SPE-184868-MS A Causation Investigation For Observed Casing Failures Occurring During Fracturing Operations
SPE-184868-MS A Causation Investigation For Observed Casing Failures Occurring During Fracturing Operations
SPE-184868-MS A Causation Investigation For Observed Casing Failures Occurring During Fracturing Operations
N. J. Adams, Neal Adams Services; R. F. Mitchell, Well Complete; A. W. Eustes and J. H. Sampaio, Colorado
School of Mines; A. O. Antonio, Neal Adams Services
This paper was prepared for presentation at the SPE Hydraulic Fracturing Technology Conference and Exhibition held in The Woodlands, Texas, USA, 24-26 January
2017.
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 have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper 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 SPE copyright.
Abstract
The objective of this paper is to present findings from a causation investigation of casing failures occurring
during fracturing operations. The study analyzed case histories and examined post-failure laboratory results.
Outcomes from casing failures include blowouts, pollution, injuries/fatalities, and loss of the well with
associated costs. This work has not been previously published nor is this type of work available in the public
domain literature.
This study cataloged pertinent data from fourteen (14) land-based case histories of casing failures while
fracturing. The case histories were analyzed with the use of photographic evidence of recovered casing,
well reports, fracturing treatment data, drilling records, and laboratory testing of recovered casing when
available. Failure causes were identified. Preliminary guidelines are provided to avoid casing failures and
mitigate the damages. Recovery guidelines are presented.
A model was developed to evaluate axial loads under several considerations associated with fracturing.
The model accounted for the casing's in-air load, buoyancy effects from wellbore fluids, piston effect,
bending, erosion, and temperature decreases. The combined loads were compared to the casing's yield
strength.
The failures were not systemic but included cracking related to fatigue, hydrogen and sulfide stress
embrittlement, and erosion at ERW weld lines. The failures were observed at various well depths, both in the
cemented and uncemented hole sections. Results address common casing sizes and couplings involved with
each failure, weight and grade, failure location and causation identification of other factors contributing to
the failure. A goal of this investigation and on-going work is to develop a thorough understanding of casing
failures and the myriad of contributing factors in order to develop a comprehensive predictive model for
land and offshore fracturing operations.
Overview
Fracture/stimulation for enhanced reservoir production has become commonplace. One source (EPA 2015)
estimated that 90% of new US land-based wells are hydraulically fractured. The combination of multi-
stage fracturing and horizontal drilling technology has opened new frontiers, particularly for oil and/or gas
2 SPE-184868-MS
shales. As with many advances in science and technology, hydraulic fracturing has presented new problems
and challenges that must be addressed. One of these challenges involves casing failures occurring during
fracturing operations.
Incidents of casing failures occurring during fracture stimulation operations are increasing at an alarming
rate. Observed outcomes from land-based failure incidents include injuries and fatalities, blowouts, oil
pollution and environmental damage, loss of the wellbore and portions of the reservoir that had been
fractured. Costs often reach millions of dollars from categories that have included blowout control, pollution
cleanup and remediation, post-incident wellbore diagnostics, fishing, reestablishment of wellbore integrity
where possible, partial or complete loss of the wellbore, plug and abandonment, and re-drills. These
outcomes affect environmental issues with the public scrutiny associated with them. The impact on public
perception affects the social license to operate so critical in many US locations. These outcomes and their
related costs and environmental impacts warrant further study.
Histories of fourteen (14) land-based casing failures were collected. Public domain information and
private files were used to augment the case histories. Failures unrelated to fracturing have been excluded.
Also excluded are those incidents reported as casing failures where the pertinent issue involved the cement
and its related integrity. The term casing failure will be used herein to address those fracturing-related
failures that occur in the pipe body, the coupling or at the pipe body-coupling interface.
Literature Review
The issue of casing failures has been widely recognized and reported in the literature for decades. In the
mid-1950s, Texter (1955) covered casing and tubing problems. The publication indicated that casing design
should address tension, collapse and burst. At that time, frequency in which wells were fractured was
inconsequential such that Texter didn't have an opportunity to observed fracture-related casing failures. It is
worthwhile to note that current casing design philosophies show little advancement since the 1950s despite
the industry's awareness of many factors that should be addressed.
Texter introduced the failure condition of the last engaged thread failure. Causes were reported as shock
loading or vibration fatigue. This observation is consistent with the findings identified in this paper although
Texter indicated that the magnitude of the problem was small. Texter's observations on the last engaged
thread failure follows:
1. All failures occurred above the top of the cement;
2. Straighter holes produced more failures than deviated wells; and
3. The failures involved various grades of pipe and both seamless and ERW types.
Clark (1987) addressed casing design by including considerations for the piston effect, decreasing
temperatures, ballooning and bending. He noted that the most common error in casing design under fracture
treatments was casing selection on the basis of burst alone. Design and field guidelines were generated that
proved successful in avoiding failures in the fracture treatment of 26 wells. Clark commented that the burst
and tensile ratings for casing could be substantially reduced by partial erosion although his work didn't
address the effects from erosion. Buckling was not considered.
