International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322

Contents lists available at ScienceDirect

International Journal of
Rock Mechanics and Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms

Wellbore breakout prediction in transversely isotropic rocks using true- T

triaxial failure criteria
⁎
N.B. Setiawan , R.W. Zimmerman
Department of Earth Science and Engineering, Imperial College London, London SW7 2AZ, UK

A R T I C LE I N FO A B S T R A C T

Keywords: This paper presents a unified approach through which the influence of the elastic and strength anisotropy on
Borehole stability wellbore instability can be thoroughly examined. The stresses at the wellbore wall are first calculated using the
Borehole breakouts Lekhnitskii-Amadei solution, which accounts for elastic anisotropy. Then, shear failure is treated by combining
Anisotropy the Mogi-Coulomb criterion for intact rock, with the Jaeger plane of weakness concept. The developed model
Mogi-Coulomb
accounts for all three principal stresses in predicting the onset of shear failure.
Lekhnitskii-Amadei
Plane of weakness
The results of the specific case investigated show that rock elastic anisotropy induce higher stress con-
centrations. The difference, compared with the stresses found using the isotropic elastic model, could reach as
high as 25% for the highest degree of anisotropy that might be expected for rocks of practical interest. The
strengthening effect of the intermediate stress, as reflected in the Mogi-Coulomb criterion, reduces the required
mud weight density by approximately 1.0 pounds-per-gallon (ppg). Furthermore, it is demonstrated that the risk
posed by bedding slippage, for a wellbore with an inclination between 15° and 50° from the vertical, is masked
when an isotropic elastic stress model is used. In contrast, the fully anisotropic model shows that an extra mud
weight of approximately 4.5 ppg would be required, in order to avoid bedding plane slippage for the case under
investigation. Although these results apply for a particular choice of strength properties and elastic properties,
they give an indication of the implications of fully accounting for anisotropy and the effect of the intermediate
stress when doing borehole stability analysis.

1. Introduction been acknowledged by many researchers. Several authors1–7 reported

Mechanically–induced instabilities are related to the drilling pro-
cess, which perturbs the local stress equilibrium. During the drilling
process, the pre-existing stress will be redistributed around the well-
bore, through which the stresses become concentrated in the rock ad-
jacent to the wellbore wall. In cases where the stress concentration is
too large for the rock to withstand, the rock around the wellbore will
fail. To overcome this, drilling fluid must be designed to provide suf-
ficient hydrostatic pressure to carry some of the load previously carried
by the excavated rock. Thus, the stress concentration around the
wellbore could be reduced, depending on the amount of hydrostatic
pressure exerted by the drilling fluid in the wellbore. Therefore, it is
possible to model the minimum hydrostatic pressure required to mini-
mise, if not to avoid, the risk of wellbore collapse. This can be achieved
by computing the stresses acting around the wellbore, and examining
their stability using appropriate failure criteria for a given type of rock
and a given in-situ stress state.
The role of geomechanics in minimising excessive drilling costs has
the benefit of having a pre-drilling geomechanical model. Their re-
spective drilling challenges have been addressed by means of, but not
limited to, the construction of a robust stress profile, the estimation of
the rock elastic and strength properties, drilling trajectory evaluation,
drilling fluid properties optimization, and casing seat design. For in-
stance, Hamid et al.5 used a geomechanical approach to aid their hor-
izontal well drilling and hydraulic fracturing in a tight carbonate re-
servoir. They reported that the non-productive time (NPT) could be
reduced to as low as 2% by implementing pre-drill geomechanical
modelling and real-time drilling geomechanics surveillance. Caicedo
et al.2 also reported that the stuck pipe incidents could be eliminated
after introducing a geomechanical approach into their drilling cam-
paign. Their statistic reflects evidence that NPT was primarily asso-
ciated with series of mud loss and pack-offs; however, the root cause of
these incidents was initially unidentified. Through detailed analyses of
drilling performance, combined with integrated geoscience data, they
were able to uncover the factors triggering the incidents, and improved
their drilling performance in the subsequent drilling campaign. By

⁎
Corresponding author.
E-mail address: i.setiawan16@imperial.ac.uk (N.B. Setiawan).

