Está en la página 1de 19

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/222929538

Structural integrity of aging buried pipelines


having cathodic protection

Article in Engineering Failure Analysis October 2006


DOI: 10.1016/j.engfailanal.2005.07.008

CITATIONS READS

48 162

2 authors:

Sergei Shipilov Iain Le May


Oak Ridge National Laboratory 114 PUBLICATIONS 822 CITATIONS
26 PUBLICATIONS 133 CITATIONS
SEE PROFILE

SEE PROFILE

All content following this page was uploaded by Sergei Shipilov on 21 January 2016.

The user has requested enhancement of the downloaded file.


Engineering Failure Analysis 13 (2006) 11591176
www.elsevier.com/locate/engfailanal

Structural integrity of aging buried pipelines having


cathodic protection
a,* b,1
Sergei A. Shipilov , Iain Le May
a
Department of Mechanical and Manufacturing Engineering, University of Calgary, 2500 University Drive N.W.,
Calgary, Alta., Canada T2N 1N4
b
Metallurgical Consulting Services Ltd., P.O. Box 5006, Saskatoon, Sask., Canada S7K 4E3

Received 8 July 2005; accepted 10 July 2005


Available online 15 September 2005

Abstract

Some features of the problem of pipeline materials degradation and failure under in-service conditions, including
environmentally assisted cracking, are examined. As the practice of cathodic protection results in hydrogen charging
of carbon steels used as components of oil and natural gas transmission pipelines, close attention is paid to the role
of hydrogen embrittlement (hydrogen-induced cracking) in pipeline safety and reliability.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Cathodic protection; Corrosion fatigue; Hydrogen-assisted cracking; Stress-corrosion cracking; Pipeline failures

1. Introduction

1.1. Safety and structural reliability of pipelines

Maintaining safe and reliable high-pressure pipeline systems is important because the products trans-
ported through pipelines hydrocarbons in either gas or liquid form including natural gas, crude oil, high
vapour pressure products such as propane and rened products such as gasoline or jet fuel are hazardous
substances [1,2]. There is always the chance that pipelines could leak or rupture and a pipeline failure can
cause serious human, environmental and nancial losses [2,3]. Most, if not all, buried pipelines that have
been in service for ve or more years experience numerous corrosion and metallurgical defects, in particular

*
Corresponding author. Tel.: +1 403 220 4149; fax: +1 403 282 8406.
E-mail addresses: shipilov@enme.ucalgary.ca (S.A. Shipilov), lemayi@metallurgicalconsulting.net (I. Le May).
1
Tel.: +1 306 934 9191; fax: +1 306 933 1814.

1350-6307/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2005.07.008
1160 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

cracks (Fig. 1) [2,48]. The sources of these cracks can be due to randomly distributed defects induced by
the manufacturing process or the degradation of carbon steels used as components of pipelines. The com-
bined action of stress (e.g., hoop and/or residual) and natural soil environment containing varying amounts
of moisture and oxygen further facilitates the initiation of the crack(s) and accelerates their propagation
through the pipe thickness by a factor that may range up to many times [9]. During pipeline operation,
the cracks can grow from an initial size to a critical size large enough to cause leakage (especially in
thin-walled pipelines) or spontaneous fracture (especially in thick-walled pipelines) [10]. This means that
the lifetime of a pipeline containing cracks or colonies of cracks (Fig. 1) on its surface is determined by
the time needed for a subcritical crack(s) to grow from an initial small size to the critical size. In other
words, the lifetime of such a pipeline is determined by the rate of crack propagation.

1.2. Stress corrosion cracking in pipelines

On March 4, 1965, a 610-mm natural gas pipeline owned by Tennessee Gas Pipeline ruptured near
Natchitoches, Louisiana, USA killing 17 people [11]. This incident was the rst publicly documented pipe-
line failure attributed to stress corrosion cracking (SCC). Prior to that time, it was generally assumed that
soil environments would not be capable of producing SCC [12]. But in the early 1970s SCC was identied as
an industry-wide concern [11,12]. In the USA between 1965 and 1986, SCC was reported to have caused
over 250 pipeline failures [6]. When the pipes were removed from the ground, the following common char-
acteristics were noted in the appearance of SCC: cracks initiated from the soil side; cracks ran parallel to the
longitudinal axis of the pipes, i.e. normal to the direction of maximum applied stress; SCC took place in the
absence of general or obvious pitting corrosion; cracks were not preferentially associated with the welded
seams; cracks occurred most frequently on the bottom portion of the pipe circumference; average failure
times were between about 15 and 20 years, but the rst failures occurred in pipelines that had been in ser-
vice for more than 40 years [4,1214]. One of the most unique characteristics was the appearance of SCC on
pipelines without corrosion protection as well as on pipelines with good protective coating and cathodic
protection (CP) [4,12,14]. Pipelines that have failed because of SCC, whether coated or bare, have done
so at times when CP was being applied. The great majority of failures occurred on coated pipelines with
potentials near the regions of failures ranging from 770 to 1920 mVSCE (vs. a saturated calomel elec-
trode, SCE). The few bare pipelines that have failed were in operation for 9 to 16 years without CP and
failed in 1 to 17 years after CP was applied [14].

Fig. 1. Colony of stress corrosion cracks on the pipeline surface exposed to a near-neutral-pH soil environment [8]. (Reprinted by
permission, Marr Associates.)
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1161

SCC has been also recognized as the cause of pipeline failures in Argentina, Australia, Canada, Iran,
Iraq, Italy, Pakistan, Saudi Arabia, The Netherlands, and countries of the former Soviet Union [2,15
24]. For example, in 1994, about 22% of the failures of gas pipelines in Russia were caused by SCC. Be-
tween 1995 and 1997, SCC caused nearly half of the failures on Russias gas pipelines of 1240 to 1420
mm diameter; and at present SCC is the cause of nearly 75% of the pipeline failures due to corrosion
[20,23]. As an example, the corrosion-induced failure of the nearly 20-year-old Komineft pipeline near
Usinsk in the Komi region of Russia in 1994 spilled enough oil to contaminate about 100 sq km and
the cost of replacing the worn out and leak-prone sections of this pipeline was expected to exceed $100 mil-
lion [25,26].
In 1994, the Canadian Energy Pipeline Association (CEPA), which represents the interests of 11 of
Canadas major oil and natural gas transmission pipeline companies, recognized SCC as a major challenge
facing the pipeline industry and established its SCC Working Group to share SCC experience and create a
proactive program to investigate, mitigate and prevent SCC on Canadian pipelines [27]. Two years later,
in 1996, Canadas National Energy Board (NEB) released a 158-page report on SCC in pipelines that said,
Stress corrosion cracking remains a serious concern for the pipeline industry [2]. According to this re-
port, since 1977, SCC has been the cause of 22 pipeline failures including 12 ruptures and 10 leaks in both
natural gas and liquid pipeline systems. As an example, Fig. 2 shows the site of the natural gas pipeline
failure that took place in 1994 near Williamstown, Ontario [2]. The Transportation Safety Board of Canada
testied [28,29] that between 1994 and 1996 there were other in-service SCC-related ruptures in pipelines
that the 1996 NEB report [2] did not mention. For example, on February 15, 1994 a rupture occurred
on the 1067-mm natural gas pipeline 35 km north of Maple Creek, Saskatchewan. The rupture was caused
by hydrogen-induced cracking (HIC) as the result of the diusion of atomic hydrogen at inclusions in the
pipe steel during normal pipeline operations [28]. Damage to the pipeline consisted of almost 22 m of rup-
tured pipe that had split open in the longitudinal direction. Associated with the failure was a re that
burned a pasture of approximately 8.5 hec. SCC also drew the blame for seven explosions on the TransCa-
nada PipeLines system between 1985 and 1995 [30]. The rupture, followed by an explosion and re, which
occurred on the 864-mm natural gas pipeline 10 km southwest of Winnipeg, Manitoba on April 15, 1996

