SPE 103232

A Field Study Optimizing Completion Strategies for Fracture Initiation in Barnett Shale
Horizontal Wells
A.A. Ketter, Devon Energy; J.L. Daniels, Schlumberger; J.R. Heinze, Devon Energy; and G. Waters, Schlumberger

Copyright 2006, Society of Petroleum Engineers

This paper was prepared for presentation at the 2006 SPE Annual Technical Conference and
Exhibition held in San Antonio, Texas, U.S.A., 2427 September 2006.
This paper was selected for presentation 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. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than
300 words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
The Barnett shale is an unconventional gas reservoir that is
currently estimated to extend over 54,000 square miles. In an
effort to improve well economics and to reduce the number of
surface locations in populated areas, there has been a rapid
increase in the number of horizontal wells being drilled and
completed. With this change in development strategy,
operators and service companies alike have had to search for
innovative solutions to overcome challenges faced in
horizontal completions.
Inefficient fracture initiation is the largest reoccurring
problem seen when completing horizontal Barnett shale wells.
These difficulties are manifested as high fracture initiation and
propagation pressures, which lead to low injection rates and
high treating pressures. These losses reduce the efficiency of
proppant placement and stimulation. As drilling activity has
increased over the past couple of years, fracture initiation
problems have represented a substantial source of expense and
downtime.
This field study examines 256 horizontal Barnett shale
wells in an effort to identify the causes of these near-wellbore
issues and to offer corrective solutions for future completions.
The goal of this study is to recommend an optimized
completion strategy to minimize these near-wellbore
problems, increase stimulation coverage and decrease
unplanned completion expenses.
In 2005, 19% of the stages in horizontal wells examined
encountered near-wellbore difficulties. This field study
inspects the major contributors to fracture initiation,
specifically focusing on cemented versus uncemented laterals,
horizontal stress anisotropy, perforation strategy, cementing
strategy and stimulation design.
The paper offers statistics on which changes have had the
greatest effect on stimulation placement. These problems can
cost operators upwards of an additional 25% per stage. Using

these optimized strategies has reduced the number of stages

where fracture initiation difficulties have been encountered by
74%.
Introduction
The Barnett shale is a Mississippian-age marine shelf deposit
that unconformably lies on the Ordovician-age Viola
Limestone/Ellenburger group and is conformably overlain by
the Pennsylvanian-age Marble Falls Limestone. The Barnett
shale is located within the Forth Worth basin and the focus of
our study will concentrate on wells within Denton, Wise, and
Tarrant counties, the core area. The Barnett in the core area
ranges from 300 to 500 ft in thickness. Permeabilities range
from 0.00007 to 0.0005 md with porosities ranging from 3 to
5%. The Barnett is believed to be its own source rock and is
abnormally pressured in this area. Commercial production is
only achieved through hydraulic fracture treatments.
Before 1997, Barnett wells were completed with massive
hydraulic fracture treatments consisting of crosslinked gelled
fluids and large amounts of proppant. Because of difficulties
with effectively cleaning up fracture damage from the
crosslinked gel and the high cost of these massive stimulation
treatments, the wells were not as economical as desired. In
1997, large volume, high rate slickwater fracture stimulation
treatments were sought as a less expensive alternative. While
well performance was not drastically increased by using
slickwater, completion costs were reduced by approximately
65%. In 2002, horizontal wells were experimented with in an
effort to increase the wellbores exposure to the reservoir. The
results of the first horizontal wells compared to vertical wells
were three times the estimated ultimate recovery (EUR) at
twice the well cost. Horizontal wells offered an economic
solution to areas outside the core and reduced the number of
surface locations needed near populated areas.
In the early stages of horizontal completions the wells
were divided equally between uncemented and cemented
laterals. Shorter laterals that required single stimulations were
uncemented and cemented laterals were implemented when
the stimulation design required multiple stages because of an
increased lateral length. Composite bridge plugs were used
for zonal isolation. Fractures in uncemented laterals are prone
to grow in such a way that unstimulated volumes, or gaps
are often left in the reservoir, which can equate to a smaller
overall fracture area and reduced productivity1, this is
illustrated in Fig. 1.

