Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331

https://doi.org/10.1007/s13202-020-00919-z

REVIEW PAPER--PRODUCTION ENGINEERING

Recent comprehensive review for extended finite element method

(XFEM) based on hydraulic fracturing models for unconventional
hydrocarbon reservoirs
Belladonna Maulianda1 · Cindy Dhevayani Savitri1 · Aruvin Prakasan1 · Eziz Atdayev2 · Twon Wai Yan1 ·
Yew Kwang Yong1 · Khaled Abdalla Elrais1 · Reza Barati3

Received: 22 March 2020 / Accepted: 25 May 2020 / Published online: 8 June 2020
© The Author(s) 2020

Abstract
Hydraulic fracturing has been around for several decades since 1860s. It is one of the methods used to recover unconventional
gas reservoirs. Hydraulic fracturing design is a challenging task due to the reservoir heterogeneity, complicated geological
setting and in situ stress field. Hence, there are plenty of fracture modelling available to simulate the fracture initiation and
propagation. The purpose of this paper is to provide a review on hydraulic fracturing modelling based on current hydraulic
fracturing literature. Fundamental theory of hydraulic fracturing modelling is elaborated. Effort is made to cover the analyti-
cal and numerical modelling, while focusing on eXtended Finite Element Modelling (XFEM).

Keywords Hydraulic fracturing · Unconventional reservoirs · Analytical method · Numerical method · eXtended Finite
Element Modelling (XFEM)

Introduction The shale revolution is referred as the bloom of shale pro-

duction due to the technological improvement of horizontal
According to BP Statistical Review (2017), the global drilling and hydraulic fracturing. The improvements include
demand for natural gas increased by 1.5% from 2016. There- hydraulic fracturing in horizontal well reduced from 200 m
fore, the production of unconventional gas reservoir must be apart to 50 m while having over 60 stages per well (Aguil-
optimized to cover the decreasing of conventional gas reser- era and Ripple 2012; Ignatyev et al. 2011) the USA is the
voir production (Bentley 2002; Hughes 2013). The uncon- world’s largest dry natural gas producer because of the shale
ventional gas reservoir is commonly defined as reservoir gas revolution. It contributes to 20% of the world consump-
with poor permeability (less than 0.1 mD) (Gordon 2012; tion, in which 40% of the supply is from shale (Sieminski
Holditch 2006). There are four categories of unconventional 2014). The revolution mainly impacts the USA due to the
gas that are becoming important as future source of energy, advancement in technologies in current stage; however, it is
they are shale gas, tight gas sand, coal bed methane and expected to spread globally to countries with abundant shale
hydrates (Abdelfattah et al. 2015). Production of unconven- reserves, such as Canada, China and Argentina (Aguilera
tional resources is expected to thrive and grow sixfold from and Radetzki 2014).
2011 to 2030 (Ruehl and Giljum 2011). Hydraulic fracturing is a well stimulation method used
to improve the permeability of tight gas reservoir for eco-
nomical production (Britt 2012; King 2012). As the frac-
* Cindy Dhevayani Savitri ture grows, it follows the direction set by in situ stresses in
cindy_17006880@utp.edu.my
the rock, which can be termed as “fracture direction”. The
1
Petroleum Engineering Department, Universiti Teknologi propagation of fracture is perpendicular to plane of least
Petronas, Seri Iskandar, Malaysia effective stress (Britt 2012; King 2012). Microseismic moni-
2
Civil Engineering Department, University of Calgary, toring is one of the methods to study the characteristics of
Calgary, Canada hydraulic fracturing (Shapiro et al. 2006). Microseismic-
3
Chemical and Petroleum Engineering Department, Kansas ity is the occurring microearthquakes events caused by the
University, Lawrence, USA injection of fluid into the borehole (Shapiro et al. 2006). All

13
Vol.:(0123456789)
3320 Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331

hydraulic fracturing operations stimulated different magni- steam which subsequently drives a turbine to produce elec-
tude of microseismicity. The cloud of microseismic events tricity. In another method, the heated fluid exchanges heat
can be translated into Stimulated Reservoir Volume (SRV), with a working fluid, which will turn into vapour during
which serves as crucial parameter to study the characteri- the heat exchange process and it will drive the turbine. The
zation of reservoir and hydraulic fracturing (Shapiro et al. cooled fluids are then re-injected into the thermal reservoir
2006). Microseismic monitoring is a passive measurement of or cooled further in a secondary recovery system (White
microseismic event and provides crucial information such as et al. 2017).
magnitude, location and time of the event (Maulianda 2016). To improve heat extraction from geothermal reservoir
It is proven to be useful in showing the fracture geometries with low permeability, multiple-induced fractures are cre-
such as fracture length, fracture height, fracture width and ated to establish multiple flow paths between the injection
fracture azimuth with high confidence (Warpinski and Wol- well and production well, and to create multiple contact
hart 2016). Other application of microseismic monitoring in surfaces for heat exchange between hot rock and cold fluid
hydraulic fracturing includes observation of activated natu- (Bataille et al. 2006; Vik et al. 2018).
ral faults and faults, permeability (Shapiro et al. 1997), and
fracturing characterization using Moment Tensor Inversion Block cave mining
(MTI) (Nolen-Hoeksema and Ruff 2001).
Stimulated reservoir volume is defined as a collection of Hydraulic fracturing started to be applied in mining indus-
fluid-induced microseismic events that reflects the volume of try only in recent decades. The application initially was for
fracture network created in the reservoir (Mayerhofer et al. methane extraction in coal mining in the 1970s in the USA
2010). Stimulated Reservoir Volume (SRV) was introduced and to control hard roof rockburst (He et al. 2015). Follow-
to act as a correlation parameter for well performance (May- ing that, it was applied in cave mining.
erhofer et al. 2010). SRV compromised by a complex frac- Block Cave Mining refers to a method of underground
ture network with different geometries ranging from curved, mass mining method in which ore extraction relies on gravity
planar, slanted and of different lengths (Vera and Shadravan action (Adams and Rowe 2013). The method to induce cav-
2015). In horizontal well, SRV and fracture network size can ing involves undercutting of block by the means of blasting
be improved with longer lateral lengths and increasing stages it in order to destroy its ability to support the overlying rock.
of simulation. The study of SRV characteristics optimizes Gravity then acts to fracture the block (Eklind et al. 2007).
the hydraulic fracturing design. Analytical modelling and Pre-conditioning is required in the event there are massive,
numerical modelling are used for SRV modelling. unfractured ore body to initiate caving and to reduce caving
material size. One of the favoured pre-conditioning methods
Other uses of hydraulic fracturing is intensive hydraulic fracturing in boreholes drilled into
the ore body (Van and Jeffrey 2000). The hydraulic fractur-
Enhanced geothermal system (EGS) ing pressure can reach up to 10,000 psi, and pumped pure
water volume in the range of 4000–5000 litres per fracture,
Geothermal energy is the energy stored within the Earth’s however, the figures can be larger based on pump size and
crust and it is one of the promising clean renewable energy pressure response (Adams and Rowe 2013).
resources (MIT 2006; Vik et al. 2018). In 2016, it was esti- Block cave mining objective is to create horizontal radial
mated that only 6-7% of its total global potential had been hydraulic fractures (HFHFs) that are able to propagate across
extracted (GEA 2016; Vik et al. 2018). The common chal- existing vertical or inclined natural fractures, or to create
lenge to exploit geothermal reservoir is the low permeability inclines hydraulic fractures in horizontal or sub-horizontal
of the reservoir. To enhance permeability, hydraulic fractur- natural fractures dominated region (He et al. 2015).
ing is performed.
Enhanced Geothermal System (EGS) or Hot Dry Rock Geologic carbon sequestration (GCS)
(HDR) geothermal system as formerly known was pio-
neered at Los Alamos National Laboratory, New Mexico, In geologic carbon sequestration (GCS), large volume of
USA in the 1970s (Barbier 2002; Xia et al. 2017). In EGS, ­CO2 are injected into deep geological formations to prevent
a pair of wells are drilled to a depth where rock tempera- it from releasing to the atmosphere (Fu et al. 2017; Hasze-
tures approach 300C (White et al. 2017). The wells are then ldine 2009; Orr 2009). The targeted storage is typically
hydraulically fractured, thus creating a connection between saline aquifers or depleted hydrocarbon reservoirs which are
the wells. Fluid such a brines, compressed ­CO2 or liquid overlain by caprocks with low permeability. Consideration
mixtures are then circulated from one well to the other in to be taken in the C­ O2 storage design is to ensure caprock
a closed system through the created hydraulic connection. integrity to prevent ­CO2 leakage. However, it is known that
At ambient surface pressure, the heated fluid converts to hydraulic fractures may be initiated and propagated in the

