcellus shale
formation can be identified from gamma ray logs which is to be 3.1. Hydraulic fracture containment by the minimum horizontal
demonstrated in the following section. Fig. 3 displays the outcrop stress
and drilling core pictures of the Marcellus shale formation and its
overlying and underlying formations. The outcrop of the Lower The in-situ stresses are very important parameters to control
Marcellus shale, its overlying Cherry Valley limestone, and its hydraulic fracturing operations. It normally includes three
 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46 39

directions. Therefore, in normal and strike-slip faulting stress re-

gimes (most formations belong to these two regimes), the hydraulic
fractures should be narrow vertical fractures and propagate in the
vertical and the maximum horizontal stress directions. Therefore,
the minimum stress controls primarily where the fracture is initi-
ated, and the other two stresses dominate to which directions the
fracture propagates. However, in the reserve faulting stress regime,
the hydraulic fractures would be horizontal and propagate in di-
rections of the two horizontal stresses. Field observations of the
hydraulic fracturing operations indicate that the difference of two
horizontal stresses is an important factor in affecting hydraulic
fracture propagation. The general rule is that the formation with
small horizontal stress difference generates shorter and more
complicated hydraulic fractures, and the formation with large
horizontal stress difference creates longer and more planar hy-
draulic fractures. Therefore, horizontal stresses are important pa-
Fig. 4. Two sets of joints (J1 and J2) in outcrop of the Marcellus shale (Harper and rameters to control hydraulic fracturing performance. The
Kostrlnik, 2010), photo courtesy of Gary Lash. J1 is parallel to the maximum hori-
minimum horizontal stress in an isotropic rock formation can be
zontal stress (sH) direction.
estimated by the following equation:

orthogonal principal stresses, i.e. the vertical stress (sV), the min- n
sh ¼ ðsV  pp Þ þ pp þ bsV (1)
imum horizontal stress (sh), and the maximum horizontal stress 1n
(sH). According to the different magnitudes of these three principal
where sh is the minimum horizontal stress; sV is the overburden
stresses, it can be classified into three stress regimes: the normal
stress, which can be calculated from an integration of the density
faulting stress regime (sV  sH  sh), the strike-slip faulting stress
log data; pp is the pore pressure; b is the minimum horizontal stress
regime (sH  sV  sh), and the reverse faulting stress regime
coefficient and can be back-calculated from in-situ stress mea-
(sH  sh  sV). The three principal stress magnitudes control how
surements; e.g., leak-off tests, mini-frac tests, and diagnostic frac-
easy the hydraulic fracture is propagating, and the stress regime
turing injection tests; n is Poisson’s ratio and can be calculated from
division depends on the fracture propagation directions. The
sonic logs using the following equation:
pioneer work on the hydraulic fracture initiation and propagation
behaviors in different stress regimes conducted by Hubbert and
1ðV

2
Willis (1957) are still valid. Their conclusions include that in-situ Vs Þ  1
p
n¼2
2 (2)
stresses control hydraulic fracture initiation, fracture breakdown, ðVp Vs Þ  1
and fracture propagation. The hydraulically induced fractures
should be formed approximately perpendicular to the minimum where Vp, Vs are the compressional and shear velocities,
stress direction and extend along the other two principal stress respectively.

Fig. 5. Map of Pennsylvania showing general orientations of structure axes, maximum horizontal stress, and J1 and J2 joint sets (Engelder, 2008; Carter et al., 2011).
40 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46

Fig. 6. (a) A ternary plot showing the mineral composition as a percentage of three major components (i.e., quartz plus feldspar, calcite plus dolomite, and illite plus chlorite) and
TOC. The boundaries between the Marcellus shale lithofacies are indicated by lines based on clay percentage (green dashed line), the ratio of quartz to carbonate (two blue long
dashed lines), and the TOC content (filled color of circles). Yellow area ¼ quartz rich; gray area ¼ clay rich; blue area ¼ carbonate rich; pink area ¼ mixture of quartz and carbonate.
(b) Crossplot showing that a higher quartz concentration corresponds to a higher TOC content (Wang and Carr, 2013). (For interpretation of the references to color in this figure
legend, the reader is referred to the Web version of this article.)