Fig. 2. Pipeline failure site near Williamstown, Ontario, Canada, October 6, 1994 [2]. (Reprinted by permission, National Energy
Board.)
1162 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

was assisted by the existence of an environmentally assisted crack [29]. In 1998, the Alberta Energy and
Utilities Board (EUB) issued an Information Letter wherein it stated that the EUB expects all pipelines
licensees to evaluate the extent of SCC on their pipelines, take appropriate measures to deal with SCC,
and collect relevant data [31]. In 2002, an addendum to the SCC Recommended Practices [32] was pre-
pared and made public by CEPA; the addendum was based in large part on the SCC report issued by CEPA
in 1997 [27]. According to the above reports, most of the SCC-related failures in Canada were found in
pipelines that were installed between 1968 and 1973. These failures were associated with transgranular
SCC, resulting from the presence of dilute ground waters containing free carbon dioxide (CO2) with a
pH in the range of 6.57.5 [2,14]. But intergranular SCC in pipelines can also occur in relatively concen-
trated bicarbonate HCO  2
3 or bicarbonate HCO3 carbonate CO3 solutions with a pH in the range
of 9.512.5 and has been experienced in several countries throughout the world since 1965 [4,11,12,1416].
A signicant amount of money is being spent on trying to keep pipelines operable in face of the problem
identied above. In 1996, for example, 13 Canadian companies, all members of CEPA, planned to spend
$4.8 million on applied research and over $30 million on pipeline maintenance related to the SCC problem.
Between 1985 and 1996, TransCanada PipeLines alone spent $202 million on its SCC management pro-
gram [2]. Thus, it is apparent that the problem costs the industry at large many millions of dollars per an-
num. The large sum of money is due to the fact that the average direct cost of a rupture on a large diameter
natural gas pipeline is approximately $1.5 million. The cost of selective pipe replacements is about $1400
per meter for 508-mm diameter pipes and about $3200 per meter for 914-mm diameter pipes [2]. The aver-
age length of the rupture of natural gas pipelines is from about 20 m for a 720-mm diameter pipe to about
60 m for a 1420-mm diameter pipe [19]. It was estimated [33] that direct nancial loss from oil leaked from
the pipeline rupture is only 1.5% of the total loss, which also includes losses due to forced outages of the
pipeline, eld and petroleum processing equipment, as well as damage to agriculture, forests, the environ-
ment, etc.
Despite considerable eort and material expenditures spent in the last four decades on pipeline failure
prevention and on research, there are always new reports of pipeline failures due to SCC [7,3446]. Such
reports testify that the problem necessitates a re-evaluation of current approaches and a shift to new ones
if a practical and workable solution(s) is to be found. In this work, undertaken as a step towards changing
the current practice in pipeline failure prevention and in research, some features of the environmentally as-
sisted cracking (EAC) problem as related to the pipeline industry including the eect of CP on SCC and
corrosion fatigue in pipeline steels are considered in outline. Each of the following paragraphs discusses
one of the principal EAC-related features that were not considered in ocial reports released in the past
decade, including the 1996 NEB report [2]. These features were also not addressed during recent pipeline
industry technical meetings, including the Stress Corrosion Cracking Workshop [45] that was organized
to address the problems posed by SCC in operating gas and hazardous liquid pipelines by the Oce
of Pipeline Safety of the US Department of Transportation, the National Association of Pipeline Safety
Representatives (NAPSR) and major pipeline industry trade groups in December 2003 and the Ban/
2005 Pipeline Workshop on Managing Pipeline Integrity Removing the Hurdles that took place in
April 2005 in Ban, Canada.

2. Eectiveness of CP under disbonded coatings

The practice of using CP, alone or in conjunction with organic protective coatings, is generally recog-
nized as the most eective and technically appropriate corrosion prevention methodology for buried pipe-
lines [47]. However, there are a number of problems with this practice [48]. One is that CP combined with a
protective coating could be eective but only if the coating fails. Conversely, if the coating remains intact,
CP has little eect and even does not make much sense as there can be no contact between the soil (which
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1163

must be moist) and the surface of the pipe [49]. Observations made in eld studies at many locations indi-
cated that most if not all EAC failures on pipelines occurred with coated pipes and cracks were initiated
under disbonded coating areas [2,4,5,15,3436,5052], i.e. at the same place where the positive (i.e., protec-
tive) eect of CP was expected; all these failures occurred on buried pipelines having CP. Marr et al. [8]
wrote in 2003, the failure of a coating system is a primary factor in the initiation and propagation of
SCC. In most cases where it was possible to study the condition of the coating surrounding cracks, the
coating apparently was not bonded to the pipe in a small area around the crack but the coating did not
necessary appear to be broken. In a few rare cases, cracking has been found under a coating that did
not appear to have disbonded areas. In such cases, groundwater must have penetrated the coating through
microscopic pores [4]. Field inspections [52] showed also that usual CP monitoring measurements do not
detect true potentials beneath disbonded coatings or at holidays in pipe coatings and that the potentials
of the pipe surface exposed through defects and holidays in the coatings vary signicantly [8].