13
Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331 3321

caprock when injected fluid pressure exceeds the minimum The study of SRV behaviour will enable better optimization
in situ principal stress of the caprock. Fluid flow along open of the hydraulic fracturing procedures. Analytical model-
hydraulic fracture is significantly more efficient as compared ling of SRV applies simplified mathematical equations to
to fluid flow in porous medium to deliver the fluid to far-field model the fracture propagation within the rocks (Table 1 and
reservoir; hence, the injection through hydraulic fracturing Fig. 1). There have been many 2D and 3D models developed
could improve both storage capacity and ­CO2 injection. The to study the fracture propagation under different conditions
most desirable scenario in GCS is when the hydraulic frac- and assumptions (Table 2).
ture can be contained within reservoir rock without fractur-
ing the caprock, at injection pressure level lower than the Two‑dimensional models
minimum principal stress of the caprock (Fu et al. 2017).
The 2D models were developed in the early 1960s as a sim-
ple approach for fracture design of industry requirement
Analytical modelling (Gidley 1990). The models are Kristinaovic-Geertsma-de
Klerk (KGD) model (Geertsma and De Klerk 1969; Zhel-
The SRV modelling is of the highest interest in the hydrau- tov 1955). Perkins-Kem-Nordgren (PKN) model (Nordgren
lic fracturing designs to predict the facture growth in terms 1972; Perkins and Kern 1961) and Radial model (Abe et al.
of length, height and width (Rahman and Rahman 2010). 1976). The 2D models have been reasonably successful in
Table 1  Summary of the available analytical models (Rahman and Rahman 2010; Xiang 2011)
2D models
Assumptions Illustration

KGD model
)1∕ Constant fracture height
6 2∕3
(
8GQ3
L = 0.48 (1−V)𝛍 ⋅t Fracture tip is a pointed shape tip
Fracture shape is rectangular
)1∕ Fracture is positioned at a plane strain condition in horizontal plane
6 1∕3
( 3
W o = 1.32 8(1−v)Q
G
𝛍
⋅t
( 3 )1∕
2G Q𝛍 4
Pw = 𝝈 min + 0.96 (1−v)3 2
L

PKN model
( ) 1 Constant fracture height
L = 𝝅cQh ⋅ t ∕2
l Fracture toughness does not affect fracture geometry
)1∕ Fracture shape is elliptical
4 1∕8 Fracture is positioned at a plane strain condition in vertical plane
( 2
W o = 4 2(1−v)Q
𝝅 3 Gc h
𝛍
⋅t
l

( 4 2 )1∕
G Q 𝛍 5 1∕5
Pw = 2.5 (1−v) 4 6
h
⋅t

Radial model
4(1−v2 )
𝝎(R, 𝜽) = E [G + G cos 𝜽] Fracture propagates in each plane
Fracture geometry is symmetrical to wellbore
Fracture shape is circular

Terms: L = fracture length 𝜎min = minimum in situ stress v = Poisson ratio, Wo = Crack width cl = fuid leak-off coefficient R = fracture radius,
Pw = Wellbore pressure 𝜇 = fluid viscosity h = fracture height, G = shear modulus t = injection time 𝜔 = fracture aperture, Q = flow rate 𝜃 = Fadian
E = Young’s modulus