Eq. (2) demonstrates that if Poisson’s ratios differ in different shale is quite thick with a thickness of 27 m and has a much higher
formations, the minimum horizontal stresses should be different minimum horizontal stress; therefore, it acts as a hydraulic frac-
correspondingly. This illustrates that the minimum horizontal turing barrier to block out the hydraulic fractures propagation
stress is lithology-dependent (Zhang and Zhang, 2017). This (Zhang, 2019). Further up from the Tully limestone, it is again a
lithology-dependent stress behavior is of critical importance to shale formation with a much lower minimum horizontal stress;
hydraulic fracturing planning and operations, since the minimum therefore, the hydraulic fracturing operation should avoid the hy-
horizontal stress controls hydraulic fracture initiation. For a reser- draulic fractures propagation into this shale formation. It should be
voir rock with a lower minimum horizontal stress (e.g., a sand- mentioned that a limestone in some case belongs to a formation of
stone), the formation breakdown pressure is much smaller, as lower Poisson’s ratio and lower minimum stress, but here the Tully
indicated in Eq. (3). Therefore, the hydraulic fracture is more prone limestone is very tight and has a high Poisson’s ratio and conse-
to be initialized in the low minimum stress formation. However, quently a high minimum horizontal stress (Fig. 7). This makes the
formation breakdown pressure is much higher for the formation Tully limestone a barrier to contain the hydraulic fractures. Addi-
with a larger minimum stress (e.g., a shale), and this formation will tionally, a thin limestone with a higher minimum horizontal stress
block the development of the hydraulic fractures and act as a hy- also exists below the Lower Marcellus shale and acts as the fracture
draulic fracture barrier to contain fracture propagation. barrier to prevent the fracture from developing to the Onondaga
limestone, which has a very low minimum horizontal stress. In
pb ¼ 3sh  sH  pp þ kT0 (3) order to fracture both Upper and Lower Marcellus shale reservoirs,
the horizontal well can be landed at the top of the Lower Marcellus
where pb is the formation breakdown pressure for a vertical shale; by doing so, the hydraulic fractures can propagate way up
perforation; sH is the maximum pffiffiffi horizontal stress; k is a coefficient, through the very thin Cherry Valley limestone (a thickness of 6 m)
and the default value is k ¼ 2. It can be obtained from formation to the Upper Marcellus shale reservoir. These analyses are verified
breakdown tests; e.g., k ¼ 2.85 in a sandstone in the western by microseismic measurements.
Sichuan Basin (Kang et al., 2020); T0 is the tensile strength of the Microseismic events measured in the Marcellus reservoirs in
rock. central Pennsylvania show that hydraulic fractures extend 60 m
Fig. 7 presents an example to illustrate how to calculate the high above the treatment reservoir (Hakala and Guthrie, 2013). It
minimum horizontal stress and how the stress affects hydraulic indicates that the hydraulic fractures propagate from the Marcellus
fracture development. In the figure, Poisson’s ratio, the minimum reservoir upward into the Hamilton shale formation and terminate
horizontal stress, and vertical stress are computed from well log- at the Tully limestone. That is due to the fact that the Tully lime-
ging data in Well MIP-3H at the MSEEL test site (MSEEL, 2015). In stone has a much higher horizontal minimum stress than the
this well, numerous well logging measurements were conducted Hamilton shale and the Marcellus reservoirs, as shown in Fig. 7.
and are available for downloading. We use the measured sonic Below the Lower Marcellus shale, a thin limestone formation with a
compressional and shear transit time to calculate Poisson’s ratios high minimum horizontal stress blocks out some hydraulic fracture
based on Eq. (2). Then, Eq. (1) is used to compute the minimum propagation downward, and the microseismic events demonstrate
horizontal stress, as shown in Fig. 7. The figure shows that the that the hydraulic fractures only penetrate a very short distance
minimum horizontal stress varies markedly with rock types, and into the Onondaga limestone.
this is particularly true for limestones. The minimum horizontal
stress in Fig. 7 is quite low from the upper Marcellus shale to the top
of the Hamilton shale. Therefore, if hydraulic fracturing is per- 3.2. Impacts of horizontal well orientations on productivity
formed in the upper Marcellus shale, the fractures will propagate to
the Hamilton shale, because there is no hydraulic fracture Certain limitations may lead to the drilling azimuth of a hori-
containment between the Upper Marcellus shale and the Hamilton zontal well deviating from the minimum horizontal stress direc-
shale. However, the Tully limestone located above the Hamilton tion. A horizontal well drilled with an angle to the minimum
 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46 41

Fig. 7. The minimum horizontal stress, vertical stress and Poisson’s ratio calculated by well log data in Well MIP-3H at the MSEEL site. The first track presents gamma ray (GR) and
sonic compressional transit time (DTCO), the second track displays the minimum horizontal stress (Sh) and vertical stress (Sv), and the third track shows Poisson’s ratio calculated
from sonic logs. The annotations in each plot are the formation tops. In the geologic column, SH, LS, M, and L.M. represent shale, limestone, Upper Marcellus shale and Lower
Marcellus shale, respectively.