https://doi.org/10.1016/j.ijrmms.2018.10.033
Received 25 May 2018; Received in revised form 12 October 2018; Accepted 27 October 2018
1365-1609/ © 2018 Elsevier Ltd. All rights reserved.
N.B. Setiawan, R.W. Zimmerman
optimizing their mud weight design, Naidu et al.6 were not just able to
drill a horizontal well safely in the highly stresses depleted tight gas
sandstone, but also successfully obtained a good production rate due to
minimum skin effect in the reservoir. Zadvornov et al.7 reported that
adjustments to the mud weight design in their initial drilling program
helped them to avoid potential drilling issues associated with narrow
drilling window, which, if not identified, could have led to severe
problems such as wellbore collapse, blowout, stuck pipe, or even loss of
the bottom-hole assembly (BHA) and the wellbore. Moreover, a 40%
NPT reduction was achieved, and the well was completed twenty days
ahead of schedule. Bradley8 summarised that, apart from the cost and
operational efficiency, the benefit of having a proper wellbore stability
analysis from production and reservoir quality perspectives are im-
provement in cement quality and, therefore, sand control performance,
reduced perforation problems due to thick cement sheaths, and better
quality of well-log data.
In wellbore stability modelling, it is common practice to assume that
the rock is isotropic, even though about 75% of drilling footage must
deal with shale formations that are inherently anisotropic.9 This in-
trinsic characteristic of anisotropy may result from various processes
such as sedimentation, compaction, diagenesis, and weathering. Elastic
properties such as Poisson's ratio and Young's modulus are reported to
show directional properties.10–12 The elastic moduli could differ by
about 50% when measured normal and parallel to the bedding.11 In
tight gas shale, the variation of the elastic modulus can be up to 400%
between the vertical and horizontal directions.13 Several studies have
also shown that rock strength varies relative to the orientation of the
planes of anisotropy and the orientation of the major principal
stress.14–17 Ong and Roegiers18,19 investigated the influence of aniso-
tropy on wellbore stability, and revealed its significant impact on col-
lapse pressure calculation.
Okland and Cook20 experienced wellbore instability problems in
several highly inclined wellbores, which had not been experienced at
lower wellbore inclination. They acknowledged that bedding plane
splitting had caused the instability, rather than a conventional shear
failure of the intact rock. Recommendation was made to avoid drilling
parallel to the bedding plane. Similarly, Zhang21 investigated the effect
of bedding planes on wellbore instability, and concluded that stability
of the wellbore behaves differently with the impact of weak bedding
planes. The maximum slip failure locations and directions do not follow
those of the conventional breakout, but rather depend on the intrinsic
strength and orientation of the bedding planes.21 From their experience
with problematic wells, Wu and Tan22 reported that the bedding planes
mainly affect wellbores having an inclination greater than 60° and
drilled towards the minimum horizontal stress. Therefore, an additional
mud weight of approximately 0.03 sg (0.25 ppg) is necessary to coun-
teract the bedding plane failure against the case without considering
the weak planes. They also concluded that, by considering bedding
plane failure, the wellbore stability modelling of high-angle wells be-
comes less sensitive to wellbore azimuth.
The intrinsic anisotropic behaviour of the rocks can influence rock
failure in two ways: either by affecting the stress redistribution, or
through the rock strength anisotropy.23 Consequently, in view to de-
velop a more general solution, it is necessary for the modelling of
mechanical wellbore stability in anisotropic rocks to involve the fol-
lowing two steps: (1) the stress calculation for a borehole in an aniso-
tropic rock mass, and (2) determination of appropriate stress–related
failure criteria for anisotropic rocks to predict the breakout pressure.9
Thus, this paper aims to present a unified approach through which
the influence of the elastic and strength anisotropy on wellbore in-
stability can be thoroughly examined. In this paper, the stresses at the
wellbore wall are first calculated using the Lekhnitskii-Amadei solution,
which accounts for elastic anisotropy. Then, shear failure is treated by
combining the Mogi-Coulomb criterion for intact rock, with the Jaeger
plane of weakness concept.
314
International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322
2. Wellbore failure analysis—breakout pressure calculation
The workflow for breakout pressure calculation of an arbitrary
wellbore orientation in transversely isotropic rocks with a vertical axis
of symmetry (TIV) can be divided into three key steps: (1) the trans-
formation of the in-situ stress components and the rock elastic tensor
from the Earth coordinate system to the wellbore coordinate system, (2)
calculation of the near-wellbore stress using Lekhnitskii-Amadei solu-
tion, and (3) calculation and examination of the shear failure of the
intact rock and bedding plane around the wellbore circumference. The
following sub-sections discuss the key steps in more detail.
2.1. In-situ stress and rock property coordinate transformation
2.1.1. In-situ stress configuration and transformation
In wellbore stability modelling, the vertical stress σv is commonly
assumed to be one of the principal stresses, although a more generalised
assumption is sometimes necessary, e.g., when drilling in a geologically
faulted structure, a tilted principal stress is usually expected. In this
paper, the orientation of the minimum horizontal stress σh , attached
with the earth coordinate system x - y -z , is defined by their angle with
respect to the north along the x-axis (Fig. 1). The wellbore can take an
arbitrary orientation, and ties up with the Earth coordinate system by
its azimuth α w , measured from north, and inclination angle βw mea-
sured with respect to the vertical axis as such a horizontal wellbore is
defined by βw = 90°. Similarly, the rock property orientation is also
attached to the Earth coordinate system by defining their dip azimuth
and dip angle, αbp and βbp , respectively (Fig. 2). Since the goal is to
evaluate the onset of breakout at the wellbore wall, it is convenient to
transform the in-situ stress components σo to the wellbore coordinate
system xb - yb -z b .
The transformation of the in-situ stress components σo ,
⎡ σh 0 0 ⎤
σo = ⎢ 0 σH 0 ⎥
⎢ 0 0 σv ⎥
⎣ ⎦ (1)
from the Earth coordinate system x - y -z into the wellbore coordinate
system xb - yb -z b , can be achieved using the following matrix operation:
w w w
⎡ σxx τxy τxz ⎤
σwcs = T
Rw R xyz σo Rxyz RwT ⎢ w w
= τyx σyy τyz w⎥
⎢ ⎥
⎢ τzxw τzyw σzzw ⎥ (2)
⎣ ⎦
in which the rotational matrices Rxyz and Rw are defined as
⎡ cosα sinα 0 ⎤
Rxyz = ⎢− sinα cosα 0 ⎥
⎣ 0 0 1⎦ (3)
⎡ cosβw cosα w cosβwsinα w sinβw ⎤
Rw = ⎢ − sinα w cosα w 0 ⎥
⎢ ⎥
⎣ − sinβw cosα w − sinβw sinα w 0 cosβw ⎦ (4)
where α is the azimuth of the minimum horizontal stress measured from
Fig. 1. The Earth coordinate system (left) and the wellbore coordinate systems
(right).
N.B. Setiawan, R.W. Zimmerman International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322

2 2 2
⎡ l11 l12 l13 2l12 l13 2l13 l11 2l12 l11 ⎤
⎢ l2 l 2
l 2
2l l 2l l 2l22 l21 ⎥
23 22 23 21
⎢ 21 22 23 ⎥
2 2 2
Nσ = ⎢
⎢
l31 l32 l33 2l33 l32 2l33 l31 2l32 l31 ⎥
⎥
⎢ l11 l21 l32 l22 l33 l23 l33 l22 + l32 l23 l33 l21 + l31 l23 l31 l22 + l32 l21⎥
⎢ l11 l31 l32 l12 l33 l13 l33 l13 + l32 l13 l33 l11 + l31 l13 l31 l12 + l32 l11 ⎥
⎢l l l l l l l l + l l l l + l l l l + l l ⎥
⎣ 21 31 12 22 13 23 13 22 12 23 13 21 11 23 11 22 12 21 ⎦ (8)

The components of the strain transformation matrix Mε with respect

to wellbore coordinate system are:
l11 = cosβw cosα w l12 = cosβwsinα w l13 = sinβw
Fig. 2. The bedding plane local axes and the stress components around the
l21 = −sinα w l22 = cosα w l23 = 0
wellbore.
l31 = −sinβw cosα w l32 = −sinβw sinα w l33 = cosβw (9)