13
3322 Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331

Fig. 1  Comparison of analytical models (Nordgren 1972; Peirce et al. 2010; Perkins and Kern 1961; Settari and Cleary 1984; Simonson et al.
1978; Warpinski and Smith 1989; Yew and Weng 2014; Yousefzadeh et al. 2015; Yu and Aguilera 2012; Zheltov 1955)

practical simulation; however, the assumption of fracture and reservoir properties. Vertical fracture propagation is lim-
shape and fracture height need to be specified to perform the ited by change in material property and minimum horizontal
models limits the practicality of 2D models. in situ stress (Yousefzadeh et al. 2015). Fracture deformation
KGD model assumes plane strain in horizontal direction. is linear plastic process in both models. FracCADE is a fractur-
It represents a fracture with horizontal penetration that is ing design commercial software from Schlumberger for PKN
lower than the vertical penetration. The model has cusp- and KGD models.
shaped fracture tip and the fracture width is uniform in ver-
tical direction. The model is most suitable to be applied for Three‑dimensional models
fractures with proportional length-to-height ratio.
In the PKN model, fracture planes are perpendicular to The pseudo-3D (P3D) model is proposed to idealize the frac-
the vertical plane strain. The model has an elliptical-shaped ture propagation in multi-layered formations. The constant
cross section and the fracture toughness is assumed to be not fracture height assumption is removed in this model (Settari
affecting the fracture geometry (Yew and Weng 2014). The and Cleary 1984; Simonson et al. 1978; Warpinski and Smith
pressure is uniform as there is the absence of fluid flow in 1989). The model is “pseudo” because variation of fracture
vertical section. The model is most suitable to be applied for geometry in three-dimensional space is not considered. The
fracture height that is higher than fracture length. model accounts for height variation by considering the in situ
Both 2D constant height models consider fracture width and stresses contrast, rock toughness and local net fluid pressure.
length as a function of fracture height, treatment parameters Some of the commercial softwares available in the market

13
Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331 3323

Table 2  Summary of researches conducted on studying analytical models (Grechka et al. 2010; Shapiro et al. 1997, 2006; Yousefzadeh et al.
2016; Yu and Aguilera 2012)
Authors Objectives Results

Shapiro et al. (1997) To determine permeability by using diffusivity Permeability estimation using fluid-induced
model and microseismic events in isotropic seismicity
reservoir Good agreement with previous permeability
studies
Unable to be applied in SRV determination due
to assumption of isotropy
Shapiro et al. (2006) To develop nonlinear diffusivity model for Suitable for shale reservoir characterization
reservoir characterization using microseismic events
Grechka et al. (2010) To compare and determine permeabilities Good agreement of permeabilities obtained
using diffusivity model and inversion model from both models
Yu and Aguilera (2012) To analyse SRV orientation and geometry by Good agreement with microseismic events of 2
using 3D diffusivity model in anisotrophic case studies
reservoir The method is suitable for other SRVs predic-
tion of similar size
The method is sensitive to spatio-temporal
distribution of microseismic events
Yousefzadeh et al. (2016) To determine and compare the fracture length The diffusivity model is more accurate
from PKN, KGD, P3D and diffusivity model
with microseismic fracture length

are FLAC3D by Itasca, FracPro by CARBO and FracMan by Equation 2 shows the equation derived from continuity
Golder (Hou and Zhou 2011). equation and Darcy law with the incorporation of real gas
The most recent development by Yu and Aguilera (2012) is equation and formation compressibility equation.
to use mass balance derived 3D model to account for fracture ( )
geometry variation. The diffusivity model accounts for the spa- 𝜕 kp 𝜕 ( 2 ) 1 1 𝜕 ( 2)
P = 𝜑𝜇ct P (2)
tio-temporal distribution of fluid-induced seismicity. In their 𝜕x 2𝜇z 𝜕x 0.000264 2𝜇z 𝜕t
study, the front of microseismic events represents the front of
pore pressure diffusion. Hydraulic diffusivity coefficient can p2
be determined from the slope of microseismic events versus u= (3)
𝜇z
time plot. The model can determine fracture geometry under
different stimulations case after the determination of hydrau-
(4)
√
P= u𝜇z
lic diffusivity coefficient from three-dimensional microseismic
events. Diffusivity model can be solved using MATLAB Par-
tial Differential Equation Toolbox (Peirce et al. 2010). dp √ 1 − 1 du
= 𝜇z u 2 (5)
dx 2 dx
Diffusivity model
dp √ 1 − 1 du
The 3D model is developed based on linear diffusion equa- = 𝜇z u 2 (6)
dy 2 dy
tion, for which the approximate solution is shown by Eq. 1.
( )
ΔPres P(x, y, t) − Pi 𝜕
(
𝜕u
)
𝜕 𝜕u 𝜑𝜇ct du
= kx + ky = (7)
ΔPinj Pinj − Pi 𝜕x 𝜕x 𝜕y 𝜕y 0.000264 dt
� � � � � �
x y z The porosity as a function of pressure is defined below
= erfc √ erfc √ erfc √
4𝜂x t 4𝜂y t 4𝜂z t (Abe et al. 1976).
(1)
(8)
( ( ))
Another variation of 3D model based on nonlinear dif- 𝜑 = 𝜑i 1 − ct P0 − P
fusion equation can be summarized as below, where it is
The permeability as a function of porosity and specific
derived from continuity equation and Darcy law.
surface is based on the Kozeny–Carman equation is defined

13
 Table 3  Continuum methods (Arndt et al. 2015; Costabel 1987; Hsiao 2006; Lee et al. 2015; Lei et al. 2017; Li et al. 2015; Lucia and Fogg 1990; McClure and Horne 2013; Postnikov et al.
3324

2017)
Continuum methods
Method Strength Limitation References

Boundary element modelling 1. Only the domain boundary is discretized Simulation becomes more complex in the pres- Costabel (1987), Hsiao (2006), Li et al. (2015) and
2. Simplified data representation by reduced ence of several layers of rock, edge or cornered McClure and Horne (2013)
number of space dimensions boundaries. Due to this the boundary condition

13
3. Able to simulate infinite domain problems. becomes more complex
Displacement discontinuity method 1. Single boundary conditions introduced Applicable for homogeneous formations only Li et al. (2015), McClure and Horne (2013)
2. Reduced computational time while most of the formations possess heteroge-
neous properties
Finite element method 1. Discretized model allows to include property 1. Hydraulic fracture path has to be pre-defined Arndt et al. (2015), Costabel (1987), Lei et al.
variation 2. Remeshing rules has to be introduced (2017), Li et al. (2015) and Sukumar and Prévost
2. Availability to analyse the fracture behaviour (2003)
at fine level
Finite volume method 1. Simple 1. Crack can propagate only along element edges Lee et al. (2015) and Lei et al. (2017)
2. Conservative 2. Interpolation is required to generate compatible
3. Memory efficient meshes between FVM and FEM by splitting
nodes to create new fracture

Table 4  Discontinuum methods (Costabel 1987; Lei et al. 2017; Li et al. 2015; Marina et al. 2014; McCluer and Home 2013; Sukumar and Prévost 2003)