horizontal stress direction is named as off-azimuth well, while a 2018). Although the geologic conditions and completion methods
horizontal well drilled in the direction parallel to the minimum are similar in the two wells, the productivity in the on-azimuth well
horizontal stress is known as the on-azimuth well. Hydraulic increases 40%e50% compared with the 45 off-azimuth well.
fracture propagation behaviors differ greatly in the off-azimuth and The most likely reason for the productivity difference is that the
on-azimuth wells. Analyses of well-productivity in 475 wells (Zinn off-azimuth well is drilled with an angle to the minimum hori-
et al., 2011) reveal that the on-azimuth wells are the best in pro- zontal stress direction. This will generate shear stresses around the
ductivity. The gas production decreases as the deviation of the wellbore and the perforation tunnels, causing hydraulic fracture
horizontal wells away from the minimum horizontal stress direc- shear slips during hydraulic fracturing. Therefore, the hydraulic
tion increases. For example, the normalized estimated ultimate fractures will kink and propagate away from the perforation di-
recovery (EUR) of a 40 off-azimuth well reduces 2.5
107 m3 or rection which causes fracture curvatures, increases fracture tortu-
27.2% compared with the on-azimuth well. When the horizontal osity and friction, and in turn affects production. For the on-
well is 90 away from the minimum horizontal stress direction, the azimuth well, the perforation tunnels are parallel to the
EUR is the lowest and reduces 5.7
107 m3 or 92.6% compared with maximum horizontal direction and have no shear stresses; there-
the on-azimuth well (Fig. 8). A similar behavior is found in other fore, it is more likely to generate planar tensile fractures. This may
shale plays. For instance, in the Eagle Ford shale oil play most wells be the reason for the high productivity of the on-azimuth well. It
are on-azimuth wells, having the highest production; however, for can be observed that the in-situ stresses play a very important role
the off-azimuth wells the productivity reduces markedly, and the in well design and well completion.
90 off-azimuth well has the lowest production. Another example
is the Duvernay shale oil play in Canada, where two wells on the
3.3. Non-uniform propagation of hydraulic fractures
same pad (one is on-azimuth and the other is 45 off-azimuth)
were drilled and produced at the same time (Stephenson et al.,
Non-uniformly distributed hydraulic fractures with low pro-
ductive or non-productive stages are common problems in shale
gas wells. These may relate to reservoir heterogeneity but could
also be caused by the distribution of natural fractures, inefficient
hydraulic fracturing job, and stage/cluster interference (stress
shadow). Such problems have occurred in the Utica and Marcellus
shale plays (Dohmen et al., 2014). Fig. 9 shows the microseismic
events and production log relationship in a horizontal well (Well D)
in the Marcellus shale gas play at a research site in Washington
County, Pennsylvania. It can be seen that gas production is closely
correlated to the microseismic events. The toe stages of the hori-
zontal well have much more microseismic events which corre-
spond to a high production, implying more effective hydraulic
fracturing operations or presence of more natural fractures. It can
also be observed from Fig. 9 that the gas production is very non-
uniform: very high near the toe but rather low at other stages.
Fig. 8. The EUR reduction amount and percentage caused by the wellbore azimuth
After 13 months of production there is almost no productivity at the
diverted away from the minimum horizontal stress direction (plotted from the data middle stages, and this may be caused by bad reservoir quality or
presented by Zinn et al., 2011). the closures of the hydraulic fractures. To solve this problem,
42 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46

are net thickness of the reservoir, number of stages, easting-end

point, cluster per stage, proppant concentration, lateral length,
and perforation shot density (Mohaghegh et al., 2017). Productivity
increases with the increase of net thickness, proppant concentra-
tion, number of stages, and lateral length, as shown in Table 1. The
reservoir thickness cannot be changed, but other parameters can be
optimized to increase productivity. For instance, the statistical re-
sults of 140 Marcellus wells indicate that the wells with a proppant
concentration of greater than 2.53 t/m (1700 lbs/in) have the
highest productivity (good wells). The proppant concentration
currently used in some shale gas plays in China is less than 2.53 t/
m; therefore, increasing proppant concentration in those shale
plays may have potentials to increase productivity.

