Accepted Manuscript

Wellbore Stability Analysis Based on a New Strength Criterion

Xiangchao Shi, Wuqiang Cai, Yingfeng Meng, Gao Li, Jiaxue Li

PII: S1875-5100(15)30183-9
DOI: 10.1016/j.jngse.2015.09.050
Reference: JNGSE 1032

To appear in: Journal of Natural Gas Science and Engineering

Received Date: 4 June 2015

Revised Date: 17 September 2015
Accepted Date: 19 September 2015

Please cite this article as: Shi, X., Cai, W., Meng, Y., Li, G., Li, J., Wellbore Stability Analysis Based
on a New Strength Criterion, Journal of Natural Gas Science & Engineering (2015), doi: 10.1016/
j.jngse.2015.09.050.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Highlights

A single-parameter parabolic strength criterion was developed.

The new criterion has a higher accuracy than Mohr-Coulomb criterion for
describing rock strength.

PT

A collapse pressure model was established based on the new criterion.

RI
Wellbore Stability Analysis Based on a New Strength Criterion

SC
Xiangchao Shi a∗∗, Wuqiang Cai a, Yingfeng Meng a, Gao Li, Jiaxue Lib

(a. State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University,
Chengdu 610500, China; b. Tarim oilfield CNPC, Korla, Xinjiang, China)

U
AN
Abstract: Wellbore stability is critical to the overall safety of drilling. The adoption of a better strength
criterion greatly improves the accuracy of wellbore stability prediction. Although the Mohr–Coulomb
strength criterion is widely used for wellbore stability analysis, it is difficult to use its linear form to
M

accurately describe the nonlinear characteristics of the rock strength under different confining pressures. In
the present study, we developed a new criterion, which we refer to as the single-parameter parabolic criterion,
for wellbore stability analysis based on the Mohr strength criterion with a parabolic failure envelope. The
D

fitting precisions of the Mohr–Coulomb criterion, the Hoek–Brown criterion, and the developed
single-parameter parabolic criterion were statistically analyzed. It was found that the single-parameter
TE

parabolic criterion accurately describes the relation between the rock strength and confining pressure. Based
on the single-parameter parabolic strength criterion, a collapse pressure calculation model for vertical and
inclined wells was developed and applied to a field case, and the new criterion was further demonstrated to
EP

be superior to the Mohr–Coulomb criterion.

Keywords: wellbore stability; single-parameter parabolic criterion; collapse pressure; fitting precision
C
AC

∗
Corresponding author. Tel.:+86-28-83035459. Fax: +86-28-83032901. E-mail: sxcdream@163.com (X. Shi)
 ACCEPTED MANUSCRIPT

Introduction

The calculation accuracy of the collapse pressure is directly determined by the rock strength
criterion, and an optimized strength criterion improves the precision in predicting the wellbore
stability[1-6]. The strength criteria that are most commonly used for wellbore stability analysis are the
Mohr–Coulomb criterion and the Drucker–Prager criterion[1, 3, 7, 8]
. Recently, Al-Ajmi and

PT
Zimmerman proposed the use of the 3D Mogi–Coulomb strength criterion for brittle rock wellbore
stability analysis[3, 4, 9, 10]. It has also been shown by substantial experimental data[11-16] that the rock

RI
strength and the confining pressure are not linearly correlated. Thus, the linear Mohr–Coulomb
strength criterion does not accurately describe the non-linear strength characteristics of rocks.

SC
Al-Ajmi, Zimmerman, and Ewy[1, 3, 5, 9]
concluded that the Mohr–Coulomb criterion was too
conservative for predicting the collapse pressure and that the Drucker–Prager criterion
underestimated the collapse pressure. Incidentally, the prediction of the collapse pressure should be

U
as accurate as possible because a high prediction suggests a greater likelihood of the formation’s
AN
being crushed, leading to undue degradation of the wellbore stability[17-19], whereas a low prediction
suggests the likelihood of formation collapse. It is crucial to maintain the stability of a borehole,
which is primarily achieved using a reasonable configuration of the drilling fluid density and the
M

well trajectory[3, 10, 20-22] based on the establishment of an accurate collapse pressure profile along the
wellbore extension using a reasonable rock strength criterion.
D

Using a statistical index that describe the correlation between the experimental data and
TE

theoretical value, we analyzed the fitting accuracies of the Mohr–Coulomb criterion, the Hoek–
Brown criterion, and the newly developed single-parameter parabolic criterion to investigate the
relative accuracy of the new criterion for evaluating the rock strength under different confining
EP

pressures. We further used the new criterion to develop a model for predicting the collapse pressure
of vertical and inclined wells and demonstrated that the collapse pressure model was more suitable
C

for the wellbore stability analysis compared to the Mohr–Coulomb criterion through a case
AC

application.

1 Strength Criterion

A rock strength criterion can be expressed in two forms, namely, the implicit shear stress and
normal stress expression ( τ = g (σ ) ) and the explicit principal stress expression ( σ 1 = f (σ 3 ) ).

However, the mechanical meaning of the implicit expression has not been thoroughly elucidated, and
the parameters are relatively difficult to determine[23]. Therefore, this paper primarily discusses the
explicit strength criterion as expressed in terms of the principal stress.
 ACCEPTED MANUSCRIPT

1.1 Mohr-Coulomb criterion

The implicit stress can be expressed as follows (Coulomb (1776)):
τ =c + σ n tan ϕ (1)

By coordinate transformation, equation (1) can be expressed in the form of the explicit principal
stress:

PT
2c cos ϕ 1 + sin ϕ
σ1 = + σ3 (2)
1 − sin ϕ 1 − sin ϕ

Equation (2) can also be written in terms of the maximum shear stress τ max and the mean

RI
maximum principal stress σ m,2 [24]:

SC
τ max = c cosϕ + sinϕσ m,2 (3)

where τ max = (σ 1 − σ 3 ) / 2 and σ m,2 = (σ1 + σ3 ) / 2 .

