Productivity-Maximized Horizontal-Well

Design With Multiple Acute-Angle

Transverse Fractures
Nadav Sorek, Quantum Reservoir Impact; Jose A. Moreno, Anadarko; Ryan N. Rice, Rice Energy; and
Guofan Luo and Christine Ehlig-Economides, University of Houston

Summary
Hydraulic fractures propagate perpendicular to the horizontal-well axis whenever the drilling direction is parallel to the
minimum- principal-stress direction. However, operators frequently drill horizontal wells parallel to lease boundaries, re sulting in
hydraulic- fracture vertical planes slanted at angles less than 90○ from the well axis.
The stimulated-rock-volume (SRV) dimensions are defined by fracture height, well length, and fracture length multiplied by
the sine of the angle between fracture planes and the horizontal-well axis (fracture angle). The well productivity index (PI) under
boundary- dominated flow (BDF) is given by the PI for one fully penetrating fracture multiplied by the number of fractures. An
extension of the unified-fracture-design (UFD) approach for rectangular drainage areas enables determination of the unique
number of fractures that will maximize well productivity under BDF conditions given the formation permeability, proppant
mass, fracture angle, and well spacing. Fracture length and width vary depending on the fracture angle, but the total -propped-
fracture volume remains constant.
Because the likely reason for drilling at an angle to the minimum-stress direction is to better cover a lease area with
north/south and east/west boundaries, the smallest fracture angle will be 45 ○, corresponding to northwest/southeast or
northeast/southwest minimum- stress direction. This results in the need to lengthen fractures by at most 40% to preserve the SRV
for a given horizontal-well length and spacing. For the same sufficiently large proppant mass, this will reduce fracture
conductivity by the same factor. However, because the flow area has increased, the result will be greater well productivity.
This study shows a simple strategy for designing wells to maximize productivity even when not drilled in the minimu m-
stress direction.

Introduction
Industry common perception suggests that horizontal-shale development is enhanced by drilling in the direction parallel to the
local minimum-principal-horizontal stress, rH;min (Wutherich et al. 2013). However, operators often drill parallel to the lease
boundaries to optimize well coverage of a lease. In other cases, the well-trajectory direction may correspond with the need to
bridge lease gaps or cover odd lease-area shapes (Zinn et al. 2011). These practices lead to the creation of hydraulic-fracture
planes with a fracture angle at less than 90 ○ to the well axis.
Extensive research has investigated the optimal design of multiple wells with multiple fractures in shale reservoirs. Many of
the pub- lished works performed numerical optimization of shale-well design (e.g., fracture spacing) by considering only fracture
angles at 90○ to the well axis (Ma 2013; Yu and Sepehrnoori 2013; Plaksina 2015; Rammay and Awotunde 2016; Yang et al.
2017). However, only a few publications also simulated the effects of acute-fracture angles on well productivity (Olorode et al.
2013; Wutherich et al. 2013; Saputelli and Soliman 2014). All these previous numerical experiments compared cases with
constant-fracture-length/width dimen- sions, leading to a decrease in the SRV with a decrease in the fracture angle. As a direct
result, the reported ultimate recovery was supe- rior for a 90○ fracture angle. However, the literature lacks a comparison for
different fracture angles that preserve the same SRV and propped volume. To obtain the condition of maintaining the same SRV,
hydraulic fractures with smaller angles should be longer. In addition, preserving the same proppant mass requires reducing the
fracture width (Figs. 1 and 2, Table 1).

Fracture
Ye Spacing Ye Ye Ye
Well Spacing, Xe

Xf

Fracture Streamlines
Half-Length, Xf No-Flow
Boundaries
No-Flow Streamlines
3D View
Boundaries

Fig. 1—Reservoir and fracture geometry. The pseudosteady-state flow-regime model, modeled after Song and Ehlig-
Economides (2011).

Copyright V
C 2018 Society of Petroleum Engineers

This paper (SPE 170965) was accepted for presentation at the SPE Annual Technical Conference and Exhibition, Amsterdam, 27–29 October 2014, and revised for publication. Original
manuscript received for review 10 January 2017. Revised manuscript received for review 26 December 2017. Paper peer approved 16 February 2018.

2018 SPE 1
Journal
 90°
Xe
Ye θ Ye
Xe

Ye
xf
xf
w 
w
Xe
Ye Xe Ye

Legend: Well Fracture

Well-drainage area

Fig. 2—Model settings and parameters for different drilling scenarios, with the same SRVs and same propped-fracture volumes for
perpendicular and acute-angle transverse fractures. Well spacing Xe and distance between fracture-creation points Ye stay the
same in both cases. In the perpendicular case, drainage-area length Xe0 and width Ye0 are equal to Xe and Ye, respectively. Parame-
ters are defined in Table 1.

Property Perpendicular Fractures Acute-Angle Fractures Comments

Well-fracture angle 90° θ θ ≤ 90°, also well to σH,max angle
Well spacing Xe Spacing between laterals
Fracture spacing Ye Spacing between fracture-creation points
Net-pay thickness h Equal to fracture height hf
Drainage-area length Xe Xe Xe = Xe/sin(θ)
Drainage-area width  Y  = Y /sin(θ)
Y Y
e e e e
Effective-fracture spacing
SRV per well SRV = X × Y × h SRV = X  × Y  × h Same SRV
e e e e
SRV = SRV
xf = xf/sin(θ)
Fracture half-length xf xf Fully penetrating fractures:

 X e  2xf , X e  2xf

Fracture width w w w = wsin(θ)

Fracture propped volume V=x×w ×h V  = x  × w × h Same economic investment in proppant:
f f f f V = V 
f f

Drainage-area aspect Ar = Ar sin2(θ)

ratio Ar = Xe/Ye Ar = Xe /Ye Xe Y e
θ arcsin
Ar

Table 1—Model-parameter definitions and relations.

Hydraulic-Fracture Angle. Geomechanics and Completion Considerations. To obtain the largest reservoir contact (for an
assumed maximum-achievable hydraulic-fracture length), many operators drill horizontal wells in unconventional reservoirs
along the direction of the minimum horizontal stress. However, we note several geomechanics and completion considerations
that favor drilling in a direc- tion that reduces the angle between the well axis and fracture planes.
Drilling a horizontal well changes the stress regime locally near the wellbore, where the resulting conditions may favor
starting a longitudinal hydraulic fracture along the wellbore axis (Daneshy 1973). When the horizontal-drilling direction is
aligned with the maxi- mum horizontal stress, the fracture continues propagating longitudinally farther from the well. In contrast,
when the drilling direction has some angle with respect to the maximum horizontal stress, the fracture reorients toward the f ar
field in-situ stress where the most- severe reorientation occurs at an angle of 90 ○ (Abass et al. 1996). Such geometry is
susceptible to both premature screenouts (Abass et al. 2009) and enhanced tortuosity (Lecampion et al. 2013). The former
prevents significant fracture growth, and the latter hinders the flow toward the wellbore, and both impede well productivity.
Though axial creation is more common in openhole completion (Daneshy 2011), the tendency still exists in cased -hole
completions because of the so-called “pseudo-openhole” effect (Daneshy 2011, 2013). In addition, as the angle between the well
and the largest horizontal stress increases, the required wellbore pressure to break the rock (crea- tion pressure) increases (Hsiao
1988; El Rabaa 1989; Lecampion et al. 2013), leading to additional operational and equipment costs. Furthermore, a smaller
fracture angle increases the contact area with the wellbore, which in turn reduces the choke-skin effect (Wei and Economides
2005).
Field Observations. Two field cases in the literature offer mixed conclusions regarding well performance for different fracture
angles. Cui et al. (2013) addressed the effects of drilling at an acute angle to the rH,min direction on production performance. They
investigated field-production data from several unconventional formations in the Western Canada Sedimentary Basin, where
numerous wells were not drilled in the commonly assumed optimal rH,min direction. In turn, they discovered superior
performances for wells drilled at an angle with the minimum-stress direction. They attributed these results mainly to strong (in-
situ) permeability anisotropy induced by the pre-existing natural fractures and to the near-wellbore tortuosity effect.

