Sunteți pe pagina 1din 13

SPE-184821-MS

Stress Analysis of Tubular Failures During Hydraulic Fracturing: Cases and


Lessons Learned

A. Haghshenas, J. E. Hess, and A. J. Cuthbert, Halliburton

Copyright 2017, Society of Petroleum Engineers

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
Wellbore integrity during the full life cycle of a well has always been a focus of the petroleum
industry, particularly during extreme drilling conditions and throughout complex well operations. Pumping
compound fluids during hydraulic fracturing operations affects stresses on all tubular components, which
can lead to component failure. During casing and tubing design, expected stresses are determined. This
includes kick tolerance, drilling and production operations, and complete wellbore evacuation. However, a
series of tubular failures indicates a more detailed analysis is necessary to help prevent well integrity loss.
This paper presents selected case studies of tubing failure during the hydraulic fracturing process. A
transient program is used to determine the precise magnitude of stress on both the tubular and the connection.
The analysis results are provided and discussed in detail. It is concluded that high localized dogleg severity
and the sequence of fluids in the wellbore should be studied carefully because the stress on the tubing
changes with the position of fluids and fluid density.

Introduction
Well design and operations are based on well integrity. The objective of proper well design is to protect the
surrounding environment, protect the well from external issues, and protect the well from geological forces
(King 2012). Compromised well integrity can result in subsurface damage, surface spills, and potential loss
of life and production. Therefore, wells are designed with multiple barriers to help prevent catastrophic
incidents. In the event of an uncontrolled hydrocarbon emission, remedial action is necessary to attempt
to resolve the situation. Although managing well integrity varies during drilling, completion, production,
and abandonment operations, there are guidelines for maintaining well integrity during different operational
phases (NORSOK D-10 2013; ISO/TS 16530:2 2014).
Completion and workover operations apply different techniques to enhance production rates. Operations
can be performed rigless or by using a workover rig, and various fluids can be pumped to accomplish
the operation objectives. During hydraulic fracturing operations, the well is exposed to extreme pressure
and flow rates to create pathways for formation fluid to flow into the wellbore. Wellbore integrity failures
(tubular and cement) during hydraulic fracturing operations have been reported worldwide, and several
authors have reviewed the failure causes.
2 SPE-184821-MS

King and Valencia (2016) reviewed causes of well integrity failure during hydraulic fracturing and
refracturing operations. According to the study, the failures in highly deviated wells are caused by design,
drilling issues, mechanical failure, human error, cement failure, geological forces, corrosion, and erosion.
King and Valencia (2016) also acknowledge that the failure sources were unknown because of a lack of
information. Well integrity begins with a proper well design that identifies stresses on the tubulars during the
life of the well. Generally, a design factor (safety factor) for selecting appropriate equipment is considered
to help ensure well integrity during the life of the well.
Daneshy (2005) proposed off-balance fracturing as the cause of tubular failure because the effect of
opened fractures around the casing reduces the confining forces, which could result in casing failure. The
study also indicated that the formation deformation caused by creating fractures could impose extra shear
forces on the casing and result in failure. Excessive stresses in the casing can cause a rupture at the casing
connections or parting of the casing in the perforated section.
Researchers have provided a list of parameters that can cause well integrity loss during hydraulic
fracturing operations. However, they do not provide any well operation specifications or detailed analysis
of operational conditions during the hydraulic fracturing process. The majority of researchers have not
considered the effect of casing connections on the integrity of well tubulars. Payne et al. (1993) studied the
effect of fatigue on 8-round casing. Schwind et al. (2001) provided comparisons of failure modes. These
authors claimed that up to 90% of all tubular failures are related to connections that could be attributed to
improper design, improper makeup, damage to the connection, etc. The operational limits of a connection
depend on geometry, material, treatment, exposure to fluid, temperature, pressure cycling, and dogleg
severity. Tubular failure at the connections is well known in the industry. However, many engineers either
overlook it or do not apply proper techniques to evaluate stress loads at the connections. Failures at the
connections are partly attributed to field-handling procedures or failure to use the correct torque value. A
proper and detailed analysis provides sufficient information on the failure risks at the connections.
This paper discusses tubular failure causes during hydraulic fracturing operations. Transient and detailed
operational sequence modeling is crucial for evaluating wellbore integrity during complex operations. The
study also shows that high localized doglegs affect the tubular and coupling pressure ratings. This ultimately
contributes significantly to casing string failure. In addition, excessive pressure during hydraulic fracturing
operations can cause casing or tubing failure by either exceeding working pressure or through cyclic fatigue.
Three examples of production liner failure during hydraulic fracturing operations were selected for analysis
from 52 documented cases. All drilling and hydraulic fracturing data were collected, and analyses were
performed using casing design software. The dynamics of every hydraulic fracturing operation stage were
modeled. The stresses on the body of the casing, liner, and tubing were monitored during the operation to
identify the failure cause.
Multiple tubular failures recorded by several operators suggested that a design flaw existed in the tubular
selection. Well design, trajectory, and hydraulic fracturing data were used to simulate the stresses on the
casing, liner, and tubing using the design software. Initial observations indicated that in all cases, the casing
was acceptable for initial conditions, shut-in conditions, and production. A full stress analysis during the
hydraulic fracturing operations indicated potential critical stress would be generated on the couplings. The
connections used for the casing were API connections, and their connection yield strength rating was less
than the pipe body. Therefore, even though the strength of the casing body was sufficient for the fracturing
operation, the couplings still failed. The well trajectory indicated high localized doglegs caused localized
tension and consequently reduced the safety factor.
After reviewing the analyses, it was concluded that the tubing connection tensile strength was insufficient
to overcome the stresses imposed during the fracturing operation. In addition, the localized dogleg severity
was beyond the accepted couplings performance rating. It was suggested that a review of the stresses on
the casing during the fracturing operation could aid appropriate casing components selection.
SPE-184821-MS 3