1.2 Hoek–Brown criterion

U
AN
The Hoek–Brown criterion has the original form and the generalized form. The original
empirical Hoek–Brown criterion was established in 1980 by E. Hoek and E. T. Brown based on the
M

study of the brittle failure of undisturbed rocks and the deformation model of fractured rock mass
[25]
:
D

σ1 = σ 3 + mσ cσ 3 + sσ c 2 (4)
TE

Considering the limitations and deficiencies of this original criterion, in 1995, Hoek[26]
proposed the generalized empirical Hoek–Brown criterion, and the expression is suitable for various
types of rock:
EP

a
 σ3 
σ1 = σ 3 + σ c  m + s (5)
 σ c 
C

where a is a material-related constant. Equation (5) reduces to the original Hoek–Brown criterion for
AC

a = 0.5.
1.3 Single-parameter parabolic criterion
What we refer to as the single-parameter parabolic criterion was originally developed from the
Mohr strength criterion with a parabolic failure envelope. This criterion can be expressed as
follows[27, 28]:

τ2
σ= − T0 (6)
aT0
 ACCEPTED MANUSCRIPT

The normal and shear stresses at different inclined angles (θ) in a rock sample respectively
satisfy the following equations:
 1 1
σ = σ 1 cos θ + σ 3 sin θ = 2 (σ 1 + σ 3 ) + 2 (σ 1 − σ 3 ) cos 2θ
2 2

 (7)
τ = (σ − σ ) sin θ cos θ = 1 (σ − σ ) sin 2θ
 1 3
2
1 3

PT
The Mohr strength criterion in equation (6) and the Mohr stress circles are depicted in Figure 1.

RI
τ=g(σ)

(σ, τ) β=arctan[g′(σ)]

SC
2θ

σ3 σc

U σ1 σ
AN
Figure 1. Mohr strength criterion and stress circles.

Here, β is the angle between the σ axis and the tangent of the rock failure envelope (implicit
M

strength criterion) at (σ, τ), and θ is the rock failure angle (i.e., the angle between the failure surface
and the maximum principal stress plane at failure). The relationship between β and θ is expressed by
D

equation(8):
β + 90o =2θ , tan β = dτ / dσ (8)
TE

The simplified explicit expression of the Mohr strength criterion with a parabolic failure
envelope on the principal stress plane can be obtained by substituting equations (7) and (8) into
EP

equation(6):
(σ1 − σ 3 )2 = 2aT0 (σ1 + σ 3 ) + 4aT02 − a 2T02 (9)
C

If we let A = 2aT0 and C = 4aT02 − a 2T 2 , equation (9) can be rewritten as

AC

(σ1 − σ 3 )2 = A(σ1 + σ 3 ) + C (10)

Simultaneously, the maximum principal stress is equal to the uniaxial compressive strength
when the confining pressure is zero. Hence, it is necessary for equation (10) to satisfy the constraint
σ1 σ =0
= σc (11)
3

According to a fitting analysis, conducted by Mingqing Y[23], of seven groups of rock strength
data obtained from 6 different types of rock, the parameter A exhibits a good linear fitting
 ACCEPTED MANUSCRIPT

performance with the uniaxial compressive strength σ c . Therefore, we can set A = 2(σ c − σ d ) to
further determine the parameter A, where σ d is a constant that is unrelated to σ c . Furthermore, by
substituting equation (11) into equation (10) and letting A = 2(σ c − σ d ) , we obtain

(σ1 − σ 3 )2 = 2(σ c − σ d )(σ1 + σ 3 ) + σ c2 − 2σ c (σ c − σ d ) (12)

Equation (12) can be simplified as

PT
σ1 = σ 3 + σ c − σ d + 2 (σ c − σ d )σ 3 + σ d2 (13)

Utilizing the explicit parabolic Mohr strength criterion in equation (13), Mingqing Y[23]

RI
obtained the strength criterion parameters σ c and σ d through a fitting analysis of the 7 groups of
the rock strength data. Five out of the seven fitting results for the parameter σ d were zero.

SC
Therefore, if σ d is set to 0, the strength criterion in equation (13) would be reduced to equation

(14), thereby becoming the single-parameter parabolic criterion.

U
(σ 1 − σ 3 )=σ c +2 σ c ⋅ σ 3 (14)
AN
In petroleum engineering, it is extremely difficult to drill a sufficient number of cores, for
which reason existing cores are precious. It is thus important to develop a method to accurately
M

determine the parameters of the applied strength criterion using a limited number of cores. The
single-parameter parabolic criterion has the important advantage of using only one parameter, and its
D

specific form can be expressed as follows[29]:

σ1 σ 4σ 3
= 1+ 3 +
TE

(15)
σc σc σc

σ 3 σ1
In the ( , ) coordinate system shown in Figure 2, the vertex of the single-parameter
σc σc
EP

1 1 σ σ
parabolic criterion in equation (15) is ( , ), the symmetry axis is 1 = 3 , the directrix passes
4 4 σc σc
σ1
C

through the origin, and the parabolic criterion is tangent to the axis.
σc
AC
 ACCEPTED MANUSCRIPT

31⁄



4

 1






Parabolic symme tr y axis



0



1

PT
Parabola vertex
(0.25, 0.25)

0
31⁄


RI
Parabolic directri x



0




Figure 2. Single-parameter parabolic criterion.

SC
2 Fitting Precision: Comparative Analyses of Three Strength Criteria

U
Here, we further investigate the precision of the single-parameter parabolic criterion by using it
to fit eight sets of the rock strength data obtained from previous works as well as the present
AN
experimental results.
2.1 Selection of the fitting index
M

Because of the effects of the rock heterogeneity, experimental errors, and the high dispersion of
rock strength data, some strength data inevitably contain significant errors and deviate from the
D

majority of normal data. If the least squares method (lsm), which requires the sum of the squares of
the deviations of all the test data to be a minimum (equation (16)), is used to obtain the
TE

undetermined strength criterion parameters, the fitting curve would be closer to the individual
abnormal data because of the squared distance term, and this will produce fitting errors.
EP

Target: min{∑ [σ 1 − f (σ 3 )] / N }
2
(16)

The fitting index used to assess the fitting precision is the mean misfit (mf), which is given by
C

equation (17). The only parameter of the single-parameter parabolic criterion, σc, is obtained to
AC

determine the least mf (see equation (18)). The rationality is illustrated in Figure 3.
mf =∑ abs[σ1 − f (σ 3 )] / N (17)

Target: min{mf } (18)

 ACCEPTED MANUSCRIPT

y6
60

y5
50 Abnormal data 2

y4
40

Abnormal data 1
30

y3

PT
y2
20 data
mf fitting index
lsm fitting index
y1
10

RI
y0

0
0x0 x201 x2
40 x3
60 x804

Figure 3. Curves fitted by the mean misfit index and the least square method index[23].