north, and α w and βw are the wellbore azimuth and wellbore inclina- while the components for the stress transformation Nσ can be obtained
tion, respectively. The superscript T denotes the matrix transpose. from the bedding dip azimuth (αbp ) and dip angle (βbp ), as such
βbp = 90° describes horizontal plane (see Fig. 2), as follow,
2.1.2. Material property definition and transformation l11 = cos(90°−βbp)cosαbp l12 = cos(90°−βbp)sinαbp l13 = sin(90°−βbp)
In elasticity, a constitutive equation relates the in-situ stress σij and l21 = −sinαbp l22 = cosαbp l23 = 0
strain εkl components for anisotropic rocks, is described by the fol-
l31 = −sin(90°−βbp)cosαbp l32 = −sin(90°−βbp)sinαbp l33 = cos(90°−βbp)
lowing generalised Hooke's law:
(10)
εij = Aijkl σkl (5)
Making use of the two transformation matrices in Eq. (8), the
where Aijkl is the fourth-rank elastic compliance tensor that, using Voigt transformation process for a compliance tensor A can be achieved by
notation, can be written as a 6 × 6 matrix with 21 independent stiffness
coefficients. However, due to physical symmetry, the number of in- At = Mε NσT ANσ MεT (11)
dependent coefficients can be reduced. For instance, in TIV rocks,
which implies that the material possesses one axis of symmetry, five
2.2. Calculation of the breakout pressure and examination of the shear
independent elastic constants form the compliance matrix A , i.e. two
failure limit
Young's modulus (Eh and Ev ), two Poisson's ratio (νv and νh ) and one
shear modulus (Gv ). In matrix form, the relationship in Eq. (5) for a TIV
By using the transformed in-situ stress σwcs in Eq. (2), and the
rocks can be written as:
transformed rock material property At in Eq. (11), the near-wellbore
1 h ν
v ν stress is then calculated around the wellbore circumference. To in-
⎡ Eh − Eh − Ev 0 0 0⎤
⎢ ⎥ corporate the anisotropic characteristic of the rocks, the Lekhnitskii-
1 ν
⎢ − Ev 0 0 0⎥ Amadei solution is used to obtain the three-dimensional near-wellbore
Eh v
⎢ ⎥ stress components, after which, by means of Newton-Raphson tech-
1
⎢ 0 0 0⎥
A=⎢
Ev ⎥ nique, the solution of the minimum mud pressure and therefore the
⎢ 1
0 0⎥ principal stress components that satisfy the failure criteria are obtained.
⎢ Gv ⎥
⎢ The highest breakout pressure around the wellbore circumference is
1
0⎥
⎢ Gv ⎥ then taken as the final answer. Similarly, to determine the slippage
⎢ 1 ⎥
onset of the bedding plane, the near wellbore stress was projected onto
⎢
⎣ Gh ⎥
⎦ (6)
the bedding orientation, after which the minimum mud pressure that
where the horizontal shear modulus Gh can be expressed in terms of the satisfies the Jaeger plane of weakness criterion is determined. The
other two horizontal elastic moduli using the following relationship,24 overall breakout limit is then obtained by contrasting the breakout for
the intact rock with that of the bedding plane. The higher of the two
Eh was chosen as the overall breakout limit. The general workflow de-
Gh =
2(1+νh) (7) scribing the procedure is shown in Fig. 3.
The following sections discuss the Lekhnitskii-Amadei solution and
In Eq. (6), only one half of compliance matrix A is written due to
the failure criteria that are used to assess the stability of the wellbore
symmetry. The subscripts v and h denote the normal and in-plane di-
wall in this paper, i.e. Mohr-Coulomb and Mogi-Coulomb for the intact
rection of the plane of isotropy, respectively. Thus νv is the vertical
rock, and the Jaeger plane of weakness for the bedding plane.
Poisson's ratio characterizing transverse contraction in the plane of
isotropy due to stress acting in the direction normal to it, and νh is the
horizontal Poisson's ratio corresponds to transverse contraction in the 3. The near-wellbore stress in anisotropic rocks
plane of isotropy due to stress applied also in the plane of isotropy.26
The transformation process of the compliance tensor into wellbore The Kirsch equation is commonly used to calculate the stress dis-
coordinate system requires two 6 × 6 transformation matrices Mε and tribution around a circular opening in elastic isotropic homogeneous
Nσ , written as, rocks. Using this equation, the stress components are calculated as a
function of the far-field in-situ stresses, wellbore geometry, and well-
2 2 2
⎡ l11 l12 l13 l12 l13 l13 l11 l12 l11 ⎤ bore pressure (mud weight). The Kirsch equation was derived based on
⎢ l2 l 2
l 2
l l l l l l ⎥ the classical plane strain assumption, which deals with a situation
23 22 23 21 22 21
⎢ 21 22 23 ⎥
2 2 2 where the dimension of the body in the longitudinal direction, i.e. along
Mε = ⎢
⎢
l31 l32 l33 l33 l32 l33 l31 l32 l31 ⎥
⎥ the wellbore axis, is extremely large in comparison with the dimension
⎢ 2l11 l21 2l32 l22 2l33 l23 l33 l22 + l32 l23 l33 l21 + l31 l23 l31 l22 + l32 l21 ⎥ of the other two directions, i.e. the wellbore cross section. Furthermore,
⎢ 2l11 l31 2l32 l12 2l33 l13 l33 l13 + l32 l13 l33 l11 + l31 l13 l31 l12 + l32 l11 ⎥
⎢ 2l l 2l l 2l l l l + l l l l + l l l l + l l ⎥ in the classical plane strain assumption, the displacement components u
⎣ 21 31 12 22 13 23 13 22 12 23 13 21 11 23 11 22 12 21 ⎦
and v , those are perpendicular to the wellbore axis, must be such that u

315
N.B. Setiawan, R.W. Zimmerman International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322

identical in all planes perpendicular to the hole axis,24 such that

∂u ∂v ∂w
= = =0
∂z ∂z ∂z
∂w
εzz = − =0
∂z
∂u ∂v
εxx = − ; εyy = −
∂x ∂y
∂u ∂v ⎞ ∂u ∂w ⎞ ∂w ∂v ⎞
γxy = −⎛ ⎜ + ; γxz = −⎛
⎟ + ; γyz = −⎛ + ⎜ ⎟

⎝ ∂y ∂x ⎠ ⎝ ∂z ∂x ⎠ ⎝ ∂y ∂z⎠ (12)
and hence, only the axial strain (εzz ) vanishes; unlike the classical plane
strain in which the anti-plane strain γxz and γyz are also zero. To obtain
the exact solution of the stress concentration in anisotropic rocks, two
stress functions F (x , y ) and ψ (x , y ) are defined, which are related to the
stress components by
∂2F ∂2F ∂2F
σxx = ; σyy = ;τ = − ∂x ∂y
∂y 2 ∂x 2 xy
∂ψ ∂ψ
τxz = − ∂y ; τyz = − ∂x
(13)
Inserting Eq. (13) into the equation of equilibrium, constitutive
equations, and strain compatibility equations, two coupled differential
equations, referred as to the Beltrami-Michell equations, are obtained.
The exact solution of the near-wellbore stress components can be ob-
tained by simultaneously solving the following Beltrami-Michell equa-
tions:
L4 F + L 3 ψ = 0
L3 F + L 2 ψ = 0 (14)
in which, the linear differential operators L2 , L3 and L4 , which are as-
Fig. 3. The workflow to determine breakout pressure limit at the wellbore wall.
sociated with the elastic properties of the rocks, can be expressed as:

and v are functions of x and y alone, and that the axial displacement w ∂2 ∂2 ∂2
L2 = β44 2
−2β45 + β55 2
vanishes. For a deviated wellbore, the solution for stress distribution ∂x ∂x ∂y ∂y
have been found by Hiramatsu and Oka.38 ∂3 ∂3 ∂3 ∂3
L3 = − β24 3 + (β25 + β46) 2 − (β14 + β56) 2 + β15 3
Consequently, with regards to anisotropic rocks, this solution limits ∂x ∂x ∂y ∂y ∂x ∂y
its application to the following conditions: (1) the wellbore axis is ∂4 ∂4 ∂4 ∂4 ∂4
L4 = β22 4 −2β26 3 −2β16 3 + (2β12 + β66) 2 2 + β11 4
parallel to the principal stress direction and (2) the material has a ∂x ∂x ∂y ∂y ∂x ∂x ∂y ∂y
maximum of one plane of elastic symmetry that is perpendicular to the (15)
wellbore axis. In practice, the solution is restricted only for the case of a ai3 aj3
vertical or horizontal wellbore in isotropic rocks or a wellbore in where βij = aij − a , and aij are the components of the compliance
33
transversely isotropic rocks (TIV) where the bedding is orthogonal to tensor A which corresponds to the rocks elastic property as in Eq. (6).
the hole axis.25 Because of the restricted application of the classical From the above equations, unlike the Kirsch equation that is in-
plane strain assumption for anisotropic rocks, generalised plane strain dependent of the rock elastic properties, it is obvious that the Lekh-
has been proposed which served as the basis for the development of the nitskii-Amadei solution explicitly accounts for the elastic properties of
Leknitskii-Amadei solution. Lekhnitskii26 and Amadei24 provide a de- the medium under investigation. Simplifying Eq. (14) in terms of F
tailed discussion on the generalised plane strain formulation, which will through substitution, one can easily obtain
not be repeated in this paper. (L4 L2 − L32) F = 0 (16)
and by using the following notation:
3.1. The Lekhnitskii-Amadei solution
l2 (μ) = β44 −2β45 μ + β55 μ2
To incorporate the anisotropic deformability behaviour of the rocks, l3 (μ) = − β24 + (β25 + β46) μ − (β14 + β56) μ2 + β15 μ3
Amadei24 presented a method to calculate the stress concentration
l4 (μ) = β22−2β26 μ−2β16 μ3 + (2β12 + β66) μ2 + β11 μ4 (17)
around arbitrarily oriented boreholes in anisotropic rocks of arbitrary
orientation, adopted from the pioneering work by Lekhnitskii. His Lekhnitskii showed that the operator in Eq. (16) can be written as,
model addresses the problem of an infinitely long internally-pressurised
l4 (μ) l2 (μ) − l32 (μ) = 0 (18)
wellbore that can take an arbitrary orientation penetrating an aniso-
tropic rock subjected to arbitrary external stresses. Only the most re- which is a sixth-order differential equation that that has six roots of μk
levant key equations of the Lekhnitskii-Amadei solution will be pre- (k = 1, …, 6) . The six roots are always complex or purely imaginary and
sented in this paper; please refer to their original works for a more three of them are the complex conjugate of the others.24 Therefore, in
detail derivation. solving the equation, one can ignore the conjugate roots and only
In the Lekhnitskii-Amadei solution, the three-dimensional stress consider the other three roots as the solution of the equation. It is im-
state around the wellbore is calculated by assuming that the wellbore is portant to also note that one must use the roots that have positive
infinitely long, and the anisotropic body is subjected to three arbitrary imaginary parts in the calculation. Finally, the total stress components
external principal stresses. A generalised plane strain concept is in- in an anisotropic body bounded internally with circular hole of infinite
troduced, which states that all components of displacement are to be length, in Cartesian coordinates, are given by