3. Service pressure uctuations and EAC-related terminology

All engineering structures including oil and natural gas pipelines are exposed to service stresses of vary-
ing amplitudes that usually are a mixture of stochastic and deterministic components [53,54]. In pipelines,
cyclic loading arises from two principal sources: (1) approximate daily pressure uctuations during normal
operation that are of the order of 10% of the nominal operating pressure; and (2) shutdowns and startups
for regular service or as a result of an upset condition in which the pressure decreases to zero or practically
zero (i.e., complete unloading) and then is increased to the nominal operating pressure [55]. As illustrated in
Fig. 3, daily pressure uctuations in a liquid pipeline correspond to cyclic loading with the stress ratio
(R = minimum/maximum load) of about 0.60.8 at a cyclic frequency from about 25 cycles per day
[54]. Under other circumstances, the stress ratio can vary from zero to 0.9 and a cyclic frequency can be
in the range between 0.7 and 1.3 cycles per day [53]. This means that corrosion fatigue (or, more correctly,
environmentally assisted fatigue cracking [56]), rather than SCC, should be considered as the failure mode
responsible for the subcritical crack growth in pipelines. SCC cannot be the cause of pipeline failure under
in-service conditions and as a phenomenon is only a special case of corrosion fatigue crack growth (FCG)
at R = 1. SCC can be obtained only in a laboratory when a specimen is tested under sustained or

Fig. 3. A 20-day pressure prole for a liquid pipeline [54]. (Reprinted by permission, National Energy Board.)
1164 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

monotonically increasing load. Despite this obvious fact, much eort and material expenditures have been
directed at studying SCC. Much less is known about corrosion FCG behaviour in pipeline steels.
Also, it is necessary to note that the terms SCC and corrosion fatigue are not semantically exact as
many environments that are not corrodents in the ordinary sense can induce intensive cracking of metals
[57]. Contrastingly, other environments that are strong corrodents may be more benecial to engineering
structures than ambiguous atmospheres [56]. Since the 1960s, many authors have questioned the terminol-
ogy used in EAC-related research and publications, such as the use of the term SCC to identify EAC phe-
nomena [56,58,59]. Technical journals normally use SCC as a general term to classify all modes of materials
fracture due to environmental (usually aqueous solutions of various compositions) factors. It was only as
recently as 2001 and 2002 [42,43] that, for the rst time, corrosion fatigue (not SCC) was mentioned as the
probable cause of pipeline failures.

4. Eect of CP on corrosion fatigue strength, SN curves

When smooth and notched (unawed) steel specimens were tested, cathodic polarization to 730 mVSCE
nullied the negative eect of aqueous solutions on the fatigue strength and practically restored the fatigue
limit observed in air [6064]. This potential is slightly positive (noble) to the generally referenced value to
mitigate general corrosion in natural waters, 780 mVSCE or 850 mVCSE (vs. a CuCuSO4 electrode,
CSE). Also, it was shown that the fatigue strength could be improved by polarization to more negative
potentials. For example, cathodic potentials in the range between 1100 and 1250 mVSCE increased
the fatigue strength of 1018 steel in seawater to approximately 100% above the fatigue limit observed in
air [62,63]. (Note: The representative chemical compositions and mechanical properties of the steels dis-
cussed in this article are given in Tables 1 and 2.) Fig. 4(a) shows that the lifetime of E36Z steel in seawater
increased as the applied cathodic potential became more negative than the open-circuit potential (OCP).
When the applied potential reached 850 mVSCE, the lifetime was equal to that obtained in air (i.e., the
steel was completely protected against corrosive action) [64,65]. Similar results were obtained in many

Table 1
Chemical composition of the steels (wt%)
Steel C Si Mn P S Cr Ni Mo Cu V Nb A1 Others Ref.
a
X-42 0.26 0.014 0.82 0.02 0.028 0.022 0.004 [55]
Fe 510 D 0.12 0.43 1.31 0.015 0.03 0.28 0.010 0.018 0.029 0.008 N [72]
E36Z 0.145 0.292 1.4 0.006 0.003 0.075 0.42 0.031 0.019 0.012 N [64,65]
X-52a 0.065 0.133 0.97 0.009 0.003 0.142 0.325 [81,92]
X-52a 0.145 0.345 1.3 0.014 0.0042 [83]
X-52a 0.16 0.35 0.99 0.019 0.015 0.01 0.01 <0.01 0.07 [85]
X-52 0.30 1.35 0.04 0.05 [80]
X-60 0.18 1.30 0.017 0.014 0.055 0.005 0.005 Ti [18]
X-65a 0.08 0.24 1.15 0.013 0.011 0.035 0.018 0.024 0.031 0.024 Ti [77,96]
X-65a 0.16 0.33 1.34 0.006 0.009 0.046 0.031 [9]
X-65a 0.26 1.4 0.03 0.03 [99]
X-70a 0.04 0.32 1.71 0.008 <0.005 0.05 0.12 0.02 0.30 0.09 0.054 0.034 0.016 Ti [70]
0.002 B
58 ppm N
X-70 0.06 0.1 1.7 0.024 0.01 0.01 0.2 0.30 0.15 0.04 0.01 [69]
X-70a 0.06 0.28 1.9 0.002 0.009 0.25 0.39 0.058 0.073 [68]
X-80a 0.035 1.74 0.010 0.005 0.111 0.254 [81,92]
X-80 0.18 0.20 0.30 0.018 0.013 1.68 2.99 0.41 0.005 [76]
a
Manufactured to API (American Petroleum Institute) specication 5L.
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1165

Table 2
Mechanical properties of the steels
Steel Yield strength (MPa) Tensile strength (MPa) Elongation (%) Reduction of area (%) Hardness (HRB) Ref.
a
X-42 311 490 21 52 [55]
Fe 510 D 362 516 35 [72]
E36Z 370 530 33 75 [64,65]
X-52a 400 548 20 72 87 [83]
X-52 358 455 [80]
X-65a 450 550 17 77 [77,96]
X-65a 458.5 570 32.3 [9]
X-70a 442 525 81.4 [70]
X-70a 484 556 78.9 [70]
X-70 475 645 27 [69]
X-70a 527.4 672.9 35.2 [68]
X-80 655 780 22 70 [76]
1015 473 522 77 [61]
a
Manufactured to API (American Petroleum Institute) specication 5L.

10-5
Fatigue crack growth rate, da/dN, m/cycle

0
E 36 Z
freely Rotating bending 10-6
corroding
Synthetic seawater
-500 pH = 8 6000 rpm
Applied potential, mVSCE

= 260 MPa
10-7
cathodic protection E 36 Z
-1000 Synthetic seawater
R = 0.1
0.2 Hz
10-8
-1500 open circuit potential
-1000 mVSCE
Air -1500 mV SCE

-2000 10-9
5 6 7 8
10 10 10 10 10 20 40 60 80 100
a Number of cycles b Stress intensity factor range, K, MPam

Fig. 4. The eect of cathodic potential on: (a) the corrosion fatigue lifetime; and (b) on the corrosion FCG rates for E36Z steel [65].
(Reprinted by permission, La Revue de Metallurgie.)

laboratories throughout the world and these results are still the basis for concluding that the application of
CP is an eective corrosion-prevention methodology for metallic structures (including oil and natural gas
pipelines [66]).
The experimental data showing the benecial eect of CP on the corrosion fatigue strength of carbon
steels such as those shown in Fig. 4(a) were obtained from reverse bend fatigue tests of smooth and/or
notched cylindrical specimens with a rotating speed from about 2006000 rpm. The type of loading and
its frequency are both not typical for realistic operating conditions. Moreover, tests of this type do not al-
low for separating the crack growth stage, which is of special practical interest the higher the rate of sub-
critical crack growth, the shorter the lifetime of the pipeline from the total corrosion fatigue lifetime that
includes the ve stages of material failure: nucleation of crack/defect(s); initiation of crack(s) from the nu-
cleus/defects; growth of small mutual cracks; growth of subcritical crack(s); and the nal (brittle) fracture
of material.
1166 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