Payne (1993) explored fatigue failures of in-service API 8-round casing while drilling through casing.
This mode of failure was described as unusual, unexpected and difficult to design for, an observation still
valid in 2016. Payne recognized the potential for well control issues if such a failure should occur.
Four detailed field cases of 8-round casing failures were presented. The findings, conclusions and
recommendations included the following:
1. Minimize the uncemented interval;
2. Use API guidelines for all mill-end make-ups;
SPE-184868-MS 3
Figure 1A fatigue crack at the wellhead-casing interface resulted in a blowout. The fatigue was vibration-induced.
OBrien (1984) used two case histories to describe buckled casing failures in an uncemented interval
between two cemented segments. This scenario is applicable while fracturing uncemented or poorly
cemented casing in the lateral section of horizontal wells. OBrien concluded that effective cementing was
the only method available to maintain casing in a stable condition which should avoid buckling.
Land-based fracturing candidates and field operations typically have shared failure characteristics
pertinent to this investigation. Most unconventional wells encounter normal or slightly sub normal formation
pressures although a few over-pressured exceptions exist. Horizontal drilling through oil/gas shales appears
to be more common than with vertical wells. Length of the lateral section often exceeds the vertical depth.
True vertical depths rarely exceed 10,000 ft. Directional drilling through the build section often created dog-
legs exceeding 15-20, values that were considered unacceptable a few years ago. Cementing faced several
challenges that included (1) cement displacement efficiency in hole/casing sizes observed with the recent
failures and (2) incidents of loss of circulation experienced while drilling, running casing or cementing.
Casual conversations with numerous sources indicated a wide-spread awareness of instances involving
casing failures. As seemingly abundant as they may be, information on most incidents is not available. The
information is under the control of the operator desiring to protect its assets, an insurance company, claims
adjustors, or involved in litigation. Two representative case histories were summarized as a means to present
and discuss characteristics associated with common failures (Appendix A).
Data items believed to be pertinent to each failure were collected and summarized (Table 1). Several
observations were identified as follows:
1. Failures occurred in vertical and horizontal wells;
2. For horizontal wells, failures were observed in the vertical (71.4%) and build (28.6%) sections;
3. Shallow failures above top-of-cement (64.3%);
4. P110 grade pipe was used in 85.7% of the failures;
5. High collapse pressure rated casing (HCP) was common to the failures (85.7%);
6. The typical failure involved the coupling and/or the last engaged thread on either end of the coupling
or the pipe tube within 1-2 ft of the coupling;
7. Testing reports identified fatigue in most failures;
8. ERW casing was used on 92.8% of the case histories;
9. Erosion on the ERW weld line (14.3%); and
10. Surface pressures typically exceeded 6,000 psi with pump rates in the range of 50-90 bbl/min at the
time of failure.
8 V 5.5/17.0 P110-HCP-ERW LTC 3,410 Y Pressured-up to 8,000 psi to open sleeve for
stage 4, wellhead blew off, flowed oil to the
cellar, coupling failure at 1,430 ft, fracture
origin coincided with tong bite marks.
9 D 5.5/20.0 L80 ERW LTC 560 Y Coupling pull-out at 560 ft, stuck while
running casing, pipe did not meet API specs
for yield and tensile strength.
10 H 5.5/17.0 P110-HCP-ERW LTC 4,096 Y Brittle failure in pipe body at 4,096 ft,
failure located 10 ft from coupling, API
yield strength is 140,000 psi but failure
had 175,000 psi, exceeded API hardness,
untempered martensite, HSCf cracks
observed.
11 H 5.5/20.0 P110-HCP-SMLS DQX* 8,555 N in. hole at 8,555 ft, casing not retrieved.
12 H 7.0/ P110-HCP-ERW 100 Y 7 in. casing parted at 100 ft, surface casing
also parted just below wellhead, casing-
wellhead ejected, minor injuries.
13 H 7.0/29.0 P110-HCP-ERW GB-CD* Failed during 1st or 2nd stage, well flowed
oil, gas and water.