SC
In Figure 3, eight data points are fitted using the two fitting indexes, and the abnormal data 1
and 2 greatly deviate from the most normal data points. The lsm-fitted curve is closer to the

U
abnormal data points to minimize the sum of the squares of the deviations, consequently amplifying
AN
the effects of the abnormal data. The mf-fitted curve accurately reflects the data trend of the normal
data, and the eight data points are averagely distributed on both sides of it. Hence, the mf-fitted curve
is not sensitive to the abnormal data, and the trend and distribution of most of the normal data are
M

more accurately reflected[23].

2.2 Fitting precisions of the three strength criteria
D

A comparative analysis was used to examine the fitting precisions of the three strength criteria.
TE

The rock strength data of Longchang sandstone and Donghai sandstone were obtained from triaxial
compressive experiments conducted by the authors. Previous works were consulted for data on M.
trachyte, D. dolomite, Dunham dolomite, and Mizuho trachyte[11]; Tyndall limestone[30, 31]
; and
EP

Jinping sandstone[32, 33]. The fitting results for the eight groups of data are shown in Figure 4.
180 300
C

150
250
/MPa

/MPa
AC

120
200
90
σ1 σ3

σ1 σ3

Experimental data 150

60 Experimental data
Mohr-Coulomb Mohr-Coulomb
Hoek-Brown 100 Hoek-Brown
30
Single-parameter parabolic Single-parameter parabolic
0 50
0 25 50 75 0 20 40 60
σ3 /MPa σ3 /MPa
(a) Longchang sandstone (b) Donghai sandstone
 ACCEPTED MANUSCRIPT

360
700

300 600
/MPa

/MPa
500
240
400
σ1 σ3

σ1 σ3
180 Experimental data
300 Experimental data
Mohr-Coulomb

PT
Mohr-Coulomb
120 Hoek-Brown
200 Hoek-Brown
Single-parameter parabolic Single-parameter parabolic
60 100

RI
0 30 60 90 120 0 30 60 90 120 150
σ3 /MPa σ3 /MPa

(c) M. Trachyte (d) D. Dolomite

SC
700 360

600 300

U
/MPa

500 240
/MPa
AN
400 180
σ1 σ3

Experimental data Experimental data

σ1 σ3

300 Mohr-Coulomb 120 Mohr-Coulomb

Hoek-Brown Hoek-Brown
M

200 Single-parameter parabolic 60

Single-parameter parabolic
100 0
0 40 80 120 160 0 30 60 90 120
D

σ3 /MPa σ3 /MPa
TE

(e) Dunham dolomite (f) Mizuho trachyte

180 300

150 250
EP

/MPa
/MPa

120 200

90 150
C

σ1 σ3
σ1 σ3

Experimental data Experimental data

60 100
AC

Mohr-Coulomb Mohr-Coulomb
Hoek-Brown Hoek-Brown
30 50
Single-parameter parabolic Single-parameter parabolic
0 0
0 15 30 45 0 30 60 90
σ3 /MPa σ3 /MPa
(g) Tyndall limestone (h) Jinping sandstone
Figure 4. Fitting curves obtained by the three strength criteria for eight groups of rock strength data.

The results of the fitting curves (Figure 4) are summarized in Table 1.

Table 1. Comparative analyses of the fitting accuracies of the three strength criteria applied to rock strength data.
 ACCEPTED MANUSCRIPT

Number Mohr–Coulomb Hoek–Brown Single-parameter

of rock σ0 criterion criterion parabolic criterion
Rock name
strength /MPa σc mf σc mf σc mf
data /MPa /MPa /MPa /MPa /MPa /MPa

Longchang sandstone 9 38 70 6.8 58 4.2 49 4.7

Donghai sandstone 8 97 125 11.2 116 7.5 113 10.7

PT
M. Trachyte 7 154 9.9 140 4.4 115 5.3

D. Dolomite 8 341 18.6 285 9.2 250 11.6

RI
Dunham Dolomite 7 257 313 20.2 280 11 247 9.4

Mizuho Trachyte 7 100 154 16 132 10.2 114 7.8

Tyndall limestone 9 52 73 4.2 68 3.3 53 3.7

SC
Jinping sandstone 9 62 107 11.8 83 6.6 84 9.7

From the fitting data in Table 3, it can be observed that the fitting by the linear Mohr–Coulomb

U
criterion significantly deviates from the experimental data and that the fitting precision is
AN
considerably poorer than those of the Hoek–Brown and single-parameter parabolic criteria, and the
fitting index mf from the Hoek–Brown criterion is close to or slightly lower than that from the
single-parameter parabolic criterion. This result shows that the single-parameter parabolic criterion
M

produces a satisfactory fitting accuracy among the three strength criteria.

The value of the uniaxial compressive strength σc predicted by the single-parameter parabolic
D

criterion is the closest to the actual rock uniaxial compressive strength (the experimentally
TE

determined value) among those predicted by the three strength criteria, as illustrated in Figure 5, and
the values of σc predicted by the Mohr–Coulomb criterion and Hoek–Brown criterion are
considerably higher than the actual rock uniaxial compressive strength. The advantage and desirable
EP

fitting performance of the single-parameter parabolic criterion enable direct determination of the
rock triaxial compressive strength using the uniaxial compressive strength according to equation (15).
C

In contrast, the Hoek–Brown criterion, a two-parameter criterion, is affected by its empirical

parameters, which are calculated using substantial rock strength data. It is therefore comparatively
AC

difficult to obtain the parameters and rock strength profile for wellbore stability analysis along the
borehole extension.
 ACCEPTED MANUSCRIPT

Experimental value Predicted by single-parameter parabolic criterion

Predicted by Mohr–Coulomb criterion Predicted by Hoek–Brown criterion
320
Uniaxial compressive strength

280
240
200
160

PT
120
80

RI
40
0
Longchang Donghai Dunham Mizuho Tyndall Jinping

SC
sandstone sandstone Dolomite Trachyte limestone sandstone

Figure 5. Comparison of the experimentally determined uniaxial compressive strength and those determined by fitting
using different strength criteria.

