Energies 12 00942 PDF

energies
Article
Uncertainty Evaluation of Safe Mud Weight Window
Utilizing the Reliability Assessment Method
Tianshou Ma 1,2, * , Tao Tang 1 , Ping Chen 1, * and Chunhe Yang 2
1 State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University,
Chengdu 610500, China; 201721000630@stu.swpu.edu.cn
2 State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics,
Chinese Academy of Sciences, Wuhan 430071, China; chyang@whrsm.ac.cn
* Correspondence: matianshou@swpu.edu.cn (T.M.); chenping@swpu.edu.cn (P.C.);
Tel.: +86-28-8303-2458 (T.M.); +86-28-8303-2707 (P.C.)

Received: 8 January 2019; Accepted: 8 March 2019; Published: 12 March 2019
Abstract: Due to the uncertainty of formation properties and improper wellbore stability analysis
methods, the input parameters are often uncertain and the required mud weight to prevent
wellbore collapse is too large, which might cause an incorrect result. However, the uncertainty
evaluation of input parameters and their influence on safe mud weight window (SMWW) is seldom
investigated. Therefore, the present paper aims to propose an uncertain evaluation method to
evaluate the uncertainty of SMWW. The reliability assessment theory was introduced, and the
uncertain SMWW model was proposed by involving the tolerable breakout, the Mogi-Coulomb
(MG-C) criterion and the reliability assessment theory. The influence of uncertain parameters on
wellbore collapse, wellbore fracture and SMWW were systematically simulated and investigated by
utilizing Monte Carlo simulation. Finally, the field observation of well SC-101X was reported and
discussed. The results indicated that the MG-C criterion and tolerable breakout is recommended
for wellbore stability analysis. The higher the coefficient of variance is, the higher the level of
uncertainty is, the larger the impact on SMWW will be, and the higher the risk of well kick,
wellbore collapse and fracture will be. The uncertainty of basic parameters has a very significant
impact on SMWW, and it cannot be ignored. For well SC-101X, the SMWW predicted by analytical
solution is 0.9921–1.6020 g/cm3 , compared to the SMWW estimated by the reliability assessment
method, the reliability assessment method tends to give a very narrow SMWW of 1.0756–1.0935 g/cm3
and its probability is only 80%, and the field observation for well kick and wellbore fracture verified
the analysis results. For narrow SMWW formation drilling, some kinds of advanced technology,
such as the underbalanced drilling (UBD), managed pressure drilling (MPD), micro-flow drilling
(MFD) and wider the SMWW, can be utilized to maintain drilling safety.
Keywords: safe mud weight window; wellbore stability; reliability assessment method; uncertainty
evaluation; collapse pressure; fracture pressure; pore pressure
1. Introduction
Due to the exhaustion of conventional petroleum resources, much attention is being paid to
some ultra-deep, unconventional and deep-water petroleum resources. To exploit these kinds of
petroleum resources, more and more deep wells and ultra-deep wells are utilized [1,2]. The initial
deep drilling was just made for petroleum exploration and production, but now is being applied to
exploit deep geothermal energy and geo-resources, and to conduct international continental scientific
drilling [1]. The initial oil well was only drilled to a depth of 19.8 m, while current oil wells can be
drilled to more than 10,000 m, and the deepest well is the SG-3 well that was drilled to a depth of
Energies 2019, 12, 942; doi:10.3390/en12050942 www.mdpi.com/journal/energies

Energies 2019, 12, 942 2 of 31
Energies 2019, 12, x FOR PEER REVIEW 2 of 29
12,262 m
deeper thanon 6000–8000
the Kola Peninsula
m. Deep and in Russia
ultra-deep in 1984. Currently,
drilling usually most oil wells
encounters somecanchallenges,
be drilled such
deeper as
the high-temperature and high-pressure (HTHP), wellbore collapse, wellbore fracture,as lost
than 6000–8000 m. Deep and ultra-deep drilling usually encounters some challenges, such the
high-temperature
circulation, andand
gas kick, high-pressure
blowout [1].(HTHP),The reasons wellbore
maycollapse, wellbore fracture,
be the uncertainty lost circulation,
of formation pressure,
gas kick, and blowout [1]. The reasons may be the uncertainty
the uncertainty of formation lithology, the uncertainty of the reservoir interface of formation pressure, the uncertainty
depth, the
of formationof
uncertainty lithology,
completion the uncertainty
depth, and of the theinteraction
reservoir interface
of multiple depth, the uncertainty
factors. In order toofminimize
completion or
depth, and the interaction of multiple factors. In order to minimize
avoid the above challenges that encountered in deep and ultra-deep drilling, drilling engineers or avoid the above challenges that
encountered
need to keep in the deep and ultra-deep
wellbore drilling,
pressure within a drilling engineers
safe window. If theneed
safetowindow
keep theiswellbore
converted pressure
to the
within a safe
equivalent window.
density, theIfsafe
the window
safe window can be is converted
called a safe to mud
the equivalent
weight windowdensity,(SMWW).
the safe window
The design can
be called a safe mud weight window (SMWW). The design of mud
of mud weight should follow the following principle: the safe mud weight should be higher than weight should follow the following
principle:
the lowestthe safesafemudmud weight
weight andshould
lowerbe thanhigher than thesafe
the highest lowest
mudsafe mud In
weight. weight andthe
general, lower thansafe
lowest the
highest
mud safe mud
weight weight.
depends on theIn general,
minimum thebetween
lowest safe the mud
pore weight
pressure depends on the pressure,
and collapse minimumwhile between the
the pore pressure and collapse pressure, while the highest safe
highest safe mud weight depends on the fracture pressure [3–5]. The important foundation mud weight depends on the fracture of
pressure should
SMWW [3–5]. The important foundation
be determined by wellbore of SMWW
stability.should be determined
The wellbore by wellbore
instability, stability.
a classical rock
The wellbore
mechanics instability,
problem, is onea classical
of the mostrock mechanics
complex problems problem,that is one of the mostduring
encountered complex problems
drilling and
that encountered during drilling and completion, which cost the drilling
completion, which cost the drilling industry certainly more than $100 million per year worldwide industry certainly more than
$100 million
[5,6]. Wellbore per year worldwide
instability [5,6]. when
is recognized Wellbore the instability
hole diameter is recognized
is markedly when the hole
different fromdiameter
the bit
is markedly different from the bit size and the hole does not maintain
size and the hole does not maintain its structural integrity [3]. In general, the mechanical failure its structural integrity [3].
In general, the mechanical failure occurs when wellbore stress concentrations
occurs when wellbore stress concentrations exceed the rock strength, and the mechanical failure of exceed the rock strength,
and the mechanical
wellbore failure of
can be classified wellbore
into two main can betypesclassified
(Figureinto1):two main types (Figure
(1) Compressive 1): i.e.,
failure, (1) Compressive
“tight hole”
or “stuck pipe” incidents, which are time-consuming to solve and therefore expensive,and
failure, i.e., “tight hole” or “stuck pipe” incidents, which are time-consuming to solve therefore
an increased
expensive,
borehole an increased
diameter borehole
will occur due diameter
to brittle will
failure occur
andduetheto brittle failure
subsequent andof
caving thethe
subsequent
wellbore wallcaving in
of the wellbore wall in brittle rocks, a reduced borehole diameter which
brittle rocks, a reduced borehole diameter which occurs in weak (plastic) shales, sandstones, and occurs in weak (plastic) shales,
sandstones,
salts; and (2)and salts; failure,
Tensile and (2) Tensile failure,
i.e., “lost i.e., “lostorcirculation”
circulation” “mud loss”orproblems,
“mud loss” problems,
which which are
are potentially
potentially dangerous,
dangerous, thus representingthus representing
a safety riska that safety hasrisk that
to be has to be avoided.
avoided.
Figure 1. Typical wellbore instability. (a) Stable borehole; (b) Enlarged borehole; (c) Borehole breakout;
Figure 1. Typical wellbore instability. (a) Stable borehole; (b) Enlarged borehole; (c) Borehole
(d) Restricted borehole; (e) Borehole fracture.
breakout; (d) Restricted borehole; (e) Borehole fracture.
In order to investigate the wellbore stability, a large number of analysis methods has been
In order to investigate the wellbore stability, a large number of analysis methods has been
proposed, such as the elastic model, plastic model, elastoplastic model, poro-elastic model,
proposed, such as the elastic model, plastic model, elastoplastic model, poro-elastic model,
thermo-poro-elastic model, chemo-poro-elastic model and chemo-thermo-poro-elastic model [2–13].
thermo-poro-elastic model, chemo-poro-elastic model and chemo-thermo-poro-elastic model [2–13].
However, due to the uncertainty of formation lithology, the uncertainty of formation pressure,
However, due to the uncertainty of formation lithology, the uncertainty of formation pressure, the
the uncertainty of mechanical properties of rocks, and the unstable wellbore pressure, the input
uncertainty of mechanical properties of rocks, and the unstable wellbore pressure, the input
parameters of wellbore stability analysis never can be known precisely. In other words, the input
parameters of wellbore stability analysis never can be known precisely. In other words, the input
parameters are often uncertain, which might cause an incorrect result [14–30]. In order to quantify the
parameters are often uncertain, which might cause an incorrect result [14–30]. In order to quantify
influence of uncertain parameters on wellbore stability and SMWW, it’s necessary to utilize reliability
the influence of uncertain parameters on wellbore stability and SMWW, it’s necessary to utilize
assessment method. Morita [14] conducted an uncertainty analysis of borehole stability based on a
reliability assessment method. Morita [14] conducted an uncertainty analysis of borehole stability
statistical error analysis method. McLellan and Hawkes [15] indicated that the traditional models
based on a statistical error analysis method. McLellan and Hawkes [15] indicated that the
fail to account for the inherent variability of rock properties, as well as uncertain values for input
traditional models fail to account for the inherent variability of rock properties, as well as uncertain
parameters, a probabilistic technique was used to evaluate the risks of borehole instability or sand
values for input parameters, a probabilistic technique was used to evaluate the risks of borehole
production, and a spreadsheet-based Monte Carlo simulation tool was involved. Ottesen et al. [16]
instability or sand production, and a spreadsheet-based Monte Carlo simulation tool was involved.
present a new analysis method of wellbore stability based on the quantitative risk analysis (QRA)
Ottesen et al. [16] present a new analysis method of wellbore stability based on the quantitative risk
principles. De Fontoura [17] investigated three analytical methods for evaluating the influence of
analysis (QRA) principles. De Fontoura [17] investigated three analytical methods for evaluating
parameter uncertainties on wellbore stability, such as the first order second moment, first order
the influence of parameter uncertainties on wellbore stability, such as the first order second
moment, first order reliability model and statistical error analysis method, and contrasted with the
Energies 2019, 12, 942 3 of 31
reliability model and statistical error analysis method, and contrasted with the results generated by
Monte Carlo method. Moos et al. [18] and Zoback [6] presented the use of QRA to formally account
for the uncertainty in each input parameter to assess the probability of achieving a desired degree of
wellbore stability at a given mud weight. Sheng et al. [19] indicated that because of uncertainty in
some key influential parameters, these can lead to uncertainty of wellbore stability, a Monte Carlo
uncertainty analysis technique was combined with a numerical geomechanical modelling method
to develop a geostatistical approach to determine the uncertainty of wellbore stability. Al-Ajmi and
Al-Harthy [20] calculated the value of drilling fluid pressure as a probability distribution by utilizing
a probabilistic approach captures uncertainty in input variables through running a Monte Carlo
simulation. Aadnøy [21], Aadnoy and Looyeh [4] also utilized QRA to assess the effects of the errors
or uncertainties associated with the key data on the instability analysis. Zhang et al. [22] deduced a
reliability calculation formula for collapse pressure in coal seam drilling by integration of the reliability
theory and the Hoek-Brown criterion, and the Weibull distribution and Monte Carlo simulation was
utilized in this study. Udegbunam et al. [23] investigated typical fracture and collapse models with
respect to inaccuracies in the input data with a stochastic method. Kinik et al. [24,25] presented a
mathematical model to estimate the true fracture pressure from the leak-off-test data, and the QRA was
also involved to represent the probability-density distribution of fracture pressure. Gholami et al. [26]
applied the QRA to consider the uncertainty of input parameters in determining of the mud weight
window for different failure criteria, such as the Mohr-Coulomb (M-C), Hoek-Brown (H-B) and
Mogi-Coulomb (MG-C) criterion. Plazas et al. [27] investigated a stochastic approach along with the
conventional wellbore stability analysis enables to take into account uncertainty from input data.
Although the reliability assessment method had been integrated with wellbore stability models,
but most of the above studies still have some shortcomings: (1) The required mud weight to prevent
wellbore collapse is too high [28,29], due to the fact traditional wellbore stability models assume that
the critical failure appears at the highest point of stress concentration, in other words, it’s performed to
yield no shear failure along the borehole wall [2,28]. In real drilling engineering, a tolerable breakout
or an appropriate breakout width will not cause an unbearable collapse problem, and it can help
lower the required mud weight to prevent collapse; once a tolerable breakout or an appropriate
breakout width was involved, the SMWW can be therefore widened, and it’s beneficial to the narrow
SMWW formations. However, most of the above studies did not involve a tolerable breakout or an
appropriate breakout width, only Morita [14] and Aadnoy and Looyeh [4] investigated its influence
on wellbore collapse. (2) Most of wellbore stability analysis that involved the reliability assessment
method always uses the M-C criterion to determine the shear failure, while the M-C criterion ignored
the influence of the intermediate principal stress, which makes the required mud weight to prevent
wellbore collapse too conservative. Al-Ajmi and Zimmerman [7,8] introduced an analytical solution of
collapse pressure in conjunction with a true triaxial criterion, i.e., the MG-C criterion, it overcomes
the above defect. (3) Most of the above studies just focus on the uncertainty evaluation of wellbore
collapse, and the uncertainty evaluation of SMWW is seldom investigated, and the influences of
uncertain basic parameters on SMWW are also seldom investigated. Therefore, this study takes the
tolerable breakout and the MG-C criterion into account to determine the required SMWW by using
the reliability assessment method. Firstly, the basic principle of reliability assessment theory was
introduced briefly. Secondly, the uncertain SMWW model was proposed by involving the tolerable
breakout, the MG-C criterion and the reliability assessment method. Thirdly, the uncertainty of input
parameters was investigated, and the uncertainty analysis of the equivalent mud weight of collapse
pressure (EMWCP), the equivalent mud weight of fracture pressure (EMWFP), the equivalent mud
weight of collapse pressure (EMWPP) and the SMWW were investigated. Finally, the field observation
of well SC-101X was reported and discussed.
Energies 2019, 12, 942 4 of 31
2. Reliability Assessment Theory

Reliability calculations provide a means of evaluating the combined effects of uncertainties, and a
means of distinguishing between conditions where uncertainties are particularly high or low [29].
“Reliability” as it is used in reliability theory is the probability of an event occurring or the probability
of a “positive outcome”. Based on the theory of reliability assessment, we can sort the influencing
factors into two types: (1) Loads Q, and (2) Resistances R [30]. The SMWW assessment needs to
consider the influence of well kick, wellbore collapse, and wellbore fracture:
• For well kick, the loads Q denotes the pore pressure, while the resistances R denotes the wellbore
pressure; and the basic random variables of loads and resistances of well control can be assumed

as XQK = XQK1 , XRK2 , · · · , XQKn , XRK = { XRK1 , XRK2 , · · · , XRKn }.
• For wellbore collapse, the loads Q denotes the collapse pressure that controlled by in-situ stress,
pore pressure and rock properties; while the resistances R denotes the wellbore pressure and rock
strength; and the basic random variables of loads and resistances of wellbore collapse can be

assumed as XQC = XQC1 , XRC2 , · · · , XQCn , XRC = { XRC1 , XRC2 , · · · , XRCn }.
• For wellbore fracture, the loads Q denotes the wellbore pressure; while the resistances R denotes
the collapse pressure that controlled by in-situ stress, pore pressure, rock properties and rock
strength; and the basic random variables of loads and resistances of wellbore fracture can be

assumed as XQF = XQF1 , XRF2 , · · · , XQFn , XRF = { XRF1 , XRF2 , · · · , XRFn }.
The loads and resistances can be assumed as:

(
Qk = Q XQk = Q XQk1 , XRk2 , · · · , XQkn
(1)
Rk = R(XRk ) = R( XRk1 , XRk2 , · · · , XRkn )
where Qk and Rk are the loads and resistances; k denotes the subscript, k can value for K, C and F for
well kick, wellbore collapse, and wellbore fracture, respectively.
The margin of safety, M, is the difference between the resistance and the load. The margin of
safety for well control, wellbore collapse and wellbore fracture can be expressed as [30,31]:
Mk = R k − Q k (2)
where Mk is the margin of safety for factor k.