SPE 103232

Fig 1. Microseismic of an uncemented horizontal well displaying the

unstimulated gaps. The five perforation clusters stimulated are identified
with the orange circles.

inefficient fracture initiation while Fig. 4 displays an efficient

fracture initiation and propagation.
Fig 3. This hydraulic fracture had inefficient fracture initiation. Notice that a
maximum pressure of 6000 psi was maintained throughout the completion and
that the desired rate of 80bbl/min or above was never achieved, eliminating
the possibility of placing the job as designed.

2000.00
1800.00
1600.00
1400.00
1200.00
1000.00
800.00
600.00
400.00

Observation Well

200.00
0.00

South-North (ft)

-200.00
-400.00
-600.00
-800.00
-1000.00
-1200.00
-1400.00
-1600.00
-1800.00
-2000.00

Well 1H

-2200.00
-2400.00
-2600.00
-2800.00
-3000.00
-3200.00
-3400.00
-3600.00
-3800.00

3000

2800

2600

2400

2200

2000

1800

1600

1400

800

1200

600

1000

400

200

-200

-400

-600

-800

-1000

-1200

-1400

-1600

-1800

-2000

-2200

-2400

-2600

-2800

-3000

-4000.00

West-East (ft)

Fig 4. This hydraulic fracture had sufficient rate and maximum pressure was
never a restriction. The designed proppant schedule was successfully placed.

As drilling progressed outside the core area and acreage

was readily available to accommodate longer laterals, the
number of cemented horizontals surpassed the number of
uncemented horizontals. However, the increase in cemented
laterals also yielded a higher rate of inefficient fracture
initiation than seen in uncemented laterals. In 2005, more
than 1 in 4 cemented horizontals experienced fracture
initiation problems compared to 1 in 25 for uncemented
laterals (see Fig 2). This overwhelming rate led to this
optimized completion strategy.
Fig 2. This graph shows the total number of cemented and uncemented wells
in the core area since 2002. The number of fracture initiation issues per type
of well is also identified.
Total Wells v Problem Wells
80
70
60
50

Uncemented
Cemented

40

Problem Uncemented
Problem Cemented

30
20
10
0
2002

2003

2004

2005

2006

Inefficient fracture initiation can be defined as the lack of

sufficient fluid injection rates resulting in the inability to pump
designed proppant concentrations, delivering an ineffective
fracture network. The stimulation job will typically be
characterized by high pumping pressures and occasionally
abnormal fracture gradients. Fig. 3 displays an example of an

Inefficient fracture initiation can be related to cement

design, perforation phasing, perforating lengths, cluster
spacing, formation stresses, and pad design for the stimulation
treatment. The cost incurred because of these problems is
quite significant, representing an additional 25% of a stages
total completion cost. The cost of an improperly placed stage
can also be detrimental to the productivity of the well by
reducing the overall fracture area. Each failure also provides a
logistic problem by setting the frac schedule back a day or
more, reducing the efficiency of the completion program. The
goal of this case study was to recommend an optimized
completion strategy that would reduce the completion cost of
cemented horizontals, increase stimulation coverage, and
accommodate an aggressive drilling programs need to
maintain a trouble free frac schedule.
The case study was divided into two distinct segments.
First the problem assessment segment, which evaluated 154
horizontal wells, 31 of which displayed inefficient fracture

SPE 103232

initiation issues. Correlations were developed to using field

data to recognize probable causes and possible solutions to
overcome these challenges. The second segment included 102
horizontals in which these new strategies were implemented.
This paper will discuss how fracture initiation problems were
reduced to 4.7% from 19.1%, a 74% improvement.
Barnett Stress Variability
A common misconception when characterizing the Barnett
shale is to classify and treat the rock along the wellbore as
homogenous. Fig. 5 shows five sets of perforations along a
horizontal wellbore. The well was fractured in four stages
with the first stage having two perforation clusters. A prefrac
Instantaneous Shut in Pressure (ISIP) was taken for all four
stages to illustrate how stress varies along the horizontal.
What should be observed from this figure is that the ISIPs
vary from the toe section at 0.75psi/ft to 0.63 psi/ft in the heel
section and that there is no consistent trend in the magnitude
of the ISIP from toe to heel. As the state of stress varies along
a wellbore, so will the types of problems seen with fracture
initiation, job placement and optimum job design.
Fig. 5 The ISIPs were noticeably different with a 0.12 psi/ft range.