U
Note that only six groups of data are included in the figure because of the absence of the uniaxial
AN
compressive strengths of M. Trachyte and D. Dolomite.
2.3 Applicable conditions of the single-parameter parabolic criterion
M

The slope of the linear Mohr-Coulomb strength criterion is a constant in the principal stress
plane. In the case of the single-parameter parabolic criterion, the derivative is obtained using
equation (14), resulting in the following:
D

dσ 1 σ
=1+ c (19)
TE

dσ 3 σ3

The inspection of equation (19) reveals that dσ1 / dσ 3 is infinite when σ 3 =0 . This indicates
EP

an internal friction angle of 90°, which probably explains why the rock strength predicted by the new
strength criterion overestimates the rock strength in the low effective confining pressure range,
C

especially in the cases of Donghai sandstone (Figure 4(b)) and Jinping sandstone (Figure 4(h)).
Therefore, the single-parameter parabolic criterion cannot be applied when the effective confining
AC

pressure or effective σ 3 is close or equal to zero, and the effective σ 3 cannot be tensile.

Wellbore instability problems in deep formations are particularly often encountered in

petroleum engineering practice. Generally, the formation confining pressure (minimum principal
stress) is greater than 15 MPa for formation depths greater than 1500 m. It was therefore necessary
for us to further explore the fitting performance of the single-parameter parabolic criterion for rocks
under confining pressures higher than 15 MPa. To do this, we removed the rock strength data with
confining pressures lower than 15 MPa and used the single-parameter parabolic criterion to refit the
 ACCEPTED MANUSCRIPT

remaining experimental data. The newly fitted uniaxial compressive strength σc and the fitting index
mf are shown in Figure 6.
UCS fitted by all data UCS fitted by partial data (σ3>15MPa)
mf fitted by all data mf fitted by partial data (σ3>15MPa)
Uniaxial compressive strength (UCS)

300 12

250 10

PT
Mean misfit (mf)
200 8

150 6

RI
100 4

SC
50 2

0 0

U
AN
Figure 6. Comparison of the fitting results for the complete and partial data (σ3 > 15 MPa).

Figure 6 compares the uniaxial compressive strengths determined by the complete dataset and
M

partial dataset, and there are almost no differences between the two results. This indicates that the
uniaxial compressive strength predicted by the single-parameter parabolic criterion is not sensitive to
D

changes in the confining pressure. However, the mf values fitted by the partial dataset are
TE

significantly lower, with some reduced by half, such as Longchang sandstone, Donghai sandstone,
Mizuho Trachyte, Tyndall limestone, and Jinping sandstone. Additionally, almost all the mf values
are smaller than the Hoek-Brown criterion. The single-parameter parabolic criterion thus accurately
EP

describes the rock strength performance under high confining pressures.

3 Collapse Pressure of a Vertical Well Based on the Single-Parameter Parabolic

C

Criterion
AC

In our study, we used the single-parameter parabola criterion to develop a wellbore stability
model for optimization analysis in setting the well trajectory and the drilling fluid density. The main
analysis procedure is illustrated in Figure 7.
 ACCEPTED MANUSCRIPT

Basic data

Geological data Engineering data

In - situ stress Wellbore Wellbore Drilling fluid

Formation lithology
(direction and magnitude) structure trajectory density

PT
Maximum Minimum Well azimuth Well inclination
Vertical stress
horizontal stress horizontal stress (α) (i)
( σv )
( σH ) ( σh )

RI
Stress state around wellbore

SC
( σ1, σ2, σ3 )

Wellbore stability analysis model based on

single-parameter parabolic criterion

U
(see Table 2 and Equations (38)～(42) )

Not
AN
The best optimizations ?

Yes
M

The optimized design of wellbore

trajectory and drilling fluid density
D

Figure 7. Procedure for optimizing the wellbore trajectory and drilling fluid density based on the single-parameter
parabolic criterion.
TE

3.1 Vertical wellbore stresses

For a vertical well, the borehole inclination i and the borehole azimuth α are equal to zero, and
EP

the state of the wellbore stress can be expressed as {σr, σθ, σz }. The stress state around a circular
elastic hole can be illustrated based on the classical Kirsch equations presented as equation (20), as
shown in Figure 8.
C
AC
 ACCEPTED MANUSCRIPT

σh

σθ σr
τrθ
H σθ
τrθ
Pi r σr σH
θ θ
σH
Pp

PT
ri

σh

RI
SC
Figure 8. Vertical wellbore stress calculation model[34].

σ r = pi
σ θ = (σ H +σ h ) − 2(σ H − σ h ) cos 2θ − pi +2η ( pi − p p )

U
(20)
σ z = σ v − 2ν (σ H − σ h ) cos 2θ +2η ( pi − p p )
AN
τ rθ = τ θ z = τ zr = 0

where σr, σθ, and σz are the radial stress, circumferential stress, and vertical stress around the
M

wellbore, respectively; pi is the wellbore fluid column pressure; pp is the formation pore pressure; σv,
σH, and σh are the overburden pressure, maximum, and minimum horizontal stress, respectively; η is
D

the coefficient of transmission; and ν is the Poisson’s ratio.

3.2 Calculation of the collapse pressure of a vertical wellbore
TE

The single-parameter parabolic criterion can be expressed in terms of the effective stress as
follows:
EP

(σ 1 − σ 3 )= σ c +2 σ c ⋅ σ 3 − α b p p (21)

where αb is the effective stress coefficient of the rock[35].