5. Eect of CP on corrosion FCG rates, da/dN-data

It is known that the application of traditional methods of electrochemical protection may be unsuccess-
ful if an engineering structure or its components are loaded (e.g., cyclically loaded) under in-service condi-
tions. Depending on the nature of the metal, either anodic protection or CP may result in a multifold
acceleration of crack growth rates [67]. In 1975, Vosikovsky [9] showed that CP resulted in the acceleration
of corrosion FCG rate, da/dN, in X-65 pipeline steel exposed to 0.6 M NaCl solution (Fig. 5). Vosikovsky
used pre-cracked fracture mechanics specimens and carried out his tests over a frequency range between
0.01 and 10 Hz. He found that the lower the cyclic frequency, the higher the acceleration in corrosion
FCG rates observed under CP in comparison with crack growth rates in air. The maximum increase in
da/dN at lowest frequency (0.01 Hz) was 50 times at the cathodic potential of 1040 mVSCE and 10 times
under the open-circuit potential of 680 mVSCE, relative to da/dN found in air. There was no eect of the
test solution on da/dN at the frequency of 10 Hz [9]. Similar results were reported by other researchers
[6872] who explained the acceleration in corrosion FCG under CP conditions as due to hydrogen embrit-
tlement caused by enhanced hydrogen uptake in steels at impressed cathodic potentials. Schmelzer and Sch-
mitt [69] found that the hydrogen content within the fracture surface of X-70 steel after corrosion FCG tests
at 1100 mV vs. a saturated AgAgCl electrode was six to nine times higher than the hydrogen content
obtained after the test without CP (Fig. 6).
Evidence in support of a hydrogen embrittlement mechanism for the accelerated corrosion FCG rates in
pipeline steels under CP has been accumulating rapidly in recent years. The severity of hydrogen embrittle-
ment is highly dependent on the rate of hydrogen diusion into the region of high tensile triaxiality in the
metal matrix ahead of the crack. The maximum eect of the environment on crack growth will not be seen

Fig. 5. (a) FCG rates at the cathodic potential of 1040 mVSCE and four frequencies; and (b) comparison of FCG rates at the cathodic
potential of 1040 mVSCE and OCP and four frequencies for X-65 steel. The solid lines at cathodic potential and dashed lines at OCP
through growth rate data were dened by corresponding equations [9]. (Reprinted by permission, ASME International.)
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1167

Fig. 6. Hydrogen distribution within the corrosion FCG fracture surface of compact tension specimens of X-70 steel in a synthetic
seawater at: (a) the cathodic potential of 1100 mV (AgAgCl) and (b) at OCP; R = 0.22 [69]. (Reprinted by permission,
Materialwissenschaft und Werkstotechnik.)

unless sucient hydrogen has had time to accumulate in this region so that a critical combination of stress
state and hydrogen concentration is attained [73]. It was also shown [9,65,69] that slower loading frequency,
more negative potential and higher temperatures resulted in a substantially higher plateau crack growth
rate on the da/dNDK diagrams such as those shown in Figs. 4(b) and 5. As the cyclic frequency in-
creased, the rate of supply of absorbed hydrogen becomes limited because crack growth is faster on a time
basis [74].
1168 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

6. Discrepancy between two corrosion-fatigue test methods

Results given in Figs. 4(a) and (b) were obtained when E36Z steel was tested in aerated seawater [64,65].
Two dierent test methods were applied. The rst test method [64] used smooth cylindrical specimens tested
at high-frequency rotating bending (6000 rpm) with the stress level of 260 MPa (R = 1) while the second
test method [65] used fracture mechanicspspecimens with pre-cracks cycled under tension-tension loading
over a range of DK from 15 to 55 MPa m with a frequency of 0.2 Hz at R = 0.1. Fig. 4(a) shows that
CP did not have a hydrogen embrittlement eect on the steel tested under high-frequency conditions
and the lifetime of steel at potentials less than 850 mVSCE was equal to that obtained in air. Fig. 4(b)
shows that the same cathodic potentials that provided a full CP of steel when the rst test method was used
(e.g., less than 1000 mVSCE), resulted in signicant acceleration in FCG due to hydrogen embrittlement
when the second test method was used. Crack growth rates under CP conditions were signicantly higher
than those observed at OCP. As illustrated in Fig. 4(b), the shift of cathodic potential from 1000 to 1500
mVSCE caused the additional acceleration in the crack growth rate by factors of two to three at intermediate
DK levels. The discrepancy between the results was due to the fact that the lifetime determined by endur-
ance tests includes the crack initiation period and the growth period for small cracks when the stress inten-
sity factor, K, is small or undened.

7. Eect of gaseous hydrogen on FCG in pipeline steels

Existing natural gas and oil pipelines are constructed largely from CMn steels having yield strength
in the range 290450 MPa (4265 ksi). Newer pipelines are being constructed with somewhat stronger
micro-alloyed steels having yield strength in the range 480830 MPa (70120 ksi). These steels, by vir-
tue of their relatively low strength levels, are generally considered to be fairly immune to hydrogen
embrittlement under sustained and monotonically increasing loads [75]. However, under cyclic loading,
acceleration in crack growth due to absorbed hydrogen can occur in steels with relatively low strength
levels. For example, Cialone and Holbrook [55] demonstrated that hydrogen dramatically increased
FCG in X-42 steel relative to that in a nitrogen environment to varying degrees depending on the load-
ing conditions:p crack growth in hydrogen at R = 0.1 p was approximately 15 times faster at
DK 10 MPa m and 150 times faster at DK 20 MPa m (Fig. 7). The rate of FCG in X-80 steel
exposed to a low-pressure (345 kPa) dry hydrogen gas exceeded the crack growth rate in air by factors
of 10 to 30 depending on DK [76].

8. Slow strain rate tensile (SSRT) tests

In slow strain rate tensile (SSRT) tests [70], X-65 and X-70 steels showed a loss of ductility when
tested in aerated 0.6 M NaCl at potentials lower than 800 mVSCE. For example, in the test at the strain
rate of 9.7 107 s1, the reduction in area of X-70 steel decreased from about 80% (the value in air, at
OCP and at 800 mVSCE) to 10% at the potential of 1200 mVSCE (Fig. 8). This loss of ductility was
due to hydrogen-induced cleavage cracking which occurred as a result of the enhanced entry and trans-
port of hydrogen in the steels produced by continuous straining [70,77]. The loss of steel ductility under
CP conditions was not aected by pH over the range between 4 and 12 [78]. A similar eect of CP on the
ductility of carbon steels including X-52, X-60 and X-80 pipeline steels was observed over a range of
strain rates from 1.8 107 to 105 s1 when steels were tested in carbonatebicarbonate solutions
(Fig. 9), articial brackish ground water (0.2 M Na2SO4 + 0.5% NaCl) with pH 6, (0.010.1)N NaHCO3
with pH from 6 to 8.3, and simulated ground water (containing in g/l: 0.122 KCl, 0.483 NaHCO3, 0.181
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1169

Fig. 7. Fatigue crack growth rates for X-42 steel in 6.9 MPa hydrogen and 6.9 MPa nitrogen [55]. (From H.J. Cialone and J.H.
Holbrook, Metall. Trans. 16A (1985) 115. Reprinted by permission, Metallurgical and Materials Transactions.)