In general, the loads and resistances are two independent random variables, and the reliability
and failure probability can be expressed as [30,31]:
Prk = P( Mk > 0) = P[( Rk − Qk ) > 0] (3)
Pf k = P( Mk < 0) = P[( Rk − Qk ) < 0] (4)
where Prk is the reliability of factor k; Pfk is the failure probability of factor k.
There is a relationship between the reliability and failure probability:
Prk + Pf k = 1 (5)
If we assume that the probability density function of Qk and Rk are fQ (Qk ) and fR (Rk ), respectively.
Figure 2 shows the interference of the probability density function of Qk and Rk , we just take the
assessment of wellbore collapse as an example, the well control and wellbore fracture is similar.
The overlapping zone denotes the probability of failure. The smaller overlapping zone, the wellbore
will be more reliable, i.e., the risk of wellbore collapse will be lower; the bigger overlapping zone,
Energies 2019, 12, 942 5 of 31
the wellbore will be more unreliable, i.e., the risk of wellbore collapse will be higher. The reliability
and failure probability can be expressed as [31]:
( R∞ R∞R∞
Prk = P( Mk > 0) = 0 f ( Mk )dMk = 0 0 f Q ( Mk + Qk ) f R ( Rk )dRk dMk
R0 R0 R∞ (6)
Pf k = P( Mk < 0) = −∞ f ( Mk )dMk = −∞ − M f Q ( Mk + Qk ) f R ( Rk )dRk dMk
k
where fQ (Qk )Energies

and f2019,
R (R12,
k )x are the probability density function of Qk and Rk , respectively.
FOR PEER REVIEW 5 of 29
Figure 2. Probability densities

Figure 2. Probability for the
densities resistances
for the (R
resistances (RC)C ) and
and loadsloads
(QC) of(Q C ) of collapse.
wellbore wellbore collapse.
3. Modeling 3.for
Modeling for Safe
Safe Mud Mud Weight
Weight Window
Window Utilizing Reliability
Utilizing ReliabilityAssessment
Assessment
3.1. Stress Distribution on the Borehole Wall
3.1. Stress Distribution on the Borehole Wall
In order to determine the stress distribution on the wall of the vertical borehole, several basic
In orderassumptions
to determineare made thetostress
proposedistribution
this model [3–5]:on thematerial
(1) The wall of the formation
of deep verticalisborehole,
linear elastic,several basic
homogeneous and isotropic; (2) The mechanical model can be simplified as a plane-strain problem;
assumptions are made to propose this model [3–5]: (1) The material of deep formation is linear elastic,
(3) The formation rock obeys the small deformation assumption; (4) In-situ stress state (ΗH, Ηh, Ηv),
homogeneous poreand isotropic;
pressure (2) The
(pp), strength mechanical
parameters (c, Π) and model
materialcan be simplified
parameters as aBased
(v) are known. plane-strain
on the problem;
(3) The formation rock obeys
stress distribution thethesmall
around deformation
wellbore, and considering assumption;
the effects of flow (4)induced
In-situstress,
stressthe state
stress (σH , σh , σv ),
pore pressurecomponents on the wall of the borehole (r = R) can be expressed as [5],
(pp ), strength parameters (c, ϕ) and material parameters (v) are known. Based on the
σ r = pm
stress distribution around the wellbore, °° and considering the effects of flow induced stress, the stress
®σ θ = − pm + ( 1 − 2 cos 2θ ) σ H + ( 1 + 2 cos 2θ ) σ h (7)
components on the wall of the borehole ° (r = R) can be expressed as [5]:
°̄σ z = σ v − 2 v (σ H − σ h ) cos 2θ

where Ηv, ΗH and Ηh are the vertical, maximum and minimum horizontal in-situ stresses respectively,
 σr = pm
MPa; pm is the wellbore pressure, MPa; v is the Poisson’s ratio; Ό is the angle of circumference, (°); Ηr,
ΗΌ and Ηz are the = − pm +on(1the
σθ components
stress −borehole
2 cos 2θwall, + (1 + 2 cos 2θ )σh
)σHMPa. (7)
σz =
For a vertical well,
 σv − 2vto(the
according − σh ) cos
σH Equation (1), 2θ
tangential stress (ΗΌ) and axial stress (Ηz) are
functions of the angle Ό, and the tangential and axial stresses will vary sinusoidally, as shown in
where σv , σHFigure
and σ3.hBoth
are tangential and axial stresses reach the highest magnitude when Ό = Δ/2 or 3Δ/2, i.e.,
the vertical, maximum and minimum horizontal in-situ stresses respectively,
Όmax = Δ/2 or 3Δ/2.
MPa; pm is the wellbore pressure, MPa; v is the Poisson’s ratio; θ is the angle of circumference, (◦ ); σr ,
σθ and σz are the stress components on the borehole wall, MPa.
σ
For a vertical well, accordingto the Equation (1),
σ
tangential stress (σθ ) and axial stress (σz ) are
r
Wellbore stress / MPa
functions of the angle θ, and the tangential and axialσ stresses will vary sinusoidally, as shown in z
Figure 3. Both tangential and axial stresses reach the highest magnitude when θ = π/2 or 3π/2, i.e.,
θ max = π/2 or 3π/2.

Circumference angle / (°)
Figure 3. The stress distribution on the borehole wall.

MPa; pm is the wellbore pressure, MPa; v is the Poisson’s ratio; θ is the angle of circumference, (°); σr,
σθ and σz are the stress components on the borehole wall, MPa.
For a vertical well, according to the Equation (1), tangential stress (σθ) and axial stress (σz) are
functions of the angle θ, and the tangential and axial stresses will vary sinusoidally, as shown in
Figure 3. Both
Energies 2019, tangential and axial stresses reach the highest magnitude when θ = π/2 or 3π/2,6 of
12, 942 i.e.,
31
θmax = π/2 or 3π/2.
100
σr
95
σθ
Wellbore stress / MPa

90 σz
85
80
75
48
47
0 90 180 270 360
Circumference angle / (°)
The stress
Figure 3. The
Figure stress distribution
distribution on the borehole wall.
3.2. Breakout Width Model
3.2. Breakout Width Model
As shown in Figure 4, the traditional method predicts the collapse pressure using the critical
failure
Aspoint,
shown and
in the critical
Figure failure
4, the point is method
traditional located at theA:collapse pressure using the critical
the point
predicts
failure point, and the critical failure point is located at the point A:
θc = θmax = 0.5π or θc = θmax = 1.5π (8)
θ c =θ max = 0.5π or θ c =θ max = 1.5π (8)
◦ ); θ the angle of critical failure point A, (◦ ).
where θ
where θ cc is
is the
the angle
angle of
of critical
critical failure point B,
failure point B, ((°); θmax
max the angle of critical failure point A, (°).
Figure
Figure 4.
4. The
The required
required cohesion
cohesion to
to keep
keep wellbore stability (Modified from [2]).
[2]).
However, the breakout width model assumes a tolerable breakout width (2ω) to calculate the
point B
collapse pressure, i.e., the critical failure point occurs at the point B [2,28]:
[2,28]:
θ c =θ max ± ω =0.5π ± ω or θ c =θ max ± ω =1.5π ± ω (9)
θc = θmax ± ω = 0.5π ± ω or θc = θmax ± ω = 1.5π ± ω (9)
where ω is the half of the tolerable breakout width, (°).
where is the half
Inωgeneral, theofbreakout
the tolerable breakout
occurs as a width, (◦ ).successive spalls along the direction of the
series of
In general,
minimum the horizontal
principal breakout occurs as aresults
stress and seriesfrom
of successive spalls
shear failure along the
subparallel to direction of the
the free wall of
minimum principal horizontal stress and results from shear failure subparallel
the borehole [2,6,28,32–37]. As shown in Figure 5, where ①, ② and ③ represent the breakout to the free wall of the
borehole
zones [2,6,28,32–37].
during deepening.AsFor
shown in Figure
a tolerable 5, wherethe
breakout, , 2 and
1 wellbore represent
3 collapse the breakout
accident couldzones during
not happen,
because the borehole will form a stable shape after several breakouts. In general, when the breakout
width does not exceed a tolerable value, the wellbore breakouts are usually expected to increase in
depth, but not in width, resulting in a stable borehole after several breakouts [2,6]. The measured
breakout widths also compared very well with those predicted by the simple theory [6]. Thus, the
given point B (θc = θmax ± ω) can be replaced by the traditional critical failure point A to calculate the
Energies 2019, 12, 942 7 of 31
deepening. For a tolerable breakout, the wellbore collapse accident could not happen, because the
borehole will form a stable shape after several breakouts. In general, when the breakout width does
not exceed a tolerable value, the wellbore breakouts are usually expected to increase in depth, but
not in width, resulting in a stable borehole after several breakouts [2,6]. The measured breakout
widths also compared very well with those predicted by the simple theory [6]. Thus, the given point
B (θ c = θ max ± ω) can be replaced by the traditional critical failure point A to calculate the collapse
pressure for a given tolerable breakout. Zoback [6] indicated that the tolerable breakout width (2ω)
to prevent borehole collapse can be set as 90◦ for a vertical borehole. Once the breakout width (2ω)
exceeds
Energiesthe allowable
2019, 12, x FOR breakout width, the borehole cannot maintain its stability.
PEER REVIEW 7 of 29
(a) (b)
Figure
Figure 5. Schematic
5. Schematic representation
representation of breakout
of breakout growth
growth (Reproduced
(Reproduced from
from [6]).[6]).
(a) (a)
TheThe theoretical
theoretical
model after the formation of wellbore breakouts, wellbore breakouts deepen but do not
model after the formation of wellbore breakouts, wellbore breakouts deepen but do not widen; (b) widen;
The (b)
The measured breakout widths compared very well with those predicted by the
measured breakout widths compared very well with those predicted by the simple theory. simple theory.
3.3. Collaspe Pressure

Substituting Model
Equation (9)for
intoM-C Criterion
Equation (7), the stress components of the critical failure point B
can be rewritten as:
3.3.1. M-C Criterion

 σr = pm

If considering
σθ = the
− pmconventional
+ [1 − 2 cos 2effective
(θmax + ωstress concept,
)]σH + the2(detail
[1 + 2 cos θmax +derivation
ω )]σh process (10)
of the
M-C criterion can
 refer to Appendix A, and it
σz = σv − 2v(σH − σh ) cos 2(θmax + ω ) can be expressed as [7]:
σ 1 = C + qσ 3 (11)
3.3. Collaspe Pressure Model for M-C Criterion
where C and q are given as:
3.3.1. M-C Criterion  2c cos ϕ
C = 1 − sin ϕ − ( q − 1) α pp
If considering the conventional effective  stress concept, the detail derivation process of the M-C
 (12)
criterion can refer to Appendix A, and it can q =be sin ϕ
1 +expressed as [7]:
 1 − sin ϕ
σ1 = C + qσ3 (11)
where φ is the friction angle, (°); c is the cohesive strength, MPa; σ1 and σ3 are the major and minor
principal
where C and stress, MPa;as:
q are given α is the Biot’s coefficient; pp is the pore pressure, MPa.
( 2c cos ϕ
C = 1−sin ϕ − (q − 1)αp p
3.3.2. Collapse Pressure 1+sin ϕ (12)
q = 1−sin ϕ
According to Equation (7), the radial stress is perpendicular to the borehole wall, i.e., it’s a
where ϕ is the
principal friction
stress. In angle, (◦ ); c is the
these models cohesive
there are three strength, MPa; σ1 of
permutations and are the
theσ3three major and
principal minorthat
stresses
principal stress, MPa; α is the Biot’s coefficient; p is the pore pressure, MPa.
need to be investigated in order to determinepthe required mud pressure [19]: (1) σz ≥ σθ ≥ σr, (2) σθ ≥
σz ≥ σr, and (3) σθ ≥ σr ≥ σz. In the fields, the case where σθ ≥ σz ≥ σr is the most commonly
encountered stress state corresponding to borehole collapse [19]. Substituting σ1 = σθ, σ2 = σz and σ3 =
σr into Equations (11) and (12), the collapse pressure model with a tolerable breakout width can be
expressed as:
1 − 2 cos 2 (θ max + ω )  σ H + 1 + 2 cos 2 (θ max + ω )  σ h − C
Energies 2019, 12, 942 8 of 31
3.3.2. Collapse Pressure

principal stress. In these models there are three permutations of the three principal stresses that
need to be investigated in order to determine the required mud pressure [19]: (1) σz ≥ σθ ≥ σr ,
(2) σθ ≥ σz ≥ σr , and (3) σθ ≥ σr ≥ σz . In the fields, the case where σθ ≥ σz ≥ σr is the most commonly
encountered stress state corresponding to borehole collapse [19]. Substituting σ1 = σθ , σ2 = σz and σ3
= σr into Equations (11) and (12), the collapse pressure model with a tolerable breakout width can be
expressed as:
[1 − 2 cos 2(θmax + ω )]σH + [1 + 2 cos 2(θmax + ω )]σh − C
pc = (13)
1+q
where pc is the collapse pressure, MPa.
If set the critical failure point θ c = θ max ± ω =π/2 or 3π/2, Equation (13) can be reduced to the
traditional model:
3σ − σh − C
pc = H (14)
1+q
3.3.3. Equivalent Mud Weight of Collapse Pressure (EMWCP)

The collapse pressure (pc ) can be predicted by using Equation (13) or Equation (14), and the
equivalent mud weight (EMW) of collapse pressure (EMWCP) can be expressed as:
pc
EMWCP = (15)
gTVD
where EMWCP is the equivalent mud weight of collapse pressure, g/cm3 ; g is the gravitational
acceleration, g = 9.8 m/s2 ; TVD is the true vertical depth, km.
According to Equations (13)–(15), the true vertical depth (TVD), the gravitational acceleration
(g) and the critical failure point A (θ max ) are usually certain, while the maximum horizontal stress
(σH ), minimum horizontal stress (σh ), tolerable breakout width (ω), pore pressure (pp ), strength
parameters (c, ϕ) and Biot’s coefficient (α) are usually uncertain. The basic random variables of loads
and resistances of wellbore collapse can be expressed as:
(
XQC = σH , σh , p p , α, ω
(16)
XRC = {c, ϕ, pm }
The loads and resistances of wellbore collapse can be expressed as:

(
QC = EMWCP = Q(XQC ) = Q σH , σh , p p , α, ω
(17)
RC = EMWWP = R(XRC ) = R(c, ϕ, pm )
Substituting Equations (13)–(16) into Equation (17), the reliability and failure probability can be
determined by using the reliability assessment theory.
3.4. Collaspe Pressure Model for MG-C Criterion
3.4.1. MG-C Criterion

If considering the conventional effective stress concept, the detail derivation process of the MG-C
criterion can refer to Appendix B, and the MG-C criterion can be expressed as [2,7]:
1
2
I12 − 3I2 = a0 + b0 I1 − σ2 − 2αp p

(18)
Energies 2019, 12, 942 9 of 31
where I1 , I2 , a0 and b0 are given as:

(
I1 = σ1 + σ2 + σ3
(19)
I2 = σ1 σ2 + σ2 σ3 + σ3 σ1
(
a0 = 2c cos ϕ
(20)
b0 = sin ϕ
where σ2 is the intermediate principal stress, MPa; I1 is the first stress tensor invariant, MPa; I2 is the
second stress tensor invariant, MPa.
3.4.2. Collapse Pressure

principal stress. In these models there are three permutations of the three principal stresses that need to
be investigated in order to determine the required mud pressure [19]: (1) σz ≥ σθ ≥ σr , (2) σθ ≥ σz ≥ σr ,
and (3) σθ ≥ σr ≥ σz . In the fields, the case where σθ ≥ σz ≥ σr is the most commonly encountered
stress state corresponding to borehole collapse [19]. Substituting σ1 = σθ , σ2 = σz and σ3 = σr into
Equations (18)–(20), the collapse pressure model with a tolerable breakout width can be expressed as:
1 1n 0 2 o1
− 3( A − 2B)2
2
pc = A− 12 a + b0 A − 2αp p (21)
2 6
where A, B, a0 and b0 are given as:

(
A = [1 − 2 cos 2(θmax + ω )]σH + [1 + 2 cos 2(θmax + ω )]σh
(22)
B = σv − 2v(σH − σh ) cos 2(θmax + ω )
(
a0 = 2c cos ϕ
(23)
b0 = sin ϕ
If set the critical failure point θ 0 = θ max ± ω = π/2 or 3π/2, substituting θ 0 = θ max ± ω = π/2 or
3π/2 into Equation (22), the collapse pressure model can degrade into the traditional model:
(
A = 3σH − σh
(24)
B = σv + 2v(σH − σh )
3.4.3. Equivalent Mud Weight of Collapse Pressure (EMWCP)

The collapse pressure (pc ) can be predicted by using Equation (21), and the EMWCP can be
expressed as:
pc
EMWCP = (25)
gTVD
According to Equations (21)–(25), the TVD, the gravitational acceleration (g) and the critical
failure point A (θ max ) are usually certain, while the vertical stress (σv ), maximum horizontal stress (σH ),
minimum horizontal stress (σh ), Poisson’s ratio (v), tolerable breakout width (ω), pore pressure (pp ),
strength parameters (c, ϕ) and Biot’s coefficient (α) are usually uncertain. The basic random variables
of loads and resistances of wellbore collapse can be expressed as:
(
XQC = σH , σh , σv , p p , v, α, ω
(26)
XRC = {c, ϕ, pm }
Energies 2019, 12, 942 10 of 31
The loads and resistances of wellbore collapse can be obtained:

(
QC = EMWCP = Q(XQC ) = Q σH , σh , σv , p p , v, α, ω
(27)
RC = EMWWP = R(XRC ) = R(c, ϕ, pm )
3.5. Facture Pressure Model
3.5.1. Tensile Failure Criterion

When the tensile stress exceeds the tensile strength, the wellbore fracture will occur. In general,
the maximum tensile stress criterion can be used to determine the wellbore fracture. If considering the
conventional effective stress concept, the tensile failure criterion can be written as:
σ3 − αp p + St = 0 (28)
where St is the tensile strength, in MPa.
3.5.2. Facture Pressure

According to Equation (10), only the hoop stress can obtain the negative value, and the minor
principal stress (σ3 ) occurs at θ = 0◦ or 180◦ . Substituting θ = 0◦ or 180◦ into Equation (10), the minor
principal stress (σ3 ) can be expressed as:
σ3 = 3σh − σH − pm − αp p (29)
Substituting Equation (29) into Equation (28), the fracture pressure model can be obtained:
p f = 3σh − σH − αp p + St (30)
where pf is the fracture pressure, MPa.
3.5.3. Equivalent Mud Weight of Fracture Pressure (EMWFP)

The fracture pressure (pf ) can be predicted by using Equation (30), and the EMW of fracture
pressure (EMWFP) can be expressed as:
pf
EMWFP = (31)
gTVD
where EMWFP is the equivalent mud weight of fracture pressure, g/cm3 .

According to Equations (30) and (31), the TVD and the gravitational acceleration (g) are usually
certain, while the maximum horizontal stress (σH ), minimum horizontal stress (σh ), pore pressure (pp ),
strength parameters (St ) and Biot’s coefficient (α) are usually uncertain. The basic random variables of
loads and resistances of wellbore fracture can be expressed as:
(
XQF = { pm }
(32)
XRF = σH , σh , p p , St , α
The loads and resistances of wellbore fracture can be obatined:

(
QC = EMWWP = Q(XQF ) = Q( pm )
(33)
RC = EMWFP = R(XRF ) = R σH , σh , p p , St , α
Energies 2019, 12, 942 11 of 31
3.6. Equivalent Mud Weight of Pore Pressure (EMWPP)

The equivalent mud weight of pore pressure (EMWPP) also can be obtained by:
pp
EMWPP = (34)
gTVD
where EMWPP is the equivalent mud weight of pore pressure, g/cm3 .

According to Equation (34), the TVD and the gravitational acceleration (g) are usually certain,
while the pore pressure (pp ) is usually uncertain. The basic random variables of loads and resistances
of wellbore fracture can be expressed as:
(
XQK = p p
(35)
XRK = { pm }
The loads and resistances of wellbore fracture can be obtained:

(
QC = EMWPP = Q(XQK ) = Q p p
(36)
RC = EMWWP = R(XRK ) = R( pm )
Substituting Equations (34) and (35) into Equation (36), the reliability and failure probability can
be determined by using the reliability assessment theory.
3.7. SMWW Model

In order to avoid or minimize wellbore instability problems and keep wellbore safety, the wellbore
overflow, wellbore collapse and wellbore fracture should be avoided. The design of mud weight
should follow the following principle: safe mud weight should higher than the lowest safe mud
weight and lower than the highest safe mud weight. In general, the lowest safe mud weight needs to
determine by using the pore pressure and collapse pressure, while the highest safe mud weight needs
to determine by using the fracture pressure, and the SMWW can be determined by [5]:
max{EMWPP, EMWCP}< SMWW < min{EMWFP} (37)
where SMWW is the safe mud weight window, g/cm3 .

Thus, in order to quantitatively assess the reliability of SMWW, we also should determine the
reliability of the lowest safe mud weight by integrating the reliability between EMWPP and EMWCP,
and the reliability of the highest safe mud weight is just determined by EMWFP.
4. Results and Discussion

To investigate the influence of uncertain parameters on wellbore stability and SMWW, take the
basic parameters of well SC-101X as an example, the TVD of the formation is 2008–2432 m, and the basic
parameters are shown in Table 1. For the calculation of wellbore stability and SMWW, Monte Carlo
simulation was utilized with 10,000 iterations.
Energies 2019, 12, 942 12 of 31
Table 1. Basic parameters for SMWW analysis.
No. Parameter Distribution Mean Std Dev Var P5 P95

1 Vertical stress/MPa Normal 54.80 2.74 5% 50.29 59.30
2 Maximum horizontal stress/MPa Normal 43.87 8.77 20% 29.40 58.30
3 Minimum horizontal stress/MPa Normal 30.91 3.09 10% 25.82 35.99
4 Pore pressure/MPa Normal 21.39 2.14 10% 17.86 24.90
5 Biot’s coefficient Normal 0.95 0.05 5% 0.87 1.00
6 Cohesive strength/MPa Normal 18.14 3.63 20% 12.17 24.09
7 Internal friction angle/(◦ ) Normal 35.00 7.00 20% 23.48 46.50
8 Tensile strength/MPa Normal 6.00 1.20 20% 4.02 7.97
9 Poisson’s ratio Normal 0.25 0.05 20% 0.17 0.33
10 Breakout width/(◦ ) Normal 40.00 4.00 10% 33.42 46.55
Note: P5 and P95 denote the lower and upper limit within the confidence interval of 5–95%.
4.1. Uncertainty of Basic Parameters

The basic parameters of wellbore stability analysis are uncertain, and these basic parameters
are not exact values but the average estimates [19], and the average values themselves are known
to have flaws. The real values are usually variable and following some distribution. As shown in
Table 1, the mean value, standard deviation (Std Dev), variance (Var) of each parameter are presented.
The mean of the input data were interpreted and identified from logging data. Regarding the choice
of random variable distributions, some of the probability distributions normally used in wellbore
stability analysis include the uniform distribution, triangular distribution, lognormal distribution,
Gamma distribution, Beta distribution, and Weibull distribution [30]. From an engineering perspective,
the normal distribution is the preferred choice for wellbore stability problems, due to the distribution
functions are mostly normally distributed [38,39]. It should be noticed that the random variable
distributions used in this work are based on assumption, but the choice of random variable
distributions should be revisited. Regarding the uncertainties in the input data, according to the
engineering experience of Aadnøy [21], measurement and interpretation errors are the major sources
of input parameter uncertainties [21,23]. Currently, the uncertainties are only “educated guesses”
with possibility for future improvements, and the assumed uncertainties in both measurement and
interpretation results were listed in Table 2. Although the uncertainties in the input data were based
on assumption, the orders of magnitudes of the input data were assumed correct. On this basis,
Monte Carlo simulation was utilized with 10,000 iterations. The simulating results of the probability
density histogram for basic parameters were shown in Figure 6. It is clearly shown that the uncertain
distribution of each parameter satisfies the normal distribution, the higher the coefficient of variance
is, the higher the level of uncertainty will be, and the wider the distribution of parameter values will
be. The 90% confidence interval (from 5 to 95%, i.e., P5–P95) of each parameter was also listed in
Table 1. For example, the 90% confidence interval is 36.64–51.07 MPa for the estimation of maximum
horizontal stress, while it’s just 50.29–59.30 MPa, 28.35–33.44 MPa, 17.86–24.90 MPa for the estimations
of vertical stress, minimum horizontal stress and pore pressure, respectively. It is clearly shown that
the estimation of maximum horizontal stress has a higher confidence interval of P5 and P95, due to its
high level of uncertainty.
Energies 2019, 12, 942 13 of 31

Table 2. Assumed uncertainties in both measurement and interpretation results.
Table 2. Assumed uncertainties in both measurement and interpretation results.
Uncertainty in Uncertainty in
No. No.Parameter Parameter
Uncertainty in Measurement Uncertainty in Interpretation
Measurement Interpretation
1 Vertical stress 5% 5–10%
1 Vertical stress 5% 5–10%
2 Maximum horizontal stress / 10–20%
2 Maximum horizontal stress / 10–20%
3 Minimum3 horizontal stress
Minimum horizontal stress 5% 5% 5–10%
5–10%
4 Pore
4 pressure Pore pressure 2–5% 2–5% 5–30%
5–30%
5 Biot’s
5 coefficient
Biot’s coefficient / / 5–10%
5–10%
6 Cohesive
6 strength
Cohesive strength / / 10–50%
10–50%
7 7 friction
Internal Internal
angle friction angle / / 10–30%
10–30%
8 8
Tensile Tensile strength
strength / / 10–50%
10–50%
9 9
Poisson’s ratioPoisson’s ratio / / 20–50%
20–50%
0.16 0.05
Samples 10000 Samples 10000
Probability density
Probability density
0.04
0.12
0.03
0.08
0.02
0.04
0.01
0.00 0.00
45 50 55 60 65 10 20 30 40 50 60 70 80
Vertical stress (MPa) Maximum horizontal stress (MPa)
(a) Vertical stress (b) Maximum horizontal stress
0.15 0.20
Probability density
Probability density
0.12
0.15
0.09
0.10
0.06
0.05
0.03
0.00 0.00
20 25 30 35 40 14 16 18 20 22 24 26 28
Minimum horizontal stress (MPa) Pore pressure (MPa)
(c) Minimum horizontal stress (d) Pore pressure
8 0.12
Probability density
Probability density
0.10
6
0.08
4 0.06
0.04
2
0.02
0 0.00
0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 5 10 15 20 25 30
Biot coefficient Cohesive strength (MPa)
(e) Biot’s coefficient (f) Cohesive strength
0.06 0.35
Probability density
Probability density
0.05 0.30
0.25
0.04
0.20
0.03
0.15
0.02
0.10
0.01 0.05
0.00 0.00
10 20 30 40 50 60 2 3 4 5 6 7 8 9 10
Internal friction angle / (°) Tensile strength (MPa)
(g) Internal friction angle (h) Tensile strength
Figure 6. Cont.
Energies 2019, 12, 942 14 of 31
Figure 6. Cont.
Figure 6. Cont.
8 0.12
8 Samples 10000 0.12 Samples 10000
Probability density
Probability density
Samples 10000 0.10 Samples 10000
Probability density
Probability density
6 0.10
6 0.08
0.08
4 0.06
4 0.06
0.04
2 0.04
2 0.02
0.02
0 0.00
0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.00 25 30 35 40 45 50 55
0.10 0.15 0.20Poisson's
0.25 ratio
0.30 0.35 0.40 25 30 Breakout
35 40 width
45 (°) 50 55
Poisson's ratio Breakout width (°)
(i) Poisson’s ratio (j) Breakout width
(i) Poisson’s ratio (j) Breakout width
Figure 6. Probability density histogram of basic parameters.
Probabilitydensity
Figure6.6.Probability
Figure densityhistogram
histogramofofbasic
basicparameters.
parameters.
4.2. Uncertainty Analysis of EMWCP