C

α b = 1 − cma / cb (22)
AC

For a vertical well, {σr, σθ, σz} can be determined as the wellbore principal stresses, and there
are three possible arrangements according to their magnitudes[36], (1) σ θ ≥ σ z ≥ σ r , (2) σ z ≥ σ θ ≥ σ r ,
and (3) σ θ ≥ σ r ≥ σ z . Regarding case (1), there are two corresponding relationships, σ 1 = σ θ and
σ 3 = σ r . From equation (20), we conclude that the maximum principal stress difference (σ θ − σ r )

occurs at θ = π / 2 or 3 π / 2 , which is orientated parallelly to the minimum horizontal stress, and we

obtain the maximum principal stress difference, as shown in equation (23).
(σ1 − σ 3 )= 3σ H − σ h − 2 pi +2η ( pi − p p ) (23)
 ACCEPTED MANUSCRIPT

Substituting equation (23) into equation (21) gives

3σ H − σ h − 2 pi +2η ( pi − p p ) = σ c +2 σ c ⋅ pi − α b p p (24)

If K = 1 − η and M = 3σ H − σ h − 2η p p , equation (24) can be simplified as

M − 2 Kpi = σ c +2 σ c ⋅ pi − α b p p (25)

PT
The solution of equation (25) gives

( KM + ησ c ) ± σ c [σ c (1 − 2 K ) + 2 KM − 4 K 2α b p p ]
pi = pbw =

RI
(26)
2K 2

where pbw is the collapse pressure.

SC
Generally,

( KM + ησ c ) + σ c [σ c (1 − 2 K ) + 2 KM − 4 K 2α b p p ]

U
( KM + ησ c )
pbw = 2
> > σH (27)
2K 2K 2
AN
However, the collapse pressure evaluated by this equation is too high and higher than the
formation fracture pressure. Equation (27) is therefore discarded for
M

( KM + ησ c ) − σ c [σ c (1 − 2 K ) + 2 KM − 4 K 2α b p p ]
p =
w
b (28)
2K 2
D

If the seepage effect between the formation fluids and the drilling fluids is ignored, equation (28)
TE

would reduce to

M − σ c [2M − σ c − 4α b p p ]
pbw = (29)
2
EP

The collapse pressures for cases (2) and (3) can be similarly obtained, as given in Table 2.
Table 2. Vertical well collapse pressure model based on the single-parameter parabolic criterion.
C

Expression of the collapse pressure Expression of the collapse pressure

σ1 ≥ σ 2 ≥ σ 3
AC

(considering seepage flow) (ignoring seepage flow)

σθ ≥ σ z ≥ σ r ( KM + ησ c ) − σ c [σ c (1 − 2 K ) + 2 KM − 4 K 2α b p p ]
M − σ c [2M − σ c − 4α b p p ]
2K 2
2

σ ≥ σθ ≥ σ σ c (1 + 2η ) + K ' M ' − 2 σ c [2ησ c + K ' ( M ' − K 'α b p p )]

z r σ c + M ' − 2 σ c [( M ' − α b p p )]
'2
K

σθ ≥ σr ≥ σ z M '' + (4η − 1)σ c − 2 σ c [ M ' − 2 K 'ησ c + 2η M '' − α b p p ] M '' − σ c − 2 σ c [ M ' − α b p p ]

 ACCEPTED MANUSCRIPT

K=1 η, K′ = 1 2η, M = 3σH σh 2ηpp, M′ = σv + 2ν(σH σ h) 2ηpp, M′′ = (σH + σh) + 2(1
ν)(σH σ h) σv

4 Collapse Pressure of an Inclined Wellbore Based on the Single-Parameter

Parabolic Criterion

4.1 Inclined wellbore stresses

PT
Under in-situ conditions, the formation rock is in a state of stress equilibrium under the
overburden formation pressure and the maximum and minimum horizontal stresses. The stress

RI
components around the wellbore can be expressed as[36-38]
σ r = pi
σ θ = σ xx +σ yy − 2 σ xx − σ yy cos2θ − 4τ xy sin 2θ − pi +2η ( pi − p p )

SC
σ z = σ zz −ν [2 σ xx − σ yy cos2θ + 4τ xy sin 2θ ]+2η ( pi − p p ) (30)
τ θ z = 2τ yz cos θ − 2τ xz sin θ
τ rθ = τ zr = 0

U
AN
where
σ xx = (σ H cos 2 α + σ h sin 2 α ) cos 2 i + σ v sin 2 i
σ yy = σ H sin 2 α + σ h cos 2 α
M

σ zz = (σ H cos 2 α + σ h sin 2 α ) sin 2 i + σ v cos 2 i

(31)
τ xy = 0.5(σ h − σ H ) sin 2α cos i
τ yz = 0.5(σ h − σ H ) sin 2α sin i
D

τ xz = 0.5(σ H cos 2 α + σ h sin 2 α − σ v ) sin 2i

TE

For a given wellbore inclination and azimuth, the state of wellbore stress is a function of θ. It
can be easily observed that the monotonicity of the variation of σθ with θ is consistent with σz, and
EP

the maximum and minimum values can therefore be obtained simultaneously.

For inclined and horizontal wells, the shear stress is generally not 0, and σθ and σz are not the
principal stresses. Hence, further resolution of the wellbore stresses is required to determine the
C

principal stresses.
AC
 ACCEPTED MANUSCRIPT

well-axis
z
σz
τ zθ

γ
σr τ θz

PT
σ
τ σθ

RI
SC
Figure 9. Stress distribution unit of an inclined wellbore.

U
The balance of the stresses in Figure 9 gives
σ = σ θ cos 2 γ + τ θ z sin 2γ + σ z sin 2 γ

AN
 1 (32)
τ = (σ z − σ θ ) sin 2γ + τ θ z cos 2γ
 2

where σ and τ are, respectively, the principal stress and shear stress in the inclined wellbore; γ is the
M

angle between the well axis and τ; σz and σθ are, respectively, the transformed vertical stress and
circumferential stress in the inclined wellbore; and τzθ is the shear stress between σz and σθ.
D

To obtain the principal stress, let

dσ
= 0 or τ = 0
TE

(33)
dγ

Consequently,
 1 2τ θ z
EP

γ 1 = 2 arctan σ − σ
 θ z
(34)

γ = + arctan τ θ z
π 1 2
 2 2 2 σθ − σ z
C

By substituting equation (34) into equation (32), the two principal stresses σj and σk can be
AC

obtained, where there is an angular difference of 90° between them. Hence, the three principal
stresses in an inclined wellbore are


σ i = σ r = pi

 (σ θ +σ z ) + (σθ +σ z ) + 4τ θ z 2
2

σ j = (35)
 2

σ = (σ θ +σ z ) − (σ θ +σ z ) + 4τ θ z 2
2


k
2
 ACCEPTED MANUSCRIPT

where σi, σj, and σk are the principal stresses.