2 2018 SPE
Journal
 On the other hand, Zinn et al. (2011) used proprietary and public data to address the effect of wellbore azimuth on well
performance in the Marcellus Shale. The Zinn et al. (2011) regression results revealed that wells drilled along a direction different
from the minimum- horizontal-stress direction were suboptimal, with somewhat-weak statistical significance.
Both cases discussed compared only well orientation and not completion data. Logically, the optimal configuration depends
not only on the well azimuth but also on completion and reservoir parameters such as fracture spacing, the amount of proppant
used, and reser- voir permeability.

The Scope of the Paper. The objective of this work was to apply fundamentals to evaluate the actual effect of an acute fract ure
angle on well productivity. We used the UFD methodology (Economides et al. 2002) that indicates th e hydraulic-fracturing-
treatment design that maximizes well productivity for any set of reservoir and proppant properties and an injected proppant m ass.
This methodology introduced the concept of a dimensionless proppant number (Np) as the permeability-weighted ratio of
propped-fracture volume to square well drainage volume. Daal and Economides (2006) and Sabaev et al. (2006) extended
the definition of Np to elongated- rectangular-reservoir drainage volumes.
This study applies and further extends the UFD approach. The need to consider lower reservoir permeability than envisioned
by the original UFD models explains the motivation for modeling proppant numbers larger than the maximum value of 100
found in pre- vious applications.
The outcome of this paper is a simple framework for a well design incorporating hydraulic fractures with a fracture angle les s
than 90○. Throughout the paper, we make several assumptions that render our approach relatively simple, such as homogeneous
and isotropic permeability and planar and symmetrical fractures with respect to the well axis. We emphasize that these
assumptions are reasonable during the design phase of unconventional plays, where, unlike the execution phase, more -detailed
and -accurate data are often not available. Eventually, the framework provides a rule of thumb for multiple-transverse-fracture
horizontal-well (MTFHW) designs.
This paper is structured as follows. First, we describe our methodology of relating the drainage -area aspect ratio to various
design variables, and we introduce new interpretations and extensions for the previous UFD work. Then, we describe the result s
of our numeri- cal study and their implications. Last, the applicability of those results is demonstrated and discu ssed in example
applications.

Methodology
In this section, we establish a correlation between the dimensionless proppant number (Np) and the drainage-area aspect ratio, A0r
Xe0 =Y
¼ e0 , with Xe0 being the length drained by one fracture and Ye0 being the distance between two adjacent fractures. This
correlation eventually produces an expression relating the proppant number Np and the optimal fracture spacing that corresponds
to maximum well productivity for a given proppant number and well azimuth. We continue by expanding the previous UFD
models to cover very-low- permeability applications and by describing the implemented method for doing so.

Rectangular SRV Drainage Area. Following the methodology presented by Song and Ehlig-Economides (2011), this work
models an MTFHW by considering each fracture as fully penetrating a rectangular drainage area. Fig. 1 shows schematic
streamlines during a pseudosteady-state (PSS) flow regime, where no-flow boundaries are at the fracture tips and the interference
planes between two adja- cent fractures, forming a rectangular drainage area.

Preserving Shale Stimulated Volume and Proppant Mass While Changing the Angle. Fig. 2 illustrates the base case that
describes a horizontal well drilled in a direction parallel to the minimum horizontal stress, with hydraulic fractures propagating
perpendicular to ¼the well axis (h 90○). As the horizontal-well-trajectory azimuth deviates from this base-case scenario, fracture
angles decrease (h < 90○).
When the well path deviates from the minimum horizontal stress, the angle at which the hydraulic fracture extends will begin
to decrease. To preserve the same SRV, one should change the fracture design by adjusting fracture width and half -length while
conserv- ing proppant mass. These changes are subject to fracture-design parameters (e.g., rock Young’s modulus, shear modulus,
bulk modulus, and the Poisson ratio), and thus to practical limitations such as maximum-attainable fracture half-length.
Note that in Fig. 2 we introduce two new drainage-area parameters (Ye0 and Xe0 ), which are a function of the angle at which the
frac- ture deviates (for this work, the prime symbol 0 denotes deviated-fracture-angle parameters]. Specifically, Ye0 denotes
effective fracture spacing, whereas Ye denotes distance between fracture-creation points along the wellbore (which we call in this
paper the fracture spac- ing). Also, Xe0 denotes the fracture length, and Xe denotes the well spacing. In addition, the well and
fracture spacing (Xe and Ye, respec- tively) remain constant, irrespective of the fracture angle. Next, assuming the same capital is
invested in the fracture treatment, we set the proppant-mass constant and thus maintain the same proppant number. As a result,
fracture half-length (x0f ) and width (w0 ) become a function of the fracture angle.
The geometric relations between the parameters are
Xe xf
X0 ¼ ; Y 0 ¼ Y · sinðhÞ; x0 ¼ ; w0 ¼ w · sinðhÞ: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð1Þ
e e e f
sinðhÞ sinðhÞ

Note that fracture area and stimulated area are conserved:

Xe0 · Ye0 ¼ Xe · Ye ; x0f · w0 ¼ xf · w: ............................................................. ð2Þ

Next, two aspect ratios are

defined:
Xe Well spacing ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð3Þ
A
r¼ ¼
Ye Fracture spacing
0 Drainage-area length
X
A0 ¼ e ; ¼
Fracture ¼ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð4Þ
length
r Ye0 Effective fracture Drainage-area width
spacing
where Ar and A0r are the aspect ratio of the base case (h¼ 90○), which is also the ratio of well/fracture spacing, and the modified
case (h < 90○), respectively.

2018 SPE 3
Journal
 Reorganizing Eq. 1 yields
Xe
¼ A0 sin2 ðhÞ: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð5Þ
r
Ye
From Eq. 5, the fracture angle is defined as
sffiffiffiffiffiffiffiffiffiffiffiffi!
Xe =Ye
h ¼ arcsin : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð6Þ
A0r
Table 1 summarizes the parameter definitions in our conceptual model as well as their relations.