fracture lengths and wider fracture fairways generated, similar

to the microseismic events seen above. Fracture initiation
pressures in this type of stress regime would be low.
Fig. 7 High stress anisotropy
Long, Narrow Fracture Fairway

Fig. 7 displays only transverse fractures along this section

of the horizontal well. This type of behavior is seen in areas
with high horizontal stress anisotropy. In this type of stress
environment, we would expect to see longer and narrower
fracture fairways generated, similar to the microseismic events
seen above. Fracture initiation pressures in this type of stress
regime would be moderate. High horizontal stress anisotropy
can result in high near-wellbore friction pressure losses and
difficulty in placing proppant. 8
Fig. 8 High stress
No Fractures

Image logs have recently been used to qualitatively

determine the states of stress in the wellbore 2. Faults, natural
fractures, and hydraulic fractures can be seen in these images
and placement of perforations can be optimized with this
information. These images can also provide information about
the stress anisotropy seen along the wellbore. Figures 6, 7,
and 8 show several images along a Barnett horizontal and
provide a discussion of their interpretation. These wells are all
drilled in the direction of minimum stress so that hydraulic
fractures normal to the borehole are achieved, a more
thorough discussion can be found in reference 2 (SPE 103202)
Fig. 6 Low stress anisotropy
Short, Wide Fracture Fairway

Fig. 6 displays both longitudinal and transverse fractures

along this section of the horizontal well. This type of behavior
is seen in areas with low horizontal stress anisotropy. In this
type of stress environment, we would expect to see shorter

Fig. 8 displays no fractures along this section of the

horizontal well. This type of behavior is seen in areas of high
stress. In this type of stress environment, we would expect
that it would be difficult to initiate a fracture and the fracture
initiation pressures would be high.
Before image logs were available, perforation strategy
was to evenly space the perforation clusters along the
wellbore, at the risk of locating one of these clusters in a high
stress area. Initiating a fracture in a high stress area can lead
to initiation pressures of 50% or greater than in a low stress
environment. An analysis of multiple image logs in the
Barnett indicates areas of high stress along 15% to 25 % of the
wellbore relative to the rest of the lateral. Using image logs
has allowed us to qualify stress variability along a wellbore,
and enabled us to high-grade perforation locations so we can
differentiate between high-stress areas and low-stress areas.
With multiple stress environments and multiple clusters per
stage, an understanding of the stress distribution is imperative
to optimizing the fracturing area.
Because image logs can also indicate where faults are
located along a wellbore, they can again be used in perforation
strategy to avoid perforating near them. This prevents the
fracture from extending into water-prone areas.