Discontinuum methods
Method Strength Limitation References

Discrete element method 1. Independent computation for the existing and High computational time due to discrete computation Costabel (1987), Lei et al. (2017), Li et al. (2015) and
hydraulically induced fractures for each element Marina et al. (2014)
2. Doesn’t require the convergence of coordination
and displacement discontinuity equations
Discrete fracture network 1. Preservation of natural fractures features 1. Matrix is simplified Arndt et al. (2015), Lei et al. (2017), Li et al. (2015)
2. Efficient—simulate problems involving thousands 2. New fracture cannot be introduced and McClure and Horne (2013)
of individual fractures in a few hours or days using 3. Low angle of interaction cannot be simulated
a single process 4. Can be used for 2D simulations only
Extended finite element 1. Availability to introduce additional parameters To include the uncertainty XRFEM has to be Arndt et al. (2015), Lei et al. (2017), Li et al. (2015)
affecting fracture propagation constructed. This is done to include Monte Carlo and Sukumar and Prévost (2003)
2. Doesn’t require pre-definition of the fracture path simulation for uncertainty analysis
3. Removes requirement to introduce remeshing rule
to improve model quality
Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331
Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331 3325

Fig. 2  Heaviside enrichment

function (Youn 2016)

Fig. 3  Junction function and

branch function (Youn 2016)

below (Gates 2011). Rearranging the equation leads to per- Commonly used numerical models
meability as a function of pressure.
( ( ))3 Numerical simulation is the incorporation of the analytical
𝜑i 1 − ct P0 − P solution to identify the behaviour of the events in the pres-
(9)
15
k= ))))2 10
( ( ( (
5s2 1 − 𝜑i 1 − ct P0 − P ence of more constituents in the system. There are several
types of the models available to assess the hydraulic frac-
ture propagation which are divided into two major groups,
continuum method and discontinuum method.
Numerical models Continuum method implies that the matter on which
fracture to be simulated is continuous. The pre-existing
Modelling and analysis of hydraulic fractures propagation fractures and stress distribution can be introduced within
and interaction have gained enormous interest in petroleum the continuity and no necessity to introduce the interaction
engineering area. Hydraulic fracturing is one of the key of the boundaries. This method provided the stress distri-
important drivers in the development of unconventional bution at arbitrary location of the fractures and fracture
reservoirs. Hydraulic fracturing demand has increased rap- propagation is deduced based on the stress values obtained.
idly. Modelling tools have been developed to estimate the However, this method may not show exact behaviour of the
hydraulic fracturing performance in the formation. fractures. The summary of strength and limitation of the
continuum method can be seen on Table 3.
Discontinuum method on the other hand does show the
fracture as separate entity which is introduced as boundary
or stress computation limit. These methods give the advan-
tage in simulating the cases with high frequency of cohesive

13
3326 Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331

elements within the formation. The summary of strength determine the fracture behaviour separating two elements
and limitation of the discontinuum method can be seen on or elements being separated by void (Fig. 2) (Youn 2016).
Table 4. For the cases with the interaction of two fractures or inter-
The summary of the methods based on the research con- action of fracture with void junction function is implemented.
ducted by several authors (Arndt et al. 2015; Costable 1987; Junction function also uses the LSF to define the fracture loca-
Hsiao 2006; Lee et al. 2015; Lei et al. 2017; Li et al. 2015; tion while including the factor which accounts for the second-
Lucia and Fogg 1990; Mcclure and Horne 2013; Postnikov ary fracture.
et al. 2017) and (Costable 1987; Lei et al. 2017; Li et al. The Branch enrichment function is applied to identify the
2015; Marina et al. 2014; Mcclure and Horne 2013; Suku- behaviour of the fracture tip within the element (Fig. 3).
mar and Prévost 2003) is provided in the diagram below. By incorporating the Heaviside, Branch and Junc-
tion enrichment function on the GFEM function Eq. 11 is
Extended finite element modelling (XFEM) produced.
FEM displacement field
XFEM is based on the introduction of the enrichment func- ���������
tions on the previously successfully implemented FEM ∑N Mfr
∑ { ( )}
where it is used to simulate the interaction of solid and liquid u(x) = Ni (x)̄ui + Nj (x) H(x) − H xj ā j +
during injection of fracturing fluid (Maulianda 2016). The i=1 j=1
�������������������������������������
enrichment functions allow the fracture propagation simula- Heaviside enrichment
tion to be computed without the necessity to introduce new Branch enrichment Junction enrichment
������������������������������������������������� �����������������������������������
meshes. The general form of XFEM is as given in Eq. 10 Mtip Mjct
4
below (Youn 2016). ∑ ∑ ∑
𝛷𝛾 (x) − 𝛷𝛾 xk b̄ k +
{ ( )} { ( )}
Nk (x) NK (x) J(x) − J xl c̄ l
N M1 M2 M3 k=1 𝛾=1 l=1
(11)
∑ ∑ ∑ ∑
u(x) = Ni (x)̄ui + Nj (x)f1enr ā j + Nk (x)f2enr b̄ k + Nl (x)f3enr c̄ l
w h e r e H(x) − H � x�j , H e a v i s i d e e n r i c h m e(n t),
i=1 j=1 k=1 l=1
( )
(10)
xk }, branch enrichment and J(x) − J xl ,
∑4
y=1
{𝛷 𝛾 (x) − 𝛷 𝛾
where M1, M2 and M3 are different nodes of enrichment, junction enrichment.
f1enr , f2enr and f3enr are enrichment functions, Ni , Nj , Nk and As the enrichment functions are user defined the XFEM
Nl are finite element shape functions and ā j , b̄ k and c̄ k are the gives a prospect to simulation of the fracture propagation
degrees of freedom. behaviour including any factor required. The necessity is to
The enrichment functions introduced in XFEM are define the enrichment function in terms of the FEM model
Heaviside and Branch enrichment functions. Heaviside composites. As the enrichment is applied on the nodes sur-
function is derived based on the solutions of the level rounding the fracture the definition of the equation should
set function (LSF). LSF defines the fracture location and be computed based on the node characteristics (Bordas
the fracture tip location. Heaviside function is used to and Moran 2006; Feng and Gray 2017; Sepehri et al. 2015;

Fig. 4  LSF function parameter visualization (Youn 2016)

13
Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331 3327

where x is the nodal point, x̄ is the normal projection of the

point x onto the fracture and n is the unit normal from the
fracture at x̄ .