PSS PI. The instantaneous PI is defined as

qðtÞ qðtÞ
Jð Þ¼ t ¼ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð7Þ
DpðtÞ pðtÞ— pwf ðtÞ
whereas the dimensionless PI in oilfield units is
141:2Bl
JDðtÞ¼ JðtÞ.STB Σ : ................................................................................................................................ ð8Þ
kh
D × psi

Instantaneous PI tends to decrease with time until reaching a constant value when the flow regime changes from transient
flow to PSS flow (or equivalently, BDF). We shall denote the PI at PSS when reaching a constant value as

JDPSS ¼ JDðt ≤ t PSS Þ¼ Constant;................................................................................................................................... ð9Þ

where tPSS is the time when PSS flow regime starts; i.e., the flow stabilizes and all boundaries affect the flow (Lee et al. 2003). In
this flow regime, the dimensionless PI JDPSS becomes constant (for single-phase flow).
Using JDPSS as a performance indicator is appropriate for unconventional plays. Although initially most of the wells in
unconven- tional reservoirs exhibit transient-flow behavior, BDF develops once the flow to adjacent fractures begins to interfere
(Song et al. 2011). For example, Yan et al. (2014) examined Bakken MTFHWs and found that 66% of the wells spudded in 2006
have reached BDF. In a different study, Holanda et al. (2017) calibrated physics-based decline curves to 992 Barnett fractured
horizontal wells for estimated-ultimate-recovery (EUR) estimations. According to personal communication with the author, 1 for
more than 98% of the wells, most of the EUR is attributed to the BDF period.
For MTFHWs with effectively infinite-conductivity fractures, Song et al. (2011) indicated that BDF occurs approximately
when pseudolinear flow ends. The pressure investigation for linear flow is twice the distance computed from the equation for the
radius of investigation during cylindrical flow (Palmer and Moschovidis 1990). Therefore, the pseudolinear flow ends at the time
in hours given by
2
948/lct ðY2e Þ 59:3/lct Y 2e
teplf = ¼ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð10Þ
4k k
Fully developed BDF occurs at a time equal to or less than 0.1 teplf · . Thus, for typical tight oil fluid and rock properties (/
× psi ), and commercial shale-gas and tight oil permeability ranging between 0.2 and 1.1 md (Roth
—5 —1
¼ l 1 cp;
0:06; ¼ ct 1 ¼ 10
2010), the time to BDF ranges from 32 to 178 days for a 50-ft fracture spacing reported as greater than typical spacing by Wu
and Olson (2016). Although in early wells the transient-flow regime persisted for a long period because of the large fracture
spacing, a good design will minimize the transient period to drain the stimulated area effectively and to enhance early cash flow.
The previous discussion fully justifies the use of JDPSS as a well-design criterion, even for low-permeability reservoirs.
Maximizing JDPSS is the foundation behind the UFD approach. The next subsection describes briefly the UFD approach and the
necessity to extend existing models to the lower-permeability range encountered in tight oil and shale-gas formations.

UFD: Previous Results and New Interpretations. Daal and Economides (2006) adjusted the proppant-number definition for
irregu- larly shaped drainage areas as
2kf Vf 2kf ð2xf · w · hÞ xf Xe 4xf2!. kf wΣ. Xe Σ Xe
Np ¼ kV ¼ kðXe · Ye · hÞ × . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð11Þ
Ye ¼ Ix CfD Y ;
¼ 2
res xf Xe Xe2 kxf
e

where kf and k are the proppant- and reservoir-permeability values, respectively. Vf and Vres are the fracture and reservoir
volumes, while w and Ye are the fracture and stimulated-reservoir width, respectively. xf and Xe represent fracture half-length and
stimulated- reservoir-area length, respectively. Reservoir thickness and fracture height are represented as h and are assumed
to be equal. The
penetration ratio ð2x =X Þ is represented as I . Last, the fracture dimensionless conductivity is represented as C ¼ kf w. Note that
f e x fD
kxf
Eq. 11 follows the original nomenclature in Daal and Economides (2006), which assumes a perpendicular fracture case. For a
general- angle fracture case, one might replace the appropriate variables with x0f , w0 , Ye0 , and Xe0 .
Economides et al. (2002) showed that proppant number correlates to a unique maximum PSS (BDF) PI that optimizes the
fracture half-length and width for a given proppant mass. By relating to the proppant mass, which represents most of the
negotiable investment in the well (because drilling must achieve the formation depth, with or without proppant), the UFD
approach automatically addresses economics. Daal and Economides (2006) and Sabaev et al. (2006) extended this concept from
square to rectangular drainage areas with increasingly high aspect ratios. Fig. 3 illustrates three different aspect ratios from Daal
and Economides (2006), as well as the relation between CfD, Np, and JDPSS .
1
Personal communication with R. W. Holanda. 2017. Interview, Harold Vance Department of Petroleum Engineering, Texas A&M University.

4 2018 SPE
Journal
 2.00 1,000
3.50
1.80 Ar = 1 Np = 100 3.20
Ar = 2 Np = 100 Ar = 4
Np = 60 Np = 60 Np = 100
1.60 2.90
Np = 30 Np = 60
2.60 Np = 30
1.40 Np = 10 Np = 30
Np = 6 2.30
1.20 Np = 10
PSS

PSS

PSS
Np = 3 2.00 Np = 10 lx = 1 Np = 6
100
JD

JD
lx = 1

JD
1.00 1.70 Np = 6 Np = 3
1.40 Np = 1
0.80 Np = 1 Np = 3 Np = 0.6
lx = 1 Np = 0.3
Np = 0.6 1.0
0.60 Np = 0.1
Np = 0.3 0.80 Np = 1
Np = 0.1 Np = 0.6
0.40 0.50 Np = 0.3
Np = 0.1
0.20 0.20 0.10
0.1 1 10 100 10 0.1 1 10 100 1,000 0.1 1 10 100 1,000
CfD CfD CfD

Fig. 3—UFD-type curves for aspect ratios equal to 1, 2, and 4, modeled after Daal and Economides (2006).

2xf =Xe 1 =
The dashed lines in Fig. 3 represent fully penetrating fractures, where the penetration ratio approaches unityðIx ¼ : Þ
The red lines pass through the maximum PI, JDPSS;max , for each proppant number and each aspect ratio.
Next, the JDPSS;max values, lying on the red lines shown in Fig. 3, are combined in Fig. 4a (Daal and Economides 2006) and in
Fig. 4b (Sabaev et al. 2006) to relate each combination of aspect ratio and proppant number to its maximum well productivity.
Note that the two independent works shown in Fig. 4 limited the analysis to proppant numbers not greater than 100.

10
8
Np = 100
Np = 100
Np = 60 7
Np = 30

Np = 10
6
Np = 6
JD PSS,opt
Np = 3 5 Np = 50
PSS,max

PSS,max

Np = 1 40
1 Np = 0.6 4
Np = 0.3 30
JD

JD

Np = 0.1 20
3

10
2
5

2
1

0.1 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1 10
1/Ar = Ye/Xe Ar = Xe/Ye

Fig. 4—Maximum dimensionless PSS PI for different proppant-number values ranging from 0.1 to 100. (Left) The horizontal axis is
1/A0r (Daal and Economides 2006). (Right) The horizontal axis is A0r and the green line that passes through the optimum of each
proppant number is denoted as JDPSS;opt (Sabaev et al. 2006).

