International Journal of Pressure Vessels and Piping 199 (2022) 104742

Contents lists available at ScienceDirect

International Journal of Pressure Vessels and Piping

journal homepage: www.elsevier.com/locate/ijpvp

A new method for assessment of burst pressure capacity of corroded X80

steel pipelines containing a dent
Jian Zhao a, Yunrong Lv b, Y. Frank Cheng a, *
a
Department of Mechanical & Manufacturing Engineering, University of Calgary, Calgary, Alberta, T2N 1N4, Canada
b
Guangdong Provincial Engineering Technology Research Center of Petrochemical Corrosion and Safety, Guangdong University of Petrochemical Technology, Maoming,
Guangdong, China

A R T I C L E I N F O A B S T R A C T

Keywords: This work developed a new finite element (FE)-based method to perform pressure capacity assessment on a
Pipelines corroded X80 steel pipe containing a dent, where their mutual interaction and the effect on further corrosion
Corrosion defect growth were defined. Parametric effects including corrosion depth, dent depth, corrosion length and service time
Dent
were determined. The presence of a dent at corrosion defect remarkably changes the local stress distribution,
Pressure capacity
Finite element modeling
elevating stress to plastic range. The mutual interaction between the dent and the corrosion defect depends on
both the initial corrosion depth and the initial dent depth. The interaction remarkably changes the local stress
level at the dent and must be included in pipeline burst failure assessment.

1. Introduction methods and criteria suffer from some limitations. First, validation ex­
periments were mostly performed on low-grade steels such as X52 steel
Pipelines provide a safe, effective and economic means to transport [2,28], while high-grade steels such as X80 steel are rarely tested for
natural gas to meet industry and civilian needs [1–5]. Increasing energy modeling purpose. Second, according to the mechano-electrochemical
demands today require the pipelines to operate under high internal (M-E) interaction theory [9,10,29–33], a plastic strain can remarkably
pressures [6,7]. High-strength steels such as X80 steel are increasingly affect the corrosion rate of pipeline steels. However, the available
used in gas pipeline transportation [8–11]. During operation in the field, assessment models do not consider the stress-corrosion interactions.
the integrity of pipelines can be compromised by various threats, such as Finally, most investigations to determine the burst pressure were limited
corrosion, dents and cracks [12–15]. Development of accurate defect to pipelines containing corrosion defects only [34,35]. For example, a
assessment techniques is crucial to integrity management of the pipe­ recent work investigated the effect of internal corrosion defect on burst
lines [2,16–18]. failure of pipelines [36], while other types of defects such as dent and its
After a long time of service in the field, pipelines usually contain both interaction with corrosion were not considered.
corrosion and dent defects. Corrosion defects such as pits and general Dent is a common defect present on the outer surface of pipelines due
corrosion are metal-loss damages, causing thinning of pipe wall [11,19]. to mechanical impact of rocks, excavation operations or damage of an­
The primary consequence of corrosion damage is to decrease pressure chors [4,15,37,38]. Of various assessment methods for dented pipelines
capacity of the pipelines [20,21]. Existing standards and criteria for [39,40], the strain-based and the depth-based criteria are commonly
assessment of corroded pipelines were reviewed by Qin and Cheng [2], used. The burst pressure of a dented pipeline is barely affected by plain
and most of the standards have been using in industry [22,23]. For dents [15,38,41]. At the same time, dents are usually adjacent to or
example, Mondal and Dhar [19,22] developed burst pressure formulas overlapped with corrosion defect. The dent area is also susceptible to
based on finite element (FE) modeling and experimental testing for crack initiation due to a local stress concentration [42]. The combina­
determination of burst pressure of corroded pipelines. The formulas are tions of a dent and metal-loss defects such as scratch and gouge were
acceptable to industry for pipeline failure prediction. The geometrical investigated [15,40,43]. It was found that the dent did not affect
parameters of corrosion defects were considered in assessment of the pressure-bearing capacity of the pipeline containing a gouge or a scratch
pressure capacity of corroded pipelines [16,24–27]. However, these [44]. Corrosion is also a metal-loss defect. The corrosion is unique in

* Corresponding author.
E-mail address: [email protected] (Y.F. Cheng).

https://doi.org/10.1016/j.ijpvp.2022.104742
Received 20 April 2022; Received in revised form 30 June 2022; Accepted 2 July 2022
Available online 8 July 2022
0308-0161/© 2022 Elsevier Ltd. All rights reserved.
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