Fig. 8. The eect of cathodic potential on the reduction in area of X-65 and X-70 steels in 0.6 M NaCl [70]. (This material has been
reproduced from Fatigue and Crack Growth in Oshore Structures, ISBN 085298 598 3 and taken from a paper entitled
Environmental cracking (corrosion fatigue and hydrogen embrittlement) X-70 linepipe steel pp. 101108 by M. Saenz de Santa
Maria and R.P.M. Proctor with permission of the Council of the Institution of Mechanical Engineers.)
1170 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

Fig. 9. Eect of cathodic potentials (vs. CuCu2SO4) and strain rate on SCC of carbon steel in carbonate/bicarbonate solution [79].
(Reprinted, with permission, from STP 610 Stress Corrosion, New Approaches, copyright ASTM International, 100 Barr Harbor
Drive, West Conshohocken, PA 19428.)

CaCl2 2H2O, and 0.131 MgSO4 7H2O) with pH from 5.4 to 9.5 [18,7983]. (Note: The latter solution
has been used in many programs relating to transgranular SCC of pipeline steels and is often referred to
as NS-4 solution.) The results demonstrated that at normal CP potentials, plastic straining could lead to
hydrogen-induced SCC that conrmed the HIC failure mechanism of buried pipelines having CP.

9. Hydrogen embrittlement under in-service conditions

The above results show that the application of CP when pipeline failure is imminent is not always advis-
able because CP results in the steel becoming charged with hydrogen. When sucient hydrogen concen-
trates in pipeline steel, the metal loses its ductility a phenomenon known as hydrogen embrittlement.
In reported cases, after one to two years of service, more than a sixfold increase of hydrogen content oc-
curred in the external surface of pipes and, after 15 years, there was approximately a tenfold increase of
hydrogen content [84]. As such, it is hardly a coincidence that almost all failures in cathodically protected
pipelines were encountered in pipes installed for more than 10 years [2,14,85] and hydrogen-enhanced cor-
rosion FCG (not SCC) was the dominant mechanism of these failures. In 1982, White [86] noted that CP
could be the cause of the failure of sweet-gas transmission pipelines. Four years later, in 1986, the rst dis-
cussion [87] on the negative role of CP in pipeline integrity and reliability was organized by CANMET and
C-FER Atlantic. After that, it was determined that HIC was the dominant cause of buried cathodically
protected pipeline failures in Canada, Italy, Russia, The Netherlands and the USA between 1987 and
2000 [18,19,28,41,84,8891]. Recent studies of SCC of X-52 and X-80 pipeline steels in near-neutral pH
solution showed that SCC accelerated as the applied cathodic potential became more negative [80,81,92];
secondary ion mass spectrometry (SIMS) measurements showed that hydrogen diused into the steel dur-
ing SCC and concentrated around the crack tip, proving that crack growth at cathodic potentials was due
to the presence of occluded hydrogen in steels [92]. On February 20, 2002, during a panel discussion on
Cathodic Protection Myth-Conceptions held during the NACE International Northern Area Western Con-
ference in Edmonton, the subject Cathodic Protection Causes Hydrogen Embrittlement of Steel Pipelines
was voted as one of the 12 most important subjects to be considered by the Canadian pipeline industry. It
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1171

was a dening moment in the public acknowledgement of the need to address the problem of cathodically
induced hydrogen embrittlement of pipeline steels.

10. Near-neutral-pH SCC and high-pH SCC

In the literature, there is another point of view on the cause of pipeline failures and this point of view has
been the dominant one since the early 1970s [93]. According to this viewpoint, SCC is a result of: (1) pref-
erential dissolution at grain boundaries associated with concentrated HCO 
3 or HCO3 CO3
2
solutions
(intergranular crack growth); or (2) dissolution of metal at the crack tip and sides associated with dilute
ground waters containing CO2 (transgranular crack growth) [14]. However, it is agreed that the root cause
of the failure is a carbonatebicarbonate solution which develops at the pipe surface as a result of the ap-
plied cathodic protection [94]. Intergranular SCC in slightly alkaline soil environments with pH from 9.5
to 12.5 is usually called high-pH SCC (or classical SCC) while transgranular SCC in near-neutral-pH
soil environments with pH from 6.5 to 7.5 is called near-neutral-pH SCC (or non-classical SCC) [2,45].
But these terms, used by both industrial experts and novices, are misleading and cause some diculty for
professionals. Firstly, there is nothing classical about any type of SCC, including season cracking and
caustic embrittlement, which were rst investigated almost 100 years ago. Secondly, the scientic accuracy
of such terminology is questionable since the external pH (used in the above terms) diers signicantly from
the pH inside a stress corrosion crack. Thirdly, while the morphological (inter- or transgranular) distinction
is generally regarded as useful, it is also misleading when associated with pH levels. For example, carbon
steels had up to the late 1960s been observed to sustain SCC only intergranularly while they were tested in
aqueous solutions of dierent pH. In order to avoid any misunderstanding related to the above terms, some
background information on the pH level during SCC in pipelines and the causes of the dierent morphol-
ogies of crack growth during SCC is necessary.

10.1. The level of pH

There are two aspects related to the role of pH in SCC that need to be discussed. Firstly, under operating
conditions, ground water, whatever its pH, is likely to come into contact with the pipe surface once some
coating deterioration has occurred [4]. If an appropriate amount of cathodic current reaches the pipe sur-
face, as is the case with most stress corrosion cracked pipelines, the pH value at the pipe surface increases
from 6.5 to 9.5 or higher depending on the partial pressure of CO2 [14]. The time needed for such a time-
dependent pH change decreased with increasing current density. For solutions with a higher initial pH,
reecting their lower CO2 contents [14], the pH near the pipe surface increases rapidly to 11 or higher dur-
ing CP. That is, the pH at the surface of cathodically protected pipes is always more alkaline than the pH of
the bulk solution as hydroxyl ions are produced in the oxygen reduction reaction [14]. This higher level of
pH determines near-surface conditions required for the initiation of minute cracks or colonies of cracks
(Fig. 1) on the surface of cathodically protected pipes.
Secondly, as early as 1969, Brown et al. [95] showed that the local chemistry within the stress corrosion
crack is xed by the metal reactions and diers considerably from that of the bulk solution. In particular,
the carbon steel specimens indicated pH 3.8 near the advancing crack tip while the pH values of the bulk
solution ranged from 2 to 10. The acidity was the result of hydrolysis of the metallic constituent released by
anodic dissolution in the region of the crack tip [95]. This lower pH within stress corrosion cracks (rather
than pH of the bulk soil environment and/or pH of near-surface solution) determines crack-tip conditions
required for further crack propagation. Browns data [95] show also that the crack-tip thermodynamic con-
ditions for steels are favourable for hydrogen reduction during crack growth at OCP and not only under
CP.
1172 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