Optimal-Fracture Spacing for Varying Angle. The green line in Fig. 4 (right) implies that one can derive a unique relation
between proppant number and optimal-aspect ratio. Thus, one can define an optimal-aspect ratio for a rectangular drainage area,
which we shall denote as ðA0r Þopt. Expressing Eq. 5 in terms of optimal-aspect ratio results in
. XeΣ
¼ A0r opt sin2ðhÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð12Þ
Ye opt
:
Eq. 12 is an important result of this work. Given the angle, h, between the maximum-horizontal-stress direction and the convenient
well azimuth, and given an optimal-aspect ratio, A0r opt , associated with a specific proppant number, Eq. 12 indicates the optimal-well/
. Σ
Xe
fracture-spacing ratio, , which maximizes the well productivity. Further, assuming that the operator will want to preserve the
Ye op
t
well spacing, we arrive at the following expression for optimal-fracture spacing:
Y ¼ X e
: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð13Þ
eopt
A0r opt sin2ðhÞ
Because the fracture length is simply dependent on the given well spacing, we can write the optimal fracture length as
Xe0 opt ¼ 2xf Xe
¼ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð14Þ
sinðhÞ
;
and by inserting Eq. 14 into Eq. 13, we obtain
Xe Ye0 opt
Xe
: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð15Þ
Yeopt ¼ 0 ¼
X eopt 2
Xe ¼
h sinðhÞ sinðhÞ
0
Yeopt sin ð Þ sin 2
h
ðÞ
Ye0
opt

2018 SPE 5
Journal
 Note that for the case of h ¼ 90○ , we obtain Ye opt ¼ Ye0 opt . In other words, by dividing the optimal effective fracture spacing
Þ sin h , one can obtain the optimal (actual) fracture spacing. Thus, for smaller degree angles, Eq. 15 indicates that a
(Ye0 opt )ð by
wider fracture spacing will maximize the well productivity.
We now make the essential observation that if the objective is to align wells in a north/south or east/west direction, the an gle h
should never be less than 45○. For example, if the well azimuth that aligns with the minimum-stress direction is 330○ from the
north, drilling wells in the north/south direction would result in a 60○ fracture angle, but p drffiiffilling wells in the east/west direction
would
in a 30○result
fracture angle. Eq. 14 indicates the optimal-fracture length will be 3 for a north/south well and 2X for the e
2X e= Therefore, the better choice will be to drill north/south wells with fractures east/west
well. approximately 15% longer than fractures from a
well drilled in the minimum-stress direction.
The extreme case of a 45 fracture angle, (e.g., when a minimum-stress direction is 315○ from the north) implies that wells
○
can be drilled north/south or east/west, with the need to increase fracture half-length by approximately 40% in either case. If the
convenient well azimuth is more than 45 ○ from the minimum-stress direction, the increased fracture length required to preserve
the SRV may not be operationally feasible.
A final important observation from Fig. 3 is that for high proppant numbers, the maximum productivity ( JDPSS;max ) coincides
with full penetration of the reservoir by the fractures
= (Ix 1); that is, dashed lines coincide with red lines. Therefore, for a specific
drainage area and high proppant numbers, the maximum BDF PI is associated with full-reservoir penetration. Because we
consider high proppant num- bers (100 and greater) to model low-permeability reservoirs, it is sufficient to model a fully
penetrated drainage area to obtain JDPSS;max .
Unique relations between proppant numbers and optimal-aspect ratios exist for proppant numbers up to 100. Modern MTFHWs
encounter aspect ratios and formation-permeability values that imply a need to develop analogous relations for proppant numbers
greater than 100. The following subsection extends the UFD coverage to proppant numbers reaching 1 million.

Modeling Methods. For modeling purposes, a single-phase numerical simulator (for simulating liquid and real-gas flow) was
used to simulate flow in hydraulically fractured reservoirs (Lee and Wattenbarger 1996). To reduce computation time, we used
symmetry and simulated one-fourth of the drainage area by considering one-half of the fracture length and width. Each
simulation was run under constant-pressure production and stopped when the PI reached a steady value at PSS. To simulate flow
from the matrix to the fracture and from the fracture to the well, we applied local grid refinement, both toward the fracture and into
the well (in both the x- and y-coordinate), by using equally spaced grids on a logarithmic basis (or equivalently, geometrically
spaced). Hydraulic fractures fully penetrated the drainage area. Simulations considered aspect ratios ranging from 1 to 2,000 and
proppant numbers ranging from 100 to 1 million.

Results
In this section, we present the simulation results, which extend previous UFD published works for high proppant numbers. Fig. 5
shows the relation between PI at PSS conditions and aspect ratio for different proppant numbers. For each proppant number,
there is a distinc- tive concave-like function, which exhibits an absolute optimal JDPSS;max value. To clarify, each given point on the
concave curves in Fig. 5 would be one of the maximum values on a JDPSS vs. CfD plot for a specific aspect ratio (red curve on
Fig. 3); therefore, the vertical axes in Figs. 4 and 5 read JDPSS;max . The straight line in Fig. 5 intercepts the corresponding maximal
JDPSS;max value for each proppant num- ber. We denote this straight line as a new function, JD PSS;opt A0r Np opt , which represents the
½ðÞ]
optimal JD PSS;max for a given proppant number and its corresponding optimal aspect ratio.

1,000
 ) – 0.3262
JDPSS,opt = 0.8087 (A
Np = 106
r opt
R 2 = 0.9999
Np = 105
100
PSS,max

Np = 104
JD

Np = 103
10
Np = 102

1
1 10 100 1,000
Aspect Ratio (Ar)

Fig. 5—Extended UFD simulation results, with the relation between dimensionless PSS PI and aspect ratio for varying proppant
numbers. The results are an extension of the works shown in Fig. 4 for high proppant numbers. A linear function passes through
the optimum of each proppant number and is denoted as JD PSS;opt . R 2 5 coefficient of determination.

½ ðAÞ0r ] Np opt appears to be linear with respect to A0r opt (see linear equation in Fig. 5). Note that
Interestingly, the function JD PSS;opt
A0r opt is a function of Np, and that we0 write Ar opt rather 0 than A
ðÞ r opt Np for brevity.
For Np¼ 100, the Ar value corresponding to the maximum JDPSS;max is 7. This value matches the Sabaev et al. (2006) results
shown in Fig. 4. Daal and Economides (2006) did not predict a maximum value ¼ for Np 100 because none of their simulations
considered aspect ratios between 5 and 10, thereby missing the maximum point.
We note that these PI values decrease for horizontal wells because of the choke-skin effect (Wei and Economides 2005).
However, the magnitude of the choke-skin effect is known to be negligible, and in any case, the overall trend remains the same.
Next, it would be useful to express A0r opt explicitly as a function of Np (Fig. 6). We obtained this correlation by simply taking
the pairs (A0r ; Np) of each optimal point in Fig. 5 and applying a regression. Interestingly, the obtained function appears to be
linear on a log-log plot. A practical approximation for this function would be
rffiffiffiffiffi
Np
A0 rop ¼ 0:7099 × Np0:5003 = ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð16Þ
t 2

6 2018 SPE
Journal
and by inserting Eq. 16 into Eq. 13, we obtain the empirical expression of optimal-fracture spacing that will maximize PSS PI
for a given proppant number, well spacing, and fracture angle,
sffiffiffiffiffi
2 Xe : ........................................................................................................................................... ð17Þ
Y eopt ¼
Np sin2ðhÞ

1,000
Ar p(N  0.5003
opt) = 0.7099 N p
R 2 = 0.9999

Optimal Aspect Ratio

100

10

1
100 1,000 10,000 100,000 1,000,000
Proppant Number

2
Fig. 6—Correlation between the optimal aspect ratio and proppant number. R 5 coefficient of determination.

Application Examples
In this section, we discuss the possible implications when applying the concept of optimal aspect ratio. The discussion will be
depend- ent on an arbitrary square lease with different design configurations of multiple wells with multiple fractures. We
assume for all cases the same maximum feasible fracture half-length. This is an example where aligning wells along the lease
boundaries might lead to a more-efficient well configuration than drilling along the horizontal-minimum-stress direction.