Case Studies
Several tubing failures during hydraulic operations were reported. Initial tubing integrity analysis indicated
the tubings were properly selected for the operation. Therefore, the actual failure causes were unknown
to the operators. A detailed investigation was requested to determine the failure causes and provide
recommendations to prevent future occurrences.
All data pertaining to the operation were collected to perform numerical analysis of the operations. The
hydraulic fracturing operation was a series of complex procedures at high pressure and high flow rate. A
transient numerical modeling tool was selected to properly and accurately simulate the sequence of flow in
the tubing and analyze stresses on the tubing string. The program transient heat transfer segment provided
unparalleled confidence regarding the magnitude of residual stresses on the tubing during various hydraulic
fracturing operation phases.
The modeling began by carefully simulating the initial condition of the tubing in the wellbore. This
included the temperature profile in the wellbore, fluid condition, and stresses on the tubing before beginning
the fracturing operation. The numerical investigation focused on detailed analysis of the tubing body,
connections, and operation progress.
This paper presents the analyses of the three cases to demonstrate the common failure causes. Based on
the conclusions, guidelines and recommendations are provided to help prevent future tubing failures.

Case 1
Well 1 was drilled vertically to a depth of 10,800 ft. The well was kicked off and reached the horizontal
at 11,560 ft measured depth (MD). It reached final depth at 20,770 ft. The average buildup rate for this
section was 11.8/100 ft. The 7-in. intermediate casing was set at 11,593 ft MD. The well was completed
by running a 4 1/2-in., P-110, 11.6-lbf production liner that was tied back to the surface. Fig. 1 shows the
Well 1 schematic.

Figure 1Well 1 schematic.


4 SPE-184821-MS

The hydraulic fracturing operation details were reviewed, and the operation was divided into 14 different
stages based on fluid properties and pump rate (Table 1). The operation began by pumping fresh water at a
30-bbl/min rate and followed by 15% hydrochloric (HCl) acid and slickwater pads. The hydraulic fracturing
fluid was pumped with a proppant concentration of 0.25 to 5.0 lbm/gal at a 24-bbl/min pump rate. Flush
fluid was pumped at a 30-bbl/min rate at the end of the operation.

Table 1Hydraulic fracturing operation sequence of Well 1.