316
N.B. Setiawan, R.W. Zimmerman International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322

w
σxx = σxx w
+ σxx , h = σxx +2R [μ12 Φ′1 (z1) + μ22 Φ′2 (z2) + λ3 μ32 Φ′3 (z 3)] Table 1
w w Elastic properties of Vaca Muerta rock.36
σyy = σyy + σyy, h = σyy +2R [Φ′1 (z1) + Φ′2 (z2) + λ3 Φ′3 (z 3)]
w
τxy = τxy w
+ τxy, h = τxy −2R [μ1 Φ′1 (z1) + μ 2 Φ′2 (z2) + λ3 μ3 Φ′3 (z 3)] Young's modulus Poisson's ratioa Shear modulus
w w
(MPsi) (–) (MPsi)
τxz = τxz + τxz, h = τxz +2R [λ1 μ1 Φ′1 (z1) + λ2 μ 2 Φ′2 (z2) + μ3 Φ′3 (z 3)]
w w Vertical 2.40 0.18 1.24
τyz = τyz + τyz, h = τyz −2R [λ1 Φ′1 (z1) + λ2 Φ′2 (z2) + Φ′3 (z 3)] (19)
Horizontal 3.88 0.24 1.56
where R takes only the real part of the result in the bracket and the Degree of anisotropy 1.62 1.26
w w w w w
stress components σxx , σyy , τxy , τxz , τyz are the in-situ stress components a
The vertical Poisson's ratio (νv ) characterizing transverse contraction in the
as in Eq. (2). The axial stress σzz can be obtained from Eq. (6) by re-
plane of isotropy due to stress acting in the direction normal to it, and the
calling the definition of plane strain, i.e. εzz = 0 , horizontal Poisson's ratio (νh ) corresponds to transverse contraction in the plane
1 of isotropy due to stress applied also in the plane of isotropy.
σzz = − (a31 σxx + a32 σyy + a34 τyz + a35 τxz + a36 τxy )
a33 (20)
Zaman27 presented the sensitivity analysis of several parameters, such
In Eq. (19), Φ′k (z k ) are the derivative of an analytic functions that
as the degree of anisotropy and far-field stress ratio, to study their
can be expressed as
impact on wellbore fracturing and collapse limits. They adopted the
−1 minimum normal stress theory for tensile failure, and the Drucker-
Φ′1 (z1) = Prager criterion for the collapse limit. Yan et al.28 applied the Lekh-
2
rw ⎛ z1 ⎞ −1−μ 2
⎜ ⎟ nitskii-Amadei solution for a wellbore in a shale gas reservoir. In their
1
⎝ rw ⎠ analysis, the Mohr-Coulomb and the Jaeger plane of weakness (JPW)
⎧ (a1 (μ 2 − λ2 λ3 μ3) + b1 (λ2 λ3−1) + c1 λ3 (μ3 − μ 2 )) 1 ⎫ criteria were used to calculate the collapse limit of the intact rock and
⎨ μ 2 − μ1 + λ2 λ3 (μ1 − μ3) + λ1 λ3 (μ3 − μ 2 ) ξ1 ⎬ the bedding plane, respectively. Other authors25,29,30 also used a similar
⎩ ⎭
approach for their respective problems. In all these applications, except
−1
Φ′2 (z2) =
z2 2
{
(a1 (λ1 λ3 μ3 − μ1) + b1 (1 − λ1 λ3) + c1 λ3 (μ1 − μ3)) 1
μ2 − μ1 + λ2 λ3 (μ1 − μ3) + λ1 λ3 (μ3 − μ2) ξ2 } those of Ong and Roegiers,18,19 and Gupta and Zaman,27 the analyses
rw ( ) −1−μ
rw
2
2 have been limited to the condition of conventional triaxial stress state,
i.e. the effect of the intermediate stress is not considered in the failure
Φ′3 (z 3) =
z
−1
2
rw ⎛ 3 ⎞ −1 − μ32
{ (a1 (λ2 μ1 − λ1 μ2) + b1 (λ1 − λ2) + c1 (μ2 − μ1)) 1
μ2 − μ1 + λ2 λ3 (μ1 − μ3) + λ1 λ3 (μ3 − μ2) ξ3 } criterion.
⎝ rw ⎠

(21) 4. Rock failure criteria

where the constants a̅1, b1̅ , c̅1 are obtained by from the boundary con-
The other aspect of wellbore stability analysis lies in the selection of
ditions, and can be expressed as
criteria to predict the onset of rock failure. Among other criteria that
rw have been developed by numerous researchers, the Mohr-Coulomb
a1 = (p − (σy, o − iτxy, o))
2 w criterion is perhaps the most commonly used, because of its relatively
r
b1 = w (ipw + (τxy, o − iσx , o)) simple mathematical formulation. The criterion states that when the
2 maximum shear stress τmax on a plane exceeds the shear strength of the
r
c1 = w (τyz, o − iτxz, o) rock, i.e. represented by its strength parameters cohesion (S0 ) and
2 (22)
coefficient of internal friction ( μ ), a failure occurs along that plane.
the expression of ξk in Eq. (23) are According to the criterion, the maximum shear stress that the rock can
withstand is linearly correlated with the normal stress σn acting on the
zk 2