Example 1. Let us assume a square lease in which the minimum-horizontal-stress direction is 330 ○ from the north (Fig. 7), and
with a homogeneous and anisotropic stress field that promotes a growth of hydraulic-fracture planes evenly from both sides of
the well. We shall assume that the maximum-achievable fracture half-length is 577 ft, and the target zone was found in a total
vertical depth of 5;000 ft.

0
330 30

300 σ H,min 60
3,000 ft

270 90

240 120

210 150
180
3,000 ft

Fig. 7—Lease dimension, top view. The green line represents the principal-minimum-horizontal stress (rH;min).

Drilling Parallel to Lease Boundaries vs. Drilling Parallel to the Minimum Stress. We consider three cases. In the first case , we
consider drilling north/south wells parallel to lease boundaries (Case A). In the next two cases, Cases B and C, we consider
drilling wells parallel to the minimum horizontal stress. Fig. 8 shows these well configurations for the assumed maximum -
achievable fracture length. In Case B, the goal is to cover a stimulated area that is equal to the lease area, and in Case C the goal
is to cover as much area as possible by extending each well from one boundary to the other.
Although three wells are sufficient to drain the total lease area (9 million ft 3) in Case A, three wells are not sufficient to drain
the total lease area in Case B. Note that although the total stimulated area is bigger in Case C than the total lease area of 9 million
ft2, total measured depth (MD) is considerably larger because of the additional vertical segment of the additional fourth well.
Although added SRV can lead to a greater EUR, the 17.8% increase in MD in Case C leads to only a 5.6% increase in t he SRV
compared with the other two cases. We note that operators probably would not consider implementing the two short wells in
either Case B or Case C because their short laterals would not justify drilling to the target depth of the formation. This is an
example where aligning wells along the lease boundaries might lead to a more-efficient well configuration and area coverage
than drilling along the horizontal-minimum- stress direction.
Next, we show how to obtain optimal productivity for each case by adjusting the fracture spacing. We determine the optimal
fracture spacing using the proppant number computed with the assumed reservoir and proppant properties shown in Table 2.

2018 SPE 7
Journal
 Total stimulated area: Total stimulated area: Total stimulated area:
9 million ft2 9 million ft2 9.5 million ft2
Total lateral length: Total lateral length: Total lateral length:
9,000 ft 7,794 ft 8,260 ft
Total MD: Total MD: Total MD:
24,000 ft 27,794 ft 28,260 ft

(a) (b) (c)

Fig. 8—Drilling horizontal wells and the resulting SRVs. The blue square is the lease area; the black arrows are horizontal-well seg-
ments; and the shaded orange is the SRV. (a) Three laterals parallel to the lease boundaries are sufficient to drain the lease area.
(b, c) Drilling wells parallel to the minimum horizontal stress. In (b), four wells are required to drain the lease area. In (c), four wells
drain more than the total lease area. The total MD is larger in Cases B and C. In all three cases, all fracture lengths are the same
(577 ft).

Property Description Value

k Reservoir permeability 5×10–4 md
kf Proppant permeability 10,000 md
SG Proppant specific gravity 2.65
p Proppant porosity 0.38
h Reservoir height 100 ft
Mp Proppant mass 1,000,000 lbm

Table 2—Reservoir and fracture properties for case studies.

The propped volume Vp is

Mp 1; 000; 000
Vp ¼ ¼ ¼ 9;754 ft3 ;
q pð62:4Þð1 — /p Þ 2:65ð62:4Þð1 — 0:38Þ
and the reservoir volume (VrÞ for Cases A and B is
Vr ¼ Ah ¼ 9 × 106 × 100 ¼ 9 × 108 ft3:
The proppant number for Cases A and B can be calculated from Eq. 11:
2kf Vp 2ð10; 000Þ 9;754
Np¼ ¼ ¼ 433:5
k Vr 5 × 10—4 9 × 108
Then, Eq. 16 can obtain an optimal aspect ratio:
rffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffi
Np 433:5
ðAr0 Þopt ¼
2 ¼ 2 ¼ 14:72
The well spacing in Case B is equal to the total fracture length, X
¼e 2xf ¼ 2 × 577 ¼ 1;154 ft. Then, using Eq. 17, the
optimal fracture spacing is
sffiffiffiffiffi
Y90
○
2 Xe ¼ 1 1154
¼ 78:4 ft;
eoptðCase BÞ ¼
Np 14:72 sin2ð90○Þ
sin2ðhÞ
and for Case A the optimal fracture spacing can be computed from Eq. 1,
○
○ Y90
eopt
Ye60 ¼ ¼ 90:5 ft:
optðCase
AÞ
sinð60○Þ

As a result, the fracture spacing is wider in the 60 ○ case.

8 2018 SPE
Journal
 We note that the total number of fractures stays approximately the same in both cases (99 fractures for the whole lease in ea ch
case), and the total lateral length in Case B is smaller than in Case A. However, depending on the formation depth, the total
drilled length for Case B may be longer or even much longer because this requires an additional vertical-well segment.
The stimulated area in Case C is bigger than the other cases, and using the same proppant mass leads to a smaller proppant nu mber:
¼
Np 409:06.
The optimal fracture spacing in this case is
○
rffiffiffiffiffiffiffiffiffiffiffiffiffiffi 115
Y90 2 ¼ 80:7 ft;
¼ 4
eoptðCase BÞ 409:06 sin ð90 Þ2 ○

and the total number of fractures in the entire lease is 8; 260=80:7 = 102.

Productivity Comparison. The equation found in Fig. 4 (right) facilitates the approximation of a dimensionless PI for a sing le
frac- ture. For the cases with the same proppant number, Cases A and B, the productivity per a single fracture is
rffiffiffiffiffi r ffiffiffiffiffi ffiffiffiffiffiffiffiffi
Np 433:5
JDPSS ¼ 0:8087 × — 0:3262 ¼ 0:8087 × — 0:3262 ¼ 11:58;
2 2
and the single-fracture productivity for Case C is
rffiffiffiffiffi r ffiffiffiffiffi ffiffiffiffiffiffiffiffiffiffiffi
Np 409:06
JDPSS ¼ 0:8087 × — 0:3262 ¼ 0:8087 × — 0:3262 ¼ 11:239:
2 2
To account for the whole-lease productivity, we compute JDPSS ¼ 99 × 11:58 ¼ 1; 146:4 for Cases A and B and JDPSS¼ 102 ×
11:239 ¼1; 146:4 for Case C. Therefore, the overall lease productivity is the same in all three cases, assuming the same investment
in proppant but more investment in drilling in Cases B and C.