4.2. Uncertainty Analysis of EMWCP
4.2.1. EMWCP Probability Analysis Analysis for for both
both M-C
M-C andand MG-C
MG-C Criteria
Criteria
4.2.1. EMWCP Probability Analysis for both M-C and MG-C Criteria
EMWCP means the equivalent mud weight of collapse pressure, and there are two types of
EMWCP
collapse model, means the M-C
i.e., the equivalent
model and mudMG-C weight of collapse
model. pressure,the
We compared
compared theand there areresults
simulating
simulating two types
between
betweenof
collapse
M-C model,
M-C model
model and
andi.e.,
MG-C
MG-Cthe M-C
model,
model,modelthe and
theresultsMG-C
results areare model.
shown
shown inWe compared
Figure
in Figure 7, the thestatistics
simulating
statistics
7, the bothbothforresults
M-C between
model
for M-C and
model
M-C
MG-C model and
modelmodel
and MG-C MG-C
was collected model, the results
in Tablein3.Table
was collected are shown
It is clearly
3. It isfound in
clearly Figure
that 7,
the same
found the statistics
that uncertainty both for M-C model
in basic parameters
the same uncertainty in basic
and MG-C
results
parameters model
resultswas
in different collected
answers
in different in Table
foranswers
M-C model 3. M-C
for Itand
is model
clearly
MG-C and found
model.MG-C that
The the same
The uncertainty
analytical
model. solution of
analytical in
thebasic
solution M-C of
parameters
criterion results
(0.6354 in
g/cm different answers
3 ) is obviously for
higher M-Cthanmodel
that and
of theMG-C
MG-C model.
the M-C criterion (0.6354 g/cm ) is obviously higher than that of the MG-C criterion (0.5195 g/cm ),
3 criterion The analytical
(0.5195 g/cm solution of
3 ), and the 3
the
M-C
andM-C
thecriterion
criterion
M-C tends (0.6354
criterion g/cm3)toaishigher
to calculate
tends obviously
calculatemeana higher
EMWCP
higher than
mean that
(0.6438 of
g/cm
EMWCP the 3MG-C
)(0.6438 criterion
than thatg/cmof 3the (0.5195
MG-C
) than thatg/cm 3),
criterion
of the
and
MG-Cthe criterion
(0.5538 M-C 3criterion
g/cm ). This
(0.5538 tends
g/cm3to
depends calculate
onThis
). a higher
the conservative
depends mean
nature
on the ofEMWCP
the M-Cnature
conservative (0.6438 of g/cm
criterion,
3) than that of the
dueM-C
the to the M-C criterion
criterion, due to
MG-C
ignoredcriterion
the M-Cthe (0.5538
influence
criterion g/cm 3 ). This
of the intermediate
ignored depends
the influence on
principalthe conservative
of the stress, nature
which makes
intermediate principal of the M-C
the required criterion,
stress, whichEMW makes due to
to preventthe
the M-C criterion
wellbore
required collapse
EMW tois ignored
prevent the
too conservative. influence
wellbore of theisthe
However,
collapse intermediate
tooMG-C principal
criterion
conservative. overcamestress,the
However, which
the defect
MG-C makes
of the the
M-C
criterion
required
criterion.
overcame EMW
Inthe to words,
other
defectprevent
of the wellbore
the collapse
M-C criterion.
M-C criterion In isother
too words,
under-predictsconservative.
rock However,
thestrength
M-C the
and estimates
criterion MG-C thecriterion
under-predicts EMWCP rock
overcame
too the
conservative,
strength defect
and estimates of
while the the M-C
the EMWCP criterion.
MG-C criterion In other words,
rightly predicts
too conservative, the
whilerock M-C
strength
the MG-C criterion under-predicts
and estimates
criterion the EMWCP
rightly predicts rock
rock
strength
much and
more
strength estimates
estimatesthe
andexactly. theEMWCP
EMWCPtoo conservative,
much more exactly. while the MG-C criterion rightly predicts rock
strength and estimates the EMWCP much more exactly.
2.5 100
Cumulative probability (%)
2.5 100
2.0 M-C 80
Probability density
2.0 M-C
MG-C 80
Probability density
MG-C
M-C
1.5 60
1.5 M-C
MG-C 60
MG-C
1.0 40
1.0 40
0.5 20
0.5 20
0.0 0
0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
3 1.6
Equivalent mud weight (g/cm
3
)
Equivalent mud weight (g/cm )
Figure 7.
Figure 7. Probability
Probability histogram
histogram of
of EMWCP
EMWCP both
both for
for M-C
M-C and
and MG-C
MG-C model.
model.
Figure 7. Probability histogram of EMWCP both for M-C and MG-C model.
Table 3. The
Table 3. The statistics
statistics of
of EMWCP
EMWCP both both forfor M-C
M-C and
and MG-C
MG-C model.
model.
Table 3. The statistics of EMWCP both for M-C and MG-C model.
EMWCP
EMWCP
Statistics
Statistics EMWCP
Statistics M-C
M-C MG-C MG-C
P5/(g/cm33) M-C
0.3221 MG-C
0.2317
P5/(g/cm
P5/(g/cm 3) ) 0.3221
0.3221 0.2317
P95/(g/cm 3 ))
3 0.9825 0.2317
0.9063
P95/(g/cm
P95/(g/cm 3) 0.9825
0.9825 0.9063
Mean/(g/cm 3 ))
3 0.6438 0.9063
0.5538
Mean/(g/cm 0.6438 0.5538
Mean/(g/cm
Analytical
3)
solution/(g/cm 33))
0.6438
0.6354 0.5538
0.5195
Analytical solution/(g/cm 0.6354 0.5195
Analytical
Std Dev/(g/cm33)) ) 0.6354
solution/(g/cm 3
0.2029 0.5195
0.2011
Std Dev/(g/cm 0.2029 0.2011
Std Dev/(g/cm3) 0.2029 0.2011
Energies 2019, 12, 942 15 of 31
As
Asshown
showninin Figure 8, due
Figure to the
8, due to influence of intermediate
the influence principal
of intermediate stress forstress
principal the MG-C criterion,
for the MG-C
the rock strength
criterion, the rockincrease
strengthfirstly and then
increase firstlydecrease,
and thenbut the rockbut
decrease, strength
the rock is much higher
strength than that
is much of
higher
the
thanM-C criterion.
that of the M-CMoreover, the standard
criterion. Moreover, deviation in the M-C
the standard criterion
deviation in istheslightly higher than
M-C criterion that of
is slightly
the MG-C
higher than criterion,
that of thewhich
MG-Calsocriterion,
reveals that
which thealso
level of uncertainty
reveals estimated
that the level by MG-Cestimated
of uncertainty criterion by
is
lesser than that of the M-C criterion. Furthermore, the EMWCP with probability
MG-C criterion is lesser than that of the M-C criterion. Furthermore, the EMWCP with probability (P95) denotes the
minimum possible
(P95) denotes EMW withpossible
the minimum probabilityEMW of 95%,
with which defines
probability of the
95%,likelihood of preventing
which defines wellbore
the likelihood of
collapse
preventingas awellbore
function of EMW, it
collapse ascan be obtained
a function from the
of EMW, it cumulative
can be obtained probability
from curves. The P95
the cumulative
EMWCP estimated 3 , while the P95 EMWCP estimated
probability curves. byThetheP95M-C
EMWCPcriterion is 0.9825
estimated byg/cm
the M-C criterion is 0.9825 g/cm3, while the by the
P95
MG-C criterion is 0.9063 g/cm 3 . The minimum possible EMW estimated by the reliability
EMWCP estimated by the MG-C criterion is 0.9063 g/cm 3 . The minimum possible EMWassessment
estimated
method is obviously
by the reliability higher than
assessment methodthat of
is the analytical
obviously solution.
higher In the
than that of following section,
the analytical we just
solution. Inuse
the
following
the section,towe
MG-C model just use
analyze thethe MG-C model to analyze the EMWCP.
EMWCP.
Figure 8. Plot
Figure of of
Plot strength criteria
strength in {σin
criteria 1-σ{σ
2} space (Modified
1 -σ2 } space from [40]).
(Modified from(a) M-C criterion;
[40]). (a) M-C (b) MG-C
criterion;
criterion.
(b) MG-C criterion.
4.2.2.
4.2.2. Influence
Influence of of Breakout
BreakoutWidth Width
As
As mentioned
mentioned in in the
the Introduction,
Introduction, the the required
requiredEMWCP
EMWCP predicted
predicted bybytraditional
traditional model
model isis too
too
conservative.
conservative. The traditional models assume that the critical failure appears at the highest point
The traditional models assume that the critical failure appears at the highest point of of
stress concentration, in other words, it’s performed to yield no shear failure along
stress concentration, in other words, it’s performed to yield no shear failure along the borehole wall. the borehole wall.
However,
However,the thereal drilling
real drilling engineering
engineering allows a tolerable
allows breakout
a tolerable or an or
breakout appropriate breakout
an appropriate width,
breakout
which do not cause an unbearable collapse problem. The breakout
width, which do not cause an unbearable collapse problem. The breakout width model was width model was involved to
overcome
involved this problem, but
to overcome this itproblem,
is seldombut investigated
it is seldomby traditional
investigatedmodels. The influence
by traditional of mean
models. The
breakout width was simulated, and the results were shown in Figure 9.
influence of mean breakout width was simulated, and the results were shown in Figure 9. For For different breakout situation
of ω = 0◦ ,breakout
different ω = 10◦ , ω = 20◦ , ωof=ω30
situation
◦ , ω = 40◦ and ω = 50◦ , the P95 EMWCP estimated by the MG-C
= 0°, ω = 10°, ω = 20°, ω = 30°, ω = 40° and ω = 50°, the P95 EMWCP
criterion is by
1.1471 3
g/cm criterion
, 1.1341 g/cm 3 , 1.0629 g/cm3 , 0.9755 g/cm3 , 0.9062 g/cm3 and 0.8593 g/cm3
estimated the MG-C is 1.1471 g/cm3, 1.1341 g/cm3, 1.0629 g/cm3, 0.9755 g/cm3, 0.9062 g/cm3
respectively. The relationship between
and 0.8593 g/cm3 respectively. The relationship the EMWCP and breakout
between width were
the EMWCP fitted andwidth
and breakout expressed
were
as follows:
fitted and expressed as follows:
EMWCP
EMWCPP95 = P95
4.6858 × 10−
= 4.6858 ×610ω−36 ω
−33.8040
− 3.8040 −4−ω
×10
× 10 ω ++1.6000
42 2
10−−33ωω++1.1483
1.6000××10 1.1483 ((RR22==0.9982)
0.9982)
− 6 3
EMWCPP90 = 3.1797 × 10 ω−6 −32.6884 × 10 −ω4 + − 4 2 −4 ( R2 = 0.9990)
2 3.8431 × 10 −4 ω + 1.0313
EMWCP = 3.1797 − ×10 ω − 2.6884 ×10 ω + 3.8431 ×10 ω + 1.0313 ( R 22 = 0.9990) (38)
EMWCPP85 = P90
2.5424 × 10 6 ω 3 − 2.1160 × 10−4 ω 2 − 4.6500 × 10−4 ω + 0.9596 ( R = 0.9998) (38)
EMWCP
EMWCPP80 = P85
= 2.5424
2.5092 × 10−×610 −6 3
ω3 − ω 2.2044
− 2.1160
× 10 −4 ωω
×10 −4 2
2 +−4.3316 10−−44ω
4.6500××10 ω++0.9596
0.8966 ((R 2
R2 =
=0.9998
0.9997))
EMWCPP80 = 2.5092 ×10−6 ω 3 − 2.2044 ×10−4 ω 2 + 4.3316 ×10−4 ω + 0.8966 ( R2 = 0.9997)
where EMWCPP95 is the minimum possible EMWCP with probability of 95%, g/cm3 ; EMWCPP90
iswhere EMWCPP95possible
is the minimum
EMWCP possible EMWCPof with probability of 95%, g/cm 3; EMWCPP90 is
the minimum with probability 90%, g/cm3 ; EMWCP P85 is the minimum
the minimum
possible EMWCP possible EMWCP with
with probability probability
of 85%, of 90%, g/cmis3;the
g/cm3 ; EMWCP EMWCP P85 is the minimum possible
minimum possible EMWCP with
P80
EMWCP with probability
probability of 80%, g/cm .3 of 85%, g/cm 3 ; EMWCP P80 is the minimum possible EMWCP with
probability of 80%, g/cm3.