4.2 Calculation of the collapse pressure for an inclined wellbore
The relationships among the three principal stresses in an inclined well are difficult to
determine because that requires the use of geological data and drilling decisions. Nevertheless, there
are three cases for determining these stresses, and they are respectively applicable to the following

PT
situations:
(1) If the minimum principal stress is σ r , the principal stresses σ 1 and σ 3 are as follows:

σ 3 = pi

RI

 (σθ +σ z ) + (σθ +σ z ) + 4τθ z 2
2
(36)
σ1 =
 2

SC
Substituting equation (36) into equation (21) gives

( σ θ +σ z ) + ( σ θ +σ z ) + 4τ θ z 2
2

− pi = σ c +2 σ c ⋅ pi − α b p p

U
(37)
2
AN
An analytical solution of equation (37) is difficult to obtain, and the collapse pressure can only
be determined by a numerical method.

(
 F1 ( p i ) = (σ 1 − σ 3 ) − σ c +2 σ c ⋅ σ 3 − α b p p )
M

 (38)
 F1 ( p bw ) = 0

(2) If the intermediate principal stress is σ r , the principal stresses σ 1 and σ 3 are as follows:
D


σ = ( θ
σ +σ z ) − (σ θ + σ z ) + 4τ θ z 2
2
TE

 3
2
 (39)
 (σ θ + σ z ) + (σ θ + σ z ) + 4τ θ z 2
2

σ 1 =
 2
EP

The collapse pressure in this case can also be determined by a numerical method.

(
 F2 ( pi ) = (σ 1 − σ 3 ) − σ c +2 σ c ⋅ σ 3 − α b p p )
C

 (40)
 F2 ( pb ) = 0
w
AC

(3) If the maximum principal stress is σ r , the principal stresses σ 1 and σ 3 are as follows:


σ = ( θ z ) (σθ +σ z )
σ +σ − + 4τθ z 2
2

 3
2 (41)

σ1 = pi

Here also, the collapse pressure can be determined numerically.

(
 F3 ( pi ) = (σ 1 − σ 3 ) − σ c +2 σ c ⋅ σ 3 − α b p p

) (42)
 F3 ( pb ) = 0
w
 ACCEPTED MANUSCRIPT

5 Calculation and Sample Applications

The case study considered here was obtained from a previous work[39]. Well A and well B were
drilled in an offshore field in the Arabian Gulf, where the rock properties indicated that the block
was in a state of stress heterogeneity. Wellbore collapse occurred in well A during drilling to the
shale bed at a depth of 2073 m (6800 ft), while well B was successfully drilled without any incident.

PT
The main geological parameters of the shale bed are given in Table 3, and the drilling parameters of
both wells are given in Table 4.

RI
Table 3. Petrophysics and reservoir parameters of an offshore Arabian Gulf field.
c φ ν σv σH σh αb pp
(MPa) (°) （MPa） （MPa） （MPa） （MPa）

SC
6.0 31.3 0.33 51.57 46.9 42.2 0.75 21.57

Table 4. Drilling parameters of the two wells in a shale layer at a depth of 2073 m.

U
Well ID Depth (m) α (°) ρm (g/cm3) Drilling accident
AN
A 2073 30 2.03 drilling failure

B 2073 85 1.96 None

M

Note: The well azimuth angle α is based on the maximum horizontal stress σH.

By inputting the basic parameters using the Newton–Raphson method, the numerical solutions
D

of the collapse pressure were obtained for different well inclinations and azimuth angles based on
the Mohr–Coulomb and single-parameter parabolic criteria, respectively (see Figure 10).
TE

1.60
Collapse pressure equivalent drilling

1.50
EP

1.40
fluid density g/cm3

1.30
C

α=0 α=30°
AC

1.20
α=45° α=60°
α=90
1.10

1.00
0 10 20 30 40 50 60 70 80 90
Angle of inclination i °
（a）
 ACCEPTED MANUSCRIPT

1.60

Collapse pressure equivalent drilling fluid 1.50 α=0 α=30°

α=45° α=60°
1.40 α=90°
density g/cm3

1.30

PT
1.20

RI
1.10

1.00

SC
0 10 20 30 40 50 60 70 80 90
Angle of inclination i °

（b）

U
Figure 10. Predicted collapse pressure at 2073 m in a shale layer in an Arabian Gulf offshore field using (a) the
AN
Mohr–Coulomb criterion and (b) the single-parameter parabolic criterion.

As can be observed from Figure 10, the trends of the collapse pressure densities predicted by
the two different strength criteria are similar, and the optimal trajectory corresponds to a well
M

inclination of 40° and an azimuth angle of 90° (direction of σh). This is consistent with the results
obtained on-site. However, the collapse pressure predicted by the single-parameter parabolic
D

criterion is more sensitive to the well inclination. The primary reason for the drilling failure in well A
may be the closeness of the applied drilling fluid density (2.03 g/cm3) to the minimum horizontal
TE

pressure gradient (2.08 g/cm3). The trajectory parameters of well A for this condition are given in
Table 4. The collapse pressures predicted by the Mohr–Coulomb and single-parameter parabolic
EP

criteria correspond to densities of 1.45 and 1.31 g/cm3, respectively. Hence, the prediction based on
the Mohr–Coulomb criterion is too conservative, and this leads to a narrower safe drilling fluid
C

density window. The designed drilling fluid density is thus higher than that predicted by the
single-parameter parabolic criterion and is close to the value corresponding to the minimum stress.
AC

This eventually leads to crushing of the strata. The collapse pressure prediction can be improved by
applying the single-parameter parabolic criterion.
The inclination of well B is 85°, which is close to 90°, and Figure 10 shows that the
corresponding collapse pressure may be considered as the lower limit of the parameter. The
successful completion of well B therefore indicates that the applied drilling fluid density is higher
than that used for well A. The collapse pressure density calculated by the single-parameter parabolic
criterion for well B using a wellbore inclination of 62° is 1.24 g/cm3. If subjected to this pressure
 ACCEPTED MANUSCRIPT

density, there would be the possibility of well A collapsing. It would therefore be necessary to
increase the collapse pressure density to 1.31 g/cm3 to prevent collapse. Alternatively, the inclination
may be changed to 43°. Optimization of the trajectory and maintenance of the wellbore stability can
thus be achieved using the collapse pressure model based on the single-parameter parabolic criterion.