Example 2. Let us assume now a much-larger lease area, which requires considerably more wells than in Example 1. We will
assume that the length of each side of the square lease is 15,000 ft, which leads to a reservoir area of 5,165 acres. We will assume
a proppant mass of 10 million lbm to be used in the entire lease. In addition, we assume a maximum lateral length of 7,500 ft.
Other assumptions are the same as in Example 1.
Next, we compare two cases, one with the drilling direction parallel to the lease boundaries and one drilling parallel to the
mini- mum-principal-stress direction. Like Example 1, the goal is to find a drilling strategy that would drain the lease area with
the smallest number of wells and drilling length.
Fig. 9 shows that 30 wells are required to cover the lease area when drilling parallel to lease boundaries, and Fig. 10 shows
that 35 wells are required when drilling parallel to the minimum-principal-stress direction. In Fig. 10, some of the wells might not
be worth- while to drill because of their very short lateral lengths. Thus, in this case, drilling parallel to the minimum-principal-
stress direction might not cover the entire lease area.

Fig. 9—Example 2 Case A drilling strategy, parallel to the lease boundaries. 30 wells are required to cover the lease area.

2018 SPE 9
Journal
Fig. 10—Example 2 Case B drilling strategy, parallel to the minimum-principal-stress direction. 35 wells are required to cover the lease
area. Wells with maximum lateral length are shown in red; wells with laterals smaller than maximum length are shown in yellow.

Table 3 summarizes the PI calculation steps for the two cases. To cover the entire lease area, Case B requires spudding five
more wells. As in Example 1, because both cases drain the same reservoir area and use the same proppant mass, they share the
same proppant number, and thus the optimal aspect ratio that leads to the optimal PI is also the same.

Case A Case B
Well/fracture angle (degrees) 60 90
Total lateral length (ft) 225,000 194,855
Total number of wells 30 35
Reservoir area (ft2) 2.25×101 2.25×101
Proppant number Np (dimensionless) 173.4 173.4
Well spacing (ft) 1,000.0 1,154.7
Optimal aspect ratio (dimensionless) 9.3 9.3
Effective fracture spacing (ft) 124.0 124.0
Actual fracture spacing (ft) 143.2 124.0
Number of fractures 1,571 1,571
PI for fracture (dimensionless) 7.2 7.2
PI total for lease (dimensionless) 11,317.3 11,317.3

Table 3—Results summary of Example 2.

Regarding total MD in this Example 2 case, we note that any formation depth deeper than 6,000 ft leads to a larger MD in
Case B. Any formation depth smaller than this number leads to a larger MD in Case A.
As a conclusion, the drilling strategy should be individually evaluated for each new scenario, and drilling parallel to the
minimum- principal-stress direction should not be considered as the only choice to develop an unconventional asset.

Discussion
Considering Heterogeneity. Several hydraulic fractures may be created simultaneously from several perforation clusters in a
stage. The clusters may be located in opposite intervals with similar rock-mechanics properties, and stages or clusters may target
particular formation properties. This may often lead to cluster spacing that is not constant along the well (Cipolla et al. 2011;
Rahim et al. 2015; Shahri et al. 2015). Although we assumed constant fracture spacing along the well in the given example, th e
same analysis can easily be performed individually for each stage.

10 2018 SPE
Journal
Considering Geomechanics. We refer readers concerned about the viability of well designs featuring parallel fractures to seve ral
key points. Warpinski et al. (1998) described fracture complexity observed in field data, but Daneshy (2003) considered off-
balance fractur- ing, a term used to describe various mechanisms leading to near-wellbore complexity, and explained ways to
avoid it. Further, Bunger et al. (2012) provided a work flow to suppress fracture curving caused by the stress-shadow effect by
adjusting design parameters (such as fracture spacing, fluid viscosity, and injection rate) that can promote planar - and parallel-
fracture growth. Peirce and Bunger (2015) used a 3D model that captures the transition from radia l to Perkins-Kern-Nordgren
fracture geometries (Perkins and Kern 1961; Nordgren 1972) to show that adjustment of the location of perforation clusters can
further mitigate the stress-shadow effect and create parallel and planar fractures. Wu and Olson (2016) discussed that nonuniform
growth of multiple fractures is attributed not only to the stress-shadow effect but also to the dynamic partitioning of flow rate
within the perforation cluster. They showed through a 3D displacement-discontinuity method how to design the right friction of
each perforation that promotes uniform-fracture growth.
Such technological advancements make it reasonable to assume that fracture dimensions and configurations will become
progres- sively more controllable by engineering design (up to practical limitations). This assumption facilitates the comparison
of cases with multiple planar and parallel hydraulic fractures that grow in different angles but with the same SRV and with the
same investment in proppant. In these settings, different fracture angles would offer different well -design options. On the one
hand, maintaining the same stimulated volume with the same number of fractures results in acute-angle fractures that are longer
and closer to each other than in the perpendicular case, such that the flow-interference planes between two fractures become
nearer. On the other hand, maintaining the same proppant volume requires diminishing fracture width as the angle becomes
smaller, such that the smaller the fracture angle, the smaller the fracture conductivity. Thus, finding an optimal aspect ratio to
different well/fracture angles translates into seeking an optimal tradeoff between three factors: reservoir-contact area, distance to
flow boundaries, and fracture conductivity.

Conclusions
This study applies UFD principles to determine the adjustments in fracture half-length and spacing that will maximize the well
produc- tivity when the well is not drilled in the minimum-stress direction. We found that as long as the MTFHW drains the same
SRV, a well with acute-angle fractures can perform as well as or better than a well with fracture planes perpendicular to the well
axis. We show how the work can be applied to effectively saturate a lease with multiple MTFHWs drilled parallel to the lease
boundaries.