terms of its continuous growth with time. When a dent is generated in a experimental testing. Mechanical properties of X65 and X80 steels are
corrosion area, the dent and the corrosion are evaluated separately by listed in Table 1. The R–O relationship curves of X65 and X80 steels are
the methods used today [39,45]. The dent should meet the critical depth shown in Fig. 1.
or the strain-based criterion for it to be acceptable to pipeline integrity. For modeling of the denting process, a rigid material model was used
Similarly, the corrosion defect is assessed independently using relevant for the indenter. The Poisson ratio was 0.3 for all materials used in the
criteria as mentioned above. However, a mutual interaction exists be­ work.
tween them, but was barely considered in the previous assessments. Sun
et al. [46] investigated the interaction between a corrosion defect and an 2.2. Modeling for defects on steel pipes
adjacent dent on an X46 steel pipeline, where the overlapped dent with
corrosion defect, which is more commonly observed in practice, was not Both corrosion and dent defects were considered in FE modeling to
considered. Moreover, the effect of the dent on further corrosion growth investigate the combined effect on burst failure of the X80 steel pipe. A
was not included in the existing assessment methods because the models FE model for the steel pipe containing corrosion defect only was also
did not consider the M-E interaction on pipelines. Generally, both developed to validate the model with existing testing results included in
external and internal corrosion of pipeline at a dented area are both literature. Both pressurizing the steel pipe to burst and denting processes
affected due to a local stress concentration and the resulting M-E are elastic-plastic deformations and nonlinear processes. An 8-nodes 3-D
interaction. While this work focuses on the interaction of a dent with element Solid185 was used for FE modeling in the software ANSYS for
external corrosion, for the inner surface, the steel is mainly under a both the pipe segment and the indenter. A dent was generated on a
tensile stress [36], which strongly enhances the corrosion rate. corrosion defect present on the pipe segment. Due to the symmetrical
The novelty of this work is to develop a new FE-based method to nature, a quarterly model was used to save calculational time. The
perform pressure capacity assessment on corroded pipelines containing schematic diagram of the quarterly portion of a steel pipe containing a
a dent, where the mutual interaction between the corrosion defect and corrosion defect combined with a dent are shown in Fig. 2, where D, L
the dent, as well as its effect on corrosion growth with time, are defined. and t are pipe outer diameter, length of the pipe segment and pipe wall
A three-dimensional (3-D) FE model was developed to simulate the thickness, respectively, d, l and w are depth, longitudinal length and
processes of denting and burst failure of a corroded high-strength X80 circumferential width of the corrosion defect, respectively, h is depth of
steel gas pipeline. Parametric effects, including corrosion depth, dent the dent center, which is also the center of the corrosion defect, and x, y
depth and corrosion length, on local stress distribution and burst pres­ and z are circumferential, radial and axial directions, respectively.
sure of the pipe were determined. The stress-enhanced corrosion growth The FE model included two parts where different element sizes were
due to the dent-corrosion defect interaction was quantitatively evalu­ applied. The stress-strain condition of the defect area was the target,
ated. A burst assessment program for corroded steel pipelines containing where fine elements were assigned with a size of 2 mm. Relatively coarse
a dent was proposed for integrity management of the pipelines. elements with a size of 10 mm were assigned on the remaining part of
the pipe segment. The element size of the indenter was same as the size
2. Algorithms for finite element analysis of the defect area. The rigid properties were applied for the indenter, and
the element type for the indenter was also Solid185. It is realized that
2.1. Modeling for pipeline steels and their mechanical properties the rigid surface element is also effective for modeling. However,
application of surface elements in a solid model is much complicated
The material used in this work was an X80 pipeline steel. There has and is not used in the work. Meshing of the FE model is shown in Fig. 3,
been so far limited investigation to define the mutual effect between a and similar grid size settings were also used in Ref. [18]. The grid
corrosion dent and a dent on an X80 steel pipe. To verify modeling re­ sensitivity test confirmed that the selected element sizes were proper for
sults obtained from this work, experimental measurements in the liter­ FE analysis. When considering the indentation process on the corroded
ature were used. Moreover, an X65 steel was also selected for modeling pipe, the contact/target elements were established in the contact area
to provide additional confirmation in terms of the modeling reliability. between the indenter and the pipe.
Generally, models used for FE analysis for steels include the bilinear Geometrical parameters of the model were consistent with experi­
stress-strain relationship [22], isotropic hardening model [47], true mental settings in literature [51,52]. The outer diameter (D) of X65 steel
stress-strain data [48], and Ramberg-Osgood (R–O) relationship [1]. As pipe was 762 mm and the pipe wall thickness (t) was 17.5 mm. The
suggested by API 579-1/ASME FFS-1 2016 [49], the R–O model was depth (d) of the corrosion defect were 8.75 mm, and its lengths (l) were
used for both X65 and X80 steels in this work. 100 mm and 200 mm, respectively. For X80 steel pipe, the outer
For X65 steel, the R–O model is described as [50]:
( )n ( )
σ σ σ
ε= + α (1)
E σy E

For X80 steel, the R–O model is [50]:

( )12.64
σ σ
ε = + 0.079 (2)
E σu

where ε is engineering strain, σ is engineering stress, E is Young’s

modulus, σ y and σu are yield strength and ultimate tensile strength,
respectively, and α and n are material parameters determined from

Table 1
Mechanical properties of X65 and X80 steels (Choi et al., 2016; Tian et al., 2019;
Freire et al., 2006; Sun and Cheng, 2018).
Steels E (GPa) σy (MPa) σu (MPa) α n

X65 206.7 449 531 3.686 10.64

Fig. 1. Stress-strain curves of the R–O model used for X65 and X80 pipe­
X80 200 601 684 – –
line steels.

2
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

Fig. 2. Schematic diagram of a quarterly portion of a steel pipe containing a corrosion defect combined with a dent.

pipe performance under different internal pressures. Downward and

upward displacement loads were applied on the indenter to create a dent
on the pipe. During burst testing, the pipe was filled with water and a
hydrostatic pressure continually increased until the pipe failed. End caps
were installed at both ends of the pipe segment. The effect of end caps
should be approximated in the model, and thus, an equivalent pressure,
pe, was applied on the end surface of the model. An axial tension, pe, can
be calculated by Refs. [4,5]:

(D − 2t)2
pe = p⋅ (4)
D − (D − 2t)2
2

Fig. 3. Element meshing of the FE model used for different parts of the A widely accepted failure criterion in FE modeling was selected in
pipe segment. this work [1]:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
1
diameters included 459 mm and 457 mm, with pipe wall thicknesses of σ Mises = √̅̅̅ (σ1 − σ 2 )2 + (σ 2 − σ3 )2 + (σ3 − σ 1 )2 > σ u (5)
8 mm and 8.1 mm. The corrosion defect was 40 mm and 39.6 mm in 2
length, and 3.75 mm and 5.4 mm in depth, respectively. The edges of the The von-Mises equivalent stress, σMises, is compared with the ulti­
corrosion defect were rounded into a radius of 1.5 mm to simulate actual mate tensile strength, σu, to determine pipe failure. The failure criterion
corrosion morphology. used in this work was that the von-Mises stress of all elements along the
It is noted that the modeling results can be affected by the length (L) pipe thickness reached the ultimate tensile stress.
of the pipe segment. For FE modeling, if L is too small, the boundary
conditions applied on both ends of the pipe will influence the stress and 2.3. FE modeling procedure
strain distributions at the defect area. To eliminate this effect, the
minimum length of the modeled pipe segment, Lmin, should meet the The Newton-Raphson (N-R) method was used for nonlinear analysis.
requirement [53]: The processes for modeling a corroded pipe in the absence and presence
of a dent were different. For burst of the corroded pipe without a dent,
l d √̅̅̅̅̅̅̅̅̅
Lmin = + D⋅t⋅l (3) the FE analysis was conducted by:
2 t
It is seen that Lmin is determined by geometrical parameters of the (1) The FE model was established with specific geometric parameters
pipe and of the corrosion defect. With inputs of relevant parameters used of the pipe and of the corrosion defect. The R–O material models
in this work, the pipe segment lengths of 2000 mm and 4000 mm were for X65 and X80 steels were applied on the pipe.
selected for X65 and X80 steel pipes, respectively. Based on calculations (2) Symmetric boundary conditions were applied on the model. The
in this work, the X65 steel pipe requires a minimum length of 627.39 internal pressure was applied on the inner surface of the pipe.
mm (for a corrosion length of 100 mm) and 916.55 mm (for a corrosion (3) Nonlinear large deformation and load steps were set, and the
length of 200 mm). The selected pipe segment length of 2000 mm meets pressure applied on the pipe was increased until failure occurred.
the requirement of Eq. (3). In addition, as analyzed by FE, this pipe The burst pressure was calculated.
segment is sufficiently long that the pipe segment length effect can be
eliminated. For corroded X80 steel pipe containing a dent, the indentation pro­
It was suggested [22] that the circumferential width (w) of corrosion cess was simulated to obtain a good approximation to the reality. The
defect did not affect the burst pressure of corroded pipelines. As a result, indenter was applied on the corrosion area and four steps were followed.
w was kept constant with the same value as the value in literature. The
dent was created on the corrosion defect by an indenter shaped of a (1) The FE models of the corroded pipe and the indenter were
sphere with a diameter 80 mm. The distance of downward movement of developed with proper input parameters. The contact/target el­
the indenter influenced the dent depth (h). ements were used in the contacting surface between the indenter
To approximate with the experimental settings of burst tests in and the corrosion defect.
literature, the boundary conditions of the FE model were consistent with (2) Symmetric boundary conditions and motion constraint in y di­
the tests. As shown in Fig. 2, the symmetric surface of the model was set rection were applied on the pipe. Pre-determined downward
as symmetric boundary conditions, which were marked as “s” in the displacement loads were applied on the indenter and compres­
figure. The displacement in y direction of the bottom line was restrained sion happened. At this stage, the internal pressure p was set as 0.
to avoid rigid body motion. The internal pressure, p, was applied on the (3) The indenter was removed upward and a spring-back at the dent
inner surface of the pipe, and sufficient load steps were set to analyze the area was observed.

3
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

(4) An internal pressure p was then applied on the inner surface of the
pipe until a non-convergence occurred. The burst pressure of the
pipe was determined.

3. Results

3.1. Verification of FE modeling results

To verify the FE modeling results, burst testing results obtained on

corroded pipelines in open publications were used as references [51,52].
Burst pressures of X65 and X80 steel pipes measured in literature and the
burst pressures modeled by FE analysis in this work are shown in
Table 2, where both X65 and X80 steel pipes are used to show the
applicability of the FE model on both low- and high-grades of pipe steels.
Moreover, the geometric parameters of the FE model were set as iden­
tical to those used in tests in literature. It is seen that the errors between
the test measurements and the modeling results are small, with a
maximum value of 3% only. It is thus thought that the FE model is
acceptable for pipeline burst analysis.

3.2. Burst pressures of X80 steel pipe containing corrosion defect and dent Fig. 4. Burst pressures of X80 steel pipe containing corrosion defect with
various depths and a dent with varied depths obtained by FE modeling.
The burst pressure of X80 steel pipe was analyzed with various
corrosion depths (d) and dent depths (h), while other parameters were
kept constant, i.e., D = 457 mm, t = 8.1 mm, L = 4000 mm, and w = 32
mm. Fig. 4 shows the burst pressure results obtained by FE modeling.
The dent depths are selected to reflect a continuous change from a
shallow dent to a deep dent. The corrosion depth of 2–6 mm was selected
to be representative of general corrosion defect present on pipelines.
When a corrosion defect is too shallow such as less than 2 mm, it does
not generate sufficient adverse effect on pipelines. If the corrosion defect
is too deep, the operators should repair the defect for safety reason. A
gradual increment in corrosion defect depth is to investigate the effect of
corrosion defect on pipeline burst failure. It is seen that, at specific dent
depths, the burst pressure of the pipe decreases as the corrosion depth
increases. However, the dent depth does not affect burst pressure of the
corroded steel pipe at individual corrosion depths. Particularly, when Fig. 5. An example of the von-Mises equivalent stress distribution at a corro­
the dent does not exist (i.e., a dent depth of zero), the burst pressure is sion defect with a depth of 5.4 mm (i.e., 0.67t) on an X80 steel pipe, where
greater than the dent is present on the pipe. The results are same as those there is no dent present, operating at a pressure p = 10 MPa [8].
obtained on corroded low-grade steel pipes containing metal-loss defects
[37,54], and the conclusions are thus convincing. Generally, denting is a except the corner edge where a stress concentration is developed. At the
large elasto-plastic deformation process. The dent area is strengthened bottom of the corrosion defect, the greatest equivalent stress is 460 MPa,
by a strain hardening mechanism [44]. Dent is not a metal-loss, and indicating that the defect is in elastic deformation. Even at the corner
there is a limited effect on pressure-bearing capacity of the pipelines. As edge, the local stress is up to 550 MPa, which is close to yielding but still
a comparison, corrosion causes a local thinning of pipe wall. As a result, at an elastic deformation, while the yielding strength of X80 steel is 601
the pressure capacity of the pipe decreases at the local area. MPa.
When a dent is generated on the corrosion area, the local stress
distribution is changed. A spring-back process occurs after the indenter
3.3. Effects of corrosion defect and dent on stress distribution on the pipe
is removed. To distinguish different stages during denting, various
initial corrosion depths (d0) and the maximum initial dent depths (h0), i.
Fig. 5 shows an example of the von Mises equivalent stress distri­
e., the dent depth prior to the spring-back, are selected. The instant dent
bution at a corrosion defect with a depth of 5.4 mm (i.e., 0.67t) on an
depth (h) refers to the dent depth after the spring-back occurs, which is
X80 steel pipe, where there is no dent present, operating at a pressure p
resulted from removal of the indenter and application of an internal
= 10 MPa [8]. The stress level at the corrosion area is relatively low,