Perforations
Sufficient cluster spacing will prevent individual fractures
from linking up and allow for multiple parallel fractures to
extend without restricting fracture growth.3 In past
microseismic studies, stress shadows have been shown to have
both negative and positive impacts. When perforation cluster
spacing is too close the stress shadow can restrict growth in
the middle cluster and disproportionately increase growth at
the heel and toe perforations. However, if two perforation
clusters are properly spaced and simultaneously competing,
the fracture growth is enhanced in the orthogonal direction.1
Because fracture height is the smallest fracture dimension
compared to fracture length, closure stress is greatly
influenced by the fracture height and decreases as the distance
between fractures increases.4
Fracture height is areadependent in the Barnett, but typically ranges from 300 to 400
ft. After multiple microseismic studies, the optimal cluster
spacing to reduce fracture interference is at a distance greater
than 1.5 times the fracture height.1 The strategy implemented
for the case study was to reduce the number of clusters per
stage from three to one or two. Prior to the study, the average
horizontal well had 2.7 frac stages and after transitioning to
reduced clusters, the number of stages per well has climbed to
3.2. Prior spacing was maintained at 1.5 times the frac height.
For the Barnett, the overburden is the greatest principal
stress. Using Hsaios derived equations, it can be determined
that the tensile stress is the largest when = 0o/180o.
Therefore a fracture will initiate at the top and bottom of the
wellbore.5 This is confirmed by the image logs, which show
drilling induced fractures (breakout occurs90 degrees from the
induced fractures and is associated with high stresses and
underballanced drilling) at both the top and bottom of the hole.
The optimal perforation design for this stress scenario would
be to orient 0o/180o phasing to preferentially align the
perforations with the preferred fracture plane. The other
viable option is 60o phasing to maximize the probability of
perforations being placed near the top and bottom of the hole.
Cementing strategies discussed in the next section can be used
to minimize fracturing difficulties when perforations not
aligned with the top and bottom of the hole are active. The
strategy implemented was to minimize inconsistency in
perforation phasing by shooting only oriented 0o/180o or 60o
phasing.
To reduce the probability of creating multiple competing
fractures, the perforation cluster length should be less than
four times the wellbore diameter. If the cluster length is less
than four times the wellbore diameter, a single dominant
fracture will be created is more likely to form.6 A
characteristic of competing multiple fractures will be high
pressure with low injection rates. Before the case study, 71%
of the fracture initiation issues had a perforation length greater
than four times the wellbore diameter, generally from 5 to 10
ft. The new implemented strategy was to keep perforation
cluster spacing to less than 4 ft. A comparison of the ratio of
average treating pressure and average treating rate of
fracturing treatments on wells where shorter clusters were
used, shows a 14% decrease in psi/bbl/min requirements. This
distance will allow an appropriate number of holes for limited
entry to be effective while also limiting cluster length.
Limited entry requirements for Barnett horizontals are

SPE 103232

designed to allow 2 bbl/min per perforation and a total

perforation differential pressure of greater than 500 psi.
Cementing
With cemented laterals dominating the statistics for fracture
initiations issues and the need for stage isolation becoming
more important as the lateral length increased, the cementing
strategy was closely examined to identify and address the
issues causing the stimulation troubles.
An acid soluble
cement system (ASC) was sought as a viable option to
conventional cement slurries. An ASC slurry has an increased
amount of calcium carbonate in the slurry. When in contact
with hydrochloric acid the cement will dissolve based on the
solubility and contact time. The solubility capacity is a
function of the calcium carbonate ratio and the contact time.
This can be controlled with acid volume and pump rate.
Conventional cements are also soluble in acid, but not to the
degree of systems containing a large percentage of calcium
carbonate. The main properties of the two slurries are detailed
in Table 1.
Table 1. This table displays the differences in properties between a
conventional slurry and an acid soluble slurry.
Conventional Cement
ASC
(50/50 Poz:Class H)
Weight
14.2 ppg
15.0 ppg
Yield
1.26 ft3/sk
2.51 ft3/sk
Water
5.66 gal/sk
10.98 gal/sk
Compressive Strength
2700 psi
550 psi
(72 hours)
Solubility
25%
92%

The main concern with using an ASC blend was

maintaining sufficient isolation by placing fractures at the
desired distances apart, therefore preventing gaps while
stimulating. The main objective when cementing laterals is to
provide annular isolation between the perforation clusters.
This will allow for the creation of independent hydraulic
fractures at each perforation.
By stragically placing
perforations, stimulation gaps can be avoided along the lateral.
ASC has the advantage of reducing the annular restriction
common when perforations are not aligned with the top and
bottom of the hole. When designing for limited entry with 60
degree phased perforations, fluid will exit the wellbore on the
side of the hole. This fluid must follow a tortuous path to get
to the point of the wellbore where the hydraulic fracture
initiates. Cement in the annulus can produce a choke effect
that restricts fluid flow. Dissolving the cement provides
conditions approaching open hole, while still providing
annular isolation between perforation clusters. As acid hits
the perforation clusters and the cement dissolves, surface
treating pressures are also reduced Fig. 4. A comparison of
the ratio of average treating pressure and average treating rate
of fracturing treatments on wells cemented using a standard
system and those using an ASC system show a 17% decrease
in psi/bbl/min requirements.
Chemical tracer logs and cement bond logs have been
used to validate proper zonal isolation and the lengths of the
lateral that are affected by contact with hydrochloric acid. A
chemical tracer log shown in Fig. 9 clearly displays that
sufficient isolation was attained and that ASC was dissolved