Energies 2019, 12, 942 16 of 31
100 1.2
95%
P95
ω =0°
1.1 P90
80 ω =10°
P85
ω =20°
EMWCP (g/cm )
Probability (%)
P80
3
ω =30° 1.0
60 ω =40°
ω =50° 0.9
40
0.8
20
0.7
0 0.6
0.0 0.5 1.0 1.5 2.0 0 10 20 30 40 50
3
Equivalent mud weight (g/cm ) Breakout width ω (°)
(a) (b)
Figure9.9.Influence
Figure Influence of
ofbreakout
breakouton
onEMWCP.
EMWCP.(a)
(a)The
Thecumulative
cumulativeprobability
probabilityof
ofEMWCP;
EMWCP;(b)
(b)EMWCP
EMWCP
varieswith
varies withbreakout
breakoutwidth.
width.
ItItisisclearly
clearly shown
shown that
that the
the EMWCP
EMWCPP95 P95,,EMWCP
EMWCPP90 , EMWCP
, EMWCP
P90 and
P85P85 and EMWCP
EMWCP estimated
estimated
P80 P80 by
by the
the MG-C
MG-C criterion
criterion always
always decline
decline with
with breakoutwidth
breakout widthininnonlinear.
nonlinear. In In other
other words, the the EMWCP
EMWCP
declines
declines withwith breakout width, but but not
notthe
thehigher
higherthethebreakout
breakoutwidth
widththe the better
better wellbore
wellbore stability.
stability. A
Amore
moreaccurate
accuratedefinition
definitionofofcollapse
collapseisisthat
thatwhen
when thethe breakout
breakout width or angle exceeds a critical limit,
angle exceeds a critical limit,
such
suchthatthatthetheremaining
remainingunaffected section
unaffected of the wellbore
section wall can no
of the wellbore wall longer
can uphold
no longerthe surrounding
uphold the
high stresses and
surrounding highflows inward
stresses andcausing bore enlargement,
flows inward causing bore subsequent
enlargement, collapse and complete
subsequent failure
collapse and
of the wellbore
complete [21].
failure Zoback
of the [6] indicated
wellbore that the
[21]. Zoback tolerable breakout
[6] indicated width (2ω)
that the tolerable to prevent
breakout widthborehole
(2ω) to
collapse is approximately ◦ for a vertical borehole. Once the breakout width (2ω) exceeds the
90approximately
prevent borehole collapse is 90° for a vertical borehole. Once the breakout width (2ω)
allowable
exceeds the breakout width,
allowable the borehole
breakout willborehole
width, the cannot maintain itself
will cannot stability.itself stability.
maintain
4.2.3.
4.2.3. Influence
Influenceof
ofMean
MeanValue
Value
According
According to to Equations
Equations (16),(16), (18)
(18) and
and (22),
(22), the
theEMWCP
EMWCP usually
usually depends
depends on on the
the vertical
vertical stress
stress
(σ ), the maximum horizontal
(σvv), the maximum horizontal stress (σH),stress (σ ), the minimum horizontal stress
H the minimum horizontal stress (σh), the (σ h ), the pore pressure
pore pressure (pp),
(p p ), the Biot’s coefficient (α), the strength parameters (c, ϕ), the Poisson’s ratio
the Biot’s coefficient (α), the strength parameters (c, φ), the Poisson’s ratio (v), and the tolerable (v), and the tolerable
breakout
breakout widthwidth (ω). The influence
(ω). The influence of of the
the mean
mean value
value ofof each
each basic
basic parameter
parameter is is aa key
key aspect
aspect to to
EMWCP. We simulated the influence of different mean value of each
EMWCP. We simulated the influence of different mean value of each basic parameter on thebasic parameter on the cumulative
probability
cumulativeofprobability
EMWCP, and the results
of EMWCP, andwere
the shown
results in Figure
were 10, where
shown the 10,
in Figure meanwhere valuetheofmean
each basic
value
parameter changed ±
of each basic parameter was changed ±10% from base value. It is clearly found that thepressure,
was 10% from base value. It is clearly found that the vertical stress, pore vertical
Biot’s
stress,coefficient,
pore pressure,and cohesive strength had
Biot’s coefficient, anda cohesive
very strong impacthad
strength on the cumulative
a very strong impactprobability of
on the
EMWCP, while the other parameters had a lesser impact.
cumulative probability of EMWCP, while the other parameters had a lesser impact.
As
Asshown
shownininFigure
Figure10j, it showed
10j, it showed thetheimpact
impact of mean
of mean value change
value on EMWCP,
change on EMWCP, all of all
thoseof basic
those
parameters were changed ± 10% from base value, but the corresponding
basic parameters were changed ±10% from base value, but the corresponding variation of EMWCP variation of EMWCP gave
different answers.
gave different The ranking
answers. of the influencing
The ranking degree of
of the influencing meanof
degree value
mean change
valueischange
the cohesive
is the strength
cohesive
>strength
pore pressure ≈ Biot’s coefficient > vertical stress > maximum horizontal stress
> pore pressure ≈ Biot’s coefficient > vertical stress > maximum horizontal stress > breakout > breakout width
>width
internal friction angle
> internal > minimum
friction horizontalhorizontal
angle > minimum stress > Poisson’s
stress > ratio. In general,
Poisson’s ratio. the EMWCPthe
In general, is
positively correlated with the vertical stress, maximum horizontal stress,
EMWCP is positively correlated with the vertical stress, maximum horizontal stress, pore pressure pore pressure and Biot’s
coefficient; while it is negatively
and Biot’s coefficient; correlated correlated
while it is negatively with the minimum horizontalhorizontal
with the minimum stress, cohesivestress,strength,
cohesive
internal friction angle, Poisson’s ratio and breakout width.
strength, internal friction angle, Poisson’s ratio and breakout width.
Energies 2019, 12, 942 17 of 31
100 100
80 80
Probability (%)
Probability (%)
60 σv=49.32MPa 60 σH=39.48MPa
σv=54.80MPa σH=43.87MPa
40 σv=60.28MPa 40 σH=48.26MPa
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
Equivalent mud weight (g/cm ) Equivalent mud weight (g/cm )
100 100
80 80
Probability (%)
Probability (%)
60 σh=27.82MPa 60 pp=19.25MPa
σh=30.91MPa pp=21.39MPa
40 σh=34.00MPa 40 pp=23.53MPa
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
100 100
80 80
Probability (%)
Probability (%)
60 α=0.86 60 c=16.33MPa
α=0.95 c=18.14MPa
α=1.00 c=19.95MPa
40 40
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
100 100
80 80
Probability (%)
Probability (%)
60 ϕ=31.50° 60 v=0.23
ϕ=35.00° v=0.25
ϕ=38.50° v=0.28
40 40
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
(g) Internal friction angle (h) Poisson’s ratio
100 0.66
σv σH σh
0.64 pp α ω
80 0.62 c ϕ v
EMWCP (g/cm )
3
Probability (%)
0.60
60 ω =36.00° 0.58
ω =40.00°
ω =44.00° 0.56
40 0.54
0.52
20 0.50
0.48
0 0.46
0.00 0.25 0.50 0.75 1.00 1.25 1.50 -10 -5 0 5 10
3 Change from base value (%)
(i) Breakout width (j) Impact of mean value change on EMWCP
Figure10.
Figure 10.Influence
Influence of
of mean
mean value
value on
on cumulative
cumulativeprobability
probabilityofofEMWCP.
EMWCP.
4.2.4. Influence of Variance
Due to the uncertainty of basic parameters, not only the mean value of input parameter affect
the EMWCP, but also the uncertainties also affect the EMWCP. In general, the variance represented
the degree of uncertainty for a single input parameter. The influences of variance on the cumulative
probability
Energies 2019, 12,of942 EMWCP were simulated, and the results were shown in Figure 11, where 18 ofthe
31
variance of input parameter was changed from 0.0 to 0.4. It is clearly found that the vertical stress,
maximum horizontal stress, pore pressure, Biot’s coefficient, cohesive strength had a very strong
impact on the cumulative probability of EMWCP, while the other parameters had a lesser impact.
When Duethe tovariance changes from
the uncertainty 0.0 toparameters,
of basic 0.4, all of the notcumulative
only the mean probability
valueofofEMWCP changed,
input parameter
but thethe
affect corresponding
EMWCP, butvariation also theofuncertainties
cumulative probability
also affect gave different In
the EMWCP. answers.
general, Forthe
thevariance
vertical
stress, maximum horizontal stress, minimum horizontal stress, internal
represented the degree of uncertainty for a single input parameter. The influences of variance on the friction angle and Poisson’s
ratio, with the
cumulative increaseofofEMWCP
probability variance, the simulated,
were cumulativeand probability
the resultscurves gradually
were shown moved
in Figure 11,onto
where right.
the
But for the
variance ofpore
inputpressure,
parameter Biot’s
wascoefficient,
changed from cohesive
0.0 tostrength
0.4. It isand breakout
clearly foundwidth,
that thewith the increase
vertical stress,
of variance,horizontal
maximum the 50–100% of the
stress, porecumulative probability
pressure, Biot’s curvescohesive
coefficient, gradually movedhad
strength ontoa right, while
very strong
the 0–50%
impact on ofthethe cumulative
probability ofcurves
EMWCP, gradually
while themovedother onto left. In fact,
parameters hadwe usually
a lesser focus
impact.
on thethe
When 50–100%
variance ofchanges
the cumulative
from 0.0 to probability
0.4, all of curve, thus, with
the cumulative the increase
probability of variance,
of EMWCP changed, the
cumulative probabilityvariation
but the corresponding curves gradually moved
of cumulative onto right.
probability In different
gave other words, the risk
answers. For of
thewellbore
vertical
collapse increaseshorizontal
stress, maximum with increase of variance.
stress, minimumOn the whole,
horizontal the ranking
stress, internal of the influencing
friction angle and degree
Poisson’s of
variance is the cohesive strength > Biot’s coefficient > pore pressure
ratio, with the increase of variance, the cumulative probability curves gradually moved onto right.> vertical stress > maximum
horizontal
But for the stress > breakout
pore pressure, width
Biot’s > internalcohesive
coefficient, friction strength
angle > minimum
and breakout horizontal
width, stress
with the> Poisson’s
increase
ratio. Moreover,
of variance, we usually
the 50–100% focus
of the on the probability
cumulative high cumulative curvesprobability
gradually moved of the onto
curves, andwhile
right, we just
the
extracted
0–50% of the the cumulative
P95 EMWCP to revealcurves
probability their influences,
gradually movedthe results
onto were
left. Inshown
fact, we in usually
Figure focus
11j. Iton is
clearly foundof
the 50–100% that
thethe P95 EMWCP
cumulative increasecurve,
probability with the
thus,variance,
with theand the ranking
increase of thethe
of variance, ratecumulative
of change
is the cohesive
probability curves strength
gradually> vertical
movedstress
onto >right.
Biot’sIncoefficient
other words, > pore pressure
the risk > maximum
of wellbore collapsehorizontal
increases
stressincrease
with > internal offriction
variance. angleOn> the
breakout
whole,width > minimum
the ranking of thehorizontal
influencing stress > Poisson’s
degree ratio. is the
of variance
Moreover,
cohesive strength a sensitivity analysis >was
> Biot’s coefficient porecarried
pressureout>for the EMWCP,
vertical the resultshorizontal
stress > maximum were shown in
stress
Figure 12. The results showed clearly that the influence ranking of
> breakout width > internal friction angle > minimum horizontal stress > Poisson’s ratio. Moreover, the basic parameters is the
cohesive
we usually strength
focus on > Biot’s coefficient
the high > poreprobability
cumulative pressure > of vertical stress >and
the curves, maximum
we just horizontal
extracted the stress
P95 >
breakout
EMWCP towidth reveal>their internal friction
influences, theangle
results>wereminimum
shown in horizontal
Figure 11j.stress > Poisson’s
It is clearly ratio.
found that theItP95is
consistent
EMWCP increase with the with results of Figure
the variance, and11.theFrom
rankingtheofrisk
the analysis point isofthe
rate of change view, we should
cohesive strengthbe>
concerned
vertical with
stress the top
> Biot’s five factors
coefficient > porewhich impact
pressure the EMWCP
> maximum the most;
horizontal stressand these factors
> internal frictionare the
angle
cohesive strength, Biot’s coefficient, pore pressure,
> breakout width > minimum horizontal stress > Poisson’s ratio. vertical stress and maximum horizontal stress.
100 100
80 80
Probability (%)
Probability (%)
Var=0.00 Var=0.00
60 Var=0.10 60 Var=0.10
Var=0.20 Var=0.20
Var=0.30 Var=0.30
40 Var=0.40 40 Var=0.40
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
100 100
80 80
Probability (%)
Probability (%)
Var=0.00 Var=0.00
60 Var=0.10 60 Var=0.10
Var=0.20 Var=0.20
Var=0.30 Var=0.30
40 Var=0.40 40 Var=0.40
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
Figure 11. Cont.

Energies 2019, 12, 942 19 of 31
Figure 11. Cont.
100 100
Probability (%) 80 80
Probability (%)
Var=0.00 Var=0.00
60 Var=0.10 60 Var=0.10
Var=0.20 Var=0.20
Var=0.30 Var=0.30
40 Var=0.40 40 Var=0.40
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
100 100
80 80
Probability (%)
Probability (%)
Var=0.00 Var=0.00
60 Var=0.10 60 Var=0.10
Var=0.20 Var=0.20
Var=0.30 Var=0.30
40 Var=0.40 40 Var=0.40
20 20
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.25 0.50 0.75 1.00 1.25 1.50
3 3
(g) Internal friction angle (h) Poisson’s ratio
100 1.2
σv σH σh
pp α ω
80 1.1
EMWCPP95 (g/cm )
C φ v
3
Probability (%)
Var=0.00
60 Var=0.10 1.0
Var=0.20
Var=0.30
40 Var=0.40 0.9
20 0.8
0 0.7
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.0 0.1 0.2 0.3 0.4
3
Equivalent mud weight (g/cm ) Var
(i) Breakout width (j) Impact of variance on P95 EMWCP
Figure
Figure 11.
11. Influence
Influence of
of variance
variance on
on probability
probability of
of EMWCP.
EMWCP.
Moreover, a sensitivity analysis was carried out for the EMWCP, the results were shown in
Cohesive strength
Figure 12. The results showed clearly that the influence ranking of the basic parameters is the cohesive
Biot coefficient
strength > Biot’s coefficientPore pressurepressure > vertical stress > maximum horizontal stress > breakout
> pore
width > internal friction Vertical
angle stress
> minimum horizontal stress > Poisson’s ratio. It is consistent with
the results of FigureMax11. Fromstress
horizontal the risk analysis point of view, we should be concerned with the
top five factors which impact
Breakout widthEMWCP the most; and these factors are the cohesive strength,
the
Biot’s coefficient, pore pressure, vertical stress and maximum horizontal stress.
Internal friction angle
Min horizontal stress
Poisson's ratio
0.44 0.48 0.52 0.56 0.60 0.64 0.68

3
Mean EMWCP (g/cm )
Figure 12. Sensitivity tornado for EMWCP.

20 0.8
EM
0 0.7
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.0 0.1 0.2 0.3 0.4
3
(i) Breakout width (j) Impact of variance on P95 EMWCP
Energies 2019, 12, 942 20 of 31
Figure 11. Influence of variance on probability of EMWCP.
Cohesive strength
Biot coefficient
Pore pressure
Vertical stress
Max horizontal stress
Breakout width
Internal friction angle
Poisson's ratio
0.44 0.48 0.52 0.56 0.60 0.64 0.68