Table 2
Comparison and verification of burst pressures of X65 and X80 steel pipes measured in experimental testing in literature [51,52] and modeled by FE analysis in this
work.
Steels Pipeline geometry, mm Corrosion defect geometry, mm Burst
Pressure, MPa

Diameter (D) Pipe wall thickness (t) Length (L) Length (l) Depth(d) Width (w) Testing ptest Modeling pFE Difference
%

X65 762 17.5 2000 100 8.75 260 24.68 24.94 1.0
X65 762 17.5 2000 200 8.75 260 22.48 22.11 1.7
X80 457 8.1 4000 40 5.40 32 22.70 23.45 3.2
X80 459 8 4000 40 3.75 32 24.2 25.38 4.6

4
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

pressure on the pipe. Figs. 6 and 7 show distributions of von Mises MPa as the dent depth increases from 15 mm to 40 mm. It is thus seen
equivalent stress of a corroded X80 steel pipe (operating pressure: 10 that the mutual interaction of corrosion defect and dent depends on the
MPa) after spring-back of indentation with different values of h0 at the depth of the corrosion defect. At a deep corrosion defect (e.g., 5.4 mm in
initial corrosion depths of 5.4 mm and 2.0 mm, respectively. The two depth), the stress level at the corrosion defect increases with increased
corrosion defect depths are selected because they are located at both initial dent depth. However, if a dent is generated at a shallow corrosion
ends of the investigated depth range. Generally, at least 3 cycles of in­ defect (e.g., 2.0 mm in depth), the local stress decreases as the initial
ternal pressure, i.e., equivalent to 72% of specified minimum yield dent depth increases. When the corrosion defect is relatively deep, a
strength (SMYS) of the pipe steel, are applied to produce rebounding of a tensile stress exists at the defect bottom and increases with increased
dented pipe to consider a residual dent depth at the end of the process, dent depth. The local stress further increased under the effect of internal
which is assumed not affected by further internal pressure to be applied pressure. For shallow corrosion defects, with the effect of spring-back, a
[55]. Internal pressure acts to reduce the dent depth, and this reduction compressive stress at the corrosion defect bottom gradually replaces the
is a nonlinear deformation process. A better comparison between tensile stress along with the increase in dent depth. As a result, the
different initial dent depths (h0) could be attained considering the tensile stress at the corrosion defect bottom decreases as the dent depth
rebounding process. It is seen that, at the corrosion depth of 5.4 mm, the increases. Furthermore, shallower corrosion defects tend to present a
stress at the corrosion defect increases with increased h0. For example, higher local stiffness. The dent depths after spring-back (h) are expected
the equivalent stress at the center of the corrosion defect increases from to be lower when compared with deeper corrosion defects with the same
437 MPa to 684 MPa when the initial dent depth h0 increases from 15 value of h0. Similarly, deeper corrosion defects have a lower local
mm to 40 mm. When the initial dent depth h0 is more than 30 mm (i.e., stiffness and, for the same initial indentation, have higher dent depths
0.066D), the bottom of the corrosion defect is in plastic deformation, after spring-back. Higher stresses are thus expected to be found in
although both the longitudinal and circumferential edges of the corro­ deeper corrosion defects.
sion defect are still in the elastic stage. However, at a small corrosion
depth such as 2.0 mm, the stress at the bottom of the corrosion defect
3.4. Effect of dent on corrosion defect growth on pipeline
decreases with the increasing h0. As seen in Fig. 7, the equivalent stress
at the center of the corrosion defect decreases from 680.6 MPa to 463.8
Corrosion is electrochemical in nature. The corrosion rate can be

Fig. 6. Distribution of von Mises equivalent stress of a corroded X80 steel pipe (operating pressure: 10 MPa) after spring-back of indentation with different values of
h0 at an initial corrosion depth of 5.4 mm (a) h0 = 15 mm, (b) h0 = 20 mm, (c) h0 = 25 mm, (d) h0 = 30 mm, (e) h0 = 35 mm, (f) h0 = 40 mm.

5
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

Fig. 7. Distribution of von Mises equivalent stress of a corroded X80 steel pipe (operating pressure: 10 MPa) after spring-back of indentation with different values of
h0 at an initial corrosion depth of 2.0 mm (a) h0 = 15 mm, (b) h0 = 20 mm, (c) h0 = 25 mm, (d) h0 = 30 mm, (e) h0 = 35 mm, (f) h0 = 40 mm.

expressed as anodic dissolution rate of steels. The presence of stress can that varies over time T, Y0 is the time that corrosion initiates, λ and β are
enhance corrosion reaction, and the enhancement factor, kσ, is defined proportionality and exponent constants, respectively. As verified pre­
as [14]: viously [58,59], constants λ and β are determined by Y0, pH, soil
chemistry, pipeline coatings, etc. To make the modeling corrosion
Rσct=0
kσ = (6) growth rate results comparable with published work, the values of the
Rct
parameters are selected: λ = 0.164, β = 0.780 and Y0 = 2.88, which are
identical to those in the work [59]. It is believed that variations in the
where Rσct = 0 and Rσct are charge-transfer resistance, which is inversely
parameters will change the corrosion growth model, and thus, provide
proportional to corrosion rate, in the absence and presence of stress (σ ),
different prediction results. In this work, the commonly used power
respectively. By measuring the values of Rσct = 0 and Rσct of pipe steel in a
model is used. In the further work, the authors will explore and verify
corrosive environment, the stress-enhanced corrosion can be quantita­
the reliability of other corrosion growth models. The depth of the
tively determined. Xue and Cheng [56] made electrochemical mea­
corrosion defect d0(T), (in the absence of stress enhancement effect) can
surements on X80 steel under various applied stresses in a bicarbonate
be determined by integrating the corrosion rate a0 over time T:
solution. It was found that kσ is approximately 1.8 when the stress is
elastic, representing the stress-enhanced corrosion factor. When the d0 (T) = d0 + λ(T − Y0 )β (8)
stress is in plastic stage, the corrosion enhancement becomes more
apparent. The specific value of the factor depends on the stress level. where d0 is initial corrosion depth. When a stress is applied, the corro­
Considering conservatism in engineering assessment, a factor 2.2 is used sion growth rate a and corrosion depth d(T) are calculated as
in this work.
There are some models to simulate corrosion growth, such as the a = kσ a0 = kσ λβ(T − Y0 )β− 1
(9)
single value corrosion growth rate model, linear growth model, power-
law model, etc. [57]. The power-law corrosion rate model, as showed d(T) = d0 + kσ λ(T − Y0 )β (10)
below, is widely accepted, and used in pipeline corrosion growth [58]: According to Fig. 6, for a corroded pipe containing a 5.4 mm deep
a0 = λβ(T − Y0 )β− 1
(7) corrosion defect and a dent with an initial depth more than 30 mm (i.e.,
h0 > 0.066D), the stress concentration is usually located on the bottom
where a0 is corrosion rate in the absence of the stress enhancement effect of the corrosion defect, which is also the site bearing with the plastic
stress with the assigned kσ = 2.2. The corrosion growth rates of the