10.2. Crack growth morphology

In 1982, Hinton and Procter [96] conducted a quantitative fractographic study of corrosion FCG in X-65
steel exposed to 0.6 M NaCl with CP. They found that there was a denite change in the relative propor-
tions of intergranular separation, transgranular cleavage, brittle and ductile striations depending on DK
and frequency (Fig. 10). Intergranular crack growth along ferrite grain boundaries was replaced by trans-
granular crack growth with increasing DK. It was suggested that as DK increased, it became more favour-
able for cracks to follow the transgranular path because the stress prole extending ahead of the crack
becomes larger than the grain size. Furthermore, as DK increased, the amount of fresh surface created
at the crack tip during each cycle increased also; this allowed more hydrogen to enter the steel and hence
increased the concentration of hydrogen throughout the grains [96]. The observations [9699] that the fea-
tures of fracture surfaces depend on crack growth rates suggest that fracture surfaces transgranular crack
growth or intergranular crack growth should not be identied only with the level of pH.

10.3. Eect of CP on SCC in carbonatebicarbonate solutions

The carbonatebicarbonate hypothesis of intergranular and transgranular SCC of pipelines is perhaps


debatable, but the destructive eect of CP is not easily dismissed. As demonstrated in a series of experi-
ments [7981,83], CP (e.g., at 720 mVCSE as in Fig. 9) resulted in a more than threefold decrease in
the reduction in area of pipeline steels in HCO 2
3 CO3 comparable with that obtained under open circuit.
Fig. 11 shows that X-52 steel lost its ductility when tested in NS-4 solution at cathodic potentials less than
OCP of 750 mVSCE [82]. There is no doubt that this decrease in the ductility of steels was due to the eect
of hydrogen embrittlement [7983,100]. This conclusion is supported by experiments [101] in which the
application of 100 mV of cathodic polarization in the NS-4 solution without altering the test environment,
increased the hydrogen permeability in X-65 steel by about a factor of two. In addition, as was shown in
Section 10.1, HIC in ground water can occur at potentials apparently above those required for hydrogen
discharge from the bulk solution. This is associated with cracks developing from pits or crevices within
which the local environment is of appreciably lower pH than that of the bulk solution [95].

Fig. 10. Percentages of intergranular cracking along ferrite grain boundaries as a function of DK for X-65 steel in 0.6 M NaCl at
1000 mVSCE at various frequencies [96]. (Reprinted by Permission, Materials Forum.)
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1173

Fig. 11. Eect of cathodic potentials on stress-strain curves of X-52 steel tested in NS-4 solution [82]. (Reprinted by permission, The
Minerals, Metals and Materials Society.)

11. Concluding remarks

Large-diameter, high-pressure buried pipelines play an important role worldwide as a means of trans-
porting natural gas and oil over long distances from their source to consumers. With the problem of age
degradation, maintaining safe and reliable pipeline systems becomes an important issue because there is al-
ways the chance that pipelines could leak or rupture and a pipeline failure can cause serious human, envi-
ronmental and nancial losses.
The rst publicly documented pipeline failure attributed to SCC occurred in 1965. In the early 1970s
SCC was identied as an industry-wide concern in the USA and SCC has subsequently been recognized
as the cause of pipeline failures in many other countries. In the literature, there continue to be new
reports of pipeline failures due to SCC. Such reports testify that the problem necessitates a re-evalua-
tion of current approaches and a shift to new ones if a practical and workable solution(s) is to be
found.
In the literature, there is an opinion that CP restores the fatigue limit of pipeline steels and, in some
cases, the fatigue lifetime can be even higher under CP than in air. The opinion is based on the results
obtained from reverse bend fatigue tests of smooth and notched cylindrical specimens with a rotating
speed from 200 to 6000 rpm. However, these results are in conict with those from a fracture mechan-
ics standpoint, which clearly demonstrate that CP increases the crack growth rate, at least for the inter-
mediate levels of DK. A critical combination of cathodic potential, DK and R can accelerate the crack
growth by up to a factor of 50 compared with air data. The accelerating eect due to CP is usually
explained as due to hydrogen embrittlement caused by enhanced hydrogen uptake in the steel at im-
pressed cathodic potentials.
Mitigation of pipeline failures may necessitate the use of materials that are more expensive than plain
CMn steels, modication of CP systems, and application of relatively low operating temperatures and
pressure levels. In addition, coating improvements may be needed, particularly with respect to bonding.
The development of suitable corrosion inhibitors to be added to coatings and primers could become an
important consideration.
1174 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

Acknowledgements

The authors thank the Natural Sciences and Engineering Research Council (NSERC) of Canada for
nancial support of this work and acknowledge the assistance of Enoch Ng and Krishna Panchalingam
of the University of Calgary with the preparation of gures.