SPE 103232

behind the casing for an approximate length of 150 feet. The

largest Gamma Ray readings came from within a 30 ft range
of the perforations, giving further confidence that fractures can
be placed at desired depths along the lateral.
Fig. 9 This post acid and hydraulic fracture chemical tracer log on a single
cluster ASC horizontal well shows acceptable zonal isolation.

A cement bond log was run before and after acidizing and
is shown in Fig. 10 The purpose of these logs was to verify the
quality of the cement job in the lateral and to determine the
degree of annular isolation after acidizing. The preacidize log
showed a channel of water along the low side of the hole. It is
in these sections where the annular space was small enough
that cement was not able to bond well to the casing. Sections
of high-quality cement surrounding the entire wellbore could
be found when the well deviated from the directional plan and
the formation provided adequate centralization. Echo rings
could be seen on the low side of the hole when the casing was
lying on bottom. Barnett horizontals typically have no
centralization along the lateral because of the increasing
number of problems encountered while running casing to
bottom. The preacidize log verifies that quality horizontal
cementing is difficult without proper centralization. Further
study is required to determine the effects of improper
centralization and cement slurry design, as both can be
optimized further.
Fig. 10 This log section shows the quality of the bond between the cement and
casing before acidizing. The light blue colors represent areas where no
cement is present along the bottom of the hole. A similar log was run after
acidizing, but showed very little change in the bond between casing and
cement.

After perforating two clusters 397 feet apart, the

perforations were acidized with 3,000 gallons of 15%
hydrochloric acid with non-emulsifier and corrosion inhibitor

additives at an average rate of 4 bbl/min. The post-acidize log

was run under pressure similar to the maximum pressure
experienced during the initial acid job to minimize
mircroannular effects. The post-acidizing log shows that a
channel has been created on the top of the hole for an
approximate length of 80 feet at the first perforation cluster,
but the second perforation cluster experienced very little effect
from acidizing. Cement slurry characteristics, specifically free
water and can be optimized to provide better annular isolation.
In this example, the acid was able to travel along the top of the
hole, reacting with the cement there. At low rates, the acid is
able to cover more of the cement along the lateral. However,
if higher rates are used, more annular friction pressure is
created and less cement along the lateral is dissolved. To
increase acid coverage without increasing the rate to maintain
optimum contact time, ball sealers have been added as an
effective diversion technique. Fig 4 shows a graph of a
typical fracture treatment with pressure increasing as balls
diverted acid to other perforations.
Frac Design Changes
Acid with biodegradable ball sealers has been used in
conjunction with ASC to reduce the amount of near-wellbore
pressure drop between the pipe and formation. Typical acid
volumes are between 1000 and 2000 gal per cluster
stimulated.
Some other techniques that have been employed in the
pad stages of fracs are the use of crosslinked gels and 100
mesh sand. Initiating a hydraulic fracture with crosslinked
gel increases the fracture width and subsequent rate of treating
pressure decay. This increased width has improved the ability
to place proppant through the near-wellbore region. Using
100 mesh sand in the pad stage has assisted with controlling
the leakoff into a complex fracture system. These may be
closely spaced hydraulic fractures or natural fractures that
have activated during the treatment. Their presence can
reduce the average fracture width making proppant placement
more difficult. 100 mesh sand can bridge many of these
fractures, so that fewer, wider, dominant fractures are present
near the wellbore.
Again using the image logs, where we see low stress
anisotropy, there is a greater likelihood that a wider fracture
fairway will be created. Using this information, pad designs
were changed to incorporate both traditional stimulation
techniques to get the wide fairway and the crosslinked fluid to
reduce the near-wellbore fracture complexity.
Conclusions
A two-part case study was undertaken to determine the causes
of stimulation difficulties in cemented horizontal wells. The
first part of the study examined 31 horizontal wells that
displayed difficulties with inefficient fracture initiation. The
second part of the study examined 102 horizontal wells
comprising 300 frac stages that used the best practices
discovered in the first part of the study. The strategies
implemented were successful at decreasing the probability of
having an inefficient fracture initiation from 19.1% of the
stages to the current rate of 4.7% of the stages: a 74%
reduction. The conclusions of the study are:

SPE 103232

1.

2.

3.

4.

5.

6.

7.

8.

9.

A Barnett shale horizontal well cannot be assumed

to be located in a homogenous stress environment.
ISIPs taken at multiple points in the lateral have
shown large variances in fracture gradients, with no
trend in the magnitude of ISIP from toe to heel.
Image logs have been used to successfully predict
horizontal stress anisotropy and can be used to
prevent inefficient fracture initiation by selectively
locating perforations.
Perforation clusters should be appropriately placed
to reduce stress shadow effects. By reducing the
number of clusters per stage, stress interference was
minimized and reduced the possibility of having
improper fracture propagation. The number of
clusters has been reduced to two or less per stage,
increasing the number of stages per well from 2.7 to
3.2.
Perforations should be oriented 0o/180o phasing to
align with the preferred fracture plane. Another
option is 60o phasing when used in conjunction with
ASC. Both perforating strategies have proved
effective.
Perforation cluster lengths should be no greater than
four wellbore diameters to prevent the creation of
competing multiple fractures. Before the case study,
71% of the problem wells had a perforation length
greater than four wellbore diameters.
The ASC in the lateral is dissolved based on
solubility and contact time. Logs showed that the
cement was dissolved less than 80 ft away from the
perforations with most of the fracture activity
propagating within 30 ft of the cluster.
Using ASC and/or shorter perforation clusters have
decreased the ratio between average treating
pressures and average treating rates during the
fracturing treatments 15%. This decrease is directly
related to the hydraulic horsepower required to
stimulate the wells.
Adding 100 mesh sand to the pad stage is useful in
controlling the leakoff into the complex nearwellbore fractures that may be present.
Initiating the fracture with crosslinked gel has been
beneficial in increasing the rate of pressure decay
and by providing sufficient width to allow for
placement of the desired proppant concentrations.

Acknowledgements
The authors of this paper would like to thank the management
of Devon Energy and Schlumberger Oilfield Services for the
permission to present this data and the opportunity to pursue
the development of horizontal completion processes based on
its results. We also wish to thank Larry Coggins and Christine
Purdy for their effort in collecting and maintaining the well
data.
References
1.

2.

3.

4.
5.

6.

Fisher, M.K: Optimizing Horizontal Completion Techniques in

the Barnett Shale Using Microseismic Fracture Mapping, paper
SPE 90051 presented at the 2004 SPE Annual Technical
Conference and Exhibition, Houston, Texas, 2629 September.
Waters, G, et al.: Use of Horizontal Well Image Tools to
Optimize Barnett Shale Reservoir Exploitation, paper SPE
103202 presented at the 2006 SPE Annual Technical Conference
and Exhibition, San Antonio, Texas, 2427 September.
Yew, C.H. et al.: On Fracture Design of Deviated Wells,
paper SPE 19722 presented at the 1989 Annual Technical
Conference and Exhibition, San Antonio, Texas, 811 October.
Warpinski, N.R: Altered-Stress Fracturing, JPT (September
1989) 990997.
Hsiao, C.: A Study of Horizontal Wellbore Failure, paper SPE
16927 presented at the 1987 Annual Technical Conference and
Exhibition, Dallas, Texas, 2730 September.
El Rabaa, W.: Experimental Study of Hydraulic Fracture
Geometry Initiated from Horizontal Wells, paper SPE 19720
presented at the 1989 Annual Technical Conference and
Exhibition, San Antonio, Texas, 811 October.