(13)
( )
𝛹i (𝐱) = 𝐱 − 𝐱𝐢 ⋅ ̂ti

where ̂ti is the unit vector tangent to the fracture at its tip and
xi is the location of the ith fracture tip.
Due to the dynamicity of the fracture propagation behav-
iour it is also necessary to define the algorithm to locate
where LSF update is required.
As the direction of the fracture propagation is based on
the stress intensity factor (SIF) as shown in Eq. 14 it is pos-
sible to define the LSF update location based on SIF.
�
⎛K � �2
1 KI ⎞
𝜃 = 2 arctan ⎜ I ± + 8⎟, −𝜋 < 𝜃 < 𝜋 (14)
4 ⎜ KII KII ⎟
⎝ ⎠

where 𝜃 is the angle of the fracture propagation relative to

the existing fracture direction while KI and KII are the mode
Fig. 5  Identifying the nodes for which LSF is to be updated (Youn one and two stress intensity factors, respectively.
2016)
The domain near the fracture tip is divided into two sec-
tions by one planar plane which perpendicular to the frac-
Youn and Griffiths 2014). Enrichment functions are added ture tip. Following discretization, the fracture propagation
into FEM model within the ABAQUS by incorporating the direction is computed based on the stress intensity factors.
external simulator or built-in python compiler. The latest Then, another perpendicular plane to the unit vector showing
inclusion into the XFEM is the stochastic simulation of the direction of fracture propagation is generated. The LSF
the fracture behaviour. Addition of the random field theory parameters are then updated for the nodes behind the newly
to the currently available XFEM which is also referred as generated plane (Fig. 5).
XRFEM allows the simulation of the cases with the inclu- Initial stress intensity is defined based on the analytical
sion of the uncertainty. computation as provided in Eq. 15.
The XRFEM includes Monte Carlo simulation and the √
average fracture behaviour is computed. This assists in 2
(
KI = 𝜎 sin 90 − 𝜃f
) 𝜋L
(15)
accounting for the uncertainties present within the forma- 2
tion. Modelling of the XFEM in ABAQUS software is simi-
lar to FEM. This is due to the XFEM using the FEM as base
√
𝜋L
(16)
( ) ( )
function. Therefore, the model generation is same as FEM KII = 𝜎 sin 90 − 𝜃f cos 90 − 𝜃f
2
until interaction assignment.
Level set function LSF function defines the output to where 𝜃f is the fracture angle relative to the coordinate sys-
be used in XFEM node enrichment based on the fracture tem and L is the fracture length.
geometry and spatial location. Two-level set function, sign For the dynamic fracture behaviour, the SIFs are derived
distance function and tangential level set functions are used based on Eq. 17 which is later solved to identify the SIFs for
to identify the distance of the fracture and fracture tip loca- the XFEM model provided in Eqs. 18 and 19.
tion (Youn 2016).
The distance of the node from the fracture in Eq. 12 is KI2 + KII2
G= (17)
defined as a multiple of the normal from the fracture which Eff
is visualized in Fig. 4. The location of the fracture tip is
defined as a function of fracture location and tangent unit where G is the energy release rate and Eff is the effective
vector of fracture as shown in Eq. 12 (Youn 2016). Young’s modulus.

𝜑(𝐱) = sign{𝐧 ⋅ (𝐱 − 𝐱)}min𝐱

̄ − 𝐱̄ (12) IIint Eeff
KIXFEM = (18)
2

13
3328 Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331

Fig. 6  General steps in XFEM

modelling

where KIXFEM is the mode I stress intensity factor for XFEM At the assign properties section the physical properties of
and IIint is the interaction state for mode I fracture propaga- the formation and fracture are introduced. The formation
tion in XFEM. is defined by its mechanical properties such as density,
Young’s modulus, Poison’s ratio, porosity, permeability and
IIIint Eeff damage criteria. Each part will have different properties and
KIIXFEM = (19)
2 will be named accordingly. Following that, section is cre-
ated. Section is the set which contains one of the materi-
where KIIXFEM is the mode II stress intensity factor for XFEM
als later to be used for the assignation of the materials to
and IIIint is the interaction state for mode II fracture propaga-
the respective part. The model is then assembled. Within
tion in XFEM.
assembly, all the parts are combined to produce single sys-
The assumption present in the derivation of Eqs. 18 and
tem where all constituents will be simulated. After assem-
19 are that the auxiliary SIFs are exhibiting no effect on
bling the model, three steps, namely initial step, geostatic
SIFs computed.
step and pumping step need to be created. In the initial step,
initial and boundary and conditions will be applied. The
XFEM modelling in abaqus: general steps
initial condition refers to the formation properties such as
saturation, stresses and pore pressure. The boundary condi-
The main steps in modelling the XFEM is shown by Fig. 6.
tions refers to the symmetry, restriction of displacement and
There will be a few assumptions done prior to building the
rotation over certain coordinates. In the geostatic step, loads
model. The assumptions are:
and geomechanical properties are introduced onto the forma-
tion. In the pumping step, the injection of the fracturing fluid
(i) Fracturing fluid is incompressible.
is defined. Following pumping step, interaction between the
(ii) Fracturing fluid is assumed to be liquid base and not
components of the system is then introduced. The fracture
foam base.
surface and propagation algorithm are defined within this
(iii) Proppant will not be simulated.
step. This stage also includes the introduction of the func-
The creation of XFEM model begin with the creation of tions to define the enrichment nodes within the simulation.
geometry. The geometry of the formation and existing frac- The nodes to be enriched are identified based on LSF and
tures are created at the part section. The geometry is drawn stress intensity factor (SIF). Meshing is then applied on the
initially at 2D and then extruded to form the 3D object. The model. Mesh is the gridding which discretizes the whole part
fractured formation and existing fracture are defined at this into sub-segments which will be computed separately later.
stage. There will be SRV and Non-SRV region in the model. There are several mesh types available within the software,
hex, hex-dominated, tet and wedge. For the simulation of