3
Energies 2019, 12, x FOR PEER REVIEW Mean EMWCP (g/cm ) 19 of 29
Figure12.
Figure Sensitivitytornado
12.Sensitivity tornadofor
forEMWCP.
EMWCP.
4.3. Uncertainty Analysis of EMWFP
4.3. Uncertainty Analysis of EMWFP
4.3.1. EMWFP Probability Analysis
4.3.1. EMWFP Probability Analysis
EMWFP means the equivalent mud weight of fracture pressure, the simulating results is
EMWFP
shown in Figure means
13.the
The equivalent
mean EMWFPmud weight of fracture
estimated by thepressure,
reliabilitythe simulatingmethod
assessment results is
is shown
1.6021
in 3
g/cm3, and the analytical solution (1.6020 g/cm3) is very close to the mean EMWFP estimatedg/cm
Figure 13. The mean EMWFP estimated by the reliability assessment method is 1.6021 by the,
and the analytical solution (1.6020 3
reliability assessment method. The g/cm ) is very
probability close to
decreases theincrease
with mean EMWFPof EMW, estimated
that means by the
the
reliability assessment method. The probability decreases with increase
higher the EMW, the higher the risk of wellbore fracture is. We should determine a proper of EMW, that means the
higher the EMW,
probability the the
to obtain higher the risk of wellbore
corresponding EMWFP, fracture
such as is. the We
EMWFPshould determine
with a proper
probability (P95)
probability to obtain the corresponding EMWFP, such as the EMWFP with
denotes the maximum possible EMW with probability of 95%, which defines the likelihood of probability (P95) denotes
the maximum
preventing possiblefracture
wellbore EMW with
as aprobability
function ofof EMW,
95%, which
it candefines the likelihood
be obtained from the of preventing
cumulative
wellbore
probabilityfracture as As
curves. a function
shown ofin EMW,
Figureit13,
canthe
be obtained
EMWFP from the cumulative
estimated probability
by the reliability curves.
assessment
As 3
method is 0.6120 g/cm , 0.8386 g/cm , 0.9800 g/cm and 1.0834 g/cm for the probability of 95%, 90%,,
shown in Figure 13,3 the EMWFP estimated
3 by the
3 reliability assessment
3 method is 0.6120 g/cm
0.8386 g/cm 3 0.9800 g/cm3 and 1.0834 g/cm3 for the probability of 95%, 90%, 85% and 80%
85% and 80%, respectively. The maximum possible EMW estimated by the reliability assessment
respectively. The maximum possible
method is obviously lower than that EMW estimated
of the by the
analytical reliability
solution assessment
(1.6020 g/cm3). method
In otheriswords,
obviouslythe
lower than that of the analytical solution (1.6020 g/cm 3 ). In other words, the uncertainty of basic
uncertainty of basic parameters results in a lower EMWFP, and the risk of wellbore fracture become
parameters
much higher. results in a lower EMWFP, and the risk of wellbore fracture become much higher.
0.7 100
0.6
80
Probability density
0.5
0.4 60
0.3 40
0.2
20
0.1
0.0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
3
Figure 13. Probability density and cumulative probability
probability for
for EMWFP.
EMWFP.
4.3.2.
4.3.2. Influence
Influence of
of Mean
Mean Value
Value
According
According totoEquations
Equations(30)–(33), the EMWFP
(30)–(33), the EMWFP is usually depended
is usually on the maximum
depended on the horizontal
maximum
stress
horizontal stress (σH), the minimum horizontal stress (σh), the pore pressure (pp), theand
(σ H ), the minimum horizontal stress (σ h ), the pore pressure (pp ), the Biot’s coefficient (α), the
Biot’s
tensile strength (S ). The influence of the mean value of each basic parameter is a
coefficient (α), andt the tensile strength (St). The influence of the mean value of each basic parameter key aspect to
EMWFP.
is a key We simulated
aspect the influence
to EMWFP. of differentthe
We simulated mean value ofofeach
influence basic parameter
different mean value on the
of cumulative
each basic
probability
parameter on the cumulative probability of EMWFP, and the results were shown in each
of EMWFP, and the results were shown in Figure 14, where the mean value of basic
Figure 14,
where the mean value of each basic parameter was changed ±10% from base value. It is clearly
found that the maximum horizontal stress, minimum horizontal stress, pore pressure and Biot’s
coefficient had a very strong impact on the cumulative probability of EMWFP, while the other
parameters had a lesser impact.
As shown in Figure 14f, it showed the impact of mean value change on EMWFP, all of those
Energies 2019, 12, 942 21 of 31
parameter was changed ±10% from base value. It is clearly found that the maximum horizontal
stress, minimum horizontal stress, pore pressure and Biot’s coefficient had a very strong impact on the
Energies 2019, 12,
cumulative x FOR PEERof
probability REVIEW
EMWFP, while the other parameters had a lesser impact. 20 of 29
100 100
80 80
Probability (%)
Probability (%)
60 σH=39.48MPa 60 σh=27.82MPa
σH=43.87MPa σh=30.91MPa
40 σH=48.26MPa 40 σh=34.00MPa
20 20
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
3 3
(a) Maximum horizontal stress (b) Minimum horizontal stress
100 100
80 80
Probability (%)
Probability (%)
60 pp=19.25MPa 60 α=0.86
pp=21.39MPa α=0.95
α=1.00
40 pp=23.53MPa 40
20 20
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
3 3
(c) Pore pressure (d) Biot’s coefficient
100 2.2
σH σh pp
α0 St
80 2.0
Probability (%)
EMWFP (g/cm )
1.8
3
60 St=5.40MPa
St=6.00MPa 1.6
40 St=6.60MPa
1.4
20 1.2
0 1.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 -10 -5 0 5 10
3
Equivalent mud weight (g/cm ) Change from base value (%)
(e) Tensile strength (f) Impact of mean value change on EMWFP
Figure 14.
Figure Influence of
14. Influence of mean
mean value
value on
on cumulative
probability of
of EMWFP.
EMWFP.
As shown in Figure 14f, it showed the impact of mean value change on EMWFP, all of those
basic parameters were changed ±10% from base value, but the corresponding variation of EMWFP
gaveSimilarly, not onlyThe
different answers. the ranking
mean value
of theofinfluencing
each basicdegree
parameter affects
of mean thechange
value EMWFP, butminimum
is the also the
uncertainty can also affect the EMWFP. In general, the variance represented
horizontal stress > the maximum horizontal stress > the Biot’s coefficient ≈ the pore pressure > the the degree of
uncertainty for basic
tensile strength. parameter.
The EMWFP We simulated
is positively correlatedthewith
influence of variance
the minimum on the
horizontal cumulative
stress and the
probability of EMWFP, and the results were shown in Figure 15, where the variance
tensile strength, while it is negatively correlated with the maximum horizontal stress, the Biot’s of each basic
parameter was changed from
coefficient and the pore pressure. 0.0 to 0.4. It is clearly found that the maximum horizontal stress and
the minimum horizontal stress had a very strong impact on the cumulative probability of EMWFP,
4.3.3. Influence
while the otherof Variance had a lesser impact. When the variance changed from 0.0 to 0.4, all of
parameters
the cumulative
Similarly, not only theofmean
probability EMWFPvaluechanged.
of each With
basic the increaseaffects
parameter of variance, the 50–100%
the EMWFP, of the
but also the
uncertainty can also affectcurves gradually
the EMWFP. movedthe
In general, onto left, while
variance the 0–50%
represented of the
the degree cumulative
of uncertainty
probability curves gradually moved onto right. In fact, we usually focus on the
for basic parameter. We simulated the influence of variance on the cumulative probability of EMWFP, 50–100% of the
cumulative probability curve, thus, with the increase of variance, the cumulative probability curves
gradually moved onto left. In other words, the risk of wellbore fracture increases with increase of
variance. On the whole, the ranking of the influencing degree of variance is the minimum
horizontal stress > the maximum horizontal stress > the Biot’s coefficient > the pore pressure > the
tensile strength. Moreover, we usually focus on the high cumulative probability of the curves, and
we just extracted the P95 EMWFP to reveal their influences, the results were shown in Figure 15f. It
Energies 2019, 12, 942 22 of 31
and the results were shown in Figure 15, where the variance of each basic parameter was changed
from 0.0 to 0.4. It is clearly found that the maximum horizontal stress and the minimum horizontal
stress had a very strong impact on the cumulative probability of EMWFP, while the other parameters
had a lesser impact. When the variance changed from 0.0 to 0.4, all of the cumulative probability of
EMWFP changed.
Energies 2019, 12, x FOR With the increase of variance, the 50–100% of the cumulative probability curves
PEER REVIEW 21 of 29
gradually moved onto left, while the 0–50% of the cumulative probability curves gradually moved
is clearly
onto right.found
In fact,that
wethe P95 focus
usually EMWFP decrease
on the 50–100%with ofthe
the variance,
cumulative and the ranking
probability of the
curve, rate
thus, withof
change is also minimum horizontal stress > the maximum horizontal stress >
the increase of variance, the cumulative probability curves gradually moved onto left. In other words,the Biot’s coefficient >
the pore
risk ofpressure
wellbore > the tensileincreases
fracture strength.with increase of variance. On the whole, the ranking of the
Moreover,
influencing degreea sensitivity
of varianceanalysis was carried
is the minimum out for the
horizontal EMWFP,
stress the resultshorizontal
> the maximum were shown stressin
Figure 16. The results showed clearly that the ranking of the basic parameters
> the Biot’s coefficient > the pore pressure > the tensile strength. Moreover, we usually focus on is the minimum
horizontal
the stress > the
high cumulative maximumofhorizontal
probability the curves,stress
and we > the
justBiot’s coefficient
extracted the P95 ≈ the pore pressure
EMWFP to reveal >their
the
tensile strength.
influences, It is consistent
the results were shown with the results
in Figure 15f. Itofis Figure
clearly 15. From
found thatthe
therisk
P95analysis
EMWFPpoint of view,
decrease with
we variance,
the should beand concerned withofthe
the ranking the top
ratetwo factorsiswhich
of change impact the
also minimum EMWFPstress
horizontal the most; and these
> the maximum
factors are stress
horizontal the minimum horizontal
> the Biot’s stress
coefficient andpore
> the the pressure
maximum horizontal
> the stress.
tensile strength.
100 100
80 80
Probability (%)
Probability (%)
Var=0.00 Var=0.00
60 Var=0.10 60 Var=0.10
Var=0.20 Var=0.20
Var=0.30 Var=0.30
40 Var=0.40 40 Var=0.40
20 20
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
3 3
(a) Maximum horizontal stress (b) Minimum horizontal stress
100 100
80 80
Probability (%)
Probability (%)
Var=0.00 Var=0.00
60 Var=0.10 60 Var=0.10
Var=0.20 Var=0.20
Var=0.30 Var=0.30
40 Var=0.40 40 Var=0.40
20 20
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
3 3
(c) Pore pressure (d) Biot’s coefficient
100 1.0
80 0.5
EMWFPP95 (g/cm )
Probability (%)
Var=0.00
60 Var=0.10 0.0
Var=0.20
σH
Var=0.30
40 Var=0.40 -0.5 σh
pp
α0
20 -1.0
St
0 -1.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.1 0.2 0.3 0.4
3
(e) Tensile strength (f) Impact of variance on P95 EMWFP
Figure 15. Influence

Influence of variance on cumulative probability of EMWFP.
EMWFP.
Energies 2019, 12, 942 23 of 31
Moreover, a sensitivity analysis was carried out for the EMWFP, the results were shown in
Figure 16. The results showed clearly that the ranking of the basic parameters is the minimum
horizontal stress > the maximum horizontal stress > the Biot’s coefficient ≈ the pore pressure > the
tensile strength. It is consistent with the results of Figure 15. From the risk analysis point of view,
we should be concerned with the top two factors which impact the EMWFP the most; and these factors
Energies
are the2019, 12, x FOR
minimum PEER REVIEW
horizontal stress and the maximum horizontal stress. 22 of 29
Max horizontal stress
Pore pressure
Biot coefficient
Tensile strength
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1
33
Mean EMWFP (g/cm )
Figure
Figure 16.
16. Sensitivity
Sensitivitytornado
tornado for
for EMWFP.
EMWFP.
4.4. Uncertainty
4.4. Uncertainty Analysis
Analysis of
of EMWPP
EMWPP
EMWPP means
EMWPP meansthe theequivalent
equivalentmud mudweight
weightofofpore porepressure,
pressure, thethe simulating
simulating results
results is shown
is shown in
in Figure 3,
Figure 17. 17.
TheThemean meanEMWPPEMWPP estimated
estimated by the
by the reliability
reliability assessment
assessment methodmethod is 0.9921
is 0.9921 g/cmg/cm
33, and
andanalytical
the analytical solution (0.9921 3
the solution (0.9921 g/cmg/cm ) is also
33) is also the same.
the same. The The probability
probability increases
increases with with increase
increase of
of EMW,
EMW, that
that means
means thethe higher
higher thethe EMW,
EMW, thethe lower
lower thethe risk
risk of of well
well kick
kick is.is.
WeWe should
should determine
determine a
a proper
proper probability
probability to obtain
to obtain the the corresponding
corresponding EMWPP,EMWPP, suchsuch
as the asEMWPP
the EMWPP with probability
with probability (P95)
(P95) denotes
denotes the minimum
the minimum possible
possible EMW EMWwithwith probability
probability of of
95%,95%, whichdefines
which definesthethelikelihood
likelihood of of
preventingwell
preventing wellkickkickasas a function
a function of EMW,
of EMW, it canit be
canobtained
be obtained
from the from the cumulative
probability curves.
curves.
As shownAs in shown
Figure 17,in Figure 17, theestimated
the EMWPP EMWPPby estimated by the
the reliability reliabilitymethod
assessment assessment method
is 1.1552 g/cmis 33,
1.1552 g/cm 3 , 1.1192 g/cm333 , 1.0949 g/cm3 and 1.0756 g/cm 3 for the probability of 95%, 90%, 85%
1.1192 g/cm , 1.0949 g/cm and 1.0756 g/cm for the probability of 95%, 90%, 85% and 80%
3
3 3
3
and 80% respectively.

respectively. The minimumThe minimum
possiblepossible
EMWEMW estimated
estimated by theby the reliability
reliability assessmentmethod
assessment method is is
obviously higher than that of the analytical solution (0.9921 g/cm 3 ).33 In other words, the uncertainty of
obviously higher than that of the analytical solution (0.9921 g/cm ). In other words, the uncertainty
pore
of pressure
pore pressure results in ainlower
results EMWPP,
a lower EMWPP, andand thetheriskrisk
of well kickkick
of well becomebecome muchmuchhigher.
higher.
4.5 100
4.0
3.5 80
Probability density
3.0
60
2.5
2.0
40
1.5
1.0 20
0.5
0.0 0
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
33
Figure
Figure 17.
17. Probability
Probabilitydensity
densityand
and cumulative
probability for
for EMWPP.
EMWPP.
According
According toto Equations
Equations(34)–(36),
(34)–(36),thethe
EMWPP
EMWPP only depends
only on the
depends porepore
on the pressure (ppp). The
pressure (pp ).
influence of theof mean
The influence valuevalue
the mean and variance
and varianceof pore pressure
of pore is theis key
pressure the aspect to EMWPP.
key aspect to EMWPP. We
simulated the influence of different mean value of pore pressure on the cumulative probability of
EMWFP, and the results were shown in Figure 18, where the mean value of pore pressure was
changed ±10% from base value. It is clearly found that the uncertain pore pressure had a very
strong impact on the cumulative probability of EMWPP. The EMWPP is positively correlated with
the mean value of pore pressure. Moreover, we also simulated the influence of variance on the
Energies 2019, 12, 942 24 of 31
We simulated the influence of different mean value of pore pressure on the cumulative probability of
EMWFP, and the results were shown in Figure 18, where the mean value of pore pressure was changed
±10% from base value. It is clearly found that the uncertain pore pressure had a very strong impact on
the cumulative probability of EMWPP. The EMWPP is positively correlated with the mean value of
pore pressure. Moreover, we also simulated the influence of variance on the cumulative probability of
EMWPP, and the results were shown in Figure 19, where the variance of pore pressure was changed
from 0.0
Energies to 0.4.
2019, 12, x When theREVIEW
FOR PEER variance changed from 0.0 to 0.4, the 50–100% of the cumulative probability
23 of 29
curves gradually moved onto right, while the 0–50% of the cumulative probability curves gradually
of the cumulative probability curve, thus, with the increase of variance, the cumulative probability
moved onto left. Inprobability
of the cumulative fact, we usually
curve,focus
thus, on thethe
with 50–100% of of
increase thevariance,
cumulative
the probability
cumulative curve, thus,
probability
curves gradually moved onto right. In other words, the risk of well kick linearly increased with
with the increase of variance, the cumulative probability curves gradually moved onto
curves gradually moved onto right. In other words, the risk of well kick linearly increased withright. In other
increase of variance, as shown in Figure 19b.
words,
increasetheof risk of well
variance, askick linearly
shown increased
in Figure 19b. with increase of variance, as shown in Figure 19b.
100
100
80
80
(%)
(%)
60
Probability
60
Probability
40
40
20 Pp=19.25MPa
20 Pp=19.25MPa
Pp=21.39MPa
Pp=21.39MPa
Pp=23.53MPa
0 Pp=23.53MPa
00.00 0.25 0.50 0.75 1.00 1.25 1.50
0.00 0.25 0.50 0.75 1.00 1.25
3 1.50
Equivalent mud weight (g/cm3 )
Figure 18. Influence of mean value of pore pressure on cumulative probability for EMWPP.
Figure
Figure 18.
18. Influence
Influence of
of mean
mean value
value of
of pore
pore pressure
pressure on
on cumulative
cumulative probability for EMWPP.
probability for EMWPP.
100 1.7
100 1.7
1.6 P95
80 1.6 P90
P95
1.5 P85
80
) 3)
P90
(%)
1.5 P80
(g/cm
1.4 P85
(%)
60
3
P80
(g/cm
1.4
Probability
60 Var=0.00 1.3
Probability
1.3
EMWCP
40 Var=0.10
Var=0.00 1.2
Var=0.20
EMWCP
40 Var=0.10
1.2
Var=0.30
Var=0.20 1.1
20 Var=0.40
Var=0.30 1.1
20 Var=0.40 1.0
0 1.0
0.9
00.00 0.25 0.50 0.75 1.00 1.25 1.50 0.9 0.0 0.1 0.2 0.3 0.4
0.00 0.25 0.50 0.75 1.00
(g/cm31.25 1.50
0.2 3 0.3 0.4 0.0 0.1
Equivalent mud weight ) Var
(a) (b)
(a) (b)
Figure 19.
Figure 19. Influence
Influence of
of pore
pore pressure
pressure variance
variance on
on EMWPP.
EMWPP.(a)
(a)Influence
Influenceof
ofvariance
varianceof
ofpore
porepressure;
pressure;
Figure 19. Influence of pore pressure variance on EMWPP. (a) Influence of variance of pore pressure;
(b) EMWCP
(b) EMWCP varies
varies with
with variance.
variance.
(b) EMWCP varies with variance.
4.5.
4.5. Uncertainty
Uncertainty Analysis
Analysis of
of SMWW
SMWW
4.5. Uncertainty Analysis of SMWW
On
On the
the basis
basis of
of the
the uncertainty
uncertainty analysis
analysis of
of EMWCP,
EMWCP, EMWFP EMWFP and and EMWPP,
EMWPP, the the cumulative
cumulative
On the
probability basis of the uncertainty analysis of EMWCP, EMWFP and EMWPP, the cumulative
probability of
of SMWW
SMWW can can be
be obtained,
obtained, and
and the
the results
results were
were shown
shown in in Figures
Figures 20
20 and
and 2121 and
and Table
Table 4.
4.
probability
Figure 20 of SMWW
shows the can be obtained,
interference of the and the results
probability were
density shownofinEMWCP,
function Figures 20 and 21and
EMWFP andEMWPP,
Table 4.
Figure 20 shows the interference of the probability density function of EMWCP, EMWFP and
Figure
the 20 showszone
overlapping the denotes
interference of the probability
the probability of wellboredensity function
instability. Theof EMWCP,
smaller EMWFPzone,
overlapping and
EMWPP, the overlapping zone denotes the probability of wellbore instability. The smaller
EMWPP,
the the overlapping
wellbore zone,
will be more zone denotes
reliable, the probability of wellbore instability. The smaller
overlapping the wellbore willi.e.,
be the
morerisk of wellbore
reliable, i.e., theinstability will be lower.
risk of wellbore The will
instability bigger
be
overlapping zone,
overlapping zone, the
the wellbore
wellbore will
will be
be more
more reliable, i.e.,i.e.,
unreliable, thethe
riskrisk
of wellbore
of wellboreinstability willwill
instability be
lower. The bigger overlapping zone, the wellbore will be more unreliable, i.e., the risk of wellbore
lower.
be The bigger overlapping zone, the wellbore will be more unreliable, i.e., the risk of wellbore
higher.
instability will be higher.
4.0 100
(%)
4.0 EMWPP 100