14 H 7.0/26.0 P110-HCP-ERW LTC 2,600 Y Coupling failure at 2,600 ft.
Key: = Unknown
a
Key: V = Vertical, H = Horizontal, D = Directional
b
OD-Outer Diameter; Wt.-Weight
c
Manufacturing Process; Key: HCP-High Collapse Pressure; ERW-Electric Resistance Weld; SMLS-Seamless
d
Coupling; Key:
*
= Premium coupling
e
Key: Y = Yes, N = No, = Unknown
f
Hydrogen Stress Cracking
Fluid properties include (1) temperature, (2) the type, size and quantity of proppants and (3) the amount,
type and concentration of acid. Casing and coupling geometries affect internal fluid velocities, pressures
and erosion. Hole configurations include the well's profile, dog-leg severity at depth, hole-to-casing sizes,
cement height in the annulus, environmental effects from annular fluids and corrosion. The amount and
severity of buckling, ballooning, and bending influences pipe stresses. Fatigue considerations include
vibrations, dampening effects of the wellbore and cyclic operations. Some but not all of these parameters
will be discussed briefly. Metallurgical laboratory testing on failed specimens often reported the failure
cause from one or more of these conditions.
An effective failure analysis requires that stresses associated with each of the characteristics and
conditions must be quantified, where possible, and subsequently combined. Calculation procedures and
equations currently exist to determine stresses from some factors such as ballooning, bending, buckling
and temperature changes. Other conditions such as the effect of vibrations and fatigue with its related
contributors are more difficult or impossible to quantify with existing technology. The effect of vibrations
will not be presented other than its contributions to cyclical loading. Discussions on these stress sources
and other parameters is provided.
Pump rate and pressure. Pump rates associated with hydraulic fracturing are greater than with routine
drilling operations. Common rates are in the range of 70-100 bbl/min. Vibrations occur and affect the surface
equipment, the wellhead-casing interface and the shallow sections of the casing string. These rates create
high fluid velocities inside the casing. Fig. 2 illustrates velocities for typically used casing strings and pump
rates. The impact of high velocities on erosion is not well understood.
Figure 2Fluid velocities associated with various pump rates and casing sizes
commonly used for fracturing operations. Velocities are shown in ft/sec (ft/s). A pump
rate of 86.5 bbl/min (bpm) (Fig A-3) resulted in fluid velocities greater than 90 ft/sec.
A fracture treatment plot of pump rates and pressures is shown in Figure A.3 (Appendix A). The 15th
stage was being pumped when a failure occurred. Appendix A should be consulted for details.
Temperature. Fracturing exposes the casing to dynamic temperature loads. Fracturing fluids, often in
surface tanks where the atmospheric temperature may be cold, invariably lower than downhole formation
temperatures. They are pumped at rates and velocities that can displace the fracturing string in 60-75
seconds. Changes from the fluid's cooling effect can cause substantial stress increases in the casing. Also,
SPE-184868-MS 7
rate changes and shut-down periods results in a dynamic temperature scenario that allows cooling and
heating phases observed during fracturing operations.
Thermal shock may damage casing. It occurs when thermal gradients cause different parts of casing to
expand or contract at different rates. Evaluating a failure for possible thermal shock is a challenge because
the generated crack is similar in appearance to fatigue or brittle cracks. Research on thermal shock in casing
strings is sparse.
Erosion. An undesired byproduct from fracturing operations is erosion. Fracturing fluids containing
proppants, usually some type and size of sand, are pumped at high rates and fluid velocities. As previously
shown in Fig. 2, velocities of 70-100 ft/sec are the expected norms. Factors affecting erosion include the
roughness of the inner tube wall, hardness and angularity of the proppant, brittle vs ductile pipe, fluid
velocity, straightness of the tube and many others. The complexities of the erosion process under stimulation
conditions precluded further investigation at this time.
Two case history failures were caused by erosion in similar circumstances. ERW pipe was being used. The
seam on the internal weld line was attacked by the proppant (Fig. 3). The erosion-generated hole grew until
it had penetrated the tube's outer wall. The failure indicator was lower-than-anticipated pump pressures. In
both cases, the operator successfully identified the anomaly prior to complete pipe parting.
Figure 3Internal erosion on the weld line of ERW pipe penetrated the pipe's wall thickness.
Erosion can initiate at physical anomalies such as grind marks. (Case History No. 3)
An attempt was made to evaluate the erosive conditions in the case histories to existing models. The
two models selected for this effort were API's RP 14E "Recommended Practice for Design and Installation
of Offshore Production Platform Piping Systems" and DNV's "Particle" Impact Erosion Model (DNV-RP
O501). In the case of RP 14E, the fluid velocities and sand quantities upon which it was based is orders
of magnitude lower than observed during stimulation operations. The DNV model did not contain details
and correlations used as its basis so a head-to-head comparison with conditions associated with the casing
failures was not possible.