ξk =
zk
rw
+ ( ) −1−μ
rw
2
k
; k = 1, 2, 3
plane,
1−iμk (23) τmax = S0 + μσn (25)

and the three complex numbers λi are defined by
l3 (μ1) l3 (μ 2 ) l3 (μ3)
λ1 = − ; λ2 = − ; λ3 = −
l2 (μ1) l2 (μ 2 ) l4 (μ3)
where the complex variable z k = x + yμk is the coordinate of the con-
tour of the circular hole with x = rw cosθ and y = rw sinθ ; rw and pw are
the wellbore radius and the wellbore pressure, respectively. Validation
of the Lekhnitskii-Amadei solution has been carried out by Gaede et al.9
by comparing the predicted stresses with those of finite element si-
mulations. They demonstrated that the three-dimensional stress profile
calculated from the Lekhnitskii-Amadei solution and the finite element
simulation show an excellent agreement at the wellbore wall, and thus
validate the application of the Lekhnitskii-Amadei solution for all spe-
cial cases of symmetry, i.e., transversely anisotropic, orthorhombic and,
even, isotropic.
Aadnoy10 implemented the Lekhnitskii-Amadei solution to calculate
the limit of fracturing and collapse pressure of an arbitrarily oriented
wellbore in rocks with weak bedding planes. Ong and Roegiers18,19
presented the effect of anisotropy in the wellbore collapse using the
Lekhnitskii-Amadei solution. Moreover, to account for the effect of rock
strength anisotropy, a generalised strength criterion of Tsai and Wu and
Drucker-Prager was adopted in their study. Later, they incorporated
poroelastic and thermal effects into the analysis. Similarly, Gupta and
Despite its popularity among geoscientists, the Mohr-Coulomb cri-
terion neglects the strengthening effect of the intermediate principal
stress, although subsurface rocks are often subjected to unequal in-situ
(24) stresses. This condition is referred as to true-triaxial stress, or polyaxial
stress, i.e., when σ1 > σ2 > σ3. Mogi31 conducted extensive true-triaxial
experiments on several rock types, and showed that the intermediate
stress appears to have an impact on rock strength. His experimental
results and observations led to the development of failure criteria that
show the importance of incorporating intermediate principal stress in
rock failure analysis.
4.1. Effect of intermediate principal stress—Mogi-Coulomb Criterion
Numerous researchers have proposed failure criteria that account
for all the three principal stresses. Colmenares and Zoback32 show the
applicability of several failure criteria that commonly used in wellbore
stability analysis, such as Drucker-Prager, Modified Wiebols and Cook
and Modified Lade. In the present paper, to demonstrate the sig-
nificance of the intermediate principal stress, Mogi-Coulomb criterion is
adopted. Proposed by Al-Ajmi and Zimmerman,33 the Mogi-Coulomb
criterion is based on Mogi's assumption that the critical octahedral
stress (τoct ) at which failure occurs depends on the effective mean stress
(σm,2 ). The Mogi-Coulomb criterion can accurately predict true-triaxial

317
N.B. Setiawan, R.W. Zimmerman
Fig. 4. The tangential stress (σθ ) distribution up to two wellbore radii using the Lekhnitskii-Amadei solution (left) and comparison of the stress components at the
wellbore wall calculated using the Lekhnitskii-Amadei solution (solid curve) and Kirsch equation (black dash-line) of a vertical wellbore in TIV rocks where the two
methods show an excellent agreement (top and bottom right). The azimuth of the minimum horizontal stress is N 120°E.
Fig. 5. Stress components at the wall of a 15° deviated wellbore drilled per-
pendicular to the minimum horizontal stress in a TIV rock. For an inclined
wellbore, the effect of elastic anisotropy is more prominent and hence the near-
wellbore stress components from Lekhnitskii-Amadei (solid curves) are dif-
ferent with that of isotropic (Hiramatsu-Oka) equation (black dash-line). The
azimuth of the minimum horizontal stress is N 120°E. The near-wellbore stress
has been normalised with the in-situ vertical stress σv .
behaviour from traditional triaxial datasets.33 The criterion is written as
τoct = a + bσm,2 (26)
where, as will be shown, a and b are material constants that can be
related to the cohesion and friction angle of the rocks. For a triaxial test,
i.e. when σ1 > σ2 = σ3, Al-Ajmi and Zimmerman33 found that the Mogi
criterion becomes identical to the Mohr-Coulomb criterion and hence,
the two constants a and b can be directly obtained from the Mohr-
Coulomb parameters. The final form of the Mogi-Coulomb criterion can
be expressed as
3 ⎞ ⎛ 3 ⎞
τmax = ⎛a + b σm,2
⎝ 2 2⎠ ⎝ 2 2⎠ (27)
318
International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322
1
where σm,2 = 2 (σ1 + σ3) . Comparing the above equation with the Mohr-
Coulomb criterion, i.e. τmax = Socosϕo + σmsinϕo , the parameters a and b
can be obtained directly from the Coulomb strength parameters as
follows:
2 2
a= Socosϕo
3
2 2 2 2
a= Socosϕo b = sinϕo
3 3 (28)
where So and ϕo are the cohesion and friction angle of the intact rocks,
respectively. Al-Ajmi and Zimmerman33 found that, generally, the
Mogi-Coulomb criterion is more accurate than both Mohr-Coulomb and
Drucker-Prager. Their observation was substantiated by testing the
three failure criteria against true-triaxial data sets from a range of rock
types. Similar to Mohr-Coulomb, the linear formulation in the Mogi-
Coulomb criterion makes it relatively easy to be implemented in well-
bore stability analysis, and yet the effect of the intermediate principal
stress is considered. In wellbore stability prediction, they observed that
the Drucker-Prager criterion tends to overestimate the onset of wellbore
failure, whereas the Mohr-Coulomb criterion usually underestimates
the failure limit.37 To incorporate the strength anisotropic behaviour of
the rocks, the Jaeger's plane of weakness concept is adopted in this
work, thus enabling the examination of the bedding plane slippage
propensity.
4.2. Effect of strength anisotropy—Jaeger's “Plane of Weakness” model
Some reservoir rocks, notably shales, are intrinsically anisotropic,
and their physical properties, such as strength and deformability, will
depend on the orientation of the rock with respect to the principal
stresses. Jaeger34 postulated that for layered, transversely isotropic
rocks, failure will occur either at a stress and at an angle given by the
Coulomb criterion, or along a bedding plane. Ambrose and Zim-
merman,14 who conducted extensive conventional triaxial rock me-
chanics testing on anisotropic shales, found that the data could be fit
reasonably well using Jaeger's plane of weakness criterion for failure
N.B. Setiawan, R.W. Zimmerman International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322

Fig. 6. (a) The near-wellbore major principal stress for various degree of anisotropy (η ) of a deviated wellbore with 15° inclination drilled perpendicular to the
minimum horizontal stress in a TIV rock. The difference with that of the isotropic case (black dash-line) is approximately 25% for the highest degree of anisotropy
(red curve). (b) At the wellbore wall, the difference of the anisotropic near-wellbore stress components with that of isotropic (Hiramatsu-Oka) solution is ap-
proximately 5% and 35% for tangential stress and axial stress, respectively. The case was run assuming a degree of anisotropy η = 3.0 . (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article).

Table 2
Rock strength properties.
Cohesion (MPsi) Friction angle (deg.) Bedding orientation

Intact rock 1462 30 Dip angle (βb ) (deg.) Dip azimuth (αb ) (N-E)
Bedding plane 850 18 30 0

bedding plane. The higher of the two was chosen as the overall
breakout limit.

5. Results and discussion

5.1. Anisotropic near-wellbore stress validation against isotropic approach

The workflow in Fig. 3 has been translated into a Matlab® program.