3.6. Production and stimulated reservoir volume

The stimulated reservoir volume (SRV) has been used to esti-

mate the production area created by stimulation and as a correla-
tion parameter for well performance and reservoir simulation in
the unconventional play. However, the SRV does not necessarily
represent the reservoir volume contributing to production, and the
volume that directly contributes to production is far less than the
Fig. 9. Relationship of microseismic events and gas production in horizontal well D in SRV. The SRV is normally obtained by microseismic measurements
the Marcellus shale play (Ciezobka and Salehi, 2013). Top figure: microseismic event during hydraulic fracturing operations or estimated by hydraulic
count for each fracturing stage. Middle figure: gas production from each perforation fracturing modeling. It contains not only the propped large frac-
cluster based on production log ran after 4-month production. Bottom figure: gas
tures, but also unpropped small fractures and microfractures.
production from each perforation cluster based on production log ran after 13-month
production (the scale in this figure is increased to show details). Although these unpropped fractures occupy a large volume, their
permeability is quite low. Furthermore, the unpropped fractures
optimized hydraulic fracturing design is needed to place more may close very soon as the production gets started; therefore, their
stages and clusters at the less productive stages or the stages with contribution to production is very small. In the Marcellus shale gas
fewer natural fractures present, or a diverter may be resorted to at play, the average hydraulic fracture half-length (the SRV half-
the less productive stages during hydraulic fracturing, together length) observed from microseismic measurements is
with an increment of fluid injection volume and proppant 152e305 m, but production data show that the effective fracture
concentration. half-length for production is only 46e61 m (Edwards et al., 2011).
Compared to the SRV half-length, only 15%e20% SRV contributes to
3.4. Hydraulic fracturing-variable rate pumping production. Based on this finding the well spacing can be further
reduced. A similar conclusion was obtained in the Eagle Ford shale
Conventionally, pump rate in hydraulic fracturing operations is play. Based on this, Friedrich and Milliken (2013) proposed a new
kept constant or small in fluctuation. Ciezobka (2016) proposed a term, the contributing reservoir volume (CRV) to describe the
controlled pump rate fluctuation method, i.e. the variable rate effectively stimulated reservoir volume, which directly contributes
pumping method, and tested it in the Marcellus research site. In to production. The CRV should be the volume where the hydraulic
this method the fluid pump rate during a hydraulic fracture fractures are propped. Based on this concept, the minimum well
pumping stage is rapidly changing from the predetermined spacing should be 2 times of the propped fracture half-length.
maximum rate to a much lower rate. The pump rate is then rapidly Therefore, the CRV concept can be used to optimize the well
increasing back to original maximum rate. This rapid change of the spacing to increase the EUR. However, reducing well spacing may
flow rate produces a pressure pulse that travels up and down the cause well interference. Therefore, the minimum well spacing
wellbore, which can open perforations previously unopened and should ensure that the propped fractures do not get connected in
increase fracture complexity. Microseismic measurements the two wells. In this case the unpropped small fractures and
demonstrate that rapid changes in the fluid pump rate during hy- microfractures get connected, but as production going on, these
draulic fracturing result in more microseismic events generated in unpropped fractures will close gradually with the increase of
the stimulated reservoir, which correlate with higher gas produc- effective stresses; consequently, the well interference may disap-
tion. It is also found that the introduction of rate fluctuation can pear. Some completion methods can be used to reduce well inter-
reduce the treatment pressure by about 1 MPa and decrease entry ference and stress shadow, such as pad drilling and simultaneous
friction (Ciezobka, 2016). hydraulic fracturing, zipper fracturing, etc.
Application of this technology was conducted in a horizontal
well where the odd stages adopted variable pump rate and nearby 4. Potential applications of the Marcellus shale play
even stages used constant pump rate. Production logs show that experiences to the Fuling shale play
the average productivity in the hydraulic fracturing stages with the
variable pump rate is 17.87% higher than the ones with constant 4.1. Introduction to the Fuling shale gas play
pump rate.
The Fuling shale gas play is the first commercial and the largest
3.5. Key performance indicators of productivity shale gas field in China, discovered in 2012 with geological reserves
of about 2.1
1012 m3. The core area with the most productive gas
Analyses of 140 horizontal wells in the Marcellus shale gas play shales occupies about 4000 km2 and is located in the Jiaoshiba area,
demonstrate that the key performance indicators for productivity Fuling District of Chongqing. The shale reservoir in the Fuling shale
 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46 43

Table 1
Key factors that impact productivity obtained from 140 Marcellus wells.

Average of all wells Poor wells Average wells Good wells

Net thickness of reservoir (m) 39.3 36.9 41.8 43.3

Proppant per meter(t/m) 2.28 2.15 2.4 2.53
Stages 9.1 8.5 9.6 11
Lateral length (m) 860 828 886 963

play is the Longmaxi shale formations, which are the Silurian-age 4.3. Comparison of the differences of horizontal stresses and
marine-deposited shale formations and contain abundant natural hydraulic fracturing
gas resources. The Fuling shale gas play consists of three large
sections: upper, middle and lower Longmaxi shales, and the gross The Marcellus and the Fuling shale plays are quite similar in
thickness of the Longmaxi formations is 249 m. The core reservoir is brittleness, but quite different in terms of in-situ stresses, and this
in the lower section of the Longmaxi formations with a net thick- might be the reason why the two shale plays are different in hy-
ness of about 38 m, deposited in the deep section of the Silurian- draulic fracturing efficiency. The in-situ stresses both in the Mar-
age Longmaxi shale to the Wufeng shale, the top of the Ordovi- cellus and the Fuling shale plays fall onto the strike-slip faulting
cian formations. stress regime, but the horizontal stress differences are significantly
The shale reservoir has a high TOC (up to 6.8%), high gas satu- large. Mini-frac stress measurements at the South Canisteo site
ration (66.8%e74.4%), and high maturation (Ro: 2.2%e2.9%), making adjacent to Eastern Gas Shale Project wells in western New York
it a good shale reservoir. It also contains large volumes of brittle (Evans et al., 1989) indicate that the maximum horizontal stress is
minerals, favorable for the fracture propagation in hydraulic frac- much higher than the minimum horizontal stress and vertical
turing. The Fuling shale formation is a highly overpressured stress in the Marcellus shale play in this area, i.e. sH ¼ 1.48sV,
reservoir with a pore pressure gradient of 1.55 g/cm3, which makes sH ¼ 1.77sh. The difference between the two horizontal stresses is
the Fuling shale play favorable for gas production. quite large in the Marcellus shale, i.e., sH  sh  16.8MPa, serving to
induce well oriented, long and planar hydraulic fractures.
However, the difference of the three in-situ stresses is very small
4.2. Comparison of the Marcellus and the Fuling shale gas plays in the Fuling shale play. For example, the vertical stress in the target
reservoir is 49.2e53.7 MPa, the minimum horizontal and the
Geological, petrophysical and mechanical properties in the maximum horizontal stresses are 48.6e49.9 MPa and
Marcellus and the Fuling shale gas plays are compared in Table 2. It 52.2e55.5 MPa, respectively (Liu, 2016). Obviously, the difference
shows that both the Marcellus and the Fuling shales are marine- between the two horizontal stresses is very small
deposited sediments. They are similar in rock mechanical proper- (sH  sh ¼ 3.4e5.6 MPa), much smaller than that in the Marcellus
ties (Young’s moduli and Poisson’s ratios), and both belong to shale gas play. Based on the experience in the Barnett shale gas
brittle and easily fractured formations. The main advantage in the play, the small stress difference in the Fuling shale gas play should
Fuling shale play is that it has a higher pore pressure gradient, generate shorter, more curvaceous and more complex hydraulic
1.55 g/cm3, compared with the Marcellus shale 1.3 g/cm3. The other fractures, since hydraulic fractures in this case are prone to curve to
advantage is that the Fuling shale play has a higher matrix preexisting fractures (e.g., Renshaw and Pollard, 1994). To over-
permeability compared with the Marcellus shale (Table 2). How- come these, it should reduce well spacing and increase proppant
ever, compared with the Fuling shale play, the Marcellus shale has concentration.
higher TOC, porosity and gas saturation. These differences may be Fig. 10 compares the results of microseismic measurements
one of the main factors explaining why the Marcellus shale play has during hydraulic fracturing operations in horizontal wells in two
a higher well productivity. horizontal sites, one in the Marcellus shale play and the other in the