6
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

defect bottom and the sides are determined as kσa0 and a0, respectively.
For the circumferential edge of a corrosion defect, the stress is relatively
small compared with the longitudinal edge and the bottom of the defect.
In addition, as analyzed, the effect of the defect width on the burst
pressure is negligible. Thus, the corrosion rate is fixed as a0 for simpli­
fication purpose. As a result, the ratio between the length and depth (i.e.,
l/d) of the corrosion defect experiences changes during corrosion pro­
cess. Similar analyses are also conducted on corrosion defects with
different initial depths, and the results of corrosion growth rate are listed
in Table 3, where the mutual interaction between corrosion defect and
dent on further corrosion growth at the defect associated with specific
stress levels is quantitatively determined.
The pressure capacity of corroded pipelines assessed in ASME B31G
is determined by corrosion length l and depth d. According to the power-
law corrosion rate model in Eq. (7), the time dependence of the longi­
tudinal length l(T) of the corrosion defect can be expressed as:

l(T) = l0 + kσ λ(T − Y0 )β (11)

where l0 is initial longitudinal length of corrosion defect. In this work, l0

= 39.6 mm. The results of l/d vs. time obtained by FE modeling with
Fig. 8. The ratio of longitudinal length of the corrosion defect to its depth, i.e.,
different initial corrosion defect depths are shown in Fig. 8. The initial l/d, vs. time obtained by FE modeling with different initial corrosion
corrosion depths of 2.0 mm and 5.4 mm are selected to reflect two defect depths.
typical corrosion scenarios. When the corrosion defect is deeper (such as
5.4 mm in initial depth), the deceasing rate of l/d with time is small,
⎡ ⎤
meaning that the depth growth is a bit faster than the length growth. As
the dent depth increases, the ratio reduces. However, for a shallow ⎢
⎢ ( )⎥ ⎥
corrosion defect such as 2.0 mm in initial depth, the deceasing rate of l/ 2t 2t ⎢1 − 0.85 dt ⎥
⎢
PASME = SF = Sflow ⎢ ⎥ (12)
⎥
d with time is much greater, meaning that the depth growth is acceler­ ⎢ 1 − 0.85( t ) ⎥
d
D D
⎣ ⎦
ated with time. However, as the dent depth increases, the ratio increases.
M

The dependence of the corrosion growth rate on depths of corrosion

defect and dent are well consistent with the stress modeling results in
where M is Folias factor, and Sflow (MPa) is flow stress which is equal to
Figs. 6 and 7.
(σy+69) [2]. For short corrosion defects (i.e., z < 50, where z = l2/Dt and
l, D and t have units of mm):
4. Failure assessment of pipelines containing both corrosion
defect and dent (
M = 1 + 0.6275z − 0.003375z2
)1/2
(13)

4.1. Comparison of the developed FE model with existing assessment For z > 50 (i.e., long corrosion defects),
methods for corroded pipelines
M = 0.032z + 3.3 (14)

Presently, the pressure capacity of natural gas pipelines containing In DNV RP F101 standard, the failure pressure, PDNV, is calculated as
corrosion defects is assessed by standards such as ASME B31G [60] and [2]:
DNV RP F101 [61]. The common characteristic of these methods is that ⎡ ⎤
the effect of corrosion defect is treated as a damage factor. According to ⎢ ( )⎥
the modified ASME B31G, the failure pressure, PASME (MPa), of a 2t 2t
⎢
⎢1 − dt ⎥
⎥
corroded pipelines is calculated by: PDNV = SF = Sflow ⎢
⎢ d ⎥
⎥ (15)
D− t D− t ⎢ 1 − (t) ⎥
⎣ M ⎦

where Sflow = 1.05σ u [19], and

√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
l2
M = 1 + 0.31⋅ (16)
Dt
Table 3 If the failure pressure calculated by Eq. (12) or (15) is smaller than
Calculated corrosion growth rates of X80 steel pipe containing a corrosion defect
operating pressure (p) multiplied by a safety factor SF, the pipelines are
with various initial depths and a dent with various initial depths.
required for repair or replacement. Apparently, the calculated failure
Initial Initial dent Corrosion growth rate (a, mm/y) pressure of corroded pipelines depends on length and depth of the
corrosion depth (h0,
Defect Longitudinal Circumferential corrosion defect, while the corrosion defect growth vs. time is ignored.
depth (d0) mm)
bottom site site It is generally believed [19,31] that the modified ASME B31G
d0 ≤ 0.67t h0 ≤ 30 2.2a0 2.2a0 a0 method is more conservative for estimation of the burst pressure, as
(0.066D) shown in Table 4. However, the error of the modified ASME B31G is
h0 > 30 a0 2.2a0 a0 much greater than the error of DNV RP F101. Thus, further in­
d0 > 0.67t h0 ≤ 30 a0 a0 a0
vestigations are conducted using DNV RP F101 as a base for comparison
(0.066D)
h0 > 30 2.2a0 a0 a0 to improve the assessment accuracy.