Buckling. Buckling is another condition that affects pipe stresses (Lubinski and Althouse 1962, Mitchell
2008). Buckling severity is a function of several variables including pressure, the section length and external
support from cement or the wellbore (Fig. 4). Buckling does not damage the pipe unless the increased
stresses exceed the pipe's yield properties. Permanently buckled pipe (Fig. 5) results when the pipe's yield
properties are exceeded.
8 SPE-184868-MS
Figure 4This artist's rendering shows the concept of buckling. A further increase
in internal pressures results in a pitch reduction accompanied by stress increases.
Figure 5A high pressured, fracture treatment allowed the pipe to buckle. A permanent deformation of the pipe occurs
when the buckling-induced stresses exceeds the pipe's yield strength. (Photograph courtesy of Dr. Robert Mitchell.)
Ballooning. Ballooning occurs when pump pressures are internally applied to the casing (Fig. 6)
(Clark 1987). Mechanics of ballooning and the equations that defined it are found in the literature. In an
unrestrained condition, the casing length would shorten as pressure is applied. However, cement restrains the
casing's movement which causes stress increases in the pipe. Pressure variations observed during fracturing
operations may result in a dynamic ballooning process that increases fatigue wear.
SPE-184868-MS 9
Figure 6The application of internally-applied treatment pressures forces an outward radial movement of the pipe wall. The
pipe's length will shorten if the pipe is not restrained. If annular cement restrains pipe movement, stresses are increased.
Bending. Pipe bending in wellbore dog-legs is inevitable in all wells, including vertical wells. The bending
load associated with a dog-leg causes a stress increase on one side of the pipe and an equal decrease on the
opposite side of the pipe. The stress magnitude is determined from a square root function resulting in positive
and negative solutions (Lubinski 1961). The resulting stress load is the sum of the initial axial stress and
the bending-related stress (Fig. 7). Bending doesn't damage the pipe unless yield properties are exceeded.
Figure 7Pipe bending across wellbore dog-legs creates bending stresses that are added
to the existing axial stress to arrive at total stresses at that depth. In addition to pipe
bending, other conditions will similarly create stresses additive to the existing axial stress.
Corrosion. All casing used in wells is susceptible to corrosion (Craig 2014). Laboratory reports on some
of the failed specimens indicates that corrosion played a role in the pipe failure during its design service
life. Most hydraulic fracturing fluid systems contain a stage(s) where large volumes of acid are pumped
down the casing string. The relationship between sources of corrosion and the role corrosion plays in casing
failures needs further investigation.
An additional consideration is the wellbore environment in which the casing is run and cemented.
Drilling fluids and additives may pose long-term threats. Hydrogen and/or sulfides from various sources can
embrittle the casing. Environmental failures have been difficult to diagnose as to a specific cause or source.
Fatigue. Fatigue in the coupling area appears to be associated with most casing failures even though
new casing was run in the wells (Fig. 8). Loads or deformations which will not cause fracture in a single
application can result in fracture when applied repeatedly. Fracture may occur after a few cycles or after
millions of cycles. This process of fracture under repeated loading is called fatigue. The mechanisms of
10 SPE-184868-MS
fatigue failure in downhole casing strings are not well understood. Experiments show that the alternating
stress component is the most important factor in determining the number of cycles of load a material can
withstand before fracture, while the average stress level is less important. Steels have the property that there
is a stress level where the material can withstand unlimited cycles without failure. This stress level is called
the endurance limit. Fatigue life is strongly influenced by the quality of the surface finish, residual stress,
surface or subsurface cracks, stress concentrations, the chemical environment, and the material toughness.
Figure 8The crack observed in this photograph resulted from coupling fatigue. (Case History No. 7)
Brittle Failures. The laboratory test reports often noted brittleness when discussing failures (Fig. 9).
Resistance to brittle failure is called toughness. The toughness of a material is its ability to absorb energy
and resist brittle fracture. Brittle fracture is catastrophic and can be manifested at stresses below the yield
strength of the material. Brittle materials have low toughness because they experience only small plastic
deformations before fracture. Generally, toughness decreases with increasing yield strength. Temperature
can have a significant impact on the toughness of carbon and low-alloy steels. Toughness is usually measured
using the Charpy impact test at a specified temperature. Elongation requirements are also a measure of
ductility and are used to ensure adequate toughness. Texter (1955) and Payne (1993) identified brittle failures
during their investigations.
Figure 9A visual examination of the failed casing showed a brittle fracture. Lab testing indicated
that the failed specimen's yield strength exceeded the API's acceptable ranges. The pipe was
improperly tempered during the manufacturing, as shown by harness testing. (Case History No. 10)
Crack propagation due to internal pressures can occur at less than plastic stress when an imperfection
or crack in the steel propagates to the point that the material fails. The service environment, defined by
SPE-184868-MS 11
temperature, presence of corrosive gases or fluids, pH, material properties and other factors affects when
this type of failure occurs.