Table 2
Geological, physical and mechanical properties of the Marcellus and the Fuling shale plays.

Marcellusa Fulingb

Geologic Epoch Middle Devonian Lower Silurian e Upper Ordovician

Depth (m) 700e2500 2250e3500
Net thickness (m): 15e107 55.4e89.1
Temperature ( C) 54 82
Calcite content (%) 5e40 11e14
Quartz content (%) 40e60 30e58
Clay content (%) 20e50 10.7e61.6
TOC (%) 2e20 0.3e6.8
%Ro 0.8e3.6 2.2e2.9
porosity (%) 4e15 1.2e8.0
Matrix permeability (md) 0.00013e0.002 0.002e0.004
Gas saturation (%) 90 66.8e74.4
Pore pressure gradient (g/cm3) 1e1.34 1.2e2.0
Young’s modulus (GPa) 7e50 17e46
Poisson’s ratio 0.15e0.35 0.19e0.25
Rickman brittleness (%) 10.1e94.5 52e82.5
Avg. initial production (104 m3/d/well) 70.8 32.6
Ave. EUR (108 m3/well) Core areas 4.2e11.3 Normally 1.5
a
References: Wrightstone (2008); Wang and Carr (2013); Zagorski et al., 2011.
b
References: Wang (2015); Wang (2017); Wang et al., 2017; Hu and Liu (2018); Sun and Bao (2018); Li (2017); Guo et al. (2016); Ma et al. (2018).
44 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46