References

[1] Hopkins P. Transmission pipelines: how to improve their integrity and prevent failures. In: Denys R, editor. Pipeline technology.
Proceedings of the 2nd international pipeline technology conference, vol. 1. Amsterdam: Elsevier; 1995. p. 683706.
[2] Stress corrosion cracking on Canadian oil and gas pipelines. Report of the inquiry, National Energy Board, MH-2-95, Calgary,
November 1996.
[3] Shipilov SA. Technol Law Insur 1996;1:13142.
[4] Wenk RL. Field investigation of stress-corrosion cracking. In: Proceedings of the 5th symposium on line pipe research. No.
L30174. Arlington: AGA; 1974. p. T1T22.
[5] Le May I, Justice JT, Jamieson RM. Fracture prevention in pipelines. In: Le May I, Monteiro SN, editors. Fracture prevention
in energy and transport systems, vol. 1. Cradley Heath: Engineering Materials Advisory Services; 1984. p. 1324.
[6] Martinez FH, Staord SW. Pipe Line Ind 1994;77(7):2937.
[7] Ashworth B, Uzelac N. Oil Gas J 2001;99(18):6471.
[8] Marr JE, Hardy SB, Huuskonen E. SCC integrity management liquid and gas pipeline systems. Calgary: Marr Associates;
2003.
[9] Vosikovsky O. J Eng Mater Technol, Trans ASME 1975;97H:298304.
[10] Lorenz PM. Compatibility of tankage materials with liquid propellants. Technical Report AFML-TR-69-99. Air Force
Materials Laboratory, Ohio, May 1969.
[11] Final Sta Report on Investigation of Tennessee Gas Transmission Company Pipeline No. 100-1 Near Natchitoches,
Louisiana, March 4, 1965. Docket No. CP65-267. Federal Power Commission, Bureau of Natural Gas, Washington, DC,
August 12, 1965.
[12] Vrable J. Mater Protect Perform 1972;11(10):237.
[13] Townsend HE. Mater Protect Perform 1972;11(10):337.
[14] Parkins RN. Conceptual understanding and life prediction for SCC of pipelines. In: Proceedings of the CORROSION/96:
Research topical symposia. Houston: NACE International; 1996. p. 149.
[15] Kentish PJ. Brit Corros J 1985;20:13945.
[16] Hussain K, Shaukat A, Hassan F. Mater Perform 1989;28(2):135.
[17] Antonov VG, Baldin AV, Galiullin ZT, Grigorev PA, Krivoshchanova EM, Matvienko AF, et al. The study of the conditions
and causes of stress corrosion cracking in gas transmission pipelines. Moscow: VNIIEgazprom; 1991.
[18] Punter A, Fikkers AT, Vanstaen G. Hydrogen induced stress corrosion cracking of the R.A.P.L. oil transmission pipeline as a
result of the combined eect of cathodic protection and plastic deformation. In: Tiratsoo J, editor. Proceedings of the 9th
international pipe protection conference. London: Elsevier; 1991. p. 25769.
[19] Polyakov VN, Kharionovsky VV. Statistics of transmission pipeline fractures. In: Rossmanith HP, editor. Structural failure,
product liability and technical insurance. London: E&FN Spon; 1996. p. 35361.
[20] Kharionovsky VV, Tcherni VP. Stress and strain state of a gas pipeline in conditions of stress-corrosion. In: Yoon M, Mensik M,
Mohitpour M, editors. Proceedings of the international pipeline conference, vol. 1. New York: ASME; 1996. p. 47983.
[21] NEB unveils pipeline SCC study results. Oil Gas J 1997;95(1):25.
[22] Arrigoni B, Sinigaglia E. Transverse stress corrosion cracking in a landslide area. In: Proceedings of the 11th biennial PRCI-
EPRG joint technical meeting. Arlington: AGA; 1997.
[23] Steklov OI. Protect Metals 1999;35:3059.
[24] Otegui JL, Rivas A, Manfredi C, Martins C. Eng Fail Anal 2001;8:5773.
[25] Big spill reported in Komi region of Russian Arctic. Oil Gas J 1994;92(44):23.
[26] No large area of pollution seen likely from Russias Komi oil. Oil Gas J 1995;93(28):33.
[27] The CEPA report on circumferential stress corrosion cracking. Canadian Energy Pipeline Association (CEPA), Calgary: CEPA;
1997.
[28] 1994 pipeline investigation report. Report No. P94H0003. Transportation Safety Board of Canada, Hull; 1995.
[29] 1996 pipeline investigation report. Report No. P96H0012. Transportation Safety Board of Canada, Hull; 1999.
[30] Canadas NEB plans to probe pipeline leaks. Oil Gas J 1995;93(38):35.
[31] Stress corrosion cracking on pipelines. Alberta Energy and Utilities Board. Information Letter IL 986; 1998.
S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176 1175

[32] CEPA stress corrosion cracking recommended practices addendum on circumferential SCC. Canadian Energy Pipeline
Association (CEPA), Calgary: CEPA; 2002.
[33] Metodika otsenki uscherba ot otkazov obektov magistralnogo nefteprovoda. RD-39-30-107-78. Ufa: VNIISPTneft; 1979. p. 225.
[34] Justice RH, Mackenzie JD. Progress in the control of stress corrosion cracking in a 914-mm OD gas transmission pipeline. In:
Proceedings of the NG-19/EPRG 7th biennial joint technical meeting on line pipe research. AGAs Pipeline Research
Committee. Arlington: AGA; 1988 [paper no. 28].
[35] Delanty B, OBeirne J. Oil Gas J 1992;90(24):3944.
[36] Krishnamurthy RM, MacDonald RW, Marreck PM. Stress corrosion cracking of a liquid transmission line. In: Yoon M,
Mensik M, Mohitpour M, editors. Proceedings of the international pipeline conference, vol. 1. New York: ASME; 1996. p.
4959.
[37] 1995 pipeline investigation report. Report No. P95H0036. Transportation Safety Board of Canada, Hull; 1997.
[38] Miller SE, Gardiner MA, Ward CR. Oil Gas J 1998;96(39):905.
[39] 1996 pipeline investigation report. Report No. P96H0008. Transportation Safety Board of Canada, Hull; 1999.
[40] Carroll LB, Madi MS. Oil Gas J 2001;99(19):546.
[41] 2000 pipeline investigation report. Report No. P00H0037. Transportation Safety Board of Canada, Hull; 2001.
[42] Veazey MV. Mater Perform 2001;19(9):1920.
[43] 1999 pipeline investigation report. Report No. P99H0021. Transportation Safety Board of Canada, Hull; 2002.
[44] Wang J, Atrens A. Eng Fail Anal 2004;11:318.
[45] Stress Corrosion Cracking (SCC) Workshop. Oce of Pipeline Safety. US Department of Transportation, December 2003.
Available from: http://primis.rspa.dot.gov/gasimp/mtg_120203.htm.
[46] Zhigletsova SK, Rodin VB, Rudavin VV, Rasulova GE, Alexandrova NA, Polomina GM, Kholodenko VP. Change of
physicochemical parameters of soils near to stress-corrosion defects of gas pipelines. In: Shipilov SA, Jones RH, Rebak RB,
Olive J-M, editors. Environment-induced cracking: prediction, industrial developments and evaluation. Oxford: Elsevier; 2005.
[47] Heidersbach RH. Cathodic protection. Metals handbook. 9th ed. Corrosion, vol. 13. Materials Park: ASM International; 1998.
p. 46677.
[48] Shipilov SA. Critical assessment of the role of cathodic protection in pipeline integrity and reliability. In: Flewitt PEJ et al.,
editors. Engineering structural integrity assessment: needs and provision. Sheeld: EMAS; 2002. p. 15562.
[49] Payer JH, Fink KM, Perdomo JJ, Rodriguez RE, Song I, Trautman B. Corrosion and cathodic protection of disbonded
coatings. In: Yoon M, Mensik M, Mohitpour M, editors. Proceedings of the international pipeline conference, vol. 1. New
York: ASME; 1996. p. 4717.
[50] Fessler RR. Pipe Line Ind 1976;44(3):379.
[51] Elshawesh F, El Houd A, Elshushan M. Oil Gas J 2000;98(48):504.
[52] Toncre AC. Mater Perform 1984;23(8):227.
[53] Gutman EM, Amosov BV, Khudiakov MA. Neft Khoz 1977(8):5962.
[54] CEPA. Submission to the National Energy Board. In: Proceedings of the MH-2-95, vol. 1(2); 1996 [quoted in [2]].
[55] Cialone HJ, Holbrook JH. Metall Trans A 1985;16A:11522.
[56] Shipilov SA. Corrosion fatigue. In: Varvani-Farahani A, editor. Advances in fatigue, fracture and damage assessment of
materials. Southampton: WIT Press; 2005.
[57] Speidel MO. Stress corrosion cracking and corrosion fatigue fracture mechanics. In: Speidel MO, Atrens A, editors. Corrosion in
power generating equipment. New York: Plenum Press; 1984. p. 85130.
[58] Brown BF. Mater Res Standards 1966;6:12933.
[59] Staehle RW. A point of view concerning mechanisms of environment-sensitive cracking of engineering materials. In: Swann PR,
Ford FP, Westwood ARC, editors. Mechanisms of environment sensitive cracking of materials. London: The Metals Society;
1977. p. 574601.
[60] Minami Y, Takada H. Boshoku Gitjutsu (Corros Eng) 1958;7:3367.
[61] Duquette DJ, Uhlig HH. Trans ASM 1968;61:44956.
[62] Hooper WC, Hartt WH. Corrosion 1978;34:3204.
[63] Hartt WH, Hooper WC. Corrosion 1980;36:10712.
[64] Charbonnier JC, Margot-Marette H, Truchon M. Stress corrosion and corrosion fatigue of weldable steels in marine
environment. Metallic corrosion, vol. 2. Frankfurt/Main: Dechema; 1981. p. 131520.
[65] Truchon M, Rabbe P. Mem etud Sci Rev Metall 1983;80:11730.
[66] Skritskii RR. Prot Metals 1993;29:33743.
[67] Shipilov SA. JOM 2005;57(3):3642.
[68] Vosikovsky O. J Test Eval 1980;8:6873.
[69] Schmelzer F, Schmitt FJ. Z Werkstotech 1981;12(3):906.
[70] Saenz de Santa Maria M, Procter RPM. Environmental cracking (corrosion fatigue and hydrogen embrittlement) X-70 linepipe
steel. In: Fatigue and crack growth in oshore structures, C137/86. London: IMechE; 1986. p. 1018.
1176 S.A. Shipilov, I. Le May / Engineering Failure Analysis 13 (2006) 11591176