(%)
3.5
3.5 EMWPP
EMWCP 80
probability
density
3.0 EMWCP
EMWFP 80
ulative probability
density
3.0 EMWFP
2.5 60
2.5 60
2.0
Probability
2.0
Probability
1.5 40
umulative
1.5 40
1.0
1.0 20
0.5 20
probability of SMWW can be obtained, and the results were shown in Figures 20 and 21 and Table 4.
Figure 20 shows the interference of the probability density function of EMWCP, EMWFP and
EMWPP, the overlapping zone denotes the probability of wellbore instability. The smaller
overlapping zone, the wellbore will be more reliable, i.e., the risk of wellbore instability will be
lower. The bigger overlapping zone, the wellbore will be more unreliable, i.e., the risk of wellbore
Energies 2019, 12, 942 25 of 31
4.0 100

3.5 EMWPP
EMWCP 80
Probability density
3.0 EMWFP
2.5 60
2.0
1.5 40
1.0
20
0.5
0.0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
3
Energies 2019, 12, x FOR PEER REVIEW
Figure 20. Probability density histogram of SMWW.
SMWW. 24 of 29
100
95%
90% EMWPP
85%
80 EMWCP
EMWFP
Probability (%)
60
40
20
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
3
Figure
Figure 21.
21. Cumulative
Cumulative probability
probability of
of SMWW.
SMWW.
Table 4. The statistics for EMWPP, EMWCP, EMWFP and SMWW.

Table 4. The statistics for EMWPP, EMWCP, EMWFP and SMWW.
Statistics
Statistics EMWPP
EMWPP EMWCP EMWFP
EMWCP EMWFP SMWW SMWW
P80/(g/cm
P80/(g/cm 3 ) )
3 1.0756
1.0756 0.8083
0.8083 1.0935
1.0935 1.0756–1.0935
1.0756–1.0935
P85/(g/cm
P85/(g/cm 3 ) 3) 1.0949
1.0949 0.8487
0.8487 0.9791 No SMWW
0.9791 No SMWW
P90/(g/cm 3) 3 1.1193 0.9053 0.8240 No SMWW
P90/(g/cm ) 1.1193 0.9053 0.8240 No SMWW
P95/(g/cm 3) 3 1.1552 0.9063 0.6144 No SMWW
P95/(g/cm ) 1.1552 0.9063 0.6144 No SMWW
P100/(g/cm 3) 3 1.3734 1.5160 − 0.8036 No SMWW
P100/(g/cm ) 1.3734 1.5160 −0.8036 No SMWW
Mean/(g/cm 3) 0.9921 0.5538 1.6021 0.9921–1.6021
Mean/(g/cm 3) 0.9921 0.5538 1.6021 0.9921–1.6021
3) 0.0992 0.2060 0.6013
Std Std
Dev/(g/cm
Dev/(g/cm ) 3 3 0.0992 0.2060 0.6013 / /
Analytical solution/(g/cm ) 3 0.9921 0.5195 1.6020 0.9921–1.6020
Analytical solution/(g/cm ) 0.9921 0.5195 1.6020 0.9921–1.6020
Figure
Figure 21 21showsshowsthe the cumulative
probability of EMWCP,
of EMWCP, EMWFP EMWFPand EMWPP,and EMWPP, and the
and the statistics
statistics all for EMWCP,
all for EMWCP, EMWFP and EMWFP EMWPP and were
EMWPP were in
collected collected
Table 4.inItTable 4. It found
is clearly is clearly
thatfound
whenthat the
when the uncertainty
uncertainty of basic parameters
of basic parameters is considered, is considered,
the mean EMWCP the mean EMWCP
is 0.5538 g/cm is30.5538
, and the P953,EMWCP
g/cm and the
P95 EMWCP
is 0.9063 g/cmis 3 , while
0.9063the g/cm 3, whilesolution
analytical the analytical g/cm3 ) is(0.5195
(0.5195 solution obviously g/cm too3) low
is obviously
to maintain toowellbore
low to
maintain wellbore
stability, that stability,risk
is, the collapse thatofis,
thethe collapse
wellbore risk of the
is actually much wellbore
higher thanis actually much higher
that of traditional than
analysis.
that of traditional
Similarly, when theanalysis.
uncertainty Similarly,
of basicwhen the uncertainty
parameters of basic
is considered, parameters
the mean EMWPP is is
considered,
0.9921 g/cm the3 ,
mean
and the EMWPP
P95 EMWPP is 0.9921
is 1.1552 , and3the
g/cm3g/cm P95the
, while EMWPP
analyticalis 1.1552 g/cm
solution 3, while
(0.9921 g/cm the3 )analytical
is obviously solution
lower,
(0.9921
that is, g/cm 3) isrisk
the kick obviously lower,
of the well that is, the
is actually muchkickhigher
risk ofthan
the well
that isof actually
traditional much higherMoreover,
analysis. than that
of traditional
when analysis.of
the uncertainty Moreover, when theisuncertainty
basic parameters considered,of thebasic
mean parameters
EMWFP is is 1.6021 g/cm3the
considered, mean
, and the
EMWFP
P95 EMWFP is 1.6021
is 0.6144 3 3
, and the
g/cm g/cm P95 EMWFP
, while is 0.6144
the analytical g/cm , (1.6020
solution 3 while the g/cm 3
analytical solution (1.6020
) is obviously g/cmis,
lower, that 3)
is
theobviously
fracture risklower, that
of the is, the is
wellbore fracture
actuallyrisk
much of the
higherwellbore is actually
than that muchanalysis.
of traditional higher than that of
traditional analysis.
According to the Equation (37), the lowest safe mud weight of SMWW should be depended on
According
the high of EMWCP to the
andEquation
EMWPP, (37),
andthethelowest
highest safe
safemud
mudweight
weightofofSMWW
SMWWshould shouldbe bedepended
depended on on
the high of EMWCP and EMWPP, and the highest safe mud weight of SMWW should be depended
on the EMWFP, that is, the lowest safe mud weight should be the EMWPP, while the highest safe
mud weight should be the EWMCP. As shown in Figure 21 and Table 4, when the cumulative
probability is 95%, the P95 EMWFP (0.6144 g/cm3) is obviously lower than both of EMWPP (1.1552 g/cm3)
and EMWCP (0.9063 g/cm3), there is no SMWW. When the cumulative probability is 90%, the P90
Energies 2019, 12, 942 26 of 31
the EMWFP, that is, the lowest safe mud weight should be the EMWPP, while the highest safe mud
weight should be the EWMCP. As shown in Figure 21 and Table 4, when the cumulative probability
is 95%, the P95 EMWFP (0.6144 g/cm3 ) is obviously lower than both of EMWPP (1.1552 g/cm3 ) and
EMWCP (0.9063 g/cm3 ), there is no SMWW. When the cumulative probability is 90%, the P90 EMWFP
(0.8240 g/cm3 ) is obviously lower than both of EMWPP (1.1193 g/cm3 ) and EMWCP (0.9053 g/cm3 ),
there is still no SMWW. When the cumulative probability is 85%, the P85 EMWFP (0.9791 g/cm3 )
is obviously lower than EMWPP (1.0949 g/cm3 ) but higher than EMWCP (0.8487 g/cm3 ), that is,
there is still no SMWW. When the cumulative probability is 80%, the P80 EMWFP (1.0935 g/cm3 ) is
obviously higher than both of EMWPP (1.0756 g/cm3 ) and EMWCP (0.8083 g/cm3 ), there is a SMWW
of 1.0756–1.0935 g/cm3 , but it is very narrow SMWW. The SMWW predicted by analytical solution is
0.9921–1.6020 g/cm3 , compared to the SMWW estimated by the reliability assessment method, the
SMWW predicted only by mean value is very wide to maintain wellbore stability, but the reliability
assessment method tends to give a narrow SMWW of 1.0756–1.0935 g/cm3 and its probability is only
80%, that is, it is very difficult to drill without any wellbore instability problems. In this situation,
the traditional over-balanced drilling (OBD) maybe cannot drill successfully, but there is an available
mud weight window (0.8083–1.0935 g/cm3 ) to avoid wellbore collapse and fracture, we maybe can
utilize underbalanced drilling (UBD), managed pressure drilling (MPD), or micro-flow drilling (MFD)
to conquer well kick, and we can drill successfully. Another choice should be wider the SMWW, the
only way is to prevent lost circulation.
4.6. Field Observation Report

In the design of well SC-101X, the traditional OBD was designed to drill in the 2008–2432 m
section, the density of water-based drilling mud was designed as 1.07 g/cm3 by considering the impact
of the surge and swab pressures. In real drilling operation, the actual drilling mud density was also
1.07 g/cm3 . When this well was drilled at the depth of 2022 m, the driller operates drilling rig to
pull the drill string up and down to prepare making a connection, the mud man found drilling mud
loss, the average loss rate of drilling mud was approximately 3.67 m3 /h, and the total loss volume
of drilling mud was approximately 10.5 m3 . In this process, the drilling mud was added some lost
circulation materials (LCMs) to against drilling mud loss, the LCMs formula was given as follows:
4% GZD-B + 5% GZD-C + 6% GZD-D. After multiple plugging, the pressure-bearing capacity was
tested, and the lost pressure was enhanced to 1.19 g/cm3 . The drilling mud density of 1.07 g/cm3 was
still used to drill. But when this well was drilled at the depth of 2067 m, the driller operates drilling rig
to pull the drill string up and down to prepare making a connection, the mud man found a gas kick.
The well was shut-in to prepare well kill. After testing the shut-in standpipe pressure, the killing mud
density was determined as 1.13 g/cm3 . During well killing, the drilling mud loss occurred again with
a low average loss rate of 3.5 m3 /h. After the well was killed successfully, and considering the higher
pore pressure of the lower formation, the use of MPD was considered. Before drilling using MPD,
the drilling mud was again added some lost circulation materials (LCMs), and the LCMs formula was
given as follows: 5% GZD-O + 5% GZD-A + 10% GZD-B + 10% GZD-C + 14% GZD-D + 4% FDJ-1 +
3% Nutshell + 3% JD-5. After multiple plugging, the pressure-bearing capacity was tested, and the lost
pressure was enhanced to 1.37 g/cm3 . In the process of MPD, the operating parameters were listed in
Table 5, and finally, the section of 2067–2432 m of well SC-101X was drilled successfully. To sum up,
the field observation for well kick and wellbore fracture is consistent with analysis results of SMWW
that estimated by uncertain model. Thus, the present model was therefore verified, and it is much
more accurate than the traditional analytical solution.
Energies 2019, 12, 942 27 of 31
Table 5. Operating parameters of MPD.
No. Parameters Value

1 Well depth (m) 2067–2432
2 EMWPP (g/cm3 ) 1.08
3 EMWCP (g/cm3 ) 1.09 (Enhanced to 1.37)
4 Mud displacement (L/s) 20–25
5 Mud density (g/cm3 ) 1.05–1.16
6 Pressure control during drilling (MPa) 0–1
7 Pressure control during pump-stopping (MPa) 2–3
5. Conclusions
On the basis of the reliability assessment theory, the uncertain EMWCP model was proposed
by involving the tolerable breakout and the MG-C criterion, the uncertain EMWFP and EMWPP
model was also proposed, and the uncertain SMWW model was determined by EMWCP, EMWPP and
EMWFP. Finally, taking the basic parameters of well SC-101X as an example, the influence of uncertain
basic parameter on SMWW was systematically investigated and discussed. The main conclusions can
be drawn as follows:
(1) For wellbore stability analysis, the uncertain distribution of each basic parameter satisfies the
normal distribution, the higher the coefficient of variance is, the higher the level of uncertainty
will be, and the larger the impact on wellbore stability will be.
(2) The EMWCP predicted by the M-C criterion is obviously higher than that of the MG-C criterion,
because of the MG-C criterion rightly predicts rock strength and estimates the EMWCP much
more exactly. Thus, the MG-C criterion is recommended for wellbore stability analysis.
(3) The breakout width has a very significant impact on EMWCP, and the EMWCP always decline
with breakout width in nonlinear. The tolerable breakout is recommended for wellbore stability
analysis, and the tolerable breakout width (2ω) of 45◦ is recommended for a vertical well.
(4) The EMWCP estimated by the uncertain model is obviously higher than that of analytical
model, that is, the collapse risk of the wellbore is heightened by the uncertain basic parameters.
The cumulative probability of wellbore collapse gradually increases with increasing of variance,
and the influencing degree of uncertain basic parameter is the cohesive strength > Biot’s coefficient
> pore pressure > vertical stress > maximum horizontal stress > breakout width > internal friction
angle > minimum horizontal stress > Poisson’s ratio.
(5) The EMWPP estimated by the uncertain model is obviously higher than that of mean pore
pressure, that is, the kick risk is also heightened by the uncertain pore pressure. The cumulative
probability of well kick gradually also increases with increasing of variance for pore pressure.
(6) The EMWFP estimated by the uncertain model is obviously lower than that of analytical
model, that is, the fracture risk of the wellbore heightened by the uncertain basic parameters.
The cumulative probability of wellbore fracture gradually increases with increasing of variance,
and the influencing degree of uncertain basic parameter is the minimum horizontal stress >
maximum horizontal stress > Biot’s coefficient > pore pressure > tensile strength.
(7) For the SC-101X well, the SMWW predicted by analytical solution is 0.9921–1.6020 g/cm3 and
very wide, compared to the SMWW estimated by the uncertain model, the P80 SMWW estimated
by the uncertain model is only 1.0756–1.0935 g/cm3 and very narrow, that is, there is a low
possibility to drill without any wellbore instability problems. The field observation for well kick
and wellbore fracture verified the analysis results of SMWW that estimated by uncertain model,
and both lost circulation prevention and MPD were utilized to maintain drilling safety.
(8) To drill the formation with narrow SMWW, some kinds of advanced drilling technology, such as
the UBD, MPD and MFD, can be utilized to control the wellbore pressure accurately, so that the
Energies 2019, 12, 942 28 of 31
wellbore stability can be maintained. Meanwhile, another choice should be wider the SMWW,
the only way is to prevent lost circulation.
Author Contributions: Conceptualization, T.M. and P.C. Investigation, T.M. and T.T. Methodology, T.M.
Writing—original draft, T.M. and T.T. Writing—review and editing, P.C. and C.Y.
Funding: This research was funded by the National Natural Science Foundation of China (Grant Nos. 41874216
and 51604230), the China Postdoctoral Science Foundation (Grant Nos. 2017T100592 and 2016M600626), and the
Young Elite Scientists Sponsorship Program by CAST (2017QNRC001).
Acknowledgments: This work was supported by the Program of Introducing Talents of Discipline to Chinese
Universities (111 Plan) (Grant No. D18016).
Conflicts of Interest: The authors declare no conflict of interest.
Abbreviations
HTHP High-temperature and high-pressure
QRA Quantitative risk analysis
M-C Mohr-coulomb
MG-C Mogi-coulomb
H-B Hoek-brown
TVD True vertical depth
EMWCP Equivalent mud weight of collapse pressure
EMWFP Equivalent mud weight of fracture pressure
EMWPP Equivalent mud weight of pore pressure
EMW Equivalent mud weight
Std Dev Standard deviation
Var Variance
SMWW Safe mud weight window
UBD Underbalanced drilling
MPD Managed pressure drilling
MFD Micro-flow drilling
Appendix A M-C Criterion

The M-C criterion can be expressed by using principal stress [7],
σ1 = C0 + qσ3 (A1)
where C0 and q are given as,

( 2c cos ϕ
C0 = 1−sin ϕ
1+sin ϕ (A2)
q = 1−sin ϕ
where ϕ is the friction angle, (◦ ); c is the cohesive strength, MPa; σ1 and σ3 are the major and minor principal
stress, MPa.
If considering the conventional effective stress concept, the M-C criterion can be rewritten as,

σ1 − αp p = C0 + q σ3 − αp p (A3)
where α is the Biot’s coefficient; pp is the pore pressure, MPa.