Stage Fluid Proppant Volume Rate Pressure


Concentration (gal) (bbl/min) (psi)
(lbm/gal)

1 Fresh water 0 4,300 30 7,142


2 15% HCl acid 0 1,000 15.0 6,123
3 Slickwater Pad 1 0 6,000 16.0 8,041
4 Slickwater Pad 2 0 15,000 23.0 8,141
5 Delayed borate-crosslinker 0.25 2,050 24.0 6,864
6 Slickwater Pad 3 0 14,100 24.0 6,802
7 Delayed borate-crosslinker 0.5 2,100 25 6,376
8 Slickwater Pad 4 0 15,000 24.0 6,362
9 Delayed borate-crosslinker 1 15,000 24.0 6,352
10 Delayed borate-crosslinker 2 16,400 24.0 5,981
11 Delayed borate-crosslinker 3 17,000 24.0 5,846
12 Delayed borate-crosslinker 4 16,550 24.0 5,361
13 Delayed borate-crosslinker 5 3,900 24.0 5,156
14 Flush 0 13,500 30 8,000

The stress magnitude on the tubing string during the operation was analyzed after the operation details
were collected and the numerical model was set up. Fig. 2a shows the numerical analysis results during the
hydraulic fracturing operation. The analysis indicated the tubing experienced excessive stress during the
operation. The stress exceeded the tubing rating in some instances. Fig. 2b shows the maximum stresses
observed at the connections.
SPE-184821-MS 5

Figure 2Stress analysis on the 4 1/2-in. tubing during Well 1 hydraulic fracturing operations.

The investigation results differed from the initial analysis performed internally by the operator. The two
main discrepancies were failure limits and the stress magnitude on the tubing. Three issues were identified
as contributing factors after further analysis and discussion. This included tubing connections, hydrostatic
pressure, and localized dogleg severity.
The tubing connection was a round thread that had only a 75% tubing body pressure rating. Therefore,
the tubing weak point was at the connections. The tubing was selected based on expected stresses during
6 SPE-184821-MS

the hydraulic fracturing operation. However, the connection tensile rating was not considered. Figs. 2a and
2b show the stress on the tubing exceeds the connection tensile limits.
Another source of discrepancy between the results was ignoring the proppant effect on the fracturing
fluid. The initial analysis assumed the fracturing fluid density was 8.5 lbm/gal. Adding proppant to the
fracturing fluid increased the density and consequently the fracturing fluid hydrostatic pressure. Fig. 3 shows
the addition of 5 lbm of proppant to 1 gal of fracturing fluid increased the density from 8.5 to 11.0 lbm/gal
(i.e., 30% higher hydrostatic pressure in the tubing).

Figure 3Fracturing fluid density increased by the addition of proppant with 2.65 specific gravity (SG).

The combination of 25% tensile degradation and hydraulic pressure underestimation by as much as 30%
influenced the initial tubing design. In addition to the two aforementioned parameters, the localized dogleg
severity effect should also be addressed. Each connection has specific characteristics and limitations that
should be considered for proper tubing string selection. One factor that is generally overlooked during casing
and tubing design is dogleg severity, which increases the tension on the string. Moreover, dogleg severity
affects the ability of the connections to seal correctly. Round threads have limited flexibility. There is no
guideline for dogleg severity limits of round connections. However, experts agree that there is an upper
limit of 8 or 10/100 ft for these types of connections.
Fig. 4 illustrates the local dogleg severity in Well 1. The dogleg severity radically changes below the
kickoff point (KOP). The extreme of dogleg severity reaches 18/100 ft. Numerical analysis indicated the
minimum design factor for the tubing in the build section during hydraulic fracturing operations was less
than one. The tubing failed close to the bottom of the build section, just below the high localized dogleg
section. This confirmed the investigation conclusions.
SPE-184821-MS 7

Figure 4Localized dogleg severity in Well 1 induced excessive tension on the tubing and reduced the design factor.