Fuling shale play. Although these wells have been completed completion; e.g., in the Marcellus shale play, the lateral length has
several years ago, comparisons of them are still significant in reached 4000 m, stage length has been reduced to 33e60m, in-
illustrating the impact of stress difference on hydraulic fracturing. jection volume has reached 15.7m3/m, and proppant concentration
The horizontal stress differences in the Marcellus wells are much has been increased to 3.57e4.46 t/m or 2e2.5m3/m. These technical
larger than those in the Fuling shale play, as listed in Table 3. It can progresses have greatly improved gas productivity.
be interpreted by microseismic events in which well oriented, long
and planar hydraulic fractures were generated in the Marcellus 4.4. Comparison of the minimum horizontal stresses and formation
wells (Fig. 10a), due to the large horizontal stress difference. The breakdown pressure
average half-length of the hydraulic fractures interpreted by the
microseismic events was 189e343m in the Marcellus shale play. The other difference between the Marcellus and the Fuling shale
The average half-length of the hydraulic fracture was less than plays is in the minimum horizontal stress value. The Fuling shale
150m in the Fuling shale play (Fig. 10b), while the well spacing was play has a high minimum horizontal stress under the impact of
600m at that time; therefore, the well spacing has a great potential tectonic stress, a result of the collision between the Indian Plate and
to reduce. Hence, the well spacing in the Fuling shale play needs to Eurasian Plate. The Fuling shale play has an even greater minimum
be optimized. Otherwise, infill wells have to be drilled in the future, horizontal stress (generally sh ¼ 0.93sV) compared with the Mar-
which may cause well interference between the original wells and cellus shale (generally sh ¼ 0.84sV). The formation with a higher
infill wells owing to depletion (Dohmen et al., 2017). Comparison of minimum horizontal stress has a much higher breakdown pressure,
hydraulic fracturing parameters in the two plays indicates that the as indicated in Eq. (3). Therefore, it can be concluded that at the
proppant concentration in the Marcellus shale play was 1.3e1.5 same depth the Fuling reservoir has a much higher breakdown
times higher than that in the Fuling shale play. The general rule in pressure than the Marcellus shale. For example, for a shale reservoir
the Marcellus shale play is that a higher proppant concentration at a depth of 2410 m, the vertical stress is sV ¼ 57 MPa. If the for-
corresponds to a higher productivity (Table 1). Nagel (2018) mation pore pressures and tensile strengths are similar in the
observed that more proppant per foot equates to a better well. Fuling and the Marcellus shale plays, the formation breakdown
There has been a 30% increase in proppant per foot and a 10%e50% pressure in the Fuling shale play, based on Eq. (3), is 38 MPa higher
increase in production from 2017 to 2018 in the U.S. Hence, than that in the Marcellus shale. This implies that the Fuling shale
increasing proppant concentration in the Fuling shale play may also reservoir is more difficult for the fracture initiation and for the
function to enhance productivity. formation breakdown; e.g., the maximum formation breakdown
It should be mentioned that the completion parameters in pressure in Well Jiaoye-1H reaches 86 MPa (Fig. 11). Therefore, the
Table 3 are the results obtained in 2013 or even earlier. The reason Fuling reservoir needs a much higher treatment pressure than the
to use those wells is that the two sites, one in the Marcellus shale Marcellus shale reservoir. It implies that a higher proppant con-
play and the other in the Fuling shale play, are comparable and the centration is needed to keep the stimulated fractures open because
microseismic and completion data are available for analyses. the high minimum horizontal stress is prone to close the fractures.
Nowadays, many progresses have been achieved in drilling and It can also be observed in Fig. 11 that the heterogeneity of rock

Fig. 10. Map view of microseismic events observed during hydraulic fracturing operations. (a) In two horizontal wells (well spacing of 290 m) in the Marcellus shale play
(Mayerhofer et al., 2011). (b) In three horizontal wells (well spacing of 600 m) in the Fuling shale play (Liu, 2016). The color-coded points are the measured microseismic events. (For
interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46 45

Table 3
Completion parameters and stress difference in the Marcellus and the Fuling wells.

Marcellus (Mayerhofer et al., 2011) Fuling (Liu, 2016)

Well name and location 1H. Tioga, PA Jiaoye 4Xe3HF. Fuling, Sichuan
Well spacing (m) 290 600
Horizontal stress difference (MPa) (difference coefficient) >17 (>0.43) 3.4e5.6 (0.07e0.1)
Stage length (m) 100 70
Total fluid injection (m3/stage) 1333 1922
Total proppant volume (m3/stage) 110e124 58.8
Proppant volume per meter (m3/m) 1.1e1.24 0.84
Microseismic interpreted half-length (m) 189e343 120e300

mechanical properties at different stages in the horizontal well 5. Conclusions

cause the minimum horizontal stress heterogeneity, which results
in very different breakdown pressures at different stages. This may Geological, petrophysical and mechanical properties of a shale
induce non-uniformly distributed hydraulic fractures, causing low reservoir and its overburden and underburden formations are
productive or non-productive stages (clusters) and affecting gas closely related to its successful development. Numerous progresses
production. in the drilling and completion of the Marcellus shale have been
achieved and should be applicable to other shale plays. The key
conclusions include the following aspects.
4.5. Potential implementations in the Fuling shale gas play The minimum horizontal stress controls where the fractures are
initiated, and the maximum horizontal stress dominates the di-
The successful experience in the Marcellus shale play develop- rection of the hydraulic fracture propagation.
ment, particularly in horizontal well stimulation may be applicable The minimum horizontal stress is controlled by the regional
to China’s shale gas plays. The following aspects help tap the po- stresses and rock types; therefore, fracture initiation and propa-
tentials of the Fuling shale gas development. gation behaviors are not similar in different formations. The for-
Microseismically interpreted half-length of the hydraulic frac- mation with a large minimum horizontal stress is normally a stress
tures is less than 150 m in the Fuling shale, and thus based on the barrier to block fracture propagation. In all, perforation locations
CRV concept the current well spacing can be reduced. Optimizing and stress barriers greatly affect the hydraulic fracture propagation.
stimulation parameters and reducing well spacing are recom- When a horizontal well is drilled in the direction parallel to the
mended to avoid drilling infill wells in the future, so as to prevent minimum horizontal stress (on-azimuth well), the stimulated hy-
interference to the parent wells. draulic fractures are mainly tensile fractures with high productivity.
Understanding in-situ stress magnitudes and directions prior to However, shear fractures are created in the off-azimuth wells,
drilling is critically important, and the horizontal well should be causing main fractures to kink, which will in turn increase fracture
drilled in the direction parallel to the minimum horizontal stress. tortuosity and friction. These may result in large resistance in
The horizontal well landing location also needs to be optimized to proppant transport and make it difficult for the proppants to move
obtain the best hydraulic fracturing efficiency. to the fracture front area, causing low well productivity.
The Fuling shale play needs a higher proppant concentration to The Fuling shale play has a high minimum horizontal stress
keep the stimulated fractures open because it has a high minimum (close to the overburden stress or sh ¼ 0.93sV). Because the for-
horizontal stress which is prone to close the fractures. mation breakdown pressure increases greatly with the minimum
The propped fractures are the volume directly contributing to horizontal stress, the Fuling reservoir has a much higher break-
production. Therefore, increasing proppant volume in the Fuling down pressure than that in the Marcellus shale reservoir. This may
shale play is recommended for increasing productivity. be a factor in limiting Fuling shale gas productivity. To overcome
Using variable rate pumping method may reduce treatment this barrier, we recommend to raise treatment pressure and in-
pressure, lower entry friction within the hydraulic fracturing sys- crease the proppant concentration and volume.
tem and increase well productivity. The difference of the two horizontal stresses in the Fuling shale