Manufacturing issues. A laboratory analysis on a failed casing specimen usually includes a visual
inspection and metallurgical testing. API standards such as API's Specification 5CT "Specification for
Casing and Tubing" usually provides the basis for comparison with the specimen's test-derived properties.
Results from the available case histories identified several failure causes including the following:
1. The specimen's tensile measurement was outside API's acceptable range. Low tensile ratings can cause
overload failures although none were reported in the available case histories. Increases in the pipe's
brittleness are associated with tested tensile values that exceed API's recommendations. Hardness
values exceeded API ranges.
2. The pipe was improperly quenched and tempered when manufacturing. Tempering may have used
incorrect temperatures. Water sprayed on the pipe during the quench process may not have been evenly
distributed.
Manufacture-related issues may not share characteristics but rather differ with each incident.
Post-Failure Outcomes
Casing failures and subsequent recovery operations observed with the case history data have outcomes
that fall in several categories including (1) damage resulting from the failure both downhole and at the
surface including blowouts, pollution and fatalities, (2) an assessment of downhole conditions affecting
the viability of recovery operations, and (3) the recovery operations. With respect to potential recovery
operations, observed outcomes fell in two categories (1) well integrity in some form was reestablished and
completion operations were completed, (2) wells requiring abandonment operations when well integrity
could not be established. Well control and blowouts have been experienced in both categories. Significant
costs are associated with most failure incidents.
An overview of the typical steps and conditions for a successful recovery follow:
1. The incident is quickly detected;
2. Contingency plans for fracture-related casing failures including well control, blowouts, and pollution
are initiated;
3. Diagnostic operations are conducted to locate and assess the failure. Caliper logs should be included
in the diagnostics program. Determine if the failure is a hole or the casing parted;
4. Initiate recovery operations;
a. Pull the upper casing section out of the well;
b. Retrieve several joints at the top of the casing that remained in the well;
c. Make-up and run replacement casing; and
d. Establish integrity by pressure testing or other means.
Pre-planning is essential for successful recoveries.
Recovery may not be possible for several reasons. The failed pipe could not be retrieved and replaced.
This case is typically associated where the failures are located in the cemented casing sections. In several
incidents, hole/casing misalignment prevented access to the lower sections of the well. Casing patches have
been attempted but rarely were successful. A casing patch may cause additional diameter restrictions such
that the completion operations can't be continued. Expandable casing has been used successfully in one
incident.
Blowouts have resulted from casing failures and they remain an on-going threat to future fracturing
operations both on land and offshore. Certified well control training is based on the premise that a competent
12 SPE-184868-MS
casing string is in place. This premise is not valid with casing failures. Conventional blowout control
practices are applicable if the failure was in the equipment immediately above ground. Relief wells are
implemented where the failure was in the deep section of the vertical hole or in the lateral. As the results
of this study have shown, most failures occur in the vertical section of the well above the top of cement.
Failures observed in this section tend to be closer to the surface than the top of cement. Currently existing
blowout control techniques are not immediately applicable in these cases.
A comparison was made of the number of blowouts associated with fracturing against the context of
all blowouts. The data for comparison was taken from the Texas Railroad Commission's Blowout database
(Texas RRC 2016). It contains blowout information from the early 1950s. The time period selected for
the analysis was 2011-2016. For each incident, the report had a Remarks section that provided a brief
description of the incident including causation for some cases. The database included 114 blowouts for
the selected study period. Of the 114 blowouts, 20 incidents or 17.5% were related to various aspects of
fracturing operations. Each fracture-related blowout was examined to determine if it occurred as a result of
a casing failure during pumping operations. Blowouts occurring during flowback and plug drill-out were
excluded (Table 2). Casing failures were identified in 8, or 40%, of the 20 incidents and 7% of all blowouts
during the time period. Twelve (12) of the 20 blowouts were excluded because the Remarks section did not
explicitly identify the cause as a casing failure. Nine (9) of the excluded incidents appeared to have resulted
from casing failures but was not explicitly identified as a casing failure. The case histories identified in
Table 1 do not include the blowouts shown in Table 2. Fig. 10 is from a casing failure where the wellhead
was ejected and damaged a pump truck.
Figure 10A crack at the wellhead-casing interface caused this failure (see Table 2,
number 4). The wellhead and short sections of surface and production casing were
ejected. The projectile speared the truck's cabin. (The driver was not in the truck.)