[71] Monahan CC, Hopkins RM. The relative severity of natural and synthetic seawaters on the fatigue behaviour of cathodically
protected steel. In: Scott P, editor. Environment assisted fatigue, EGF7. London: Mechanical Engineering Publications; 1990.
p. 97122.
[72] Maahn E, Nielsen HP. Inuence of cathodic protection on the initiation and growth of small cracks in steel under fatigue loading
in seawater. In: Scott P, editor. Environment assisted fatigue, EGF7. London: Mechanical Engineering Publications; 1990. p.
395414.
[73] Troiano AR. Trans ASM 1960;52:5480.
[74] Nakasa K, Takei H, Kajiwara K. Eng Fract Mech 1981;14:50717.
[75] Groeneveld TP, Elsea AR. Hydrogen-stress cracking in natural-gas-transmission pipelines. In: Bernstein IM, Thompson W,
editors. Hydrogen in metals. Metals Park: ASM; 1973. p. 72737.
[76] Clark Jr WG. The eect of hydrogen on the fatigue crack growth rate behavior of HY-80 and HY-130 steels. In: Bernstein IM,
Thompson W, editors. Hydrogen in metals. Metals Park: ASM; 1973. p. 14962.
[77] Hinton BRW, Procter RPM. Corros Sci 1983;23:10123.
[78] Kasahara K, Isowaki T, Adachi H. Study on hydrogen-stress cracking susceptibilities of line pipe steels. Metallic corrosion, vol.
1. Frankfurt/Main: Dechema; 1981. p. 3949.
[79] Payer JH, Berry WE, Boyd WK. Constant strain rate technique for assessing stress-corrosion susceptibility. Stress corrosion
new approaches, ASTM STP 610. Philadelphia: ASTM; 1976. p. 8293.
[80] Rebak RB, Xia Z, Safruddin R, Szklarska-Smialowska Z. Corrosion 1996;52:396405.
[81] Gu B, Yu WZ, Luo J, Mao X. Corrosion 1999;55:3128.
[82] Bosch C, Bayle B, Magnin T, Longaygue X. Proposal for a critical test to classify the SCC resistance of materials. In: Moody NR
et al., editors. Hydrogen eects on material behavior and corrosion deformation interactions. Warrendale: TMS; 2003. p.
58796.
[83] Gonzalez JJ, Meja L. CORROSION/2003. Houston: NACE International; 2003 [paper no. 526].
[84] Mazel AG. On stress corrosion cracking of gas pipelines. Reliability and safety of gas pipelines subject to stress corrosion
cracking, STP 24. Moscow: RAO Gazprom; 1993. p. 1122.
[85] Parry PJ. A case study of failure analysis and subsequent magnetic particle inspection of a natural gas pipeline. In: Revie RW,
Mak DK, Champion CS, Patchett BM, editors. Proceedings of the international conference on pipeline inspec-
tion. Ottawa: CANMET; 1984. p. 599612.
[86] White WE. Hydrogen induced failure and cathodic protection. In: Simpson LA, editor. Failure problems and solutions in the
energy industry. Oxford: Pergamon Press; 1982. p. 5161.
[87] Vosikovsky O, Leewis K, editors. Cathodic protection: A + or  in corrosion fatigue? Ottawa: CANMET; 1986.
[88] Kiefner JF, Eiber RJ. Oil Gas J 1987;85(13):98100.
[89] Kiefner JF, Eiber RJ. Oil Gas J 1987;85(15):3842.
[90] Chiovelli SC, Dorling DV, Glover AG, Horsley DJ. Oil Gas J 1994;92(11):91100.
[91] Cabrini M, Pistone V, Sinigaglia E, Tarenzi M. Oil Gas J 1998;98(10):615.
[92] Gu B, Luo J, Mao X. Corrosion 1999;55:96106.
[93] Fessler RR, Groeneveld TP, Elsea AR. Stress-corrosion and hydrogen-stress cracking in buried pipelines. In: Staehle RW et al.,
editors. Stress corrosion cracking and hydrogen embrittlement of iron base alloys. Houston: NACE; 1977. p. 13546.
[94] Beavers JA, Christman TK, Parkins RN. Mater Perform 1986;27(4):226.
[95] Brown BF, Fujii CT, Dahlberg EP. J Electrochem Soc 1969;116:2189.
[96] Hinton BRW, Procter RPM. Metals Forum 1982;5:8091.
[97] Frandsen JD, Marcus HL. Scripta Metall 1975;9:108994.
[98] Masuda H, Mastsuoka S, Nishijima S, Shimodaira M. Boshoku Gitjutsu (Corros Eng) 1986;35:2734.
[99] Shipilov SA. Fatigue Fract Eng Mater Struct 2002;25:24359.
[100] Delafosse D, Bayle B, Bosch C. The role of crack-tip plasticity, anodic dissolution and hydrogen on the SCC of CMn and low-
alloy steels. In: Shipilov SA, Jones RH, Rebak RB, Olive J-M, editors. Environment-induced cracking: prediction, industrial
developments, and evaluation. Oxford: Elsevier; 2005.
[101] Beavers JA, Durr CL, Garrity KC. CORROSION/2002. Houston: NACE International; 2002 [paper no. 426].

View publication stats

También podría gustarte