Nomenclature
A0ropt ¼ optimal drainage ratio, fraction
B ¼ formation volume factor, res
bbl/STB CfD ¼ fracture conductivity,
dimensionless H ¼ reservoir/fracture
thickness, ft
Ix ¼ penetration ratio, fraction
J ¼ PI, STB/D/psi
JD ¼ PI, dimensionless
JDPSS ¼ PI at PSS, dimensionless
k ¼ formation permeability,
md kf ¼ fracture permeability,
md nf ¼ number of fractures
pwf ¼ well flowing pressure, psia
p ¼ average pressure in the reservoir, psia
q ¼ flow rate, STB/D
Vf ¼ fracture volume, ft3
Vres ¼ reservoir volume, ft3
w ¼ fracture width, ft
w0 ¼ acute-angle-fracture width, ft
xf ¼ 90○-fracture half-length, ft
x0f ¼ acute-angle-fracture half-length, ft
Xe ¼ well spacing, ft
Xe0 ¼ acute-angle-fracture length, ft
Ye ¼ perforation (actual) spacing, ft
Ye0 ¼ distance between two acute-angle fractures, effective spacing, ft
h ¼ angle, degrees
l ¼ viscosity, cp
/ ¼ porosity, fraction
/p ¼ proppant porosity, a fraction

References
Abass, H. H., Soliman, M. Y., Al-Tahini, A. M. et al. 2009. Oriented Fracturing: A New Technique to Hydraulically Fracture an Openhole
Horizontal Well. Presented at the SPE Annual Technical Conference and Exhibition, New Orleans, 4–7 October. SPE-124483-MS.
https://doi.org/10.2118/ 124483-MS
Abass, H. H., Hedayati, S., and Meadows, D. L. 1996. Nonplanar Fracture Propagation From a Horizontal Wellbore: Experimental Study. SPE
Prod & Fac 11 (3): 133–137. SPE-24823-PA. https://doi.org/10.2118/24823-PA.
Bunger, A. P., Zhang, X., and Jeffrey, R. G. 2012. Parameters Affecting the Interaction Among Closely Spaced Hydraulic Fractures. SPE J. 17
(1): 292–306. SPE-140426-PA. https://doi.org/10.2118/140426-PA.
Cipolla, C., Weng, X., Onda, H. et al. 2011. New Algorithms and Integrated Workflow for Tight Gas and Shale Completions. Presented at the
SPE Annual Technical Conference and Exhibition, Denver, 30 October–2 November. SPE-146872-MS. https://doi.org/10.2118/146872-MS.
Cui, A., Brezovski, R., and Glover, K. 2013. Controls of Anisotropic In-Situ Stress and Permeability in Optimization of Wells and Hydraulic
Fractures for Unconventional Reservoirs: Examples From the Western Canada Sedimentary Basin. Presented at the 47th US Rock
Mechanics/Geomechanics Symposium, San Francisco, 23–26 June. ARMA-2013-289.

2018 SPE 11
Journal
Daal, J. A. and Economides, M. J. 2006. Optimization of Hydraulically Fractured Wells in Irregularly Shaped Drainage Areas. Presented at the
SPE International Symposium and Exhibition on Formation Damage Control, Lafayette, Louisiana, 15–17 February. SPE-98047-MS.
https://doi.org/ 10.2118/98047-MS.
Daneshy, A. A. 1973. A Study of Inclined Hydraulic Fractures. SPE J. 13 (2): 61–68. SPE-4062-PA. https://doi.org/10.2118/4062-PA.
Daneshy, A. A. 2003. Off-Balance Growth: A New Concept in Hydraulic Fracturing. J Pet Technol 55 (4): 78–85. SPE-80992-JPT.
https://doi.org/ 10.2118/80992-JPT.
Daneshy, A. A. 2011. Hydraulic Fracturing of Horizontal Wells: Issues and Insights. Presented at the SPE Hydraulic Fracturing Technology
Conference, The Woodlands, Texas, 24–26 January. SPE-140134-MS. https://doi.org/10.2118/140134-MS.
Daneshy, A. A. 2013. Horizontal Well Fracture Initiation and Extension From Cemented Cased Wells. Presented at the SPE Annual Technical
Confer- ence and Exhibition, New Orleans, 30 September–2 October. SPE-166086-MS. https://doi.org/10.2118/166086-MS.
Economides, M., Oligney, R., and Valko, P. 2002. Unified Fracture Design. Proc., CEUR Workshop, Vol. 1542, 33–36. https://doi.org/10.1017/
CBO9781107415324.004.
El Rabaa, W. 1989. Experimental Study of Hydraulic Fracture Geometry Initiated From Horizontal Wells. Presented at the SPE Annual
Technical Con- ference and Exhibition, San Antonio, Texas, 8–11 October. SPE-19720-MS. https://doi.org/10.2118/19720-MS.
Holanda, R. W. D., Gildin, E., and Valkó, P. P. 2017. Combining Physics, Statistics and Heuristics in the Decline Curve Analysis of Large
Datasets in Unconventional Reservoirs. Presented at the SPE Latin America and Caribbean Petroleum Engineering Conference, Buenos
Aires, 17–19 May. SPE- 185589-MS. https://doi.org/10.2118/185589-MS.
Hsiao, C. 1988. A Study of Horizontal-Wellbore Failure. SPE Prod Eng 3 (4): 489–494. SPE-16927-PA. https://doi.org/10.2118/16927-PA.
Lecampion, B., Abbas, S., and Prioul, R. 2013. Competition Between Transverse and Axial Hydraulic Fractures in Horizontal Wells. Presented
at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, 4–6 January. SPE-163848-MS.
https://doi.org/10.2118/163848-MS.
Lee, J. and Wattenbarger, R. A. 1996. Gas Reservoir Engineering, Vol. 5. Richardson, Texas: Textbook Series, Society of Petroleum Engineers.
Lee, J., Rollins, J. B., and Spivey, J. P. 2003. Pressure Transient Testing, Vol. 9. Richardson, Texas: Textbook Series, Society of Petroleum
Engineers. Ma, X. 2013. Integrated Hydraulic Fracture Placement and Design Optimization in Unconventional Gas Reservoirs. PhD dissertation,
Texas A&M Uni-
versity, College Station, Texas (December 2013).
Nordgren, R. P. 1972. Propagation of a Vertical Hydraulic Fracture. SPE J. 12 (4): 306–314. SPE-3009-PA. https://doi.org/10.2118/3009-PA.
Olorode, O., Freeman, C. M., Moridis, G. et al. 2013. High-Resolution Numerical Modeling of Complex and Irregular Fracture Patterns in
Shale-Gas Reservoirs and Tight Gas Reservoirs. SPE Res Eval & Eng 16 (4): 443–455. SPE-152482-PA. https://doi.org/10.2118/152482-
PA.
Palmer, I. D. and Moschovidis, Z. A. 1990. Fracture Geometries Consistent With Tiltmeter Data and Pressure and Tracer Communication in the Glisp
1-1 Stimulation. J Can Pet Technol 29 (1): 56–66. PETSOC-90-01-06. https://doi.org/10.2118/90-01-06.
Peirce, A. and Bunger, A. 2015. Interference Fracturing: Nonuniform Distributions of Perforation Clusters That Promote Simultaneous Growth of
Multi- ple Hydraulic Fractures. SPE J. 20 (2): 384–395. SPE-172500-PA. https://doi.org/10.2118/172500-PA.
Perkins, T. K. and Kern, L. R. 1961. Widths of Hydraulic Fractures. J Pet Technol 13 (9): 937–949. SPE-89-PA. https://doi.org/10.2118/89-PA.
Plaksina, T. 2015. Automated Optimization Strategies for Horizontal Wellbore and Hydraulic Fracture Stages Placement in Unconventional Gas
Reser-
voirs. PhD dissertation, Texas A&M University, College Station, Texas (May 2015).
Rahim, Z., Al-Kanaan, A., Al-Anazi, H. et al. 2015. Comparing Effectiveness Between Cemented Plug and Perf and Open Hole Ball Drop
Completion Assemblies to Ensure Optimal Multistage Fracturing Treatment and Well Performance—Field Examples. Presented at the Abu
Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, 9–12 November. SPE-177511-MS. https://doi.org/10.2118/177511-
MS.
Rammay, M. H. and Awotunde, A. A. 2016. Stochastic Optimization of Hydraulic Fracture and Horizontal Well Parameters in Shale Gas Reservoirs.
J. Nat. Gas Sci. Eng. 36A (November): 71–78. https://doi.org/10.1016/j.jngse.2016.10.002.
Roth, M. 2010. Shale Gas Reservoirs—Similar, Yet So Different. Oral presentation given at the 16th Annual 3D Seismic Symposium, Denver, 16
March. Sabaev, V. V., Mach, J. M., Wolcott, D. S. et al. 2006. Vertically Fractured Well Performance in Rectangular Drainage Area. Presented at
the SPE Rus-
sian Oil and Gas Technical Conference and Exhibition, Moscow, 3–6 October. SPE-101048-MS. https://doi.org/10.2118/101048-MS.
Saputelli, L. and Soliman, M. 2014. Design Optimization of Horizontal Wells With Multiple Hydraulic Fractures in the Bakken Shale. Presented
at the SPE/EAGE European Unconventional Resources Conference and Exhibition, Vienna, Austria, 25–27 February. SPE-167770-MS.
https://doi.org/ 10.2118/167770-MS.
Shahri, M. P., Chok, H., Safari, R. et al. 2015. Automated Hydraulic Fracturing Stage Design Based on Integrated Fracture Potential. Presented at
the Unconventional Resources Technology Conference, San Antonio, Texas, 20–22 July. URTEC-2153591-MS.
https://doi.org/10.15530/URTEC-2015- 2153591.
Song, B., Economides, M. J., and Ehlig-Economides, C. A. 2011. Design of Multiple Transverse Fracture Horizontal Wells in Shale Gas
Reservoirs. Pre- sented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, 24–26 January. SPE-140555-MS.
https://doi.org/10.2118/ 140555-MS.
Song, B. and Ehlig-Economides, C. A. 2011. Rate-Normalized Pressure Analysis for Determination of Shale Gas Well Performance. Presented at
the North American Unconventional Gas Conference and Exhibition, The Woodlands, Texas, 14–16 June. SPE-144031-MS.
https://doi.org/10.2118/ 144031-MS.
Warpinski, N. R., Branagan, P. T., Peterson, R. E. et al. 1998. An Interpretation of M-Site Hydraulic Fracture Diagnostic Results. Presented at the
SPE Rocky Mountain Regional/Low-Permeability Reservoirs Symposium, Denver, 5–8 April. SPE-39950-MS.
https://doi.org/10.2523/39950-MS.
Wei, Y. and Economides, M. J. 2005. Transverse Hydraulic Fractures From a Horizontal Well. Presented at the SPE Annual Techni cal
Conference and Exhibition, Dallas, 9–12 October. SPE-94671-MS. https://doi.org/10.2118/94671-MS.
Wu, K. and Olson, J. E. 2016. Mechanisms of Simultaneous Hydraulic-Fracture Propagation From Multiple Perforation Clusters in Horizontal Wells.
SPE J. 21 (3): 1000–1008. SPE-178925-PA. https://doi.org/10.2118/178925-PA.
Wutherich, K., Reese, S., Xu, L. et al. 2013. Quantifying the Effect of Drilling Azimuth in Shale Gas Reservoirs. Presented at the SPE Eastern
Regional Meeting, Pittsburgh, Pennsylvania, 20–22 August. SPE-165672-MS. https://doi.org/10.2118/165672-MS.
Yan, C., Luo, G., and Ehlig-Economides, C. A. 2014. Systematic Study of Bakken Well Performance Over Three Well Completion. Presented at the
SPE/ CSUR Unconventional Resources Conference–Canada, Calgary, 30 September–2 October. SPE-171566-MS. https://doi.org/10.2118/171566-
MS.
Yang, C., Vyas, A., Datta-Gupta, A. et al. 2017. Rapid Multistage Hydraulic Fracture Design and Optimization in Unconventional Reservoirs
Using a Novel Fast Marching Method. J. Pet. Sci. Eng. 156 (July): 91–101. https://doi.org/10.1016/j.petrol.2017.05.004.
Yu, W. and Sepehrnoori, K. 2013. Optimization of Multiple Hydraulically Fractured Horizontal Wells in Unconventional Gas Reservoirs. J. Pet. Eng.
2013: 1–16. https://doi.org/10.1155/2013/151898.
Zinn, C. J., Blood, D., and Morath, P. 2011. Evaluating the Impact of Wellbore Azimuth in the Marcellus Shale. Presented at the SPE Eastern
Regional Meeting, Columbus, Ohio, 17–19 August. SPE-149468-MS. https://doi.org/10.2118/149468-MS.