The bending stresses were calculated with Lubinski's equation. Fig. A-2 shows the dog-legs used for the
calculations. The results were added to the initial tension loads (Fig. 11b). As previously indicated, bending
stresses typically don't have a significant impact when evaluating the cumulative loads. The combined loads
from tension and bending at the bottom of the string are lower than combined loads at the surface.
Loads associated with pressure and temperature changes were determined. An increase in pressure and
a temperature decrease results in length changes (shorten) if the casing is not restrained. Since cement
SPE-184868-MS 15
prevents pipe movement, additional forces are applied. The combined loads from the initial pipe tension,
bending, ballooning and temperature decreases are shown in Fig. 11c. Temperature typically provides the
largest stress changes followed by pressure (ballooning).
To complete this evaluation, the pipe body's yield strength of 367,000 lbf was compared to the applied
loads (Fig. 11d). The pipe's yield strength exceeded the applied load by a factor of 2. This observation is
consistent with the modeling results from the other case histories.
A disclaimer for the API pipe body yield strength of 367,000 lbf is warranted. The subject pipe was
manufactured to provide a collapse rating greater than the API's rating. Mills use proprietary methods when
manufacturing HCP pipe. As such, the API's yield strength may not apply to HCP pipe. Nonetheless, P110
HCP pipe is likely to have a yield strength that exceeds the combined loading.
A finding from this analysis was that the observed casing failures were caused by factors other than
previously considered, perhaps acting simultaneously with the axial load and Clark's considerations. Testing
reports indicated that fatigue was a contributing cause for most failures. Since fatigue results from cyclic
loading, the focus of an on-going investigation is to identify the individual loads involved in cyclic
operations and the manner in which they interact.
An ultimate goal is to develop a representative model that will address most or all of the fracturing
conditions experienced by casing strings. The model can serve as a predictive aid for field operations, casing
design and well planning. After the model has been successfully tested for land-based incidents, its scope
should be expanded to handle fracturing conditions associated with deep water, HPHT offshore wells.
As previously identified, common failures found during fracturing consist of fractures near connectors,
or longitudinal brittle failure of the connectors, with fatigue cracks. The failure modes identified in this
investigation aren't addressed by the API or ISO.
10. Coupling design and selection is an important aspect of the well design process. The current study
indicates that the coupling and/or the last engaged thread on either side of the coupling have failed.
Causes include fatigue that increases the casing's brittleness, i.e., a reduction in the pipe's ductility.
Efforts to protect the casing's ductility should be considered;
11. Multiple sizes and types of couplings are available. Make-up stresses should be determined and
evaluated as these stresses vary widely among the couplings. Couplings that involve reduced make-
up stresses might be favored;
12. Drillers running casing should be cautioned against rapidly lowering the casing and abruptly stopping
it to avoid shock; and
13. Running casing for long horizontal wells or crooked holes often demonstrates friction loads that
exceed the pipe's hanging weight at the surface. Drillers have resorted to pipe rotation or setting block
weight on the top of the casing. These practices should be avoided. Options to address this issue
include avoidance of severe dog-legs or reaming the hole if they occur, using a larger bit to drill the
section, reduce the drilling fluid's lubricity, lower the fluid loss to develop a thin filter cake, or partially
float-in the casing. These options have been used successfully.
The findings from on-going investigations and future research efforts will be used to expand and refine
practical guidance to assist operations when planning and drilling wells to be fractured, both on land of
offshore.
Acknowledgements
Support at various stages of this investigation was provided by Mr. George Slota (CMC COMETALS
STEEL) and Mr. John Hadjioannu (EPI Materials Testing Group). Their efforts are appreciated.
References
ANSI/API TR 5C3, Technical Report on Equations and Calculations for Casing, Tubing, and Line Pipe Used as Casing
or Tubing; and Performance Properties Tables for Casing and Tubing. 2008. Washington, DC: API.
API RP 14E, Recommended Practice for Design and Installation of Offshore Production Platform Piping Systems. 2013.
Washington, DC: API.
API RP 5LW, Recommended Practice for Railroad Transportation of Line Pipe. 2009. Washington, DC: API.
API RP 5LW, Recommended Practice for Transportation of Line Pipe on Barges and Marine Vessels. 2009. Washington,
DC: API.
Clark, H. C. 1987. Mechanical Design Considerations for Fracture-Treating Down Casing Strings. SPE Drilling
Engineering 2 (2): 116-288. SPE-14370-PA. http://dx.doi.org/10.2118/14370-PA.
Craig, B. D. 2014. Oilfield Metallurgy and Corrosion, 4th edition. Denver, CO: MetCorr.
DNV RP O501, Recommended Practice Erosive Wear in Piping Systems. 2011. Oslo, Norway: DNV GL.