13
Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331 3329

the FEM tet type of mesh must be selected as it the mesh Abe H, Keer L, Mura T (1976) Growth rate of a penny-shaped crack
compatible with remeshing rule. However, for XFEM hex in hydraulic fracturing of rocks, 2. J Geophys Res Solid Earth
81(35):6292–6298. https​://doi.org/10.1029/JB081​i029p​05335​
type mesh can be implemented. Hex type mesh produced Adams J, Rowe C (2013) Differentiating applications of hydraulic
lesser number of elements for the same geometry than tet. fracturing. In: ISRM international conference for effective and
Lastly, job file is created. At this section, simulation is initi- sustainable hydraulic fracturing, Brisbane, Australia. http://dx.doi.
ated by including required input data which was defined in org/10.5772/56114
Aguilera RF, Radetzki M (2014) The shale revolution: global gas and
previous sections. Within this section, user defines the step oil markets under transformation. Mineral Econ 26:75–84. https​
size and step number to be simulated if required to differ ://doi.org/10.1007/s1356​3-013-0042-4
from default values. Aguilera RF, Ripple RD (2012) Link between rocks, hydraulic fractur-
In summary, a review has been made on hydraulic fractur- ing, economics, environment, and the global gas portfolio. In: SPE
Canadian unconventional resources conference, Calgary, Alverta,
ing modelling through analytical and numerical method. The Canada. https​://doi.org/10.2118/16271​7-MS
basic theory of the methods was presented and comparison Arndt S, Van der Zee W, Hoeink T, Nie J (2015) Hydraulic fracturing
are made between the strength and weaknesses based on simulation for fracture networks. In: Simulia community confer-
published studies. Analytical method can be classified into ence, Berlin. https​://pdfs.seman​ticsc​holar​.org/a083/e5de0​ba924​
b50f5​86244​632d8​6c439​5469c​1.pdf?_ga=2.58988​013.11538​
2D modelling and 3D modelling. Two-dimensional model- 11318​.15900​43751​-19482​69119​.15900​43751​
ling has an advantage of obtaining fast solution due to its Barbier E (2002) Geothermal energy technology and current status:
simplistic calculation, but it is limited to model constant and overview. Renew Sustain Energy Rev 6:3–65. https​://doi.
fracture height. This limitation is removed under 3D mod- org/10.1016/S1364​-0321(02)00002​-3
Bataille A, Genthon P, Rabinowicz M, Fritz B (2006) Modeling
elling; however, it come at the expenses of longer compu- the coupling between free and forced convection in a verti-
tational time due to higher complexity. Numerical method cal permeable slot: implications for the heat production of an
can be classified as continuum modelling and discontinuum enhanced geothermal system. Geothermics 35:654–682. https​
modelling, and under it, XFEM has the greatest advantage in ://doi.org/10.1016/j.geoth​ermic​s.2006.11.008
Bentley RW (2002) Global oil & gas depletion: an overview.
simulating hydraulic fracturing. The significance of XFEM Energy Policy 30:189–205. https​: //doi.org/10.1016/S0301​
is its ability to simulate arbitrarily propagating fracture, -4215(01)00144​-6
whereas other methods require the fracture trajectory to be Bordas SPA, Moran B (2006) Enriched finite elements and level sets
pre-defined. This paper has included general steps to cre- for damage tolerance assessment of complex structures. Eng
Fract Mech 73(9):1176–1201. https​://doi.org/10.1016/j.engfr​
ate XFEM model for the purpose of simulating hydraulic acmec​h.2006.01.006
fracturing. Britt L (2012) Fracture stimulation fundamentals. J Nat Gas Sci Eng
8:34–51. https​://doi.org/10.1016/j.jngse​.2012.06.006
Acknowledgements The authors would like to thank Universiti Costabel M (1987) Principles of boundary element methods. In:
Teknologi PETRONAS for their support and approval in publishing Computer Physics Reports (eds) Computer Physics Reports,
this work. This research did not receive any specific grant from funding vol 6(1-6). Elsevier, Amsterdam, pp. 243–274. https​: //doi.
agencies in the public, commercial, or not-for-profit sectors. org/10.1016/0167-7977(87)90014​-1
Eklind M, Ericsson P, Jonsson J, Lewen M, Nord G, Samuelsson
B, Potts A, Casteel K, Ericsson M (2007) Mining methods in
Compliance with ethical standards underground mining, 2nd edn. Atlas Copco Rock Drills AB,
Orebro. https​://minin​gandb​lasti​ng.files​.wordp​ress.com/2009/09/
Conflict of interest The authors declare no conflict of interest. minin​g_metho​ds_under​groun​d_minin​g.pdf
Feng Y, Gray K (2017) Modeling near-wellbore hydraulic fracture
Open Access This article is licensed under a Creative Commons Attri- complexity using coupled pore pressure extended finite element
bution 4.0 International License, which permits use, sharing, adapta- method. In: The 51st US rock mechanics/geomechanics sym-
tion, distribution and reproduction in any medium or format, as long posium, San Fransisco, California, USA. https​://www.onepe​tro.
as you give appropriate credit to the original author(s) and the source, org/confe​rence​-paper​/ARMA-2017-0352
provide a link to the Creative Commons licence, and indicate if changes Fu P, Settgast RR, Hao Y, Morris JP, Ryerson FJ (2017) The influ-
were made. The images or other third party material in this article are ence of hydraulic fracturing on carbon storage performance.
included in the article’s Creative Commons licence, unless indicated J Geophys Res Solid Earth 122:9931–9949. https ​ : //doi.
otherwise in a credit line to the material. If material is not included in org/10.1002/2017J​B0149​42
the article’s Creative Commons licence and your intended use is not Gates ID (2011) Basic reservoir engineering, 1st edn. Kendall-Hunt,
permitted by statutory regulation or exceeds the permitted use, you will Dubuque
need to obtain permission directly from the copyright holder. To view a GEA (2016) Annual U.S. global geothermal power production
copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/. report, Geothermal Energy Association (GEA). https​://www.
eesi.org/files​/ 2016_Annua​l _US_Globa​l _Geoth​e rmal​_ Power​
_Produ​ction​.pdf
Geertsma J, De Klerk F (1969) A rapid method of predicting width
References and extent of hydraulically induced fractures. J Petrol Technol
21(12):1571–571581. https​://doi.org/10.2118/2458-PA
Abdelfattah MH, Abdelalim AM, Yassin MHA (2015) Unconventional Gidley JL (1990) Recent advances in hydraulic fracturing, 12. Society
reservoir: definitions, types and Egypt’s potential. https​://doi. of Petroleum Engineers, United States
org/10.13140​/RG.2.1.3846.0880