Fig. 7. Sensitivity of breakout pressure with wellbore inclination for different
degree of anisotropy (η ). The breakout pressure was calculated using Mohr-
Coulomb and Mogi-Coulomb and has been normalised with the in-situ vertical
stress. In both cases, the risk of wellbore collapse increases with increasing
wellbore inclination.
along the bedding planes, and the Mohr-Coulomb criterion for failure
within the intact rock. The Coulomb criterion for failure along the
bedding plane is34
τ = Sw + μ w σn (29)
where Sw and μ w are the shear strength and the friction coefficient of a
weakness plane. The resultant shear stress τ acting on the plane of
weakness can be found from the two in-plane shear stress35
τ = 2
τxy 2
+ τxz
(30)
As discussed in Section 2.2, the overall shear failure limit is then
obtained by contrasting the breakout for the intact rock with that of the
Subsequently, a hypothetical case of a vertical wellbore penetrating a
TIV rock has been used to validate the program. The in-situ stress state
is assumed to be in a normal fault stress regime, where σv > σH > σh
with a considerable horizontal stress magnitude difference. At depth of
10,000 ft, the overburden stress (σv ) gradient is assumed to be 1.0 psi/ft,
and the maximum (σH ) and minimum horizontal stress (σh ) gradient are
0.85 psi/ft and 0.7 psi/ft, respectively. The azimuth of the minimum
horizontal stress is N 120°E and the pore pressure gradient is 0.43 psi/ft.
The elastic properties of the TIV rocks, listed in Table 1, are adopted
from the work of Ambrose36 who conducted an extensive triaxial test on
(
Vaca Muerta shale, indicating a degree of anisotropy η = Eh of 1.62.
v
E
)
In this special case of anisotropy, when the wellbore axis coincides
with the rock axis of symmetry, the presence of elastic anisotropy
would not affect the near-wellbore stress. This is because the two dif-
ferential Beltrami-Michell equations in Eq. (14) are no longer cou-
pled.24 Thus, the Kirsch solution can be used as a benchmark for the in-
plane stress components σθ , σr as well as σz if the Poisson's ratio in the
Kirsch equation, is replaced with the vertical Poisson's ratio (νv ) of the
anisotropic rocks. The validated result, shown in Fig. 4, shows excellent
agreement between the near-wellbore stress components modelled
using Leknitskii-Amadei solution and Kirsch equation, around the
wellbore circumference and up to two wellbore radii into the formation.
Upon validating the code and to examine the effect of anisotropic
elastic property on the near-wellbore stress distribution, the case was
generalised into an inclined wellbore penetrating a TIV rock. For a 15°
deviated wellbore drilled perpendicular to the minimum horizontal
stress direction, the effect of the elastic anisotropy embedded in the

319
N.B. Setiawan, R.W. Zimmerman
Fig. 8. The sensitivity analysis of bedding dip and bedding plane cohesion with breakout pressure.
Lekhnitskii-Amadei solution is more prominent and, hence, the differ-
ence with that of Hiramatsu-Oka solution is more pronounced (Fig. 5).
Although, for this case, the tangential stress (σθ ) difference of the two
methods is approximately 5%, there is more than 13% difference ob-
served on the axial (σz ) stresses which will alter the magnitude of the
breakout pressure. In Fig. 6, the magnitude of the stresses has been
normalised with the in-situ vertical stress σv . Detailed examination also
suggests that there is a phase shift on the peak stress location which will
affect not only the potential breakout location but also, on a broader
sense, the optimum drilling direction. The latter will be discussed in the
following section.
5.2. Effect of elastic anisotropy on near-wellbore principal stress
To demonstrate the dependency of the near-wellbore stress on
elastic anisotropy, a parametric study was carried out by varying the
degree of anisotropy (η). In Fig. 6a, it is shown that a higher degree of
anisotropy appears to induce more stress concentration compared to the
case with lower degree of anisotropy. The difference with that of the
isotropic case could reach approximately 30% for the highest degree of
anisotropy compared with that of the isotropic case. Fig. 6b illustrates
the difference of the other near-wellbore stress components with that of
the isotropic. Recalling that the intermediate principal stress has an
influence on rock strength, the magnitude of the breakout pressure
would be altered if one ignores the presence of the intermediate prin-
cipal stress. Moreover, the case shown in Fig. 6b emphasises not only
the importance of considering the intermediate stress, but more im-
portantly, neglecting elastic anisotropy would also affect the breakout
pressure attributed to the reduction of the tangential stress and axial
stress by approximately 5% and 35%, respectively.
5.3. Effect of elastic and strength anisotropy on breakout pressure
To predict the onset of the wellbore wall instability, upon obtaining
the near-wellbore stress, the principal stress must then be calculated
and inserted into the failure criterion. Acting on a free surface at the
wellbore wall, the radial stress σr , exerted by the hydrostatic pressure of
the drilling mud, is always one of the principal stresses, while the other
two must act on any surface such that the three are perpendicular to
each other. For instance, as clearly shown in Fig. 5, the anti-plane shear
stress τθz is non-zero, thus σθ and σz are not the principal stresses. Hence,
320
International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322
the other two principal stress components must be obtained from the
axial stress σz , tangential stress σθ and the shear stress τθz using the
following relationship:
σθ + σz 2 (σθ + σz )2
σP1,2 = ± τθz +
2 4 (31)
The internal wellbore pressure pw affects the near-wellbore stress as
such, for example, the circumferential stress σθ decreases with the in-
creasing pw according to Eqs. (19)–(24). The underlining concept in any
breakout limit calculation is, therefore, to determine the necessary
amount of pw to avoid shear failure that satisfies the chosen failure
criterion for a given rock strength and given in-situ stress configuration,
as illustrated in Fig. 3. To demonstrate the importance of incorporating
the anisotropic behaviour of the rocks in wellbore stability modelling,
in addition to the in-situ stress configuration and the rock elastic
properties listed in Table 1, the following rock strength was also as-
sumed (Table 2).
For a highly inclined wellbore with βw = 80°, the profile of the
breakout pressure as a function of the degree of anisotropy (η) is shown
in Fig. 7. As shown, for several elastic anisotropy ratios, it is obvious
that the profile of breakout pressure as a function of the wellbore in-
clination using Mohr-Coulomb is less sensitive compared to that of
Mogi-Coulomb criterion for a wellbore with an inclination higher than
45°. With Mogi-Coulomb, as the wellbore approaches horizontal, i.e.
βw = 90°, the gap between the lower degree of anisotropy with that of a
higher degree of anisotropy widens. In other words, since the near
wellbore stress for both cases were calculated using Lekhnitskii-Amadei
solution, the inclusion of the intermediate principal stress in the Mogi-
Coulomb criterion seems to be responsible for reducing the breakout
pressure by 0.5 ppg, i.e. the difference of the normalised breakout
pressure of a horizontal wellbore between that with η = 1.1 and 4.0 , in
Fig. 7.
For a given weak bedding plane characteristic (Table 2), the sensi-
tivity for various bedding orientations and strengths as a function of
wellbore inclination is given in Fig. 8. When the angle of attack is low,
i.e. the relative angle between the wellbore axis and the bedding
normal, the influence of weak bedding plane diminishes and, when
coincide, it vanishes. In practical engineering design, this suggests that
wells drilled sub-parallel through a fracture or fault are less stable than
those drilled at an oblique angle.
The following section elaborates the impact of both elastic and
N.B. Setiawan, R.W. Zimmerman International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322

a) Hiramatsu-Oka equation and Mohr-Coulomb failure criterion b) Hiramatsu-Oka equation and Mogi-Coulomb failure
criterion

c) Lekhnitskii-Amadei solution and Mohr-Coulomb failure d) Lekhnitskii-Amadei solution and Mogi-Coulomb failure
criterion criterion