Fig. 11. Formation breakdown pressures in the horizontal well of the Jiaoye-1H in the Fuling shale gas field (based on Sinopec Exploration Co., 2015).
46 H. Zheng et al. / Energy Geoscience 1 (2020) 36e46

play is very small, inducing short and complex hydraulic fractures. Sichuan Basin. Petroleum Geol. Exp 40 (1), 20e24.
Hubbert, M.K., Willis, D.G., 1957. Mechanics of hydraulic fracturing. Trans. Soc.
Therefore, well spacing should be reduced in order to sufficiently
Petrol. Eng. AIME. 210, 153e168.
fracture the reservoir. Optimizing well spacing and stimulation Kang, H., Fan, X., Huang, Z., Zhang, J., 2020. Investigating the effect of cyclic loading
parameters is also recommended to avoid drilling infill wells in the on the breakdown pressure of the Xujiahe sandstone. In: ARMA 20e1127 Pre-
future. sented at 54th US Rock Mechanics/Geomechanics Symp. Colorado, USA.
Li, J.L., 2017. Pressure prediction of Jiaoshiba shale gas reservoir in the fuling shale
gasfield. Geophys. Prospect. Pet. 56 (4), 567e574.
Declaration of competing interests Liu, Y.W., Liao, R.G., Zhang, Y., Gao, D.W., Zhang, H.L., Li, T., Zhang, C., 2016. Appli-
cation of surface-downhole combined microseismic monitoring technology in
the Fuling shale gas field and its enlightenment. Nat. Gas. Ind. 36 (10), 56e62.
The authors declare that they have no known competing Ma, Y.S., Cai, X.Y., Zhao, P.R., 2018. China’s shale gas exploration and development:
financial interests or personal relationships that could have understanding and practice. Petrol. Explor. Dev. 45 (4), 561e574.
appeared to influence the work reported in this paper. Mayerhofer, M.J., Stegent, N.A., Barth, J.O., Ryan, K.M., 2011. Integrating fracture
diagnostics and engineering data in the Marcellus Shale. SPE 145463 presented
at SPE Ann. In: Tech. Conf. Exhib. On Held in Denver, Colorado, USA.
References Mohaghegh, S.D., et al., 2017. Shale analytics: making production and operational
decisions based on facts: a case study in Marcellus Shale. In: SPE-184822,
Carter, K., Harper, J., Schmid, K., Kostelnik, J., 2011. Unconventional natural gas re- Presented at SPE Hydraulic Fracturing Technology Conference and Exhibition,
sources in Pennsylvania: the backstory of the model Marcellus shale play. En- 24e26 January. The Woodlands, Texas, USA.
viron. Geosci. 18 (4), 217e257. Nagel, N., 2018. Completion engineer for a day: how geology and geomechanics can
Ciezobka, J., 2016. Variable pump rate fracturing leads to improved production in influence completion designs in unconventionals. Presentation of URTeC
the Marcellus shale. In: SPE-179107 Presented at SPE Hydraulic Fracturing 3044072.
Technology Conference, 9-11 February, the Woodlands, Texas, USA. Neuhaus, C.W., Blair, K., Telker, C., Ellison, M., 2013. Hydrocarbon production and
Ciezobka, J., Salehi, I., 2013. Controlled hydraulic fracturing of naturally fractured microseismic monitoring - treatment optimization in the Marcellus Shale. In:
shalesea case study in the Marcellus Shale examining how to identify and 75th EAGE Conference and Exhibition Incorporating SPE EUROPEC 2013, SPE-
exploit natural fractures. In: SPE-164524 Presented at SPE Unconventional 164807.
Resources Conference-USA, 10-12 April, the Woodlands, Texas, USA. Renshaw, C.E., Pollard, D.D., 1994. Are large differential stresses required for straight
Dohmen, T., Zhang, J., Barker, L., Blangy, J.P., 2017. Microseismic magnitudes and b fracture propagation paths? J. Struct. Geol. 16, 817e822.
values for delineating hydraulic fracturing and depletion. SPE J. 22 (5), Stephenson, B., et al., 2018. Geometry and failure mechanisms from microseismic in
1624e1634. the Duvernay shale to explain changes in well performance with drilling azi-
Dohmen, T., Zhang, J., Blangy, J.P., 2014. Measurement and analysis of 3D stress muth. In: SPE-189863, Presented at SPE Hydraulic Fracturing Technology