Case 2
Well 2 was drilled vertically to a depth of 10,700 ft. The well was kicked off and reached the horizontal at
11,370 ft MD, with a final depth of 20,800 ft. The average buildup rate for this section was 13.4/100 ft.
The last part of the lateral section was drilled uphill with a 93 inclination. The 7-in. intermediate casing
was set at 11,397 ft MD, and the well was completed by running a 4 1/2-in., P-110, 11.6-lbf production liner
that was tied back to the surface. Fig. 5 shows the Well 2 schematic.

Figure 5Well 2 schematic.

The hydraulic fracturing operation details were reviewed, and the operation was divided into 15 different
stages based on fluid properties and pump rate (Table 2). The operation began by pumping fresh water at
8 SPE-184821-MS

a 20-bbl/min rate and followed by 15% HCl acid at a 12-bbl/min rate. Slickwater pads were then pumped
at a 27-bbl/min rate. The hydraulic fracturing fluid was pumped with a proppant concentration of 0.25 to
5.0 lbm/gal at a 27-bbl/min pumping rate.

Table 2Hydraulic fracturing operation sequence for Well 2.

Stage Fluid Proppant Volume Rate Pressure


Concentration (gal) (bbl/min) (psi)
(lbm/gal)

1 Flush: fresh water Sleeve 0 15,000 20 7,243


displacement 13,174 gal
2 Prepad: fresh water 0 400 12 5,785
3 Spearhead: 15% HCl acid 0 1,000 12 5,770
4 Pad: waterfrac 0 14,900 27 7,148
5 Pad: delayed borate-crosslinker 0 14,000 27 6,807
6 Delayed borate-crosslinker 0.25 2,000 27 6,789
7 Pad: delayed borate-crosslinker 0 14,000 27 6,748
8 Delayed borate-crosslinker 0.5 2,000 27 6,699
9 Pad: delayed borate-crosslinker 0 13,000 27 6,679
10 Delayed borate-crosslinker 1 14,650 27 6,662
11 Delayed borate-crosslinker 2 14,150 27 6,348
12 Delayed borate-crosslinker 3 17,000 27 6,102
13 Delayed borate-crosslinker 4 15,300 27 5,928
14 Delayed borate-crosslinker 5 6,500 27 5,812
15 Preflush: delayed borate-crosslinker 0 1,900 27 5,906

Fig. 6 shows the stress analysis on the tubing for the operation's selected stages. The tensile stress
exceeded the tubing connection limits, and failure was expected. Fig. 7 shows the localized dogleg severity
in the buildup section reaches 26/100 ft, which is approximately threefold higher than the accepted value for
round connections. The minimum normalized design factor for the tubing at the end of the buildup section
was less than one. Therefore, tubing integrity loss during hydraulic fracturing operations was expected.

Figure 6Stress analysis of the 4 1/2-in. tubing during hydraulic fracturing operations in Well 2.
SPE-184821-MS 9

Figure 7Localized dogleg severity in Well 2 induced excessive tension on the tubing and reduced the design factor.

Case 3
Well 3 was drilled vertically to a depth of 10,800 ft. The well was built up to the horizontal at 11,560 ft
MD and reached final depth at 21,500 ft. The average buildup rate for the built-up section was 11.8/100
ft. The well was completed by running a 4 1/2-in., P-110, 11.6-lbf production liner that was tied back to
the surface. Fig. 8 shows the Well 3 schematic.

Figure 8Well 3 schematic.

The hydraulic fracturing operation was divided into 18 different stages based on fluid properties and pump
rate (Table 3). The operation began by pumping fresh water at a 15-bbl/min rate and followed by 15% HCl
10 SPE-184821-MS

acid at a 10-bbl/min rate. The pads were then pumped at a 20- to 30-bbl/min rate. The hydraulic fracturing
fluid was pumped with a proppant concentration of 0.25 to 5.0 lbm/gal at a 30-bbl/min pumping rate.

Table 3Hydraulic fracturing operation sequence for Well 3.