ISO/TR 10400:2007, Petroleum and natural gas industries Formulas and calculation for the properties of casing, tubing,
drill pipe and line pipe used as casing or tubing. 2007. Geneva, Switzerland: ISO.
Lubinski, A. 1950. A Study on the Buckling of Rotary Strings. API Drilling and Production Practice, New York, New
York, 1 January. API-50-178.
Lubinski, A. 1961. Maximum Permissible Dog-Legs in Rotary Boreholes. SPE Journal of Petroleum Technology 13 (2):
175-194. SPE-1543-G-PA. http://dx.doi.org/10.2118/1543-G-PA.
Lubinski, A., Althouse, W.S., and Logan, J. L. 1962. Helical Buckling of Tubing Sealed in Packers. SPE Journal of
Petroleum Technology 14 (6): 655-670. SPE-178-PA. http://dx.doi.org/10.2118/178-PA.
Mitchell, R. F. 2008. Tubing BucklingThe State of the Art. SPE Drilling & Completion 23 (4): 361-370. SPE-104267-
PA. http://dx.doi.org/10.2118/104267-PA.
OBrian, T. B. 1984. Why Some casing Failures Happen. World Oil 199 (1): 113-114, 116, 118.
Payne, M. L., Leturno, R. E. and Harder, C. A. 1993. Fatigue Failure of API 8-Round Casings in Drilling
Service. SPE Annual Technical Conference and Exhibition, Houston, Texas, 3-6 October. SPE-26321-MS. http://
dx.doi.org/10.2118/26321-MS.
Texas RRC. 2016. Online Research Queries. Railroad Commission of Texas, http://www.rrc.state.tx.us/about-us/resource-
center/research/online-research-queries/.
SPE-184868-MS 19
Texter, H. G. 1955. Oil-Well Casing and Tubing Troubles. API Drilling and Production Practice, New York, New York,
1 January. API-55-007.
Weiner P. D. and True, M. E. 1969. A Method of Obtaining Leak Proof API Threaded Connections in High- Pressure Gas
Service. Drilling and Production Practice, Washington, D.C., 1 January. API-69-040.
Tiemann, M. and Vann, A. Hydraulic Fracturing and Safe Drinking Water Act Regulatory Issues. R41760, Congressional
Research Service, Washington, DC (July 2015).
20 SPE-184868-MS
Appendix A
th
Figure A-1A failure occurred at ~7,550 ft while pumping the 15 stage.
The operator successfully recovered the wellbore and produced the well.
The production section of the well was drilled to 12,068 ft measured depth (md). The 4.5 in. casing string
used for fracturing was run to 12,068 ft and cemented. The estimated TOC was 5,091 ft. The casing string
included a dual float shoe, 288 joints of 4.5 in. OD, 11.6 pounds per foot (ppf), P110 grade, HCP, LT&C
coupled, electric resistance welded (ERW) pipe. Attached to the string as it was being run were 48 solid
body, 5-bladed turbolizers with the first being placed 10 ft above the shoe and the remainder placed every
third joint.
Drilling operations create dog-legs in the wellbore. Survey data was used to determine the bending
severity from the dog-legs (Fig. A-2). The magnitudes of these dog-legs were characteristic of other case
histories involving build sections. Although the severities are initially alarming, preliminary conclusions
from this investigation indicates that dog-legs may not play a significant role in casing failures.
SPE-184868-MS 21
Fig. A-3 shows the surface treating (pump) pressures and slurry rates for the 15th stage in which the failure
is believed to have occurred. Pumping operations started at ~2309 hours with a maximum rate of ~86.5 bbl/
min and pressure of ~8,550 psi at 0005 hours. It appears the pumps were equipped with an overpressure
(auto-stop) feature which was set at 8,500 psi and triggered when surface pressures reached ~8,550 psi. After
a short shut down period of 77 seconds, operations continued with a pump rate of ~84 bbl/min accompanied
by pump pressure fluctuations.
th
Figure A-3The fracture treatment report for the 15 stage recorded pump rates and pressures. The failure
occurred shortly after midnight. Pressures after the failure were substantially lower than previously estimated.
Fracturing operations were summarized and shown in Table A-1. The cumulative pump time was 2,595
minutes or over 45 hours. Each stage started with a volume of acid. Maximum pump rates routinely exceeded
80 bbl/min while pressures exceeded 8,000 psi.
22 SPE-184868-MS
No. Pump Time (mins) Treated FE Acid (gals) Sand (lbs) Max Pressure (psi) Max. Rate
Water (gals) (bbl/min)
A summary of the fracturing job through the 16th stage provides insight as to the severity of fracturing
operations. Over the 2,595 minutes (43.25 hours) of pumping, the following were pumped:
1. 6,894,503 gallons of treated water;
2. 156,746 gallons of acid; and
3. 2,565,639 pounds of sand.
