Petroleum xxx (2017) 1e6

Contents lists available at ScienceDirect

Petroleum
journal homepage: www.keaipublishing.com/en/journals/petlm

Revisiting fracture gradient: Comments on A new approaching

method to estimate fracture gradient by correcting
MattheweKelly and Eaton's stress ratio
Farizal Hakiki a, *, Muizzuddin Shidqi b
a
King Abdullah University of Science and Technology (KAUST), Ali I. Al-Naimi Petroleum Engineering Research Center (ANPERC), Thuwal 23955,
Saudi Arabia
b
Shell Upstream Indonesia Services B. V., Jakarta Representative Ofce, Talavera Ofce Park 23rd Floor, Jl. Letjen TB Simatupang Kav. 22 e 26, Jakarta
12430, Indonesia

a r t i c l e i n f o a b s t r a c t

Article history: A study performed by Marbun et al. [1] claimed that A new methodology to predict fracture
Received 22 March 2017 pressure from former calculations, MattheweKelly and Eaton are proposed. Also, Marbun et al.'s
Received in revised form paper stated that A new value of Poisson's and a stress ratio of the formation were generated and
27 May 2017
the accuracy of fracture gradient was improved. We found those all statements are incorrect and
Accepted 21 July 2017
some misleading concepts are revealed. An attempt to expose the method of fracture gradient
determination from industry practice also appears to solidify that our arguments are acceptable to
Keywords:
against improper Marbun et al.'s claims.
Fracture gradient
Overburden pressure
Copyright 2017, Southwest Petroleum University. Production and hosting by Elsevier B.V. on
Pore pressure behalf of KeAi Communications Co., Ltd. This is an open access article under the CC BY-NC-ND
MattheweKelly's stress ratio license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Eaton's stress ratio
Poisson's ratio
Pressure safety drilling window
Geomechanics

1. The physics of fracture gradient

v [C-2]
jj
We would show the physics of fracture gradient equation as
an introduction. The relations of strain and stress which are
1h 0  i
derived from Hooke's Law (Eq. C-1) and Poisson's ratio (PR) x s  v s0y s0z [C-3]
denition (Eq. C-2) appears as Eq. C-3 through C-5 [2] (See E x
Appendix). Biot's Theory does not consider earth stress differ-
ence and interfacial effect [3]. 1h 0  i
y sy  v s0x s0z [C-4]
E
s0 E [C-1]
1h 0  i
z s  v s0x s0y [C-5]
E z
The simplied subsurface case assumes a homogeneous in
* Corresponding author. lateral strain x y h and equal lateral effective stress s0x and
E-mail address: farizal.hakiki@kaust.edu.sa (F. Hakiki). s0y . The s0x and s0y could be considered the same i.e. horizontal
Peer review under responsibility of Southwest Petroleum University. effective stress s0h for simplicity. Eq. C-3 can, therefore, be
transformed into Eq. C-6. Each s0x and s0y are, in fact, different and
may be attributed to maximum horizontal principal stress SHmax
Production and Hosting by Elsevier on behalf of KeAi and minimum horizontal principal stress Shmin . The focus re-
mains on lateral and vertical stress-strain relation.

http://dx.doi.org/10.1016/j.petlm.2017.07.001
2405-6561/Copyright 2017, Southwest Petroleum University. Production and hosting by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: F. Hakiki, M. Shidqi, Revisiting fracture gradient: Comments on A new approaching method to estimate
fracture gradient by correcting MattheweKelly and Eaton's stress ratio, Petroleum (2017), http://dx.doi.org/10.1016/j.petlm.2017.07.001
2 F. Hakiki, M. Shidqi / Petroleum xxx (2017) 1e6
1  information abstracted from their paper. Detailed ndings arise
h 1  vs0h  vs0z [C-6]
E
The model considers a critical point where minimum effec-
tive lateral stress s0h f Shmin reaches zero lateral strain h 0
(See Eq. C-7).
 v 
s0h s0 [C-7]
1v z
Eq. C-8 is the extension of Eq. C-7 by substituting Terzaghi's
Principle s0h sf  u Shmin  u and s0z sOB  u . The over-
burden pressure sOB contributes to vertical stress and the mini-
mum pressure to withstand on lateral direction is named
fracture pressure sf . The sf is usually considered to be equal with
Shmin as the least principle stress S3 in practical aspect. Detailed
discussion may refer to the Coulomb Faulting Theory and the
Anderson's Stress and Faulting Classication System [4].
 v 
sf sOB  u u [C-8]
1v
Term v=1  v mentioned as Eaton's stress ratio is a material
or lithology dependence. Density is subject to porosity f hence
for the same lithology it must have different Poisson's ratio at
various compaction level [5e11]. A rock temperature also in-
uences Poisson's ratio [12]. We denote this term as h in Eq. C-9
to generalize the case.
sf hsOB  u u [C-9]
Once Eq. C-9 is divided by depth D; this is what being called
fracture gradient sf =D. Some existing equations mostly play with
the stress ratio h. The classic model assumes that v 1/4
(equivalent with h 1/3) and sOB =D 1 psi/ft at Gulf Coast area
resulting Eq. C-10 [13]. Other model proposed a term called
matrix stress coefcient or stress ratio Ki which is exactly equal
to h (See Eq. C-11) [14]. Another model came with an equation
that exact with Eq. C-8 divided by depth D. Eaton preferred to
present it in Poison's ratio term [15].
sf 1 2u
[C-10]
D 3 3D
sf s  u u
Ki OB [C-11]
D D D
Those proposed equations of fracture gradient could come up
from Hooke's Law (C-1) and Poisson's ratio denition (C-2) and
by adopting Terzaghi's Principle. This principle lets Biot's con-
stant a 1 for relating effective stress s0 , total stress s, and pore
pressure u (See Eq. C-12) and by neglecting rock grain de-
formations [16].
s0 s  au [C-12]
A fracture gradient versus depth provides a guideline for
casing design in which the lowest fracture gradients addressed to
casing shoe location [17,18].
2. Comments on Marbun et al.'s paper (2015)
The latest study by Marbun et al. [1] claimed that they pro-
posed a new methodology to predict fracture gradient by
modifying the constant value (stress ratio) in Eaton's and Mat-
theweKelly's (MK) equation with limited eld data. Those con-
stants are named v=1  v and Ki , respectively for Eaton's and
MK's stress ratio. That statement sounds an exaggeration of
saying that their method is a new one. There are eight misleading
Please cite this article in press as: F. Hakiki, M. Shidqi, Revisiting fracture gradient: Comments on A new approaching method to estimate
fracture gradient by correcting MattheweKelly and Eaton's stress ratio, Petroleum (2017), http://dx.doi.org/10.1016/j.petlm.2017.07.001
as follow.
(a) A questionable thought addresses on how to achieve
overburden pressure and pore pressure in detail. Marbun
et al.'s paper stated that overburden is computed using
neutron and density log. How it was performed is ques-
tionable. Pore pressure prediction method referred to
Stephani's Bachelor Thesis in 2013 that is not open to
public thus it is not citable and not discussable.
(b) Marbun et al. used Gulf Coast data to generate stress ratio
Ki and Poisson's ratio v. The processed data came with a
justication that the conventional methods do not meet
Formation Integrity Test (FIT) and Leak of Test (LOT) value.
This is, absolutely, not an apple-to-apple comparison as
the selected well data have their own clear rock stratig-
raphy (See Fig. 2 of Marbun et al.'s paper). No justication
should be made under such fallacy while ill-matched data
is inevitably expected.
(c) Poisson's ratio-porosity empirical model has an expo-
nential decay where vo and q are positive empirical con-
stants (See Eq. C-13) [5e11].
v vo exp  qf [C-13]
Porosity of sedimentary rock has a decaying relationship with
depth, f f z [19e22]. Effective vertical stress s0z affects
porosity f or mathematically written as f f s0z . Effective
stress is also depth-dependent s0z f z [23] which then
implies f f z. Poisson's ratio, therefore, does have a trend
with a depth v f z. A direct digitation by citing the classic
results [14,15] are performed (See Fig. C-1). It shows that the
direct citation compared to Eqs. (6) and (7) of Marbun et al.'s
papers are different.
Ki 2:84  109 D2 1:1102903  104 D
 0:106214796336 [6]
v 4:618272  1014 D3  2:03019285297  109 D2
3:27436946581803  105 D 0:281550687489
[7]
The data of MK's stress ratio Ki is restricted to 2e20 thousand
feet and it has a distinct difference with Eq. (6) at shallower
than 4000 ft depth. Marbun et al.'s paper covers 0e4000 ft.
Some FIT and LOT data are even shallower i.e. 1000 ft in which
are not covered by the original data in the literature [14,15].
The eld case is not even deeper than 5000 ft. This mistake is
pointed out through Fig. C-1. The model for extrapolating
original MK's stress ratio using Eq. (6) is wrong for shallower
depth.
(d) A simple regression of 2-degree polynomial they per-
formed is subject to initial data of stress ratio Ki and
Poisson's ratio v (See Eqs. (10) and (11) of Marbun et al.'s
paper). Innite options of stress and Poisson's ratio from
around the world are able to produce the constants C1 , C2 ,
and C3 in Eqs. (10) and (11). A considerable claim should be
a curve tting with a Gulf Coast's data initialization on the
local well data i.e. LOT and FIT. This method could be fed
up with an innite degree of freedom of initialization data
to select h. It can be from Gulf Coast or somewhere else. It
does not mean the paper is exible in various eld as they
claimed. It is more into letting a puzzling way whether or
 F. Hakiki, M. Shidqi / Petroleum xxx (2017) 1e6 3

Figure C-1. Matthew-Kelly's (MK) stress ratio and Poisson's ratio (PR) versus depth. Part (a): Comparing literature data with Marbun et al.'s Eqs. (6) and (7). Data is directly
cited from literature [14,15] with asterisk * in the legend). The yellow-dashed area is zoomed in (b) covering 0e4000 ft. Part (b): It is to emphasize that MK's stress ratio is
different with the literature which only provides restricted initial depth, i.e. 1000s and 2000s ft for Poisson's ratio (PR) and MK's stress ratio respectively.

not the input (lithology, stress ratio, overburden and pore
pressure) is correct.
Kinew C1 D2 C2 D C3
vnew
i C1 D2 C2 D C3
(e) An attempt carried out by Marbun et al.'s paper is multi-
plying MK's stress ratio and Poisson's ratio with a and then
adding with b (See Eq. C-14 and C-15, see Marbun et al.'s
Eqs. (8) and (9) for explicit declarations).
 
Kinew Ki a b aK D2 bk D gK a b
 
vnew va b av D2 bv D gv a b
Marbun et al.'s version for Eq. C-14 and C-15 are Eqs. (10) and
(11) respectively. Table 1 of Marbun et al.'s paper declares the
value of C1 , C2 , and C3 . A relation indicates that C1 ai a,
C2 bi a, C3 gi a b where i can be K or v. The value of a and
b follow the rock strata. There is no rm science to claim that
Marbun et al.'s paper modied MK's and Eaton's ratio. That
paper ends to be a method that employed MK's and Eaton's
stress ratio in the form of Eqs. (10) and (11) within Gulf
Coast's data as its initialization. Thus, the claim by authors
that they did modify the constant in Eaton's and MK's are not
proper. Their results are conventional curve tting with a 2-
degree polynomial tting model.
(f) Their works through Table 2 reveal that the easiest way to
perform a curve tting is shifting the curve which means
a 1 and b is arbitrary. Selecting both arbitrary a and b are
less preferable than holding a at 1. A constant bis the key to
handling to shift fracture gradient curve. The FIT and LOT is
a single point data, not a line hence it is easy to drag initial
fracture gradient curve to cross the desired point. Marbun
[10] et al.'s method is just to drag passing through a single
point even if there are more than FIT or LOT data in each
well because the rock strata are in a compartment of ve
[11] layers. The method comes out to assume FIT or LOT data
are in different layer for each point as if that method is
robust. Changing a means amplifying a curve that might
alter the pattern (curvature). The option to change a is
innite degree of freedom. The method performed in
Marbun et al.'s paper will change the effective vertical
stress contribution to somewhat mess up the results by
imposing as1. Consequently, the attempt to utilize
[C-14] neutron and density log for overburden pressure estima-
tion and to deploy mud logging data required in d-exp
equation for pore pressure prediction would be useless if a
[C-15] is more various than b.
(g) The absence of seismic data of p-wave velocity VP and s-
wave velocity VS might be a reason not to be able to pro-
vide Poisson's ratio because Poisson's ratio analytically
comes up as Eq. C-16 [4].
 2
1 VP
1
2 VS
v  2 [C-16]
VP
VS 1
Marbun et al.'s works have, however, published the lithology
of the formation as presented in Sections 2.4 and 2.5 (See
Fig. 2 of Marbun et al.'s paper). It is possible to develop a local
Poisson's ratio model following typical of rock Poisson's ratio
and will be adjusted to local porosity if needed. The typical of
Poisson's ratio of rocks are around [24]: 0.05e0.40 (sand-
stone), 0.10e0.33 (limestone), 0.25e0.40 (claystone) and

Please cite this article in press as: F. Hakiki, M. Shidqi, Revisiting fracture gradient: Comments on A new approaching method to estimate
fracture gradient by correcting MattheweKelly and Eaton's stress ratio, Petroleum (2017), http://dx.doi.org/10.1016/j.petlm.2017.07.001
4 F. Hakiki, M. Shidqi / Petroleum xxx (2017) 1e6

Table C-1
A comparison of typical Poisson's ratio (PR) of a certain rock from published data on Marbun et al.'s Eq. (11) (Equation's constants in Table 1).

Group Typical of PRa Poisson's ratio

Well 1b Well 3b Well 5b Data from Figure 13

1 (Claystone) 0.250 [24]-0.400 [24] 0.445e0.456 0.445e0.455 0.445e0.456 0.445e0.481

2 (Limestone) 0.100 [24]-0.365 [25] 0.337e0.379 0.335e0.369 0.337e0.377 0.421e0.463
3 (Claystone) 0.250 [24]-0.400 [24] 0.389e0.395 0.382e0.389 0.388e0.395 0.420e0.424
4 (Sandstone) 0.050 [24]-0.420 [25] 0.356e0.358 0.351e0.352 0.356e0.358 0.435e0.438
c
5 (Shale) 0.050 [24]-0.417 [25] 0.391e0.392 0.388e0.390 0.391e0.392
a
Refer to Gercek [24] and Condessa [25]. All PR presented by Gercek [24] have smaller lower-boundary. Some PR presented by Condessa [25] have higher upper-
boundary.
b
Computed using Eq. (11) of Marbun et al.'s paper [1]. The writing format is associated with top to bottom depth boundary of stratigraphy clustering which is
presented in Table C-2.
c
Data is not necessarily compared to ve lithology groups of Well 1, 3, and 5 because Fig. 13 only provides four groups. We assume the lithology clustering is from
Group 1 to 4.

0.05e0.32 (shale). Saturating uid also inuences Poisson's
ratio [25]. Dry/gas saturated sandstone is on 0.10e0.17 and
water saturated sandstone is on 0.32e0.42. Oil saturated
limestone is on 0.18e0.22 and water saturated limestone is on
0.32e0.37. Assigning constant b means shifting the curve
hence it will be similar to the local formation. As if that paper
came up with a novel result that is b has a correlation with
lithology. It has no meaning to correlate b with respect to the
lithology. The value could considerably be in normal trend
because it is in the range of typical values, but the whole
curve and lithology grouping is questionable (See Tables C-1
and C-2).
(h) We found mismatched depths presented by Marbun et al.'s
paper. Figure 5 through 9 of Marbun et al.'s paper do not
show the corresponding wells. It is not acceptable because
Figure 2 indicates that the stratigraphy has their depth
interval for each of three wells. It is necessary to have a
correlation of Poisson's ratio or MK's stress ratio for not
only each lithology but also each well hence there should
be 15 correlations to reveal. Figure 13 of Marbun et al.'s
paper only shows four groups of lithology whereas they
claimed there are ve groups (See Tables C-1 and C-2).
3. Practice in industry
Pore pressure, three principal stresses (sOB , Shmin , and SHmax ),
pressure safety drilling window, and drilling trajectory are the
typical work output of geologist, interpreter geophysicist, pet-
rophysicist, and structural geologist or geomechanicist (if any,
depends on company policy). They will then inform drilling
engineers to develop a drilling program: drilling rig power, hole
and casing size, bit, hydraulic program (include mud design), and
well control.
3.1. Pore pressure
Best practice is to avoid predicting pore pressure by looking
into data acquired in nearby wells. The procedure usually sets a
range of pore pressure, which acquired from formation water
gradient, variations in PVT, mud gradients, owing BHP and
historical pressure records. It is possible to transform seismic
properties into pore pressure provided that sufcient spatial
resolution exists [26]. Once seismic-interpreted Poisson's ratio
and three principal stresses match nearby-wells' pore pressure
data or drilling mud weight used, seismic velocities could be
transformed into pore pressure for targeted well [26].
3.2. Overburden pressure
The preferred procedure is to look at density data provided by
density log. However, a poorly-drilled borehole or highly-varied
lithologies (interbedded/high impurity facies) within a drilled
formation may cause scatter in density log recording. A quick
data QC may be carried out in the drilling site by overlaying
density log recording by formation lithology sequence acquired
from cutting analysis. Further analysis may involve laboratory
study on mineralogy by XRD and EDS, pore shape by SEM image,
pore size distribution by mercury intrusion, porosity-
permeability relationships, and resistivity (also m, cementation
factor; n, saturation exponent) [27]. A single regression line of
density versus depth function is created from post-processed
density log then transformed into overburden pressure.
3.3. Poisson's ratio and pressure safety drilling window

Table C-2 The preferable equation is Eq. C-8 because it has more
A comparison of lithology classication versus depth interval. physical meaning to predict a fracture gradient. A range of
Group Depth interval (TVD in meter)
pressure from pore pressure to fracture pressure is named
pressure safety drilling window. This lead to where a casing
Well 1a Well 3a Well 5a Data from
should be addressed. If there is high condence that the for-
Figure 13
mation will withstand a kick while drilling (e.g. competent
1 (Claystone) 0e1331 0e1257 0e1358 0e1980 reservoir, low likelihood of gas encounter, development well in a
2 (Limestone) 1331e3560 1257e3009 1358e3451 1980e3725
3 (Claystone) 3560e4049 3009e3553 3451e4048 3725e4025
well-dense eld), top-down method will be used as it exerts the
4 (Sandstone) 4049e4239 3553e3716 4048e4198 4025e4355 lightest mud possible up to the deepest possible TVD. When
b
5 (Shale) 4239e4470 3716e3988 4198e4402 drilling a less-certain well such as exploration or there is an
a
Data from Fig. 2 of Marbun et al.'s paper [1]. agreement that a formation would not stand pressure exerted by
b
Data is not necessarily compared to ve lithology groups of Well 1, 3, and 5 a kick, bottom-up approach will be used as the same principle
because Fig. 13 only provides four groups. We assume the lithology clustering is applies. It has also been widely applied that well segmentation is
from Group 1 to 4. necessary for similar lithology [28,29]. Thus, Marbun et al.'s

Please cite this article in press as: F. Hakiki, M. Shidqi, Revisiting fracture gradient: Comments on A new approaching method to estimate
fracture gradient by correcting MattheweKelly and Eaton's stress ratio, Petroleum (2017), http://dx.doi.org/10.1016/j.petlm.2017.07.001
 F. Hakiki, M. Shidqi / Petroleum xxx (2017) 1e6 5

method for rock grouping is not novel. Pressure While Drilling

(PWD) data could give a meaningful approach to predict mini-
mum horizontal stress by improving LOT measurement accuracy,
lost circulation identication, and ballooning incidents identi-
cation because PWD continuously measures annular pressure
while drilling process is running [4].

4. Concluding remarks

A serious action to restate the claims of Marbun et al.'s paper

novelty must be taken. An erratum on the data and results is a
must. Summary of the eight faulty contentions are:

(a) Questionable considerations exist in obtaining overburden Figure A-1. Illustration of compression on z-direction. Blue box is initial condition
pressure and pore pressure extensively. and red box is nal condition after being pressed through z.
(b) Comparing local well with Gulf Coast data is not a proper
appeal. Statements according to Figure A-1 and expanding to the
(c) The governed equations to predict MK's stress ratio and understanding of x and y-direction:
Poisson's ratio are distorted tting from original Gulf Coast
data through a direct digitation. 1. Effective stress in z-direction s0z will affect a longitudinal
(d) The manner to select data initialization is fuzzy. strain jj Dz=zo and normal or transversal strains Dx=xo
(e) The published works did not develop a newer equation; and Dy=yo .
only a curve tting without a discovery and doubtful 2. Effective stress in y-direction s0y will affect a longitudinal
curvature. strain jj Dy=yo and normal or transversal strains
(f) The introduced constants a and b deceive the readers as if Dx=xo and Dz=zo .
those constants are the novelty. 3. Effective stress in x-direction s0x will affect a longitudinal
(g) The summarized typical Poisson's ratio of rocks discloses strain jj Dx=xo and normal or transversal strains Dy=yo
that the Marbun et al.'s prediction is not correct. and Dz=zo .
(h) The depth data between stratigraphy and results emerge a
disparity. Thus, in term of Dz=zo , it will follow statement (2) and (3)
above and be as follow.
The presented discussion and correlation of Marbun et al.'s
!  
paper have no contribution to existing methods in particular and z z
science of rock mechanics in general. It is just a case study in v    [A-3]
jj x s0x y s0y
which the Marbun et al.'s proposed methods have no solid sci- s0jj
ence background and have no any novelties. Best practice in the
Let consider the net strain due to stress on all directions.
industry has already exposed an accountable direction to
develop a pressure safety drilling window.   
Dz Dz Dz
net
z [A-4]
zo s0x zo s0y zo s0z
Acknowledgments

s0 s0 s0
The authors appreciate Mr. Andika Perbawa at KAUST, Saudi net
z z z z
x y z
[A-5]
Arabia and Dr. Henry Pasaribu at Shell Global Solutions,
Input Eq. A-3 into Eq. A-5.
Netherlands for their fruitful discussion.
s0 s0 s0
net
z vx  vy z
x y z
[A-6]
Appendix. Derivation of constitutive stressestrain
relationships Input Eq. A-1 into Eq. A-6.

sy s0 0
Hooke's Law (Eq. A-1) is a law stating that the strain in a s0x
0
solid is proportional to the applied effective stress s within the
net
z v v z [A-7]
E E E
elastic limit of that solid. The fact is a rock undergoes non-elastic
behavior before it gets cracked. 1h 0  i
net
z sz  v s0x s0y [A-8]
E
s0 E [A-1]
By similar understanding, we will obtain net net
x and y , similar
Poisson's ratio (Eq. A-2) is the ratio of transverse strain to like Eq. A-8.
longitudinal strain jj in the direction of compressing or
stretching force. The compression will yield >0 and jj <0 while 1h 0  i
net
x sx  v s0z s0y [A-9]
the stretching force will result <0 and jj >0. Therefore, Pois- E
son's ratio of rock and granular material is kept in positive value.
1h 0  i
net
y s  v s0z s0x [A-10]
v [A-2] E y
jj

Please cite this article in press as: F. Hakiki, M. Shidqi, Revisiting fracture gradient: Comments on A new approaching method to estimate
fracture gradient by correcting MattheweKelly and Eaton's stress ratio, Petroleum (2017), http://dx.doi.org/10.1016/j.petlm.2017.07.001
6 F. Hakiki, M. Shidqi / Petroleum xxx (2017) 1e6

References [16] T. Hantschel, A.I. Kauerauf, Fundamentals of Basin and Petroleum Systems
Modeling, Springer, Heidelberg, 2009, http://dx.doi.org/10.1007/978-3-
540-72318-9.
[1] B.T.H. Marbun, A.N. Corina, G.V. Arimbawa, R. Aristya, S. Purwito,
[17] A.T. Bourgoyne, K.K. Millheim, M.E. Chenevert, F.S. Young, Applied drilling
A.F. Hardama, A new approaching method to estimate fracture gradient by
engineering, in: Soc. Petrol. Eng. Textbook Series, vol. 2, SPE, Texas, 1991,
correcting Matthew-Kelly and Eaton's stress ratio, J. Pet. Sci. Eng. 135
ISBN 978-1-55563-001-0.
(2015) 261e267, http://dx.doi.org/10.1016/j.petrol.2015.09.006.
[18] V.A. Akinbinu, Prediction of fracture gradient from formation pressures
[2] M.A. Biot, General theory of three-dimensional consolidation, J. Appl. Phys.
and depth using correlation and stepwise multiple regression tech-
12 (2) (1941) 155e164, http://dx.doi.org/10.1063/1.1712886.
niques, J. Pet. Sci. Eng. 72 (2010) 10e17, http://dx.doi.org/10.1016/
[3] H. Zhao, M. Chen, Extending behavior of hydraulic fracture when reaching
j.petrol.2010.02.003.
formation interface, J. Pet. Sci. Eng. 74 (2010) 26e30, http://dx.doi.org/
[19] L.F. Athy, Density, porosity, and compaction of sedimentary rocks, AAPG
10.1016/j.petrol.2010.08.003.
Bull. 14 (1) (1930) 1e24.
[4] M.D. Zoback, Reservoir Geomechanics, Cambridge University Press, New
[20] K. Magara, Comparison of porosity-depth relationships of shale and
York, USA, 2007.
sandstone, J. Pet. Geol. 3 (2) (1980) 175e185, http://dx.doi.org/10.1111/
[5] D.H. Han, A. Nur, D. Morgan, You have access effects of porosity and clay
j.1747-5457.1980.tb00981.x.
content on wave velocities in sandstones, Geophysics 51 (11) (1986)
[21] J. Zheng, L. Zheng, H.H. Liu, Y. Ju, Relationships between permeability,
2093e2107, http://dx.doi.org/10.1190/1.1442062.
porosity and effective stress for low-permeability sedimentary rock, Int. J
[6] T. Adachi, S. Sakka, Dependence of the elastic moduli of porous silica gel
Rock Mech Min. Sci. 78 (2015) 304e318, http://dx.doi.org/10.1016/
prepared by the sol-gel method on heat-treatment, J. Mater. Sci. 25 (11)
j.ijrmms.2015.04.025.
(1989) 4732e4737, http://dx.doi.org/10.1007/BF01129933.
[22] Y.Z. Sun, L.Z. Xie, B. He, C. Gao, J. Wang, Effects of effective stress and
[7] J.P. Panakkal, H. Willems, W. Arnold, Nondestructive evaluation of elastic
temperature on permeability of sandstone from CO2-plume geothermal
parameters of sintered iron powder compacts, J. Mater. Sci. 25 (2) (1990)
reservoir, J. Rock Mech. Geotech. Eng. 8 (6) (2016) 819e827, http://
1397e1402, http://dx.doi.org/10.1007/BF00585456.
dx.doi.org/10.1016/j.jrmge.2016.07.004.
[8] P.A. Berge, B.P. Bonner, J.G. Berryman, Ultrasonic velocityeporosity re-
[23] J.E. Smith, The dynamics of shale compaction and evolution of pore-uid
lationships for sandstone analogs made from fused glass beads, Geophysics
pressures, Math. Geol. 3 (1970) 239e263, http://dx.doi.org/10.1007/
60 (1) (1995) 108e119, http://dx.doi.org/10.1190/1.1443738.
BF02045794.
[9] M. Asmani, C. Kermel, A. Leriche, M. Ourak, Inuence of porosity on Young's
[24] H. Gercek, Poisson's ratio values for rocks, Int. J. Rock. Mech. 44 (1) (2007)
modulus and Poisson's ratio in alumina ceramics, J. Eur. Ceram. Soc. 21 (8)
1e13, http://dx.doi.org/10.1016/j.ijrmms.2006.04.011.
(2001) 1081e1086, http://dx.doi.org/10.1016/S0955-2219(00)00314-9.
[25] L.G. Condessa, Hydrocarbon identication in fresh-water bearing reser-
[10] O. Yehezkiel, M. Shokhat, M. Ratzker, M.P. Dariel, Elastic constants of
voirs using dynamic Poisson's ratio: a case study, in: The SPWLA 36th
porous silver compacts after acid assisted consolidation at room temper-
Annual Logging Symposium, 26e29 June, Paris, France, 1995.
ature, J. Mater. Sci. 36 (5) (2001) 1219e1225, http://dx.doi.org/10.1023/A:
[26] C.M. Sayers, G.M. Johnson, G. Denyer, Predrill pore-pressure prediction
1004846212506.
using seismic data, Geophysics 67 (4) (2002) 1286e1292.
[11] C. Yu, S. Ji, Q. Li, Effects of porosity on seismic velocities, elastic moduli and
[27] F. Hakiki, A.T. Wibowo, Formulation of rock type prediction in cored well
Poissons ratios of solid materials and rocks, J. Rock. Mech. Geotech. Eng. 8
using fuzzy subtractive clustering algorithm, in: Proceeding of the 38th
(2016) 35e49, http://dx.doi.org/10.1016/j.jrmge.2015.07.004.
Indonesian Petroleum Association Conference and Exhibition, Paper
[12] Z. Zhang, Rock Fracture and Blasting: Theory and Application, Elsevier,
IPA14-SE-118, Jakarta, Indonesia, 2014.
Oxford, 2016.
[28] K.J. Walker, K. Wutherich, I. Terry, J.E. Shreves, J. Caplan, Improving pro-
[13] M.K. Hubbert, D.G. Willis, Mechanics of hydraulic fracturing, Trans. AIME
duction in the Marcellus Shale using an engineered completion design: a
210 (1957) 153e168.
case study, in: SPE Annual Technical Conference and Exhibition, Paper SPE-
[14] W.R. Matthews, J. Kelly, How to predict formation pressure and fracture
159666-MS, 8e10 October, San Antonio, USA, 2012, http://dx.doi.org/
gradient, Oil Gas J. (1967) 92e106.
10.2118/159666-MS.
[15] B.A. Eaton, Fracture gradient prediction and its application in oileld op-
[29] B. Ajayi, et al., Stimulation design for unconventional resources, Schlumb.
erations, in: Transactions of SPE 43rd Annual Fall Meeting of SPE-AIME,
Oileld Rev. 25 (2) (Summer 2013) 34e46.
vol. 46, 1969, pp. 1353e1360, http://dx.doi.org/10.2118/2163-PA.

Please cite this article in press as: F. Hakiki, M. Shidqi, Revisiting fracture gradient: Comments on A new approaching method to estimate
fracture gradient by correcting MattheweKelly and Eaton's stress ratio, Petroleum (2017), http://dx.doi.org/10.1016/j.petlm.2017.07.001