13
3330 Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331

Gordon D (2012) Understanding unconventional oil. Energy and Cli- McClure MW, Horne RN (2013) Discrete fracture network modeling of
mate. https​://carne​gieen​dowme​nt.org/2012/05/03/under​stand​ing- hydraulic stimulation: coupling flow and geomechanics. Springer,
uncon​venti​onal-oil-pub-48007​ Berlin. https​://doi.org/10.1007/978-3-319-00383​-2
Grechka V, Mazumdar P, Shapiro SA (2010) Predicting permeability MIT (2006) The future of geothermal energy: impact of enhanced geo-
and gas production of hydraulically fractured tight sands from thermal systems (EGS) on the United States in the 21st Century,
microseismic data. Geophysics 75(1):B1–B10. https ​ : //doi. Massachusetts Institute of Technology. https​://www1.eere.energ​
org/10.1190/1.32787​24 y.gov/geoth​ermal​/pdfs/futur​e_geo_energ​y.pdf
Haszeldine RS (2009) Carbon capture and storage: how green can Nolen-Hoeksema RC, Ruff LJ (2001) Moment tensor inversion of
black be? Science 325:1647–1652. https​://doi.org/10.1126/scien​ microseisms from the B-sand propped hydrofracture, M-site, Col-
ce.11722​46 orado. Tectonophysics 336(1):162–181. https​://doi.org/10.1016/
He Q, Suorineni F, Oh J (2015) Modeling interaction between natural S0040​-1951(01)00100​-7
fractures and hydraulic fractures in block cave mining. American Nordgren R (1972) Propagation of a vertical hydraulic fracture. Soc
Rock Mechanics Association, San Fransisco, California. https​:// Petrol Eng J 12(04):306–314. https​://doi.org/10.2118/3009-PA
www.onepe​tro.org/confe​rence​-paper​/ARMA-2015-842 Orr FM (2009) Onshore geologic storage of CO2. Science 325:1656–
Holditch SA (2006) Tight gas sands. J Petrol Technol 58(06):86–93. 1658. https​://doi.org/10.1126/scien​ce.11756​77
https​://doi.org/10.2118/10335​6-MS Peirce A, Detournay E, Adachi J (2010) colP3D: a MATLAB code
Hou MZ, Zhou L (2011) Modelling and optimization of multiple frac- for simulating a pseudo-3d hydraulic fracture with equilibrium
turing along horizontal wellbores in tight gas reservoirs, harmo- height growth across stress barriers. https​://open.libra​ry.ubc.ca/
nising rock engineering and the environment. Taylor & Francis cIRcl​e/colle​ction​s/facul​tyres​earch​andpu​blica​tions​/52383​/items​
Group, London. https​://doi.org/10.1201/b1164​6-249 /1.00793​37, 2019
Hsiao GC (2006) Boundary element methods—an overview. Appl Perkins T, Kern L (1961) Widths of hydraulic fractures. J Petrol Tech-
Numer Math 56(10):1356–1369. https​://doi.org/10.1016/j.apnum​ nol 13(09):937–949. https​://doi.org/10.2118/89-PA
.2006.03.030 Postnikov A, Gutman I, Postnikova O, Olenova KY, Khasanov I,
Hughes JD (2013) Energy: a reality check on the shale revolution. Kuznetsov A (2017) Potemkin G (2017) Different-scale investiga-
Nature 494(7437):307–308. https​://doi.org/10.1038/49430​7a tions of geological heterogeneity of bazhenov formation in terms
Ignatyev A, Mukminov I, Vikulova E, Pepelyayev R (2011) Multistage of hydrocarbon potential evaluation (Russian). Oil Ind J 03:8–11
hydraulic fracturing in horizontal wells as a method for the effec- Rahman M, Rahman M (2010) A Review of Hydraulic Fracture Models
tive development of gas-condensate fields in the arctic region. In: and Development of an Improved Pseudo-3D Model for Stimulat-
SPE arctic and extreme environments conference and exhibition, ing Tight Oil/Gas Sand. Energy Sources A Recov Util Environ
Moscow, Russia. https://doi.org/10.2118/149925-MS Effects 32(15):1416–1436. https​://doi.org/10.1080/15567​03090​
King G (2012) Hydraulic fracturing 101: what every representa- 30605​23
tive, environmentalist, regulator, reporter, investor, university Ruehl C, Giljum J (2011) BP energy outlook 2030. https​://www.
researcher, neighbor and engineer should know about estimating bp.com/conte​nt/dam/bp/busin​ess-sites​/en/globa​l/corpo​rate/pdfs/
frac risk and improving frac performance in unconventional gas energ​y-econo​mics/energ​y-outlo​ok/bp-energ​y-outlo​ok-2011.pdf
and oil wells. In: SPE hydraulic 489 fracturing technology confer- Sepehri J, Soliman MY, Morse SM (2015) Application of extended
ence, The Woodlands, Texas, USA. https:​ //doi.org/10.2118/15259​ finite element method to simulate hydraulic fracture propaga-
6-MS tion from oriented perforations. In: SPE hydraulic fracturing
Lee D, Cardiff P, Bryant EC, Manchanda R, Wang H, Sharma MM technology conference, The Woodlands, Texas, USA. https://doi.
(2015) A new model for hydraulic fracture growth in unconsoli- org/10.2118/173342-MS
dated sands with plasticity and leak-off. In: SPE annual techni- Settari A, Cleary MP (1984) Three-dimensional simulation of hydrau-
cal conference and exhibition, Houston, Texas, USA. https​://doi. lic fracturing. J Petrol Technol 36(07):1177–171190. https​://doi.
org/10.2118/17481​8-MS org/10.2118/10504​-PA
Lei Q, Latham J-P, Tsang C-F (2017) The use of discrete fracture net- Shapiro SA, Huenges E, Borm G (1997) Estimating the crust perme-
works for modelling coupled geomechanical and hydrological ability from fluid-injection-induced seismic emission at the KTB
behaviour of fractured rocks. Comput Geotech 85:151–176. https​ site. Geophys J Int 131(2):F15–F18. https​://doi.org/10.1111/
://doi.org/10.1016/j.compg​eo.2016.12.024 j.1365-246X.1997.tb012​15.x
Li Q, Xing H, Liu J, Liu X (2015) A review on hydraulic fracturing Shapiro S, Dinske C, Rothert E (2006) Hydraulic-fracturing controlled
of unconventional reservoir. Petroleum 1(1):8–15. https​://doi. dynamics of microseismic clouds. Geophys Res Lett. https​://doi.
org/10.1016/j.petlm​.2015.03.008 org/10.1029/2006G​L0263​65
Lucia FJ, Fogg GE (1990) Geological/stochastic mapping of hetero- Sieminski A (2014) International energy outlook. Energy Informa-
geneity in a carbonate reservoir. J Petrol Technol 42(10):1298– tion Administration (EIA). https​://www.eia.gov/outlo​oks/ieo/
291303. https​://doi.org/10.2118/19597​-PA pdf/0484(2014).pdf
Marina S, Imo-Imo EK, Derek I, Mohamed P, Yong S (2014) Model- Simonson E, Abou-Sayed A, Clifton R (1978) Containment of massive
ling of hydraulic fracturing process by coupled discrete element hydraulic fractures. Soc Petrol Eng J 18(01):27–32. https​://doi.
and fluid dynamic methods. Environ Earth Sci 72(9):3383–3399. org/10.2118/6089-PA
https​://doi.org/10.1007/s1266​5-014-3244-3 Sukumar N, Prévost J-H (2003) Modeling quasi-static crack growth
Martin CD, Chandler NA (1993) Stress heterogeneity and geological with the extended finite element method part I: computer imple-
structures. Int J Rock Mech Min Sci Geomech Abstr Madison mentation. Int J Solids Struct 40(26):7513–7537. https​://doi.
Wisconsin. https​://doi.org/10.1016/0148-9062(93)90059​-M org/10.1016/j.ijsol​str.2003.08.002
Maulianda B (2016) On hydraulic fracturing of tight gas reservoir rock, Van AA, Jeffrey RG (2000) Hydraulic fracturing as a cave induce-
Degree of Doctor Philosophy, Petroleum Engineering Depart- ment technique at Northparkes mine. In: MassMin 2000 pro-
ment, University of Calgary, Calgary, Alberta, Canada. https​:// ceedings, Brisbane, Australia. https​://www.tib.eu/en/searc​h/id/
doi.org/10.11575​/PRISM​/27177​ BLCP%3ACN0​48376​496/Hydra​ulic-Fract​uring​-as-a-Cave-Induc​
Mayerhofer MJ, Lolon E, Warpinski NR, Cipolla CL, Walser DW, ement​-Techn​ique/
Rightmire CM (2010) What is stimulated reservoir volume?. https​
://doi.org/10.2118/11989​0-PA