This fluid summary clearly indicates that fracturing conditions impose aggressive loading conditions on
the casing string that aren't associated with normal, non-fracturing conditions.
Observations from Table A-1 and Fig. A-2 include the following:
1. The maximum treating pressure of ~8,550 psi was lower than the pipe's burst rating of 10,690 psi for
the 4.5 in., 11.6 pounds per foot (ppf), P110 grade, LTC pipe;
2. Applied annular collapse pressures were negligible because the annulus valve was left open;
3. The service provider's treatment reports indicate that breakdown occurred at 2309 hours. The pressure
drop observed from 2321-2330 hours was not addressed in the available information even though the
pump rate was relatively constant at 39 bbl/min; and
4. Although the observed pump rates appeared relatively constant for long periods of time, the associated
treating pressures frequently varied.
The failure in the build section was similar to Magill's split failures near the heel of the lateral. This well
used P110 pipe as was the case in Magill's publication.
The operator initiated diagnostic operations after it became clear that a failure had occurred. Tubing with
a test packer was run in the well, set at various depths and pressures were applied. After several iterations,
SPE-184868-MS 23
the failure was located at ~7,550 ft, which is in the build, cemented section of the hole. These conditions
prevented the retrieval of the failure specimen for laboratory testing.
The operator was able to produce the well with a pump jack even though the heel failure was not squeezed.
Costs associated with the 16 fracture stages were not lost. Drilling costs to the time that fracturing started
was estimated to be ~$1,500,000. A replacement well was planned to capture the reserves originally planned
for stages 16-34.
Stage Notes/Comments
1 (Lower Buda) Shut down after pumping 44 bbl of 28% HCl acid due to leak on pump suction hole. Flushed lines with
54 bbl of water. Shut down after pumping additional 36 bbl of 28% HCl acid due to coupling on a hose pulling off. Over
flushed the casing and shut down for the night. Operations on the next day (10/25) was pumped as designed. Ran 122 bbl
of 28% HCl and 110 bbl of 19% HCl acid. (1508-1647 hours)
2 (Upper Buda) Pumped as designed. Saw a ~1,000 psi drop during the 1.25 ppg stage.(1648-1827 hours)
3 (Eagle Ford Shale) Pumped as designed, except for the acid volume being short. Saw a ~1,000 psi drop during the pad
stage. Did not observe pressure increase when Ball 4 should have been on seat. (1828-2007 hours)
4 Pumped as designed, except did not pump acid. Lost prime during the 1.5 ppg stage and had to sweep/flush sand at 45 bbl/
min. Stated back sand at 1.5 ppg and pumped remainder of sand. (2008-2147 hours)
Recovery efforts successfully retrieved the upper section of the failed pipe and the top two joints of the
lower section. The coupling connector on the top of the lower fish was retrieved. The specimens were tested
with standard API testing protocols. All test results indicated the pipe was within satisfactory API ranges.
The testing report concluded the failure mode was likely due to initiation and propagation of a fatigue crack
near the last engaged thread (first exposed thread) (Fig. A-4 and A-5, Case History No. 8). Other comments
contained in the report include the following:
1. The fracture surface is smooth, flat and oriented perpendicular to the axis of the casing which is typical
of fatigue cracking;
2. The fracture is located near the last engaged thread of the pin connection on the casing. The last
engaged threads of a connection experience higher stresses and stress concentrations compared to the
24 SPE-184868-MS
rest of the connection, making these threads susceptible to the initiation and propagation of fatigue
cracks;
3. Washing damage was present on the fracture surface of the pin connection. A fatigue crack (or multiple
cracks) likely propagated through the wall thickness of the pin, created a fluid path directly from the
string bore to the wellbore annulus. Subsequent erosion of the pin created characteristic "washing"
in the pin.
4. The remaining portion of the fracture was oriented 45 to the casing longitudinal axis, which is typical
of ductile overload. The fatigue cracks propagated and reduced the cross-sectional area. When the
applied load finally exceeded the load carrying capacity of the remaining cross-sectional area after
erosion, the connection failed due to ductile overload.
The characteristics observed in this incident are similar to other observed failures. A fatigue crack is
observed in the last engaged thread (first exposed thread). Crack propagation leads to a failure. Cracks
have been observed in the field and mill ends of the coupling. The failure appears to have occurred without
warning signs.
Figure A-4The failure resulted from a fatigue crack in the last engaged thread (first
observed thread). The photograph shows characteristics shared with other failures.
Figure A-5Erosion occurred as the crack propagated around the casing's circumference. Erosion did not cause the failure.