Note: a0 is corrosion growth rate of X80 steel in near-neutral pH solutions

without considering the stress enhancement effect.

7
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

Table 4 time for a corroded X80 steel pipe with various initial corrosion depths
Comparison of the burst pressures obtained by FE model developed in this work and dent depths. It is seen that the failure pressure decreases with time,
as compared with the results from the modified ASME B31G and DNV RP F101 as expected. The dent depth affects remarkably the failure pressure. For
methods. deep corrosion defects such as 5.4 mm in depth, the failure pressure is
d (mm) l Burst pressure, P (MPa) lower at specific times, compared with the pipe containing a corrosion
(mm)
FE Modified Difference DNV Difference defect of 2 mm in depth. Particularly, for pipelines containing deep
model ASME (%) RP (%) corrosion defects, the co-presence of a deeper dent (i.e., over 30 mm in
B31G F101 depth) makes the failure pressure decreases much more rapidly with
2 39.6 24.438 23.071 5.6 25.420 3.9 time, as compared with the pipe containing a dent with the depth
3 39.6 24.310 22.598 7.0 25.037 2.9 smaller than 30 mm. However, as shown in Fig. 9, an opposite tendency
4 39.6 23.917 21.986 8.1 24.489 2.3 is observed for the pipelines containing shallow corrosion defects due to
5 39.6 23.471 21.159 9.9 23.638 0.7
different stress-enhanced corrosion rates as presented in Table 3.
6 39.6 22.993 19.981 13.1 22.135 3.9
After the steel pipe is dented in the absence of an internal pressure
Note: The input parameters of the corroded pipeline used for FE modeling and the indenter is removed, a spring-back process happens due to
include D = 457 mm, t = 8.1 mm. The pipe steel is X80 steel. elasticity of the material. When an internal pressure is applied on the
pre-dented pipe, a further reduction of the dent depth will occur. Thus,
4.2. A new assessment method for burst pressure of corroded pipelines the measured dent depth is deviated from the initial depth. Although the
considering the effect of dent and service time spring-back behavior of dented pipelines has been investigated [15,18,
44], there has been rare work conducted on corroded pipelines. Ac­
As stated, the mutual interaction between corrosion defect and dent cording to the FE model in this work, the initial depth h0 and instant
must be included in failure assessment for an improved pipeline burst depth h of the dent can be obtained. Fig. 10 shows the relationship be­
prediction accuracy. A dent, when it is located at a corrosion defect, can tween the dent depth (h) under internal pressure (p) and the initial dent
obviously influence the stress condition at the defect area and then the depth (h0) of a corroded X80 steel pipe with various initial corrosion
corrosion growth. Thus, when the dent occurs at corrosion area on depths. It is seen that, as the initial corrosion depth increases, the dent
pipelines, the failure assessment should include both the pressure- depth decreases at specific initial dent depth h0. This means that the
bearing capability of the pipelines and the remaining service life due dent, upon spring-back, will become shallower if it is created at a deeper
to dent-enhanced corrosion growth on the pipelines. A new criterion is corrosion defect. Furthermore, the dent depth h increases linearly with
proposed for assessment of burst pressure capability and remaining the initial depth h0 by:
service life of corroded pipelines containing dent. For a corroded X80
steel pipe containing a dent, the pressure capacity can be determined by h = 0.435h0 − 1.115d0 − 1.209 (19)
the criterion considering length of corrosion, dent depth and service It is noted that the derived Eq. (19) is based on the FE modeling
time. The failure pressure, P(T), of the corroded pipe at a certain time, T, defined in this work. When the pipeline and defect parameters change,
is calculated by: the relationship will be re-derived.
⎡ ⎤
d(T)
2⋅t ⎢ 1− ⎥
P(T) = ⋅Sflow ⋅⎢
⎣1 −
t ⎥
d(T) ⎦
(17) 4.3. Implication on failure assessment of corroded pipelines containing
D− t t⋅M(T) dent

√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ In practice, most surface defects such as dents and corrosion on

l2 (T) pipeline are generally acceptable for continuous operation, while the
M(T) = 1 + 0.31⋅ (18)
Dt dent-corrosion interaction is usually ignored, especially the corrosion
Fig. 9 shows the relationship between failure pressure and service defect growth with time. Both the interaction and the corrosion growth

Fig. 9. Relationship between failure pressure P(T) and service time T for a Fig. 10. Relationship between dent depth (h) under internal pressure (p) and
corroded X80 steel pipe with various initial corrosion depths and dent depths. the initial dent depth (h0) on a corroded X80 steel pipe with various initial
The initial corrosion defect length is l0 = 39.6 mm. corrosion depths.