Fig. 9. Sensitivity analysis of breakout pressure for an inclined wellbore with α w = 130° and βw = 60° using (a) Hiramatsu-Oka equation for the near-wellbore stress
and Mohr-Coulomb failure criterion, (b) Hiramatsu-Oka equation for the near-wellbore stress and Mogi-Coulomb failure criterion, (c) Leknitskii-Amadei solution for
the near-wellbore stress and Mohr-Coulomb failure criterion and (d) Leknitskii-Amadei solution for the near-wellbore stress and Mogi-Coulomb failure criterion. The
white dot and the white triangle represent the wellbore and the bedding orientations, respectively. The black curve in the wellbore inclination and azimuth sensitivity
plot is the breakout pressure of the intact rocks while red curve represents that of the bedding plane and the white dot is the assumed mud weight. The overall
breakout limit is determined by taking the highest of the two and shaded as a blue area in the sensitivity plot. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article).

strength anisotropy on breakout pressure for optimizing the drilling
direction.
5.4. Sensitivity analysis of breakout pressure—comparison of anisotropic
approach and true-triaxial failure criteria
To demonstrate the importance of incorporating elastic anisotropy
and rock strength anisotropy in predicting breakout pressure, four in-
dividual figures have been generated illustrating the variation of the
breakout pressure of an inclined wellbore with α w = 130° and βw = 60°.
It is important to note that although the rock properties in Table 1 and
Table 2 characterise an anisotropic material, here, for illustration pur-
poses, the Hiramatsu-Oka equation was also used by treating the rock as
an elastically isotropic material, apart from the Lekhnitskii-Amadei
solution, to calculate the near-wellbore stress. All results are tabulated
in Fig. 9.
When the rock was elastically treated as an isotropic material, i.e.
the Hiramatsu-Oka solution was deployed, the preferable drilling
direction was found to be at around N 120°E (Fig. 9a). A similar trend
was observed when the Mogi-Coulomb true-triaxial failure criterion
was used, as shown in Fig. 9b. However, when the rock's elastic ani-
sotropic behaviour was considered, in which the near-wellbore stress
was calculated using Leknitskii-Amadei solution, the preferable drilling
orientation shift towards around N 160°E (Fig. 9c and Fig. 9d).
For intact rock breakout pressure, shown as a black curve in the
wellbore inclination and azimuth sensitivity plot, the inclusion of the
intermediate stress embedded in the Mogi-Coulomb criterion is re-
sponsible for reducing the breakout limit by approximately 1.0 ppg; see
for example Fig. 9b in which the intact rock breakout is 1.0 ppg less
than that of Fig. 9a. It is also obvious that not just the overall minimum
breakout pressure is changing but the preferable drilling direction, re-
presented by the dark blue area, is also shifted azimuthally.
Another interesting observation that can be made from the figures is
that the risk posed by the weak bedding plane is masked by the iso-
tropic assumption. For instance, the blue shaded area in Fig. 9b is much
less than that of Fig. 9d as such the risk posed by the bedding slippage