shadowing related to the spacing of hydraulic fracturing in unconventional Conference and Exhibition, 23-25 January. The Woodlands, Texas, USA.
reservoirs. In: SPE-170924, Presented at SPE Annual Technical Conference and Sun, J., Bao, H.Y., 2018. Comprehensive characterization of shale gas reservoirs: a
Exhibition, 27-29 October, Amsterdam, The Netherlands. case study from Fuling shale gas field. Petroleum Geol & Exp 40 (1), 1e12.
Edwards, K.L., Weissert, S., Jackson, J., Marcotte, D., 2011. Marcellus Shale hydraulic Wang, G.C., Carr, T.R., 2013. Organic-rich Marcellus Shale lithofacies modeling and
fracturing and optimal well spacing to maximize recovery and control costs. In: distribution pattern analysis in the Appalachian Basin. AAPG (Am. Assoc. Pet.
SPE-140463, Presented at SPE Hydraulic Fracturing Technology Conference, 24- Geol.) Bull. 97, 2173e2205.
26 January. The Woodlands, Texas, USA. Wang, H., Guo, Y.T., Wang, L., Hou, Z.K., Xu, F., 2017. An experimental study on
EIA, 2017. Marcellus Shale Play Geology Review. U.S. Energy Information Adminis- mechanical anisotropy of shale reservoirs at different depths. Rock Soil Mech.
tration, U.S. Department of Energy. 38 (9), 2496e2505.
Engelder, T., 2008. Structural Geology of the Marcellus and Other Devonian Gas Wang, M.F., Chen, C., Qu, D.P., Miao, Z.W., Xiao, J.L., Gao, H.X., 2015. The geophysical
Shales: Geological Conundrums Involving Joints, Layer-Parallel Shortening characteristics of shale gas reservoir from Wufeng member to Longmaxi
Strain, and the Contemporary Tectonic Stress Field: Field Trip Guidebook. member in Jiaoshiba block of Fuling shale gasfield. Geophys. Prospect. Pet. 54
AAPGeSociety of Exploration Geophysicists Eastern Section Meeting, Pitts- (5), 613e620.
burgh, Pennsylvania, p. 88. Wang, Y.F., Zhai, G.Y., Wang, J.Z., Zhang, Y.X., 2017. Factors influencing gas produc-
Evans, K.F., Engelder, T., Plumb, R.A., 1989. Appalachian stress study 1. A detailed tion effectiveness of Longmaxi formation shale in Sichuan basin and adjacent
description of in situ stress variations in Devonian shales of the Appalachian areas. J. Geomechanics 23 (4), 540e547.
Plateau. J. Geophys. Res. 94 (B6), 129e7154. Wrightstone, G., 2008. Marcellus Shale Geologic Controls on Production: Presen-
Friedrich, M., Milliken, M., 2013. Determining the contributing reservoir volume tation. American Association Petroleum Geologists Eastern Section meeting,
from hydraulically fractured horizontal wells in the Wolfcamp formation in the Pittsburgh, PA, p. 49.
Midland Basin. In: URTeC-1582170, Presented at SPE/AAPG/SEG Unconventional Zagorski, W., Bowman, D.C., Emery, M., Weightstone, G., 2011. An overview of some
Resources Technology Conference, 12-14 August, Denver, Colorado, USA. key factors controlling well productivity in core areas of the Appalachian Basin
Guo, X.S., Hu, D.F., Wei, Z.H., Li, Y.P., Wei, X.F., 2016. Discovery and exploration of Marcellus Shale Play. AAPG Search and Discovery Article #110147 22.
Fuling shale gas field. China Petrol Explor 21 (3), 24e37. Zhang, Y.S., Zhang, J., 2017. Lithology-dependent minimum horizontal stress and in-
Hakala, A., Guthrie, G., 2013. Unconventional Resources Portfolio Annual Report. situ stress estimate. Tectonophysics 703e704, 1e8.
NETL-TRS-UCR-2013, p. 96. Zhang, J., 2019. Applied Petroleum Geomechanics, first ed. Gulf Professional Pub-
Harper, J., Kostrlnik, J., 2010. The Marcellus Shale Play in Pennsylvania, Part 2 Basic lisher of Elsevier, USA, p. 515.
Geology (Photo Courtesy of Gary Lash). Report of Department of Conservation Zinn, C.J., Blood, D., Morath, P., 2011. Evaluating the impact of wellbore azimuth in
and Natural Resources of Pennsylvania, p. 21. the Marcellus Shale. In: SPE-149468, Presented at SPE Eastern Regional
Hu, D.G., Liu, C., 2018. Geological factors of well fracability in Fuling shale gas field, Meeting, 17-19 August, Columbus, Ohio, USA.

View publication stats