6 Conclusions

PT
(1) Based on the Mohr strength criterion with a parabolic failure envelope, we developed a
single-parameter parabolic criterion for wellbore stability analysis. We demonstrated the criterion to

RI
be considerably accurate with only one parameter. From the analyses of eight sets of rock test data
using the fitting index mf, it was found that the single-parameter parabolic criterion is applicable to

SC
most rocks compared to the Mohr–Coulomb criterion with higher accuracy. In addition, compared to
the Hoek-Brown criterion, it is easier to determine the only parameter σc, which has a clear physical

U
meaning.
(2) The complete dataset and the partial dataset with σ3 higher than 15 MPa were separately
AN
fitted using the new criterion. The uniaxial compressive strengths separately from the complete
dataset and the partial data showed almost no differences, and the mf values from the criterion was
M

significantly smaller, with some of them decreased by half.

(3) We developed a collapse pressure calculation model for vertical and inclined wells based on
D

the single-parameter parabolic criterion. Through the case analysis using collapse pressure models
based on the Mohr-Coulomb strength criterion and the new criterion, the superiority of the
TE

single-parameter parabolic criterion was adequately explained when determining a reasonable

configuration of the drilling fluid density and optimization of the well trajectory.
EP

Acknowledgments

This work was financially supported by the Open Fund (Number: PLN1421) of the State Key
C

Laboratory of Oil and Gas Reservoir Geology and Exploitation (Southwest Petroleum University),
AC

the SWPU Science and Technology Fund (Number: 2013XJZ029), the Scientific Fund of the
Sichuan Provincial Education Department (Number: 14ZB0060), and the CNPC Key Laboratory of
Drilling Engineering.

References：
：
[1] Ewy R T. 1999. Wellbore-stability predictions by use of a modified Lade criterion. SPE
Drilling & Completion 14(02): 85-91. SPE-87895-PA. http://dx.doi.org/10.2118/87895-PA.
[2] Islam M A., Skalle P., Al-Ajmi A M. et al. 2010. Stability analysis in shale through deviated
boreholes using the Mohr and Mogi-Coulomb failure criteria. Paper ARMA-10-432 presented
 ACCEPTED MANUSCRIPT

at the 44th U.S. Rock Mechanics Symposium and 5th U.S.-Canada Rock Mechanics
Symposium, Salt Lake City, Utah 27-30, June.
[3] Al-Ajmi A M. and Zimmerman R W. 2006. Stability analysis of deviated boreholes using the
Mogi-Coulomb failure criterion, with applications to some oil and gas reservoirs. Paper
SPE-104035 presented at the IADC/SPE Asia Pacific Drilling Technology Conference and
Exhibition, Bangkok, Thailand, 13-15 November. http://dx.doi.org/10.2118/104035-MS.
[4] Al-Ajmi A M. 2006. Wellbore stability analysis based on a new true-triaxial failure criterion.

PT
PhD dissertation, University of KTH, Stockholm.
[5] Al-Ajmi A. and Zimmerman R. 2006. A new 3D stability model for the design of non-vertical
wellbores. Paper ARMA-06-961 presented at the Golden Rocks, the 41st U.S. Symposium on

RI
Rock Mechanics (USRMS), Golden, Colorado17-21, June.
[6] Maleki S., Gholami R., Rasouli V. et al. 2014. Comparison of different failure criteria in
prediction of safe mud weigh window in drilling practice. Earth-Science Reviews 136: 36-58.

SC
http://dx.doi.org/10.1016/j.earscirev.2014.05.010.
[7] McLean M R. and Addis M A. 1990. Wellbore stability analysis: a review of current methods
of analysis and their field application. Paper SPE-19941 presented at the SPE/IADC Drilling

U
Conference, Houston, Texas, 27 February-2 March http://dx.doi.org/10.2118/19941-MS.
[8] McLean M R. and Addis M A. 1990. Wellbore stability: the effect of strength criteria on mud
AN
weight recommendations. Paper SPE-20405 presented at the SPE Annual Technical Conference
and Exhibition, New Orleans, Louisiana, 23-26 September.
http://dx.doi.org/10.2118/20405-MS.
M

[9] Al-Ajmi A M. and Zimmerman R W. 2005. Relation between the Mogi and the Coulomb
failure criteria. International Journal of Rock Mechanics and Mining Sciences 42(3): 431-439.
D

http://dx.doi.org/10.1016/j.ijrmms.2004.11.004.
[10] Al-Ajmi A M. and Zimmerman R W. 2009. A new well path optimization model for increased
TE

mechanical borehole stability. Journal of Petroleum Science and Engineering 69(1): 53-62.
http://dx.doi.org/10.1016/j.petrol.2009.05.018.
[11] Mogi K. 2007. Experimental rock mechanics. Taylor & Francis. 361.pp.
EP

[12] Yang Y., Gao F., and Lai Y. 2013. Modified Hoek–Brown criterion for nonlinear strength of
frozen soil. Cold Regions Science and Technology 86: 98-103.
http://dx.doi.org/10.1016/j.coldregions.2012.10.010.
C

[13] Singh M., Raj A., and Singh B. 2011. Modified Mohr–Coulomb criterion for non-linear triaxial
and polyaxial strength of intact rocks. International Journal of Rock Mechanics and Mining
AC

Sciences 48(4): 546-555. http://dx.doi.org/10.1016/j.ijrmms.2011.02.004.