8
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

affect the pipeline’s pressure capability and remaining service life. Ac­ Declaration of competing interest
cording to the new assessment method developed in this work, the
following steps should be followed: The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
(1) Measure geometric parameters of corrosion defect and dent, and the work reported in this paper.
determine their initial depths based on empirical equation such
as Eq. (19). Data availability
(2) Estimate the initial pressure capacity, P(0), of the pipeline using
Eq. (15) before the corrosion defect starts to grow. If P(0) < SF × Data will be made available on request.
p, the pipeline may fail during service. The segment containing
the defects should be repaired or replaced. Acknowledgements
(3) If P(0) > SF × p, the pipeline can bear the operating pressure and
continue the service. However, the corrosion defect growth over This work was supported by Natural Science and Engineering
service time should be considered and estimated by the devel­ Research Council (NSERC) of Canada.
oped model. If the pressure capacity decreases and becomes
smaller than p after a certain time T, the pipeline should be References
properly maintained.
[1] G.J. Qin, Y.F. Cheng, Failure pressure prediction by defect assessment and finite
element modelling on natural gas pipelines under cyclic loading, J. Nat. Gas Sci.
Following the assessment steps (1)–(3), the failure pressure of the Eng. 81 (2020), 103445.
corroded pipelines containing a dent can be determined. Moreover, the [2] G.J. Qin, Y.F. Cheng, A review on defect assessment of pipelines: principles,
remaining service time can be estimated by determining the corrosion numerical solutions, and applications, Int. J. Pres. Ves. Pip. 191 (2021), 104329.
[3] C. Kim, L. Chen, H. Wang, H. Castaneda, Global and local parameters for
growth rate. It is realized that the remaining life of a pipeline is not only characterizing and modeling external corrosion in underground coated steel
affected by corrosion and dent, but also by other threats such as fatigue pipelines: a review of critical factors, J. Pipeline Sci. Eng. 1 (2021) 17–35.
and cracking. Thus, a proper pipeline integrity management program [4] Y. Shuai, D. Zhou, X. Wang, H. Yin, S. Zhu, J. Li, Y.F. Cheng, Local buckling failure
analysis of high strength pipelines containing a plain dent under bending moment,
should be composed of multiple modeling and assessment of various J. Nat. Gas Sci. Eng. 77 (2020), 103266.
threats present on the pipeline. [5] Y. Shuai, X. Wang, Y.F. Cheng, Buckling resistance of an X80 steel pipeline at
corrosion defect under bending moment, J. Nat. Gas Sci. Eng. 93 (2021), 104016.
[6] M. Witek, Possibilities of using X80, X100, X120 high-strength steels for onshore
5. Conclusions
gas transmission pipelines, J. Nat. Gas Sci. Eng. 27 (2015) 374–384.
[7] X.K. Zhu, A comparative study of burst failure models for assessing remaining
A new FE-based method is developed for assessment of the burst strength of corroded pipelines, J. Pipeline Sci. Eng. 1 (2021) 36–50.
pressure capability and remaining service life of corroded pipelines [8] Y.Z. Jia, J.Q. Wang, E.H. Han, W. Ke, Stress corrosion cracking of X80 pipeline steel
in near-neutral ph environment under constant load tests with and without
containing a dent, while considering mutual interaction of the corrosion preload, J. Mater. Sci. Technol. 27 (2011) 1039–1046.
defect and the dent. An assessment program containing specific steps is [9] L.Y. Xu, Y.F. Cheng, An experimental investigation of corrosion of X100 pipeline
proposed for practical use. steel under uniaxial elastic stress in a near-neutral pH solution, Corrosion Sci. 59
(2012) 103–109.
The presence of a dent at corrosion defect remarkably changes the [10] L.Y. Xu, Y.F. Cheng, Corrosion of X100 pipeline steel under plastic strain in a
stress distribution, elevating the local stress to plastic range. The mutual neutral pH bicarbonate solution, Corrosion Sci. 64 (2012) 145–152.
interaction between the dent and the corrosion defect depends on the [11] S.H. Zhang, Y.F. Cheng, X. Feng, Performance characteristics and technical
challenges of X80 pipeline steel, Oil Gas Store. Trans. 38 (2019) 481–495.
corrosion depth. At a deep corrosion defect (e.g., 5.4 mm in depth), the [12] Y.F. Cheng, Stress Corrosion Cracking of Pipelines, Wiley, New York, NY, US, 2013.
stress level at the corrosion defect increases with increased initial dent [13] J.L. Sun, Y.F. Cheng, Assessment by finite element modeling of the interaction of
depth. However, for a shallow corrosion defect (e.g., 2.0 mm in depth), multiple corrosion defects and the effect on failure pressure of corroded pipelines,
Eng. Struct. 165 (2018) 278–286.
the local stress decreases as the initial dent depth increases. [14] X. Tang, Y.F. Cheng, Micro-electrochemical characterization of the effect of applied
The mutual interaction between corrosion defect and dent affects the stress on local anodic dissolution behavior of pipeline steel under near-neutral pH
corrosion defect growth, and the effect depends on both the initial condition, Electrochim. Acta 54 (2009) 1499–1505.
[15] X. Tian, M. Lu, Y. Chen, Numerical simulation method for failure pressure of buried
corrosion depth and the initial dent depth. The corrosion grows more
dented pipeline, J. Fail. Anal. Prev. 20 (2020) 944–949.
rapidly along the depth direction than the longitudinal direction, espe­ [16] B.N. Leis, Evolution of metal-loss severity criteria: gaps and a path forward,
cially for shallow corrosion defects, e.g., 2.0 mm in initial depth. J. Pipeline Sci. Eng. 1 (2021) 51–62.
The mutual interaction between corrosion defect and dent must be [17] M.M. Shahzamanian, M. Kainat, N. Yoosef-Ghodsi, S. Adeeb, Systematic literature
review of the application of extended finite element method in failure prediction of
included in pipeline burst failure assessment. The failure assessment pipelines, J. Pipeline Sci. Eng. 1 (2021) 241–251.
should include both the pressure-bearing capability of the pipelines and [18] J. Zhao, Y.F. Cheng, A new criterion based on strain determination for dent
the remaining service life due to the dent-enhanced corrosion growth. assessment of pipelines, J. Pipeline Sci. Eng. 2 (2022) 18–28.
[19] B.C. Mondal, A.S. Dhar, Finite-element evaluation of burst pressure models for
The presence of a deep dent (i.e., >30 mm in depth) makes the burst corroded pipelines, J. Pressure Vessel Technol. 139 (2017), 021702.
pressure decrease more rapidly with time. [20] R. Amaya-Gómez, M. Sánchez-Silva, E. Bastidas-Arteaga, F. Schoefs, F. Muñoz,
Reliability assessments of corroded pipelines based on internal pressure - a review,
Eng. Fail. Anal. 98 (2019) 190–214.
Author statement [21] M. Xie, Z. Tian, A review on pipeline integrity management utilizing in-line
inspection data, Eng. Fail. Anal. 92 (2018) 222–239.
The first author, Jian Zhao, is a PhD student. He conducted all nu­ [22] B.C. Mondal, A.S. Dhar, Improved Folias factor and burst pressure models for
corroded pipelines, J. Pressure Vessel Technol. 140 (2018), 011702.
merical modeling work and performed the result analysis. He also [23] K.J. Yeom, Y. Lee, K.H. Oh, W.S. Kim, Integrity assessment of a corroded API X70
drafted the manuscript. The second author, Dr. Yunrong Lv, participates pipe with a single defect by burst pressure analysis, Eng. Fail. Anal. 57 (2015)
in result analysis and the manuscript revision. The corresponding 553–561.
[24] J. Bao, W. Zhou, Influence of depth thresholds and interaction rules on the burst
author, Dr. Frank Cheng, is the principal investigator of the project. He
capacity evaluation of naturally corroded pipelines, J. Pipeline Sci. Eng. 1 (2021)
planned the scope of work and designed the research plan. He worked 148–165.
with the first author to analyze the results and revised the manuscript for [25] B. Ma, J. Shuai, D. Liu, K. Xu, Assessment on failure pressure of high strength
submission. pipeline with corrosion defects, Eng. Fail. Anal. 32 (2013) 209–219.
[26] S. Zhang, W. Zhou, Assessment of effects of idealized defect shape and width on the
burst capacity of corroded pipelines, Thin Wall. Struct. 154 (2020), 106806.