13
Journal of Petroleum Exploration and Production Technology (2020) 10:3319–3331 3331

Vera F, Shadravan A (2015) Stimulated reservoir volume 101: SRV Colorado School of Mines. Arthur Lakes Library, Colorado, USA.
in a Nutshell. In: International petroleum technology conference, http://hdl.handl​e.net/11124​/17032​4
Doha, Qatar. https​://doi.org/10.2523/IPTC-18413​-MS Youn D, Griffiths D (2014) Hydro-mechanical coupled model of
Vik HS, Salimzadeh S, Nick HM (2018) Heat recovery from multiple- hydraulic fractures using the eXtended finite element method. In:
fracture enhanced geothermal systems: the effect of thermoelas- Shale energy engineering 2014: technical challenges, environmen-
tic fracture interactions. Renew Energy 121:606–622. https​://doi. tal issues, and Public Policy Pittsburgh, Pennsylvania, USA. https​
org/10.1016/j.renen​e.2018.01.039 ://doi.org/10.1061/97807​84413​654.024
Warpinski N, Wolhart S (2016) A validation assessment of microseis- Yousefzadeh A, Li Q, Aguilera R (2015) Microseismic 101: monitor-
mic monitoring. In: SPE hydraulic fracturing technology confer- ing and evaluating hydraulic fracturing to improve the efficiency
ence, The Woodlands, Texas, USA. https:​ //doi.org/10.2118/17915​ of oil and gas recovery from unconventional reservoirs. In: SPE
0-MS Latin American and caribbean petroleum engineering conference,
White M, Fu P, McClure M, Danko G, Elsworth D, Sonnenthal E, Quito, Ecuador. https​://doi.org/10.2118/17727​7-MS
Kelkar S, Podgorney R (2017) A suite of benchmark and chal- Yousefzadeh A, Li Q, Virues C, Aguilera R (2016) Identification of
lenge problems for enhanced geothermal system. Geomech Geo- activated fracture networks using microseismic spatial anoma-
phys Geo-Energy Geo-Resour 4:79–117. https​://doi.org/10.1007/ lies, b-values, and magnitude analyses in horn river basin. In:
s4094​8-017-0076-0 SPE hydraulic fracturing technology conference, The Woodland,
Xia Y, Plummer M, Mattson E, Podgorney R, Ghassemi A (2017) Texas, USA. https​://doi.org/10.2118/17915​3-MS
Design, modeling, and evaluation of a doublet heat extraction Yousefzadeh A, Li Q, Virues C, Aguilera R (2017) Comparison of
model in enhanced geothermal systems. Renew Energy 105:232– PKN, KGD, Pseudo3D, and diffusivity models for hydraulic frac-
247. https​://doi.org/10.1016/j.renen​e.2016.12.064 turing of the horn river basin shale gas formations using micro-
Xiang J (2011) A PKN hydraulic fracture model study and formation seismic data. In: SPE unconventional resources conference, Cal-
permeability determination, Master of Science, Petroleum Engi- gary, Alberta, Canada. https​://doi.org/10.2118/18505​7-MS
neering Department, Texas A&M University, Texas, USA. https​ Yu G, Aguilera R (2012) 3D analytical modeling of hydraulic fractur-
://pdfs.seman​ticsc​holar​.org/056b/efb22​96c90​e58fc​bd7ef​48ac4​ ing stimulated reservoir volume. In: SPE Latin America and car-
3622a​2c71a​c.pdf ibbean petroleum engineering conference, Mexico City, Mexico.
Warpinski N, Smith MB (1989) Rock mechanics and fracture geometry. https​://doi.org/10.2118/15348​6-MS
In: Recent advances in hydraulic fracturing, SPE monograph, vol Zheltov AK (1955) Formation of vertical fractures by means of highly
12, pp 57–80 viscous liquid. In: The 4th world petroleum congress, Rome, Italy.
Yew CH, Weng X (2014) Mechanics of hydraulic fracturing. Gulf https​://www.onepe​tro.org/confe​rence​-paper​/WPC-6132
Professional Publishing, Houston. https​://doi.org/10.1016/C2013​
-0-12927​-3 Publisher’s Note Springer Nature remains neutral with regard to
Youn DJ (2016) Hydro-mechanical coupled simulation of hydraulic jurisdictional claims in published maps and institutional affiliations.
fracturing using the extended finite element method (XFEM),
Degree of Doctor Philosophy, Engineering—Civil Speciality,

13