[14] Singh M. and Singh B. 2012. Modified Mohr–Coulomb criterion for non-linear triaxial and
polyaxial strength of jointed rocks. International Journal of Rock Mechanics and Mining
Sciences 51: 43-52. http://dx.doi.org/10.1016/j.ijrmms.2011.12.007.
[15] Yu J., Lin C M., and Chen Y. 2009. Improved Elastic Nonlinear Normal Deformation
Constitutive Model of Rock Fractures. Paper 2009-066 presented at the ISRM International
Symposium on Rock Mechanics-SINOROCK, The University of Hong Kong, China, 19-22
May.
[16] Mahtab M A. and Goodman R E. 1968. Stresses Around Wellbores in Nonlinear Rock. Society
of Petroleum Engineers Journal 8(03): 304-312. http://dx.doi.org/10.2118/2005-PA.
 ACCEPTED MANUSCRIPT

[17] Edwards S., Matsutsuyu B. and Willson S. 2004. Imaging unstable wellbores while drilling.
Paper SPE-79846 presented at the SPE/IADC Drilling Conference, Amsterdam, Netherlands,
19-21 February. http://dx.doi.org/10.2118/79846-MS.
[18] McLellan P J. and Cormier K. 1996. Borehole instability in fissile, dipping shales, Northeastern
British Columbia. Paper SPE-35634 presented at the SPE Gas Technology Symposium,
Calgary, Alberta, Canada, 28 April-1 May. http://dx.doi.org/10.2118/35634-MS.
[19] Santarelli F J., Dardeau C., and Zurdo C. 1992. Drilling Through Highly Fractured Formations:

PT
A Problem a Model and a Cure. Paper SPE-24592 presented at the SPE Annual Technical
Conference and Exhibition, Washington, D.C, 4-7 October.
http://dx.doi.org/10.2118/24592-MS.

RI
[20] Morita N. 2004. Well orientation effect on borehole stability. Paper SPE-89896 presented at the
SPE Annual Technical Conference and Exhibition, Houston, Texas, 26-29 September.
http://dx.doi.org/10.2118/89896-MS.

SC
[21] Zhou S., Hillis R R., and Sandiford M. 1996. On the Mechanical Stability of lnclined Wellbores.
SPE Drilling & Completion 11(02): 67-73. SPE-28176-PA.
http://dx.doi.org/10.2118/28176-PA.

U
[22] Ma T., Chen P. 2015. A wellbore stability analysis model with chemical-mechanical coupling
for shale gas reservoirs. Journal of Natural Gas Science and Engineering. 26: 72-98.
AN
[23] You M Q. 2010. Study of Mathematical Equation and Parameter Determination of Strength
Criteria for Rock. Chinese Journal of Rock Mechanics and Engineering 29(11): 2172-2184.
[24] Jaeger J C. and Cook N. 1976. Fundamentals of rock mechanics. Mathuen, London: Chapman
M

and Hall.
[25] Hoek E. and Brown E T. 1980. Empirical strength criterion for rock masses. Journal of
D

Geotechnical and Geoenvironmental Engineering 106(9): 1013-1035.

[26] Hoek E, Kaiser P K, and Bawden W F. 2005. Support of underground excavations in hard rock.
TE

Boca Raton, Florida: CRC Press.

[27] Li C, Zheng H, Ge X, Wang S. 2010. Research on two-parameter parabolic Mohr strength
criterion and its damage regularity. Chinese Journal of Rock Mechanics and Engineering
EP

24(24): 4428-4433.
[28] Pengnian Z. 1979. Stress and displacement of adjacent rock of tunnel under two-parameter
parabolic form. Nonferrous Metal (1): 41-44.
C

[29] You M Q. 2011. Comparison of the accuracy of some conventional triaxial strength criteria for
intact rock. International Journal of Rock Mechanics and Mining Sciences 48(5): 852-863.
AC

http://dx.doi.org/10.1016/j.ijrmms.2011.05.006.
[30] You M Q. 2010. Three independent parameters to describe conventional triaxial compressive
strength of intact rocks. Journal of Rock Mechanics and Geotechnical Engineering 2(4):
350-356.
[31] Carter B J., Duncan E S., and Lajtai E Z. 1991. Fitting strength criteria to intact rock.
Geotechnical & Geological Engineering 9(1): 73-81. http://dx.doi.org/10.1007/BF00880985.
[32] Bin W., Jiebing Z., and Aiqing W U. 2010. Experimental validation of nonlinear strength
property of rock under high geostress. Chinese Journal of Rock Mechanics and Engineering,
29(3): 542-548.
[33] Franklin J A. and Hoeck E. 1970. Developments in triaxial testing technique. Rock Mechanics
 ACCEPTED MANUSCRIPT

and Rock Engineering 2(4): 223-228.

http://dx.doi.org/10.1007/BF01245576.
[34] Lou Y S, Jin Y Q. 2006. Rock mechanics and petroleum engineering. Beijing, China:
Petroleum Industry Press.
[35] Biot M A. and Willis D G. 1957. The elastic coefficients of the theory of consolidation. J. Appl.
Mech 24: 594-601.
[36] Fjar E., Holt R M., Raaen A M. et al. 1991. Petroleum related rock mechanics. Elsevier,

PT
Amsterdam.
[37] Hiramatsu Y. and Oka Y. 1968. Determination of the stress in rock unaffected by boreholes or
drifts, from measured strains or deformations. International Journal of Rock Mechanics and

RI
Mining Sciences & Geomechanics Abstracts 5(4): 337-353.
http://dx.doi.org/10.1016/0148-9062(68)90005-3.
[38] Aadnoy B S. 1988. Modeling of the Stability of Highly Inclined Boreholes in Anisotropic Rock

SC
Formations (includes associated papers 19213 and 19886). SPE Drilling Engineering 3(03):
259-268. SPE-16526-PA. http://dx.doi.org/10.2118/16526-PA.
[39] Awal M R., Khan M S., Mohiuddin M A. et al. 2001. A new approach to borehole trajectory

U
optimisation for increased hole stability. Paper SPE-68092 presented at the SPE Middle East
Oil Show, Bahrain, 17-20, March. http://dx.doi.org/10.2118/68092-MS.
AN
M
D
TE
C EP
AC