9
J. Zhao et al. International Journal of Pressure Vessels and Piping 199 (2022) 104742

[27] S. Zhang, W. Zhou, Development of a burst capacity model for corroded pipelines [45] H.O. Heggen, J. Bratton, D. Kemp, J. Liu, J. Austin, Fitness for service of dents
considering corrosion defect width and a revised Folias factor equation, J. Nat. Gas associated with metal loss due to corrosion, in: Proc. 2014 10th Int. Pipeline Conf.,
Sci. Eng. 88 (2021), 103812. 2014. Calgary, Alberta, Canada.
[28] K.U. Amandi, E.O. Diemuodeke, T.A. Briggs, Model for remaining strength [46] J. Sun, Y.F. Cheng, J. Woo, M. Kainat, S. Hassanien, Assessment of interaction
estimation of a corroded pipeline with interacting defects for oil and gas between a dent and an adjacent corrosion feature on pipelines and the effect on
operations, Cogent Eng. 6 (2019), 1663682. pipeline failure pressure by finite-element modeling, J. Pipeline Syst. Eng. Pract. 12
[29] J.L. Sun, Y.F. Cheng, Modeling of mechano-electrochemical interaction between (2021), 4021029.
circumferentially aligned corrosion defects on pipeline under axial tensile stresses, [47] G. Qin, Y.F. Cheng, Modeling of mechano-electrochemical interaction at a
J. Petrol. Sci. Eng. 198 (2021), 108160. corrosion defect on a suspended gas pipeline and the failure pressure prediction,
[30] Y. Wang, L.Y. Xu, J. Sun, Y.F. Cheng, Mechano-electrochemical interaction for Thin-Walled Struct. 160 (2021), 107404.
pipeline corrosion: a review, J. Pipeline Sci. Eng. 1 (2021) 1–16. [48] B.C. Mondal, A.S. Dhar, Burst pressure of corroded pipelines considering combined
[31] L.Y. Xu, Y.F. Cheng, Reliability and failure pressure prediction of various grades of axial forces and bending moments, Eng. Struct. 186 (2019) 43–51.
pipeline steel in the presence of corrosion defects and pre-strain, Int. J. Pres. Ves. [49] American Petroleum Institute/American Society of Mechanical Engineers, Fitness-
Pip. 89 (2012) 75–84. For-Service, API 579-1/ASME FFS-1, 2016. Washington DC, US.
[32] L.Y. Xu, Y.F. Cheng, Development of a finite element model for simulation and [50] W. Ramberg, W.R. Osgood, Description of stress–strain curves by three parameters,
prediction of mechanoelectrochemical effect of pipeline corrosion, Corrosion Sci. in: Technical Note No. 902, National Advisory Committee for Aeronautics, 1943.
73 (2013) 150–160. Washington DC, US.
[33] L.Y. Xu, Y.F. Cheng, A finite element-based model for prediction of corrosion defect [51] J. Freire, R.D. Vieira, J. Castro, A.C. Benjamin, Burst tests of pipeline with
growth on pipelines, Int. J. Pres. Ves. Pip. 153 (2017) 70–79. extensive longitudinal metal loss, Exp. Tech. 30 (2006) 60–65.
[34] X.H. Zhang, B.Q. Chen, C.G. Soares, Effect of non-symmetrical corrosion [52] C. Oh, Y. Kim, J. Baek, Y. Kim, W. Kim, Ductile failure analysis of API X65 pipes
imperfection on the collapse pressure of subsea pipelines, Mar. Struct. 73 (2020), with notch-type defects using a local fracture criterion, Int. J. Pres. Ves. Pip. 84
102806. (2007) 512–525.
[35] B.Q. Chen, X.H. Zhang, C.G. Soares, The effect of general and localized corrosions [53] G. Fekete, L. Varga, The effect of the width to length ratios of corrosion defects on
on the collapse pressure of subsea pipelines, Ocean Eng. 247 (2022), 110719. the burst pressures of transmission pipelines, Eng. Fail. Anal. 21 (2012) 21–30.
[36] M. Fahed, I. Barsoum, A. Alfantazi, M.D. Islam, Burst pressure prediction of pipes [54] M. Zarea, R. Batisse, B. Leis, P. Cardin, G. Vignal, Full scale experimental database
with internal corrosion defects, J. Pressure Vessel Technol. 142 (2020), 031801. of dent and gouge defects to improve burst and fatigue strength models of
[37] E.R. Lancaster, S.C. Palmer, Burst pressures of pipes containing dents and gouges, pipelines, in: Int. Gas Union Res. Conf., ASME, Washington DC, US, 2011,
J. Proc. Mech. Eng. 210 (1996) 19–27. pp. 845–861.
[38] Y. Shuai, J. Shuai, X. Zhang, Experimental and numerical investigation of the strain [55] B.K. Alexander, Methodology for Assessing the Effects of Plain Dents, Wrinkle
response of a dented API 5L X52 pipeline subjected to continuously increasing Bends, and Mechanical Damage on Pipeline Integrity, Corrosion’2007, Paper No.
internal pressure, J. Nat. Gas Sci. Eng. 56 (2018) 81–92. 07139, 2007. Nashville, Tennessee, US, