SPE

SPE 19720

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
Experimental Study of Hydraulic Fracture Geometry Initiated
From Horizontal Wells
W. El Rabaa, Halliburton Services
SPE Member
II
Copyright 1989, Society of Petroleum Engineers, Inc.

This paper was prepared for presentation at the 64th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers held in San Antonio, TX, October 8-11, 1989.

This paper was selected for pre&entation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society
of Petroleum Engineers. Permission to copy is restricted to an abstract c:lf not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment
of where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833838, Richardson, TX 75083-3836. Telex, 730989 SPEDAL.

ABSTRACT channels.
D. Interference between fractures conducted from
Horizontal wells drilled in tight formations separate perforated interval.
are likely to be stimulated by acidizing and/or
propped fracturing to increase productivity. Based on this study several recommendations are
Normally, to control the placement of induced presented to help eliminate some of these stimula-
fractures in these wells, they are completed for tion related problems.
later stimulation, i.e. , cement cased and perfo-
rated. Unlike vertical well fracturing, fractures INTRODUCTION
in horizontal wells can be induced along, or
inclined to, or crossing the horizontal section. Exploitation of hydrocarbon reservoirs by
This orientation depends on deviation of the drilling horizontal wells has increased in the last
horizontal section from the minimum stress few years. Numerous developments in drilling tech-
direction. nology, reservoir studies, logging, and stimulation
tools have made horizontal wells a feasible option
An experimental study of fracture geometry from for many reservoir developments. In tight reser-
horizontal wells at various well azimuth deviations voirs where enhancement of productivity is re-
was conducted. Experiments were performed by qUired, horizontal wells can be stimulated either
applying triaxial loading conditions to rock blocks by proppant fracturing or by acidizing. For
6 x 12 x 18 in. surrounding cased and perforated selective stimulation along the horizontal section,
boreholes. The borehole directions varied from 0 the wellbore is completed by cemented liners and
to 90 degrees from the applied minimum stress. targeted zones are perforated for multiple or
Also, length of perforated interval was the second individual treatments.
variable parameter in this study. After each
experiment, rock blocks were sawed to observe the The number of horizontal wells fractured to
shape of induced fracture. The study showed that date is small compared to vertical wells fractured.
fracture geometry near the horizontal well is Most available literature regarding induced
controlled by well deviation and length of perfo- fracture geometry has dealt mainly with vertical
rated interval. From a stimulation viewpoint, the well fracturing. Conclusively, the final geometry
combined effects of these two parameters, if not of an induced hydraulic fracture is controlled by
investigated beforehand, may cause critical prob- reservoir properties, stresses, and pumped fluids
lems during fracturing. Laboratory observations (rate, volume, properties) . Since parameters do
indicated the following stimulation related prob- not change for horizontal wells, it is expected
lems: also that for reasonable size treatments, these
reservoir properties will ultimately control
A. Creation of multiple fractures from the same fracture geometry in horizontal wells.
perforated interval.
B. Created fractures are nonplanar, with rough Usually at depths encountered in the oilfield,
walls. an induced fracture is vertical and perpendicular
c. Conununication of fluid between the perforated to the least principal horizontal stress ( aHm. ) .
section and fracture can occur through small Thus, for vertical wells, an induced fracture wafild
be aligned with the vertical well trajectory. Such
arrangement represents the ultimate fluid communi-
References and illustrations at end of paper. cation between the well and the fracture. Normal-
189
2 EXPERIMENTAL STUDY OF HYDRAULIC FRACTURE GEOMETRY INITIATED FROM HORIZONTAL WELLS
ly, at the end of a treatment, maximum fracture
height and width will be along the vertical well-
bore walls. On the other hand, a fracture induced
from a horizontal well does not extend in a plane
aligned with the well unless the well has been
drilled normal to a . . For other horizontal well
directions, initial~~cture geometry is dependent
on well deviation from a . , and type of fracture
entrance, e.g., open hol~ ~erforation, or notched
1
hole. Not far from the wellbore, eventually the
fracture is reoriented to become normal to oHm· ·
Factors such as pumping rate and ratio between tRe
two principal horizontal stresses, oHm. and oHmax'
· t a t'1on curva
have some influence on reor1en ln t ure.
This paper presents results of fracturing
experiments performed on hydrostone blocks sur-
rounding horizontal wells. The induced fracture
geometry for various fracture entrance programs ~s
discussed. Some analytical work relevant to th1s
study is also presented. After examining the
bearing of various factors that are now understood
from this work, several recommendations are offered
to eliminate possible stimulation related problems.
DEFINITION AND CONVENTION
In this paper, both experimental and analytical
work was performed for a horizontal well that was
drilled in a state of stress in which two of the
three principal stresses act horizontally and one
of them is the minimum. Thus, the wellbore is in
the plane of the two principal stresses~ oHm .and
a . · , and the induced fracture plane 1s ve'fl.1cal
n~rffi~h to oHm. . This situation seems to cover
most, if not at~, cases where horizontal wells are
drilled. In cases where the principal stresses are
not in vertical and horizontal directions, horizon-
tal wells will be inclined to the principal axes of
stresses. Such cases will not be covered under
this study. Horizontal well deviation from the
a . is the angle (9). Angle (9) is measured from
1
0 Hm n toward 0 which is perpendicular to oHm. ,
i~~~ 9 range ~Mm 0 to goo (Fig. 1). Stress lflle
to overburden weight is in vertical ov.
Compressive stresses are taken as positive.
HORIZONTAL VERSUS VERTICAL WELL FRACTURING
From a stimulation viewpoint, horizontal well
fracturing practice, starting with prepad and
ending with flush, does not differ from that used
in vertical wells. Furthermore, fluids and prop-
pants used in horizontal wells are not greatly
different from those conventionally used, however,
because of the horizontal traveling distance,
fluids with higher proppant carrying capacity
should be used'. 1
Differences in fracture geometry and fracturing
parameters stem mainly from the spatial position of
horizontal section with respect to the principal
in-situ stresses and completion program. The
following differences have been compiled from
pertinent literature in fracturing and from the
work conducted for this study.
1. The induced fracture plane does not contain the
wellbore unless the well is drilled normal to
0 direction of a . {9 = 0), a fracture is initiated
Hmin"
190
SPE 19720
2. It is possible that multiple fractures can be
created from the same open or perforated interval,
in the same pumping stage.
3. Fracture width at the wellbore may not be the
maximum for wells not normal to oHm. • Also, width
1
at the fluid entrance may be fndependent of
fracture geometry at high well deviation angles.
Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
4. The ratio between horizontal principal
stresses 2 (K = a /oHm. ) and injection rate 3 (Q)
affects the fra~~ retltientation path. Effects
of both parameters, K and Q, on the reorientation
path will be addressed later.
5. During the reorientation process, the fracture
is extended under both shearing and tensile modes
resulting in rough surfaced fracture walls. Such
extra wall roughness would increase fluid friction
and hence higher operating pressure may be
required.
6. Initial fracture geometry near the wellbore is
controlled by length and type of fluid entrance to
fracture, i.e., perforations, open hole, or slotted
hole. Depending upon well deviation and type of
entry, connecting distance or communication chan-
nels between wellbore and fracture is decided.
7. Horizontal wells are usually parallel to the
horizontal stratigraphic system, thus they are
drilled in one stress zone. Since vertical varia-
tion in stress controls the height growth, eleva-
tion of the horizontal well is a decisive factor in
induced fracture height in thick pay zones.
BREAKDOWN PRESSURE VERSUS WELL DEVIATION (Open
Hole)
Initiation of hydraulic fractures on wellbore
walls is a tensile failure process which can be
predicted by the maximum normal stress theory.
Tensile failure occurs when the least effective
principal stress exceeds the tensile strength (at)
of a formation. By pressurization of the wellbore,
tensile stresses are induced, diminishing compres-
sive stress concentration around the wellbore. At
the point where the maximum tensile stress reaches
at, hydraulic fracturing occurs, characterizing the
breakdown point. Since stress concentration around
wellbore is a function of wellbore deviation,
breakdown pressure is wellbore deviation dependent.
Complete derivation of stress concentration around
horizontal wellbore can be found in Ref. 4. For
illustration, breakdown pressure in a horizontal
well is calculated for two cases using similar
analysis to that in Ref. 4. For simplicity, the
effect of fluid penetration into formation is
ignored and pore pressure is set to null. In both
cases the state of stress is such that a > oHm >
oHm. , however in one case a = gooo psivand 1n~e
secBHd case a = 7500 psi wi~h oHmax = 12go psi and
a . = 5580 psi in both cases. -~Tfie formation has
a~B~sson•s ratio (v) = 0.25 and tensile strength
at = 500 psi. Breakdown pressure for both cases is
plotted versus wellbore deviation angle {9) in Fig.
2. It can be concluded from Fig. 2 that:
1. When the wellbore is drilled parallel to the
at the highest"Efgakdown pressure and when e =goo,
SPE 19720
the fracture initiates at the lowest breakdown
pressure. This indicates that a wellbore drilled
along a . could be more stable during drilling
than we¥.r~ndrilled in any other direction.
2. The magnitude of breakdown pressure increases
as differences between principal stresses de-
creases. In other words, breakdown is dependent on
relative magnitudes of principal stresses.
For comparison purposes, a vertical well in
this stress field will be hydraulically fractured
at a breakdown pressure of ggso psi. This value
corresponds approximately to a horizontal well
deviated at 50° from aHmin when av = gooo psi.
POINTS OF FRACTURE INITIATION (OPEN HOLE ELASTIC
CONCEPT)
There are two points at which hydraulic frac-
tures are initiated in the horizontal wellbore
walls, depending on magnitudes of av and aHmax:
when a > a initiation points are at
v o~F
where ~ is the angle measured counterclockwise
from the high side of well (Fig. 1).
The case in which a > aHm seems more likely
at depths encountered iX
the o~ field. At points
of initiation (~ = 0 and ~), the initial fracture
would extend at an angle y within the tensile
region created by borehole pressure. The angle y
can be found by Eq.1, 5
1 -1 2apz
y = 2 tan
a~ - az
where a1ll, a , a,, are the tangential, axial, and
shear stres~es ~Jt points of initiation on the
wellbore walls (Fig. 3). Since a , a , and a are
t
functions of wellbore deviation 8) , z angle *zwill
also be a function of (8).
It is reasonable to assume that the initial
fracture will extend in the tensile zone before it
turns to become normal to a . • The maximum
length of initial fracture (Lfvm~~n be calculated
from the following equation. 3
Lf = D cosec y sin (~t/2)
where D is the wellbore diameter and ljl t is the
angle encompassing the tension zone. The angle ljlt
is determined from the circumferential distribution
of tensile stresses that are induced around the
wellbore due to pressurization. For a fixed state
of stress, borehole pressure, and tensile strength
of the rock, ljlt varies also with well deviation
(8).
CRITICAL PERFORATION SPACING
As demonstrated in the previous section,
initial fracture length Lf is a function of both y
and ljlt' consequently Lf is a function of (8),
considering the simple case where two perforations
Wadood El Rabaa 3
in the tensile zone could be connected to each
other by the initial fracture. Conversely, when
the two perforations are separated by a distance
greater than Lf' two independent initial fractures
may occur. Hence, L can be used as a parameter to
estimate the critical distance between two perfora-
tions beyond which more than one initial fracture
can be created. The above procedure is applied
Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
for a horizontal well drilled in a formation which
has at = 500 psi and state of stress as a = gooo,
aHm = 72go and oHm. = 5580 psi. The initial
fra~ure length or c:fftical distance between two
perforations is normalized by borehole diameter
(Lf/D), and plotted versus the wellbore deviation
in Fig. 4.
It can be seen that for a distance separating
two perforations equal to one wellbore diameter or
less, the initial fracture would connect the two
perforations regardless of well deviation angle
(8). The value of y when 8 = oo or goo is 0,
resulting in L =infinity (Eq. 2). For a= 0, such
result physically seem unrealistic, therefore Fig.
~ = 4 shows 8 changes from 10° to goo. Even though
this analysis did not incorporate stresses induced
by perforations or the c::asing, experimental work
showed consistently that the initial fracture
connects all perforations helically spaced ~ D.
EXPERIMENTAL WORK
Fracturing experiments were carried out in 6 x
12 x 18 in. (35/100) hydrostone blocks stressed in
a polyaxial test chamber. Stress on each surface
of the block was applied by platens pushed by
hydraulic pressure. Rubber neoprene pads were
placed between platens and rock to reduce friction
and evenly distribute loads on rock. Hydraulic
pressure for fracturing was delivered by a 1000 cc
(1) intensifier conventionally controlled by servo-
valve and function generator. Fracturing fluid (10
wt motor oil) entered the blocks through a cased
wellbore of 0. 31 in. OD. In all tests the frac-
turing fluid was injected at constant rate. Injec-
tion rates used were from the range of 84 to 333
cc/min; rates are given separately for various
tests. Results of 22 experiments will be discussed
in this paper. Experiments can be divided into six
sets according to the wellbore deviation angle (a)
as follows:
Set 1: Eleven tests total, a = 0°, 10 cased per-
forated wellbore and one with an open
hole.
(2) Set 2: Two tests with a = 15° and differed in
length of perforation interval (Pi).
Set 3: Two tests with 8 = 35° and differed in
length of perfor~ted interval.
Set 4: Two tests with a = 55° and differed in
perforated interval.
Set 5: Two tests with a = 75° and differed in
perforation interval.
Set 6: Three tests with 8 = goo and differed in
type of fluid entrance to fracture,
simulating open hole, perforated casing
and slotted casing.
The experimental work has generated a fair
amount of data that is tabulated and illustrative
figures represent selected cases from each set.
191
4 EXPERIMENTAL STUDY OF HYDRAULIC FRACTURE GEOMETRY INITIATED FROM HORIZONTAL WELLS
EXPERIMENTS, DISCUSSION AND RESULTS
Experiments discussed in this section are: for
six sets arranged in the order of well deviation
(e) from a . , 0°, 15°, 35°, 55°, 75°, and 90°.
Experimenta~~ta and charac~eristics are tabulated
for each set. Perforation interval (Pt) is defined
herein as the distance between the two extreme per-
forations given in terms of borehole diameter (D).
Set 1, 8 = 0°
This set included 11 tests in which the well-
bore is drilled in the direction of OHm.1 , or 8 =
0°. Ten fracturing tests (HZ-la, b, c, nd, e and
HZ-2a, b, c, d, e) wer~ conducted through cased
perforated wellbore with D = 0.31 in. The 11th
test (H15) was run using an open hole. Normally,
in this direction, most horizontal wells are ex-
pected to be drilled. When e = 0°, the induced
fracture is normal to the wellbore, thus, it is
possible to achieve several independent treatments
along the horizontal section. Table 1 summarizes
characteristics of individual tests as well as
major observations. The test groups HZ-1 and HZ-2
are identical except for magnitudes of applied
stresses. In spite of this difference, results
produced by both groups are similar. For perfora-
tion interval > 4D more than one fracture was
observed from the same interval. However, one of
the induced fractures was larger than the other.
This may indicate that eventually one fracture will
dominate and take all fracturing fluid for
propagation. Observed fracturing pressure in Tests
HZ-1 and HZ-2 have indicated that breakdown
pressure decreased as perforation interval (Pt) got
longer and an increased number of perforations was
used. This result seemed to be in agreement with
previous studies in vertical wells. 7 However, when
extension pressure was compared after a specific
volume had been pumped in each test, it was found
that extension pressure in tests that produced two
fractures (Pt ~ 4D) was clearly higher than those
which produced one fracture (Table 2). Increase in
extension pressure (BHTP) in multiple fracturing
was mathematically found 8 to be proportional to IN
where N is the number of fractures. Figure 5 shows
a cutaway view of fractures created in Tests HZ-2
(a-e).
Area of fluid entry into one fracture when
fracture is perpendicular to wellbore is uDWf.
Therefore, a larger diameter borehole can be more
helpful to reduce entry friction. In the general
case, when a wellbore is deviated at angle 8 and
one fracture is created, the area of fracture entry
(Ain) intersecting wellbore can be found by Eq. 3:
Ain = 2DWf sec e E(K) (3)
where E(K) is the elliptical integral and K = sin
e.
Test H15 was conducted with an open hole
section equal to 4D in length. In this test,
fracture width (W£) was measured in the middle of
the open hole ana normal to the wellbore. The
fracture generated by Test H15 had three connected
planes (asH-shape). An initial fracture along the
wellbore extended vertically above and below the
wellbore with a total height of about BD, or twice
192
SPE 19720
the open hole length. Two other fractures were
nearly perpendicular to initial fracture and
separated by the open hole (4D) and slightly over-
lapped near wellbore vicinity (Fig. 6). Obviously
these two fractures constituted the main induced
fracture since they were perpendicular to a . •
Width measurements at wellbore normal to in~Hh
fracture plane showed that Wf did not increase
Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
after breakdown as injection continued and frac-
ture got larger, (Fig. 7). Such observation may
suggest that the width of initial fracture along
the wellbore in this case (8 = 0°) is independent
of the main fracture size. It is worth mentioning
that the initial fracture was only connected along
its height with the two offset fractures. Thus,
communication of fluid would be limited only to a
length of about four times the open hole section.
Bearing of both observations, nonincreasing width
and limited communication length, on fracture
treatment can cause premature screenout. The ISIPs
observed in tests of this set were within the
applied oHm. • Results of fracturing experiments
1
at wellbore zaeviation angle e = 0 are summarized
below.
1. Multiple fractures can be created when Pt ~ 4D.
One fracture would eventually become predominant.
Initially higher extension pressure is expected.
2. For small Pt < 4D one fracture is likely to be
induced.
3. Perforations made within a distance of < D are
likely to connect.
4. In an open hole (4D length) a complex (H-shape)
fracture can be induced. Width of the initial
fracture along the wellbore does not reflect frac-
ture size. Communication between wellbore fluids
and the main fracture body is done through the
initial fracture perimeter (length of communication
is about four times open hole length).
Set 2, 8 = 15°
Two tests, HZ-7b and HZ-5b, were performed in
cased horizontal wellbore making 15° with oHm. •
Conditions of injection rate and applied stres§~s
were the same in both tests. Differences between
tests were made in the perforation length and
number; P in Test HZ-7b was 2D and in Test HZ-5b
was 5D. 9characteristics of each test as well as
shape of induced fracture are listed in Table 3.
In Test HZ-7b one fracture was induced connecting
three perforations (total of six). The fracture
had a slight kink around the casing but extended
normal to a . not far from wellbore. Also, in
this test atilll.Perforations were connected by small
fractures in the wellbore wall. Test HZ-5b (Pt =
5D) has produced an interesting fracture geometry,
especially near the wellbore. The induced fracture
has extended through most perforations (within 4D
distance) as they spiral around the casing. In one
side of the wellbore, the fracture extended in two
vertical planes (two fractures) separated and
slightly overlapped. In the other side, the two
fractures were united along perforations and
extended as a fracture normal to oHm. • Fracture
geometry in Test Hz-5b is shown in-FTQ. 8. Since
this wellbore deviation is only 15°, results seemed
not very different from thdse obtained when 8 = 0°.
SPE 19720
However, created fractures were communicating with
only 10 perforations (total 16).
Set 3, a = 35°
Two tests, HZ-3b and HZ-6b, were performed in
cased holes inclined 35° with oHm. direction.
Characteristics of tests are listed ~fi~able 4. In
Test HZ-3b, eight perforations spirally made with
Pt = 3D. Distance between any two consecutive
perforations was less than D. The fracture induced
in this test was extended among all perforations,
resulting in a one continuous nonplanar fracture.
Fracture wings away from wellbore were normal to
aHm. and they .connect with each other through
perfBrations. Wings seemed to be on two vertical
planes separated by distance (Ls) approximated by
Eq. 4.
Ls = Pt sin a + 2D ( 4)
It was observed that one wing was shorter with a
rough surface containing several step-like features
in the reorientation region, Fig. g. The !SIP
recorded at end of test was g5o psi, which is more
than twice the applied minimum stress, 400 psi.
The high !SIP could be reflecting the components of
stresses acting on the S-shape fracture near the
wellbore.
In Test HZ-6b, seven pairs of perforations were
made in a distance equal to gD. Each pair was
phased goo and spaced 1.5 D from each other. Four
separate fractures were observed in this test.
Three of these fractures extended only between two
consecutive perforations around casing but they
were not connected to each other. The fourth
fracture was initiated from one pair of perfora-
tions and propagated as two wings away from the
wellbore normal to aHrnin"
Again the !SIP recorded at end of the test was
g5o psi. Figure 10 shows the fracture geometry
produced in test HZ-6b. It can be concluded from
this set that the reorientation process is accom-
panied by extra wall roughness, and in spite of
long Pt and multiple initial fractures being in-
duced, only one fracture becomes the main frac-
ture. Also, it was observed that the fracture wing
which makes the smaller angle with the wellbore is
shorter.
Set 4 - a = 55°
Two tests, HZ-3a and HZ-Ga, were performed in
cased holes making 55° with a . direction.
Characteristics of tests are liste~hnTable 5. In
Test HZ-3a, eight perforations were made spirally
around the casing with P = 3D. One fracture was
induced in this test. The fracture was initiated
by connecting most perforations which were made
nearly aligned perpendicular to a . . The frac-
ture was continuous but nonplanar,~~~ing two non-
identical wings (Fig. 11). For Test HZ-Ga, 14
perforations were made spirally around casing with
Pt = gD. One main nonplanar fracture was induced
connecting four pairs of perforations as an S-shape
fracture. Perforations connected by the main frac-
ture occupied a distance equal to 4.5D. Two short
fractures were observed extending from the other
two pairs of perforations. The main fracture re-
Wadood El Rabaa 5
oriented itself near its extreme perforations and
formed two nonidentical wings perpendicular to
a . • Again, in both tests, the wing which made
tHw~nsmaller angle with wellbore direction was
shorter of the two.
Set 5, e = 75°
Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
Two tests, HZ-5a and HZ-7a, were performed with
the cased horizontal well making a 75° angle with
aBm. direction. Characteristics of both tests are
l~st~d in Table 6. In Test HZ-5a, in which eight
perforations extended over a distance P = 3D, one
1
nonplanar S-shape fracture was induced. The frac-
ture connected most of the consecutive perforations
of those directions that were nearly normal to
aHm. • The fracture had nonidentical wings.
Fraa~ure reorientation occurred near the end
perforations (Fig. 12). Test HZ-7a, in which five
pairs of perforations were made at goo phasing, re-
sulted in creation of one continuous nonplanar
fracture. A fracture was initiated, passing
through all perforations, which were made verti-
cally. Perforation in this direction occupied a
distance equal to 3D. The fracture had two
nonidentical wings not in the same vertical plane,
and the horizontal shift also fit Eq. 4.
At this wellbore deviation, tests indicated
that it is not likely to create multiple fracture
from the same perforated interval. Fractures
seemed to be continuous and nonplanar, however,
they do not communicate with all perforations when
long P is used.
1
Set 6, a goo
When a wellbore is drilled at a deviation of e
goo, it is expected that the induced fracture
will be aligned with the wellbore trajectory.
Therefore, selection of perforation program would
not be the main parameter that controls fracture
geometry. Published work for the vertical well
case 7 can be directly applied to this wellbore
deviation. At this deviation, type of entry to
initiate fracture at wellbore walls was studied on
fracturing pressure. Three tests were performed in
the same block sample, HZ-4. Casing of the hori-
zontal well was completed at three points as open
hole, perforated section, and slotted section.
Length of each section was 1 in. , or 3D. In the
perforated section, eight spiral perforations were
made. A 1 in. diameter slot was made aligned with
the. wellbore in the third point. Three fractures
were induced individually by a 2 in. straddle
packer from each entry section. Induced fractures
are shown in Fig. 13. It was observed that the
fracturing pressure (breakdown or extension) for
each test was considerably different. Comparison
between breakdown pressures and extension pressure
recorded after injecting the same amount of fluid
are listed in Table 7. It is oovious that pressure
recorded in the slotted section was the lowest and
that for the perforated section was the highest.
It is the author's belief that the effect of fluid
entry type on fracturing pressure would be the same
regardless of wellbore deviations, i.e., a slotted
wellbore will be the easiest to fracture and to
extend. It was also observed that fractures
created from one zone can communicate with a
previous fracture created from a nearby zone (Fig.
193
6 EXPERIMENTAL STUDY OF HYDRAULIC FRACTURE GEOMETRY INITIATED FROM HORIZONTAL WELLS
1J) . This observation is more clear in a labora-
tory scale experiment because of the small size of
the model. However, communication of fractures in
field scale can occur, caused by fracture turning
due to stress field alteration induced by a nearby
fracture. Furthermore, extension pressure in later
fractures created from the same wellbore would be
increasing due to the altered state of stress.
Both phenomena are likely to occur when spacing
between zones is small or wellbore deviation is
large (9 near 90°). The normal distance (d)
between two fractures created from two zones spaced
(S) from each other in the same wellbore is found
by Eq. 5:
d =s cos e (5)
where e is well deviation from aHmin direction.
For example, where two zones to be fractured
are spaced 250 ft apart in a wellbore deviated 80°
from oHm. direction, induced fracture planes will
be sepa/J\:ed by only 43 ft. At such a short
distance from the first fracture, the stressed
field could be altered enough to make the second
fracture rotate toward ·the first fracture and
unintentional communication would take place.
Rotation of the second fracture as a function of d
is addressed in the next section.
FRACTURE ROTATION VERSUS ZONE SPACING
Stimulation of several zones along the hori-
zontal section is usually the goal of many hori-
zontal well projects. The optimum number of
fractures per horizontal section can be found based
on reservoir studies. 9 Normally, inducing a
fracture in one zone would alter the existing
stress field (magnitudes and directions) around the
fracture for a certain distance. A second fracture
created from a nearby zone would be subjected to
the new stress field, hence it might not be paral-
lel to the first fracture and could result in
communication. Severity of stress alteration
around a fracture is dependent on the created
propped width, fracture height, rock properties,
and ratios between principal stresses.
For the simple case of infinitely long 2-D
fracture (plane strain) in homogeneous, isotropic,
elastic rock, the stress perturbation at any point
around the fracture can be determined by Eq. 31 in
Ref. 6. The new state of stress at any point
around the fracture is the summation of pre-
existing stresses and stresses induced by fractur-
ing. Rotation of the second fracture can take
place when the new stress field has an orientation
different from the original one. Rotation angle of
the second fracture toward the first fracture is
calculated after determining the new direction of
aHm. . Results of fracture rotation angle versus
norffiRlized zone spacing for a specific case in
which the ratio ( K) between the horizontal
principal stresses (oHm /oHm. ) was 1.5 is shown
in Fig. 14. Three curv~ in !Re figure correspond
to three increases in the pressure in fracture (dp)
in terms of a . • More concern about fracture
rotation shoul~~ngiven for high deviation angles,
e > 60°. For example, consider the data for the
following case:
194
SPE 19720
aHmin = 2500 psi, aHmax = 3750 ft
created fracture height (h) = 200 ft
well deviation 70° from aHmin
zone spacing = 250 ft
Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
6
formation modulus =5 x 10 psi and v = 0.25.
The normal distance between prospective frac-
tures (d) can be found by Eq. 5 as 86 ft. Normal-
ized spacing (2d/h) is 0.86. Pressure increase in
a propped fracture of 0.5 in. width can be approxi-
mated by a 2-D fracture model. Using the Perkins
and Kern 10 2-D model, dp for above data is approxi-
mately 500 psi or 0.22 aHmi • Examination of Fig.
14 for normalized spacing-U1.~6 and dp = 0.25 a . ,
shows that the second fracture may rotate 10~t2-
ward the first fracture. For higher modulus rock,
or wider propped first fracture, increase in dp is
higher and consequently rotation angle of second
fracture is also higher. It is important to
mention that curves given in Fig. 13 described a
very specific case where K = 1.5. For other field
situations with different K, the rotation angle
must be recalculated for various lengths of the
second fracture. This section has shown that for
fracture rotations the magnitude of oHm should be
determined, especially in horizontai--~ells. In
most hydraulic fracture models for vertical wells,
aHmax is not usually required as an input value.
PARAMETERS CONTROLLING REORIENTATION CURVATURE
Experiments have shown that when e is not equal
to 0° or goo, induced fractures can be nonplanar
and S-shaped. Reorientation usually begins at
terminals of the pressurized interval (either
perforations or open hole) . It is reasonable to
predict that the greater the radius of reorienta-
tion curvature, the less is the resistance to fluid
flows in the reorientation region, hence less
friction and less possibility of proppant bridging.
Two factors have great control on the curvature
radius: pumping rate (Q) and ratio between maximum
to m1nlmllm horizontal principal stresses (K).
Analytical and experimental studies 3 have indicated
that radius of reorientation path increases as the
pumping rate gets higher. The effect of (K) on the
radius of the reorientation path was studied
analytically 2 for 2-D fracture (plane strain). It
was found that the radius of curvature gets smaller
as K increases or aHmax >> aHm. . Using equations
derived in Ref. 2,--Fig. 15 Ja~ constructed. As
shown in Fig. 15, as stress ratio increases, the
distance available for a reorientation path is
decreasing. In other words, the radius of curva-
ture has to become smaller. At a stress ratio of
about three or more, the reorientation process must
happen in a very short distance, or fracture would
turn goo to become normal to oHm. . Analysis shown
in Fig. 15 was done for wellborJndrilled in direc-
tion of a . . Qualitatively, the analysis can be
applied a~~Rrious well deviations using components
of stresses normal to and along the wellbore
direction. From a practical point of view, K is a
characteristic of the preexisting state of stress
and when it is > 21 it can not be easily altered.
Analysis of fracture reorientation path indicated
that knowledge of aHmax is also required.
SPE 19720 Wadood El Rabaa
SUMMARY AND CONCLUSIONS
Fracturing experiments on horizontal wells
drilled at various deviations from aHm.1 direction
were conducted. Length of perforatea.-- .Pnterval on
fracture geometry was investigated. Schematic
summary of fracture geometry versus well deviation
is shown in Fig. 16. Parameters that control
breakdown pressure, creation of multiple fractures,
fracture rotation angle and radius of reorientation
curvature were discussed. The following can be
concluded from this study.
1. Regardless of well deviation from a . direc-
tion, condensed perforations with an in~fVal Pt ~
4D are likely to allow only one major fracture to
be created.
2. Multiple fractures can be created from the same
perforation interval when p t ~ 4D and e < 75°.
However, only one main fracture is likely to extend
away from wellbore and it does not communicate with
all perforations.
3. The reorientation process when e is between 35°
and 75° creates rough fracture walls. Extra
friction can be noticed in the operating pressure.
Also, at 9 of 35° to 75°, the main induced fracture
geometry is nonplanar (S-shaped) with nonidentical
wings. The wing that extends at an acute angle
with wellbore directions is consistently shorter
(Figs. 9, 11, ·and 12).
4. Width of the initial fracture along the well-
bore might not be the maximum fracture width and
may not increase as fracture gets larger.
5. Slotted wellbore walls give the lowest break-
down and extension fracturing pressures.
6. Fractures induced from separate zones when e =
75° - 90° may communicate either by direct breaking
into each other or by fracture rotation due to
altered stresses.
7. Knowledge of a Hm in horizontal well frac-
turing is required. ax
Principal findings and observations mentioned
in this study are summarized in the diagram in Fig.
17.
Based on this study, several recommendations
are suggested to help eliminate possible problems
in stimulation of horizontal wells.
RECOMMENDATIONS
Planning Recommendations
1. The in-situ stress field must be determined
(magnitudes and directions) in the target zone.
Several techniques are available to do so. 9
2. Stresses above and below the target zone must
be determined for fracture height containment,
placement of horizontal section, and 3-D fracture
modeling. 9
3 . Wellbore diameter selected should be as large
as feasible for larger communication area.
195
7
4. Selected wellbore deviation from aHm. must be
determined with some accuracy (suggest + ~).
5. Steel goods (tubing, casing, etc.) must be
selected to withstand higher fracturing pressure.
Completion Recommendations
Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
1. Wellbore must be cased and perforation must be
condensed within a short interval. Charges that
produce deeply penetrating large diameter holes are
recommended.
2. Slotting the wellbore casing and walls is the
most effective technique to ensure low operating
pressures and excellent communication.
Treatment Recommendations
1. Injection rate must be as high as feasible to
increase the radius of reorientation curvature. A
higher rate is also recommended for transporting
sand in horizontal pipes. 11
2. Tail-in sand selected for high conductivity
near the wellbore should withstand stress greater
than aHmin. Near wellbore, initial fracture may
not be 1.nduced exactly normal to aHm. • Thus,
1
initial fracture is subjected to stress 8omponents
fi:·om both aHm. and aHm • A conservative value
for the stresl~cting on ~roppant near wellbore is
the aHmax for proppant strength selection.
3. A fracture created from a short perforation
interval can be modeled using conventional design
models assuming fracture is initiated from a point
source.
Acknowledgement
The author wishes to thank the management of
Halliburton Services for granting permission to
prepare and publish this paper. Special thanks are
extended to Mr. David Meadows for his help in the
experimental work.
Nomenclature
A.l.n area of fracture entrance at wellbore
D wellbore diameter
d normal distance between two parallel
fractures
dp delta pressure excess pressure in
fracture
h total fracture height
K ratio = (aHmax/aHmin)
maximum length of initial fracture at
the wellbore wells
L shift distance between fracture wings
s
N number of fractures accepting fluid at
the same time
length of perforated interval
8 EXPERIMENTAL STUDY OF HYDRAULIC FRACTURE GEOMETRY INITIATED FROM HORIZONTAL WELLS SPE 19720

r injection rate 8. Nolte, K. G.: "Discussion of Influence of

Geologic Discontinuities on Hydraulic Fracture
R normalized distance to crack path Propagation," Journal of Petroleum Technology,
asymptote August 1987, pp. 998-999.

s spacing between stimulation zones 9. Soliman, M. Y., Hunt, J. L., and El Rabaa, W.:
"On Fracturing Horizontal Wells," SPE 18542
fracture width

X,Y,Z coordinate system of horizontal
wellbore
0 the minimum principal horizontal stress
Hmin
the maximum principal horizontal stress
a the vertical principal stress
v
tensile strength of a formation
e the horizontal well deviation angle
from the direction of aHmin
tangential, axial and shear stresses at
a point on the wellbore walls,
respectively
Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
presented at the SPE Eastern Regional Meeting,
Charleston, WV, Nov. 1-4, 1988.
10. Perkins, T. K. and Kern, L. R.: "Widths of
Hydraulic Fractures," J. Pet. Tech. {Sept.,
1961).
11. Shah, S. N., and Lord, D. L.: "Hydraulic
Fracturing Slurry Transport in Horizontal
Pipes," SPE 18994; presented at SPE Joint Rocky
Mountain Regional/Low Permeability Reservoirs
Symposium and Exhibition, Denver, co, March
6-8, 1989.

posit~on angle on wellbore walls

measured from high side

angle span of tension zone

orientation of initial fracture with

wellbore direction

REFERENCES
1. Austin, c. E., Rose, R. E., and Schnuh, F. J.:
"Simultaneous Multiple Entry Hydraulic Fracture
Treatments of Horizontally Drilled Wells," SPE
18263 presented at the 63rd Annual Technical
Conference & Exhibition, Houston, TX, October
2-5, 1988.
2. Viola, E., and Piva, A.: "Crack Paths in Sheets
of Brittle Material," Engineering Fracture
Mechanics, Vol. 19, No. 6, pp. 1069-1084, 1984.

3. Yew, c. H. and Li, Y. : "On Fracture Design of

Deviated Wells," SP:E: Production Engineering,
pp. 428, November 1988.

4. Hsiao, c.: "A Study of Horizontal Wellbore

Failure," SPE 16927, SPE 62nd Annual Technical
Conference & Exhibition, Dallas, TX, September
27-30, 1987.
5. Daneshy, A. A.: "True and Apparent Direction of
Hydraulic Fractures," SPE 3226 presented at the
Fifth Conference on Drilling and Rock Mechan-
ics, Austin, TX, January S-6, 1970.

6. Sneddon, I. N., and Elliott, H. A.: "The

Opening of a Griffith Crack Under Internal
Pressure," QAM Journal, Vol. 4, No. 3, 1946,
pp. 262-267.

7. Daneshy, ·A. A.: "Experimental Investigation of

Hydraulic Fracturing Through Perforations,"
JPT, October 1971, pp. 1201-1206.

196
 TABLE 1

Characteristics of Set 1 Tests 8 = 0

Injection Rate in All Tests = 130 cc/min)

Applied No. of
Test Perforation Stress Induced
No. Interval {No} tesi} Fractures Observations or Fracture ShaEe

HZ-la 1/2 D (2) 1 Fracture connected perfs.

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
0 1000
v Fracture connected two perfs,
lb 1 D (4) 1
and perfs were connected along
wellbore walls.

lc 2 D (6) 0 1 As in HZ - lb.
Hmax
1000
ld 4 D (10) 0 2 One fracture was major and the
Hmin
second smaller.
200
le 6 D (12) 2 As in HZ - ld.

HZ-2a 1/2 D (2) 0 =1200 1 Fracture connected perfs.

v
2b 1 D (4) 1 Fracture connected two perfs
and perfs were connected along
wellbore walls.

2c 2 D (6) 0 1 As in HZ - 2c.
Hmax
800
2d 4 D (10) 2 One fracture was major and second
0 was very small.
Hmin
400
2e 6 D (12) 2 One inclined fracture and second
branching from fracture along the
wellbore.

Hl5 4 D 0 = 1600 3 Small initial vertical fracture

v
(open hole) oHmax 1250 along the wellbore (length = 40
(connected) normal to oHm ) connected two
800 wings normar--~ o . and
separated by openaw~~e (Fig. 6}.

* Perforations were arranged in a helical form approximately 1/2 D deep.

Distance between consecutive perforations was equal to or less than
wellbore diameter.

TABLE 2

Breakdown Pressure and Extension Pressure for

Test GrouEs HZ-1 and HZ-2

(8 = oo D = Wellbore Diameters)
Test
Group P2. 1/20 1 D 2 D 4 D 6 D
Breakdown Pressure 3850 3500 3100 2850 3600
HZ-1
Extension Pressure 1400 1600 1600 2700 2550

Breakdown Pressure 4250 3900 4100 3400 3800

HZ-2
Extension Pressure 1500 1300 1400 1900 2000

197
 TABLE 3

Characteristics of Set 2 Tests, 8 = 15°

(Perforations Were in Spiral Arrangement and Injection
Rate was 160 cc/min in Both Tests)

Perforation
Test Number Observations

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
HZ-7b 20 6 One fracture was induced with slight off-
set wings. Fracture passed by three
perforations.

HZ-Sb so 16 One fracture extended through each perfo-

ration as they spiralled around the
casing. Fracture had two overlapped
wings on the side of wellbore.

TABLE 4

Characteristics of Set 3 Tests, 8 = 35°

(Injection Rate 85 cc/min, av = aHmax = 1000 psi

aHmin = 400 psi)

Test Perforation
No. Policy Observation

HZ-3b 3D 8 perfs spiral One fracture was induced connecting all

perfs. One fracture wing had rough step
like surface.

HZ-6b 90 14 perfs, each Four fractures were induced. Three small

pair, 90° phased ones connected each consecutive perfora-
from next pair tion pairs, and one major fracture
extended from one pair of perfs. Normal
to aHmin"

TABLE 5

Characteristics of Set 4, e = 55°

Injection Rate of 80 cc/min, av = aHmax = 1000, aHmin 400 psi)

Test Perforation
No. Policy Observations

HZ-3a 3D 8 perfs spiral One nonplanar fracture was induced with

two nonidentical wings.

HZ-6a 90 14 perfs, each One main nonplanar fracture was induced

pair 90° phased with two nonidentical wings. Two short
from next pair fractures were induced from two pairs of
perfs.

198
 TABLE 6

Characteristics of Set 5, 9 = 75°

(Injection Rate 10° cc/min, ov = oHmax = 1000 psi, oHmin =.400 psi)

Test Perforation

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
No. Policy Major Observations

HZ-5a 3D 8 Spiral One nonplanar fracture was induced.

Fracture had nonidentical wings.

HZ-7a 6D 5 Pairs, goo One nonplanar fracture was induced from

Phasing 3 pair of perforations.

TABLE 7

Fracturing Pressure for Set 6, 0 = goo

Test HZ-4, Injection Rate 80 cc/min

ov oHmax = 1000 psi, oHmin = 400 psi

Entry Type
Slot Open Hole Perforated

Breakdown Pressure 2200 3000 3800

Extension Pressure g5o 1250 1600

CTHmin

Fig. 1 -Horizontal well configuration system in the in-situ stress field.

199
 BREAKDOWN PRESSURE VERSUS WELL DEVIATION
Shmin=5580 psi Shmax=7290 psi
PR=.25)
g
0
Svert is the overburden stress)
tO
..- VERTICAL STRESS
Svert=9000 psi
Svert=7500 psi

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
······
0
0
0
~--~---,----,_----r----+----~----r---~----~----~--~
0 20 40 60 80 100
Well deviation from Uhmin, deg.

Fig. 2- Breakdown pressure versus well deviation

Uv

t region

UHmin

Fig. 3 ·Initiation points and fracture orientation on openhole walls

200
 Shmin=5580 psi Shmax=7290 psi Sv=9000 psi
Lf/0 is a normalized distance between perfs
If) D - well diameter
..t

~ I

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
I I

~~~~~
~ ------+----------+----------r-:r~------~
I I
I I
I I
I I
I I
I I
I I

~~---~·r----+----------+----------r~--------~
I
I
I I

~
I
:
I

~ ,'
~~------~~~--4----------+--------.~~-+----------1
I I
I I
I I
I I
I I
I I

~~--------~----------+-----:7----r--------~ \ !
I
I
.....
~
..·
.··
N--~--------~--------~----------,_---------1
······· ..•
10 30 50 70 90
Well deviation fromUhmin ,deg. Fig. 5 ·Fracture generated in tests HZ-2 (a-e)
(PI = perforation length, D = Well dia.)

Fig. 4 • Critical distance between perforation

versus well deviation

lll lll
Q) o_
__c
u
~

TIME -

Fig. 7 - Fracture width and fracturing pressure, test H15

Fig. 6 ·Fracture shape, test H15

201
 Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
Fig. 8- Fracture shape, test HZ-Sb Fig. 9 - Fracture shape, test HZ-3b

\ - MAIN FRACTURE .

c:r~~
4
IIZ-6b 0

Fig. 10 - Multiple initial and main fractures,

Fig. 11 - Fracture shape, test HZ-3a
Test HZ-6b (slice along wellbore)

202
 Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
Fig. 12 - Fracture shape, test HZ-Sa Fig. 13 - Fractures produced in test HZ-4

ROTATION ANGLE VERSUS NORMALIZED SPACING

Second fracture rotating toward first
0 fracture at C=d 2
II')

w
0::
=>
0~4--4~~~-------+-------4------~r-------t--------1
4:
0::
u...
I-
(/)
0::
~~~-------~1.-----r------~

0 2 3 4 6
NORMALIZED SPACING , 2d/h

Fig. 14 - Fracture rotation angle

versus normalized spacing

203
 REORIENTATION DISTANCE VERSUS STRESS RATIO
K=ratio of maximum to minimum stress
N -
R- normalized reorientation distance

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/89SPE/All-89SPE/SPE-19720-MS/2014907/spe-19720-ms.pdf/1 by PT Medco E&P Indonesia, Martin Rylance on 02 December 2022
0
UHmax 55

<1Hmin ~
t
)
~ """" 6
t-R-
I{')

ci
\
_/
"'
,/

0
loo-.

2 3
K
4 5
1-------t--- 15:
Fig. 15 - Reorientation normalized distance
versus stress ratio (K)
~~~ot1-~
openhole long ~ short ~
f-*-
0'Hmin

Fig. 16 - Top view of fractures induced at various wellbore

deviation angles

·MODERATE BKDN PRESSURE

·HIGH FRICTION (rough walls)
• LIMITED COMMUNICATION
·NON-IDENTICAL WINGS
• MULTIPLE FRACTURES

• HIGH BKDN PRESSURE

• MULTIPLE FRACTURES
• LIMITED COMMUNICATION
• BHTP a:vN

( BKDN = BREAKDOWN)
~-------------------•- Hmin
0

Fig. 17 - Principal findings of horizontal

well fracturing

204