Rearranging Equation (A3), the M-C criterion can be expressed as [7],
σ1 = C + qσ3 (A4)
where C is given as,

C = C0 − (q − 1)αp p (A5)
Energies 2019, 12, 942 29 of 31
Appendix B MG-C Criterion

The MG-C criterion can be expressed as [2,7],
τoct = a + bσm,2 (A6)
where τ oct , σm,2 , a and b are given as,

 h i1
2
+ (σ2 − σ3 )2 + (σ3 − σ1 )2
1 2
3 ( σ1 − σ2 )

τoct = (A7)
 σm,2 = 12 (σ1 + σ3 )
( √
2 2
a= 3√ c cos ϕ (A8)
b = 2 3 2 sin ϕ
where a is the intersection of the line on τ oct -axis; b is the inclination of the line; σ2 is the intermediate principal
stress, MPa.
If considering the conventional effective stress concept, the MG-C criterion can be rewritten as,
1
= a0 + b0 I1 − σ2 − 2αp p
2
I12 − 3I2

(A9)
where I1 , I2 , a0 and b0 are given as,

I1 = σ1 + σ2 + σ3
(A10)
I2 = σ1 σ2 + σ2 σ3 + σ3 σ1
0
a = 2c cos ϕ
(A11)
b0 = sin ϕ
References
1. Ma, T.; Chen, P.; Zhao, J. Overview on vertical and directional drilling technologies for the exploration
and exploitation of deep petroleum resources. Geomech. Geophys. Geo-Energy Geo-Resour. 2016, 2, 365–395.
[CrossRef]
2. Ma, T.S.; Chen, P.; Yang, C.H.; Zhao, J. Wellbore stability analysis and well path optimization based on the
breakout width model and Mogi-Coulomb criterion. J. Pet. Sci. Eng. 2015, 135, 678–701. [CrossRef]
3. Fjar, E.; Holt, R.M.; Raaen, A.M.; Risnes, R. Petroleum Related Rock Mechanics, 2nd ed.; Elsevier: Amsterdam,
The Netherlands, 2008; pp. 309–339.
4. Aadnoy, B.; Looyeh, R. Petroleum Rock Mechanics: Drilling Operations and Well Design; Gulf Professional
Publishing: Oxford, UK, 2011; pp. 65–296.
5. Chen, M.; Jin, Y.; Zhang, G. Rock Mechanics of Petroleum Engineering; Science Press: Beijing, China, 2008;
pp. 55–130.
6. Zoback, M.D. Reservoir Geomechanics; Cambridge University Press: Oxford, UK, 2007; pp. 167–339.
7. Al-Ajmi, A.M.; Zimmerman, R.W. Stability analysis of vertical boreholes using the Mogi-Coulomb failure
criterion. Int. J. Rock Mech. Min. Sci. 2006, 43, 1200–1211. [CrossRef]
8. Al-Ajmi, A.M.; Zimmerman, R.W. A new well path optimization model for increased mechanical borehole
stability. J. Pet. Sci. Eng. 2009, 69, 53–62. [CrossRef]
9. Ma, T.S.; Chen, P. A wellbore stability analysis model with chemical-mechanical coupling for shale gas
reservoirs. J. Nat. Gas Sci. Eng. 2015, 26, 72–98. [CrossRef]
10. Ma, T.S.; Chen, P.; Zhang, Q.B.; Zhao, J. A novel collapse pressure model with mechanical-chemical coupling
in shale gas formations with multi-weakness planes. J. Nat. Gas Sci. Eng. 2016, 36, 1151–1177. [CrossRef]
11. Ma, T.S.; Zhang, Q.B.; Chen, P.; Yang, C.H.; Zhao, J. Fracture pressure model for inclined wells in layered
formations with anisotropic rock strengths. J. Pet. Sci. Eng. 2017, 149, 393–408. [CrossRef]
12. Ma, T.S.; Yang, Z.X.; Chen, P. Wellbore stability analysis of fractured formations based on Hoek-Brown
failure criterion. Int. J. Oil Gas Coal Technol. 2018, 17, 143–171. [CrossRef]
13. Ma, T.S.; Liu, Y.; Chen, P.; Wu, B.S.; Fu, J.H.; Guo, Z.X. Fracture-initiation pressure prediction for transversely
isotropic formations. J. Pet. Sci. Eng. 2019, 176, 821–835. [CrossRef]
Energies 2019, 12, 942 30 of 31
14. Morita, N. Uncertainty analysis of borehole stability problems. In Proceedings of the SPE Annual Technical
Conference & Exhibition, Dallas, TX, USA, 22–25 October 1995.
15. McLellan, P.J.; Hawkes, C.D. Application of probabilistic techniques for assessing sand production and
borehole instability risks. In Proceedings of the SPE/ISRM Rock Mechanics in Petroleum Engineering,
Trondheim, Norway, 8–10 July 1998.
16. Ottesen, S.; Zheng, R.H.; McCann, R.C. Wellbore stability assessment using quantitative risk analysis.
In Proceedings of the SPE/IADC Drilling Conference, Amsterdam, The Netherlands, 9–11 March 1999.
17. De Fontoura, S.A.B.; Holzberg, B.B.; Teixeira, E.C.; Frydman, M. Probabilistic analysis of wellbore
stability during drilling. In Proceedings of the SPE/ISRM Rock Mechanics Conference, Irving, TX, USA,
20–23 October 2002.
18. Moos, D.; Peska, P.; Finkbeiner, T.; Zoback, M. Comprehensive wellbore stability analysis utilizing
quantitative risk assessment. J. Pet. Sci. Eng. 2003, 38, 97–109. [CrossRef]
19. Sheng, Y.; Reddish, D.; Lu, Z. Assessment of uncertainties in wellbore stability analysis. In Modern Trends in
Geomechanics; Wu, W., Yu, H.S., Eds.; Springer: Berlin, Germany, 2006; pp. 541–557.
20. Al-Ajmi, A.M.; Al-Harthy, M.H. Probabilistic wellbore collapse analysis. J. Pet. Sci. Eng. 2010, 74, 171–177.
[CrossRef]
21. Aadnøy, B.S. Quality assurance of wellbore stability analyses. In Proceedings of the SPE/IADC Drilling
Conference and Exhibition, Amsterdam, The Netherlands, 1–3 March 2011.
22. Zhong, L.; Yan, X.; Tian, Z.; Yang, H.; Yang, X. A collapse pressure analysis in reservoir drilling based on the
reliability method. Acta Pet. Sin. 2012, 33, 477–482.
23. Udegbunam, J.E.; Aadnøy, B.S.; Fjelde, K.K. Uncertainty evaluation of wellbore stability model predictions.
J. Pet. Sci. Eng. 2014, 124, 254–263. [CrossRef]
24. Kinik, K.; Wojtanowicz, A.K.; Gumus, F. Temperature-induced uncertainty of the effective fracture pressures:
Assessment and control. In Proceedings of the SPE Deepwater Drilling and Completions Conference,
Galveston, TX, USA, 10–11 September 2014.
25. Kinik, K.; Wojtanowicz, A.K.; Gumus, F. Probabilistic assessment of the temperature-induced effective
fracture pressures. Spe Drill. Complet. 2016, 31, 40–52. [CrossRef]
26. Gholami, R.; Rabiei, M.; Rasouli, V.; Aadnoy, B.; Fakhari, N. Application of quantitative risk assessment in
wellbore stability analysis. J. Pet. Sci. Eng. 2015, 135, 185–200. [CrossRef]
27. Plazas, F.; Calderon, Z.; Quintero, Y. Wellbore stability analysis: A stochastic approach applied to a colombian
cretaceous formation. In Proceedings of the SPE Latin American and Caribbean Petroleum Engineering
Conference, Quito, Ecuador, 18–20 November 2015.
28. Kanfar, M.F.; Chen, Z.; Rahman, S.S. Risk-controlled wellbore stability analysis in anisotropic formations.
J. Pet. Sci. Eng. 2015, 134, 214–222. [CrossRef]
29. Phoon, K.K.; Ching, J. Risk and Reliability in Geotechnical Engineering; CRC Press: Boca Raton, FL, USA, 2014;
pp. 131–179.
30. Baecher, G.B.; Christian, J.T. Reliability and Statistics in Geotechnical Engineering; John Wiley & Sons: Chichester,
UK, 2003; pp. 303–432.
31. Liao, H.; Guan, Z.; Yan, Z.; Ma, G. Assessment method of casing failure risk based on structure reliability
theory. Acta Pet. Sin. 2010, 31, 161–164.
32. Li, Q.; Tang, Z. Optimization of wellbore trajectory using the initial collapse volume. J. Nat. Gas Sci. Eng.
2016, 29, 80–89. [CrossRef]
33. Zhao, K.; Han, J.; Dou, L.; Feng, Y. Moderate collapse in a shale cap of a nearly depleted reservoir. Energies
2017, 10, 1820. [CrossRef]
34. Zhang, H.; Yin, S.; Aadony, B.S. Poroelastic modeling of borehole breakouts for in-situ stress determination
by finite element method. J. Pet. Sci. Eng. 2018, 162, 674–684. [CrossRef]
35. Ma, T.S.; Peng, N.; Tang, T.; Wang, Y.H. Wellbore stability analysis by using a risk-controlled
method. In Proceedings of the 52nd US Rock Mechanics/Geomechanics Symposium, Seattle, WA, USA,
17–20 June 2018.
36. Li, X.; Jaffal, H.; Feng, Y.C.; Mohtar, C.E.; Gray, K.E. Wellbore breakouts: Mohr-Coulomb plastic rock
deformation, fluid seepage, and time-dependent mudcake buildup. J. Nat. Gas Sci. Eng. 2018, 52, 515–528.
[CrossRef]
Energies 2019, 12, 942 31 of 31
37. Li, X.; El Mohtar, C.S.; Gray, K.E. Modeling progressive breakouts in deviated wellbores. J. Pet. Sci. Eng.
2019, 175, 905–918. [CrossRef]
38. Mondal, S.; Chatterjee, R. Quantitative risk assessment for optimum mud weight window design: A case
study. J. Pet. Sci. Eng. 2019, 176, 800–810. [CrossRef]
39. Udegbunam, J.E.; Arild, Ø.; Fjelde, K.K.; Ford, E.; Lohne, H.P. Uncertainty based approach for predicting the
operating window in UBO Design. SPE Drill. Complet. 2013, 28, 326–337. [CrossRef]
40. Zhang, L.; Cao, P.; Radha, K.C. Evaluation of rock strength criteria for wellbore stability analysis. Int. J. Rock
Mech. Min. Sci. 2010, 47, 1304–1316. [CrossRef]
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Energies 12 00942 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Energies 12 00942 PDF

Uploaded by

Copyright:

Available Formats

energies

Energies 2019, 12, 942; doi:10.3390/en12050942 www.mdpi.com/journal/energies

2. Reliability Assessment Theory

The loads and resistances can be assumed as:

where Mk is the margin of safety for factor k.

Prk = P( Mk > 0) = P[( Rk − Qk ) > 0] (3)

Pf k = P( Mk < 0) = P[( Rk − Qk ) < 0] (4)

where fQ (Qk )Energies

Figure 2. Probability densities

Figure 3. The stress distribution on the borehole wall.

Wellbore stress / MPa

3.3. Collaspe Pressure

3.3.2. Collapse Pressure

3.3.3. Equivalent Mud Weight of Collapse Pressure (EMWCP)

The loads and resistances of wellbore collapse can be expressed as:

3.4. Collaspe Pressure Model for MG-C Criterion

3.4.1. MG-C Criterion

where I1 , I2 , a0 and b0 are given as:

3.4.2. Collapse Pressure

where A, B, a0 and b0 are given as:

3.4.3. Equivalent Mud Weight of Collapse Pressure (EMWCP)

The loads and resistances of wellbore collapse can be obtained:

3.5. Facture Pressure Model

3.5.1. Tensile Failure Criterion

where St is the tensile strength, in MPa.

3.5.2. Facture Pressure

where pf is the fracture pressure, MPa.

3.5.3. Equivalent Mud Weight of Fracture Pressure (EMWFP)

where EMWFP is the equivalent mud weight of fracture pressure, g/cm3 .

The loads and resistances of wellbore fracture can be obatined:

3.6. Equivalent Mud Weight of Pore Pressure (EMWPP)

where EMWPP is the equivalent mud weight of pore pressure, g/cm3 .

The loads and resistances of wellbore fracture can be obtained:

3.7. SMWW Model

max{EMWPP, EMWCP}< SMWW < min{EMWFP} (37)

where SMWW is the safe mud weight window, g/cm3 .

4. Results and Discussion

Table 1. Basic parameters for SMWW analysis.

No. Parameter Distribution Mean Std Dev Var P5 P95

4.1. Uncertainty of Basic Parameters

Energies 2019, 12, x FOR PEER REVIEW 12 of 29

4.2. Uncertainty Analysis of EMWCP

probability of 80%, g/cm3.

(i) Breakout width (j) Impact of mean value change on EMWCP

Figure 11. Cont.

0.44 0.48 0.52 0.56 0.60 0.64 0.68

Figure 12. Sensitivity tornado for EMWCP.

0.44 0.48 0.52 0.56 0.60 0.64 0.68

Figure 15. Influence

Min horizontal stress

Max horizontal stress

and 80% respectively.

4.0 EMWPP 100

Cumulative probability (%)

Table 4. The statistics for EMWPP, EMWCP, EMWFP and SMWW.

4.6. Field Observation Report

Table 5. Operating parameters of MPD.

No. Parameters Value

Appendix A M-C Criterion

where C0 and q are given as,

where α is the Biot’s coefficient; pp is the pore pressure, MPa.

where C is given as,

Appendix B MG-C Criterion

where I1 , I2 , a0 and b0 are given as,