Stage Fluid Proppant Volume Rate Pressure


Concentration (gal) (bbl/min) (psi)
(lbm/gal)

1 Fresh water 0 0 15 8,568


2 Fresh water 0 0 11 5,494
3 15% HCl acid 0 0 10 5,531
4 Waterfrac 0 0 21 8,533
5 Fresh water 0 0 30 6,571
6 Fresh water 0 0 12 5,707
7 15% HCl acid 0 0 10 5,556
8 Waterfrac 0 0 30 6,934
9 Delayed borate-crosslinker 0 0 30 7,655
10 Delayed borate-crosslinker 0.2 400 30 7,670
11 Delayed borate-crosslinker 0 0 30 7,706
12 Delayed borate-crosslinker 0.4 800 30 7,671
13 Delayed borate-crosslinker 0 0 30 7,642
14 Delayed borate-crosslinker 0.5 7,500 30 7,454
15 Delayed borate-crosslinker 1.5 21,200 30 6,738
16 Delayed borate-crosslinker 2.75 38,500 30 6,261
17 Delayed borate-crosslinker 3.8 33,000 30 5,950
18 Fresh water 0 0 30 6,552

Figs. 9 and 10 show the stress analysis, localized dogleg severity, and minimum normalized design of
the tubing for selected stages of the operation factor. The stress analysis and minimum normalized design
factor results were similar to Well 1, and the tubing failure was identified in the buildup section.

Figure 9Stress analysis on the 4 1/2-in. tubing during Well 3 hydraulic fracturing operations.
SPE-184821-MS 11

Figure 10Localized dogleg severity in Well 3 induced excessive tension on the tubing and reduced the design factor.

Discussion
The majority of casing and tubing failures occur at the connections. These failures are attributed to improper
makeup, excessive loads, threads characteristics, and design. Mitchell (2006) categorized failure at the
connections into leakage, structural failure, damaged thread during makeup, burst, failure under tensile
stress, and excessive makeup torque categories. Although field operations are considered the primary
contributing factor to leakage or failure at the connections, engineers should also consider the effects of
complex drilling trajectories and complicated hydraulic fracturing processes on the connections.
The three case studies presented show the connection stress ratings were less than the tubing body stress
rating in which the tension rating at the connection was 25% less than the tubing. In addition, the high
dogleg severity in the buildup section induced an additional localized tensile load on the connections. The
design factor for connections in the build section was less than one, which caused connection failure under
tensile force.
Sand added to the hydraulic fracturing fluid increased the density and hydrostatic pressure. For example,
5 lbm of sand added to 1 gal of fracturing fluid could increase the hydrostatic pressure by 30% and
significantly affect the tubular's stress analysis and failure criteria. The complicated hydraulic fracturing
process should be properly divided into stages to perform accurate stress analysis on the tubing and
connections. Although the tubing design might indicate a sufficient design factor at shut-in and the highest
expected pressure at the surface, failure could occur during the sequence of pumping pads and proppants.
Therefore, it is recommended that the stress analysis of the tubular during the entire operation be evaluated.
The threaded connections are designed to provide a seal and are generally the weakest point on the tubular.
In addition to makeup torque and field handling, tensile and bending forces affect the connection integrity.
Therefore, connection manufacturers should be consulted to determine the thread limitations. The general
range of dogleg severity accepted for round threads is between 8 and 10/100 ft. The well trajectory designs
for the wells necessitated an average dogleg severity greater than 8/100 ft, and localized dogleg severity
was significantly higher. Therefore, in addition to connection "jump-out" during the hydraulic fracturing
operation, the leakage risks through the connection were significantly higher in the buildup section.
12 SPE-184821-MS

Efficient hydraulic fracturing operations depend on the field-proven pump rate. Therefore, the
recommendation focuses on the drilling operation and tubing design. A reduced planned buildup rate and
suitable downhole motors for controlling dogleg severity could reduce the tension at the connections, thus
lowering tubing failure risks. However, the magnitude of uncertainty in controlling the trajectory precisely
to the planned path is high and does not eliminate the risk of connection failure on the tubing.
Premium connections with performance values equal to compression, tension, burst, and collapse of the
tubing body are recommended to prevent tubing integrity loss during the hydraulic fracturing operation.
The connection should also be capable of maintaining a seal at the anticipated dogleg severity range in the
wells. Fig. 11 shows the Well 2 stress analysis with a connection rating equal to the tubing body rating. The
expected stress during the operation is less than the tubing capability. Therefore, from the tubular design
and characteristics, the integrity is maintained.