12 2018 SPE
Journal
Nadav Sorek is a reservoir analyst at Quantum Reservoir Impact, Houston. His experience includes analyzing reservoirs around
the globe, including South and North America and the Middle East, and developing Quantum Technologies products. Previ-
ously, Sorek was a research assistant at the Reservoir Dynamics and Control Research Group at Texas A&M University, teaching
reservoir-simulation and -optimization classes. His current interest is optimizing reservoir management and developing physics-
based and data-driven practical-reservoir-flow models. Sorek has authored or coauthored more than 10 technical papers. He
holds a PhD degree in petroleum engineering from Texas A&M University, a master’s degree in energy engineering, and a bach-
elor’s degree in environmental engineering from the Technion—Israel Institute of Technology. Sorek is a member of SPE.
Jose A. Moreno is a completions engineer at Anadarko Petroleum Corporation. His current interests include hydraulic-fracturing
modeling, geomechanics, and optimization of completion operations. He holds a bachelor’s degree in mechanical engineer-
ing from Brown University and a master’s degree in petroleum engineering from Texas A&M University. Moreno is a member
of SPE.
Ryan N. Rice is a partner at Rice Energy. Previously, he worked for 4 years as a reservoir engineer for Rice Energy. Rice’s current
interests include multivariate quantitative analyses, prescriptive analytics, and empirical big data. He holds a bachelor’s degree
in petroleum engineering from Texas A&M University. Rice is a member of SPE.
Guofan Luo is a PhD degree candidate in petroleum engineering at the University of Houston. His area of research is enhanced-
oil-recovery application in tight formations and hydraulic-fracture stimulation. Luo holds a master’s degree in civil engineering
from Texas A&M University. He is a member of SPE.
Christine Ehlig-Economides is a professor and the Hugh Roy and Lillie Cranz Cullen Distinguished University Chair at the University
of Houston. Previously, she was a professor at Texas A&M University for 10 years and before that worked for Schlumberger for
20 years. Ehlig-Economides’ research interests include production and reservoir engineering of conventional and unconven-
tional reservoirs. She has authored or coauthored more than 120 technical papers and holds two patents. Ehlig-Economides
holds a PhD degree in petroleum engineering from Stanford University, a master’s degree in chemical engineering from the
University of Kansas, and a bachelor’s degree in math-science from Rice University. She is a member of SPE.

2018 SPE 13
Journal