321
N.B. Setiawan, R.W. Zimmerman
for a wellbore with an inclination between 15° and 50° is not captured
when isotropic stress model is used. In contrast, an extra mud weight of
approximately 4.5 ppg is necessary to avoid bedding slippage when the
anisotropic approach is used. In all cases in Fig. 9, the weak bedding
plane governs the breakout limit for the given wellbore orientation.
Additionally, Fig. 9a and Fig. 9b suggest that wellbore instability could
be avoided by increasing the mud weight by at least 2.5 ppg from the
current 10.0 ppg, i.e. to push the white dot in the sensitivity plots
higher than the blue shaded area. In contrast, the anisotropic model
shows that at least 14.5 ppg is needed to prevent breakout.
6. Conclusions
Some reservoir rocks, notably shales, are intrinsically anisotropic,
and their physical properties, such as strength and deformability, will
depend on the orientation of the rock with respect to the principal
stresses. To comprehensively examine the dependency of the rock an-
isotropic behaviour and the wellbore instability propensity, a hy-
pothetical case was built in which the strength anisotropic and rock
deformability characteristic are incorporated based on the Vaca Muerta
shale. For this purpose, the three-dimensional stress at the wellbore
wall was calculated using the Lekhnitskii-Amadei solution. Combined
with the Mogi-Coulomb and Jaeger plane of weakness failure criteria,
the potential of rock failure in shearing mode for the intact rock and the
bedding plane was then evaluated. The use of Mogi-Coulomb criterion
allows the onset of the intact rock shear failure to be predicted ac-
counting all the three principal stresses. It is demonstrated that the
inclusion of intermediate stress embedded in the Mogi-Coulomb cri-
terion is responsible for reducing the breakout limit by approximately
1.0 ppg.
With regards to the effect of elastic anisotropy, it was found that a
higher degree of rock elastic anisotropy appears to induce higher stress
concentration. The difference with that of the isotropic case could reach
approximately 25% for the highest degree of anisotropy assumed in the
current model. The findings emphasise not only that considering in-
termediate stress is of great importance, but more importantly, ne-
glecting elastic anisotropic characteristic would also affect the breakout
pressure attributed to the reduction of the tangential stress and axial
stress by approximately 5% and 35%, respectively.
By incorporating the anisotropic strength behaviour, it is demon-
strated that the influence of weak bedding planes diminishes when the
relative angle between the wellbore axis and the bedding normal is low.
In practical engineering design, this suggests that wells drilled sub-
parallel through a fracture or fault are less stable than those drilled at
an oblique angle. Finally, the risk posed by the weak bedding plane
might be masked by the isotropic assumption. Thus, it is of paramount
importance to have a more robust modelling approach to improve the
drilling design, for instance, by way of optimizing the mud weight
design and obtaining a more favourable drilling direction.
Acknowledgements
This study is supported by the Indonesian Endowment Fund for
Education (LPDP) of the Republic of Indonesia.
References
1. Moazzeni AR, Nabaei M, Shahbazi K, Shadravan A. Mechanical earth modelling improves
drilling efficiency and reduces non-productive time (NPT). In: Proceedings of the SPE Deep
Gas Conference and Exhibition. 24–26 January, 2010Bahrain. Paper SPE 131718.
2. Caicedo HU, Pribadi MA, Bahuguna S, Wijnands FM, Setiawan NB. Geomechanics, ECD
management and RSS to manage drilling challenges in a mature field. In: Proceedings of the
SPE Oil and Gas India Conference and Exhibition. 20–22 January 2010, Mumbai. Paper SPE
129158.
3. Xi G, Labbassen N, Al Badi BS, Lecoq TF, et al. Mud weight optimization to reduce non-
productive-time while drilling through shale/carbonate sequence of UAE: a case study. In:
Proceedings of the Abu Dhabi International Petroleum Exhibition and Conference. 9–12
November 2015. Paper SPE 177851.
322
International Journal of Rock Mechanics and Mining Sciences 112 (2018) 313–322
4. Ong SH, Power WL, Sitio A, Tanjung E Geomechanics improves drilling operations and
reduces non-productive times (NPT) in Kilo Field, Offshore Northwest Java. In: Proceedings
of the SPE/IATMI Asia Pacific Oil & Gas Conference and Exhibition. 20–22 October 2015,
Nusa Dua, Bali, Indonesia. Paper SPE 176445.
5. Hamid O, Khan K, Rahim Z. et al. Reducing drilling operational risk and non-productive
time using real-time geomechanics surveillance. In: Proceedings of the International
Petroleum Technology Conference. 14–16 November 2016. Bangkok. Paper IPTC-18793.
6. Naidu NA, Hinaai Q, Ismaili I. et al. Successful drilling to completion of a highly stressed
depleted tight gas sandstone reservoir based on geomechanical study. In: Proceedings of the
Abu Dhabi International Petroleum Exhibition and Conference. 7–10 November 2016. Abu
Dhabi. Paper SPE 183380.
7. Zadvornov DA, Grachev OV, Maximov DA. et al. Innovative geomechanical approach lead
to successful drilling of the first highly deviated well at the Kruzenshternskoe Field.
Proceedings of the SPE Russian Petroleum Technology Conference and Exhibition. 24–26
October 2016. Moscow. Paper SPE 182032-MS.
8. Bradley WB. Failure of inclined boreholes. J Energy Resourc Tech. 1979;101:232–239.
9. Gaede O, Karpfinger F, Jocker J, Prioul R. Comparison between analytical and 3d finite
element solutions for borehole stresses in anisotropic elastic rock. Int J Rock Mech Min Sci.
2012;51:53–63.
10. Aadnoy BS. Modeling of the stability of highly inclined boreholes in anisotropic rock
formations. SPE Drill Eng. 1988;3(3):259–268.
11. Ong SH, Roegiers JC. Fracture initiation from inclined wellbores in anisotropic formations.
Soc Petrol Eng J. 1995;48(7):612–619.
12. Wang Z. Seismic anisotropy in sedimentary rocks, Part 2: laboratory data. Geophysics.
2002;67:1423–1440.
13. Suarez-Rivera R, Green SJ, McLennan J, Bai M. Effect of layered heterogeneity on fracture
initiation in tight gas shales. In: Proceedings of the SPE Annual Technical Conference and
Exhibition . 24–27 September 2006. San Antonio. Paper SPE 103327.
14. Ambrose J, Zimmerman RW Failure of anisotropic shales under triaxial compression and
extension. In: Proceedings of the 13th ISRM International Congress of Rock Mechanics. 10–13
May 2015, Montreal, Canada. Paper ISRM-13CONGRESS-2015-2344.
15. Ambrose J, Zimmerman RW, Suarez-Rivera R Failure of shales under triaxial compressive
stress. In: Proceedings of the 48th U.S. Rock Mechanics Symposium. 1–4 June 2014,
Minneapolis. Paper ARMA 14-7506.
16. Shi XC, Yang X, Meng YF, Li G. An anisotropic strength model for layered rocks considering
planes of weakness. Rock Mech Rock Eng. 2016;49:3783–3792.
17. Tien YM, Kuo MC. A failure criterion for transversely isotropic rocks. Int J Rock Mech Min
Sci. 2001;38:399–412.
18. Ong SH, Roegiers JC. Influence of anisotropies in borehole stability. Int J Rock Mech Min
Sci. 1993;30:1069–1075.
19. Ong SH, Roegiers JC Horizontal wellbore collapse in an anisotropic formation. In:
Proceedings of the SPE Production Operations Symposium. 21–23 March 1993, Oklahoma
City. Paper SPE25504.
20. Okland D, Cook JM Bedding-related borehole instability in high-angle wells. In: Proceedings
of the SPE/ISRM Rock Mechanics in Petroleum Engineering Conference. 8–10 July 1998.
Trondheim, Norway. Paper SPE/ISRM 47285.
21. Zhang J. Borehole stability analysis accounting for anisotropies in drilling to weak bedding
planes. Int J Rock Mech Min Sci. 2013;60:160–170.
22. Wu B, Tan CP Effect of shale bedding plane failure on wellbore stability – example from
analyzing stuck-pipe wells. In: Proceedings of the 44th U.S. Rock Mechanics Symposium.
27–30 June 2010, Salt Lake City. Paper ARMA 10-350.
23. Gazaniol D, Forsans T, Boisson MJF, Piau JM. Wellbore failure mechanisms in shales:
prediction and prevention. J. Petrol. Technol. 1995;47(7):589–595.
24. Amadei B. Rock Anisotropy and the Theory of Stress Measurements. Berlin: Springer-Verlag;
1983.
25. Kanfar MF, Chen Z, Rahman SS. Risk-controlled wellbore stability analysis in anisotropic
formations. J Petrol Sci Eng. 2015;134:214–222.
26. Lekhnitskii SG. Theory of Elasticity of an Anisotropic Elastic Body. San Francisco: Holden-
Day; 1963.
27. Gupta D, Zaman M. Stability of boreholes in a geologic medium including the effects of
anisotropy. Appl Math Mech Engl Ed. 1999;20:837–866.
28. Yan G, Karpfinger F, Prioul R, Tang H, Jiang Y, Liu C Anisotropic wellbore stability model
and its application for drilling through challenging shale gas wells. In: Proceedings of the
International Petroleum Technology Conference. 10–12 December 2014, Kuala Lumpur. Paper
IPTC 18143.
29. Lu YH, Chen M, Jin Y, et al. Anisotropic wellbore stability model for transversely isotropic
formation and its application in drilling through shale formation. In: Proceedings of the SPE
Asia Pacific Unconventional Resources Conference and Exhibition. 9–11 November 2015,
Brisbane, Australia. Paper SPE-176922.
30. Li Q, Yousefzadeh A, Aguilera R Anisotropic wellbore stability model for naturally frac-
tured and foliated formations with multi-weakness plane failure. In: Proceedings of the SPE/
CSUR Unconventional Resources Conference. 20–22 October 2015, Calgary. Paper SPE-
175896.
31. Mogi K. Effect of the intermediate principal stress on rock failure. J Geophys Res.
1967;72:5117–5131.
32. Colmenares LB, Zoback MD. A statistical evaluation of intact rock failure criteria con-
strained by polyaxial test data for five different rocks. Int J Rock Mech Min Sci.
2002;39:695–729.
33. Al-Ajmi AM, Zimmerman RW. Relation between the Mogi and the Coulomb failure criteria.
Int J Rock Mech Min Sci. 2005;42:431–439.
34. Jaeger JC. Shear failure of anisotropic rocks. Geol Mag. 1960;97(1):65–72.
35. Lee H, Ong SH, Azeemuddin M, Goodman H. A wellbore stability model for formations
with anisotropic rock strengths. J Petrol Sci Eng. 2012;96–97:109–119.
36. Ambrose J. Failure of Anisotropic Shales Under Triaxial Stress Condition. [Ph.D. thesis].
Imperial College London; 2014.
37. Al-Ajmi AM, Zimmerman RW. Stability analysis of vertical boreholes using the
Mogi–Coulomb failure criterion. Int J Rock Mech Min Sci. 2006;43:1200–1211.
38. Hiramatsu Y, Oka Y. Stress around a shaft or level excavated in ground with a three-
dimensional stress state. Mem Fac Eng Kyoto Univ. 1962;24:56–76.