Figure 11A connection with a similar rating to the tubing body helped
prevent tubing integrity loss in Well 2 during hydraulic fracturing operations.

The well plan should consider the cost of removing the current connections if the tubing is already
purchased. Recutting to the new connections should also be considered. Other issues during changing
connections are transportation costs, recutting time, project delay, and licensed threading vendor availability.
The associated cost of rethreading could be substantial. The operator might consider modifying the well
trajectory and directional drilling practices to avoid these costs, while accepting the risk of tubular integrity
issues.

Conclusions
The analysis and discussion presented in this case study provides insight into the effect that pumping
multiple stages of fluid in the well and wellbore trajectory has on the stresses on the tubing and connections.
The analysis provides sufficient proof of tubing failures for the cases presented in this paper.
SPE-184821-MS 13

The combination of high localized dogleg severity and high stress load during hydraulic fracturing
operations can cause tubing connection failure. The operation analysis predicted the tubing failure
depth that closely matched the field data.
The connections limitations should be carefully considered during well tubulars design. It
is recommended to review data or consult with manufacturers regarding the connections
specifications. The connections tensile and pressure ratings should be considered during the design
phase to help mitigate tubular integrity failure.
Tubing failure can be prevented by reducing localized dogleg severity to less than 8/100 ft. The
localized dogleg can be controlled by closely following the well plan and/or adjusting the KOP for
a less aggressive wellbore trajectory design.
Tubing failure can be prevented by using a premium connection with 100% compression and
tension efficiency and a 100% burst and collapse pressure rating. The connection seals integrity
should not be compromised by the dogleg severity anticipated in the field.
The cost of modifying the drilling plan trajectory compared to the cost of changing tubing
connections should be evaluated to determine the most appropriate action to help mitigate the issue.

Acknowledgements
The authors thank Halliburton for an important review of this study and permission to publish this paper.

References
Daneshy, A.A. 2005. Impact of Off-Balance Fracturing on Borehole Stability and Casing Failure. Presented
at the SPE Western Regional Meeting, Irvine, California, USA, 30 March1 April. SPE-93620-MS. http://
dx.doi.org/10.2118/93620-MS.
ISO/TS 16530-2:2014, Well integrity Part 2: Well integrity for the operational phase. 2014. Geneva, Switzerland: ISO.
King, G.E. 2012. Hydraulic Fracturing 101: What Every Representative, Environmentalist, Regulator, Reporter, Investor,
University Researcher, Neighbor, and Engineer Should Know About Hydraulic Fracturing Risk. J Pet Technol 64 (04):
3442. SPE-0412-0034-JPT. http://dx.doi.org/10.2118/0412-0034-JPT.
King, G.E. and Valencia, R.L. 2016. Well Integrity for Fracturing and Re-Fracturing: What Is Needed and Why?. Presented
at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, USA, 911 February. SPE-179120-
MS. http://dx.doi.org/10.2118/179120-MS.
Mitchell, R.F. 2006. Casing Design. In Petroleum Engineering Handbook Volume II, ed. R.F. Mitchell and L.W. Lake,
Chap. 7, 287342. Richardson, Texas, USA: SPE.
NORSOK D-010:2013, Well integrity in drilling and well operations, revision 4. 2013. Oslo, Norway: NORSOK.
Payne, M.L., Leturno, R.E., and Harder, C.A. 1993. Fatigue Failure of API 8-Round Casings in Drilling Service. Presented
at the SPE Annual Technical Conference and Exhibition, Houston, Texas, USA, 36 October. SPE-26321-MS. http://
dx.doi.org/10.2118/26321-MS.
Schwind, B.E., Payne, M.L., Otten, G.K. et al. 2001. Development of Leak Resistance in Industry Standard OCTG
Connections using Finite Element Analysis and Full Scale Testing. Presented at the Offshore Technology Conference,
Houston, Texas, USA, 30 April3 May. OTC-13050-MS. http://dx.doi.org/10.4043/13050-MS.

S-ar putea să vă placă și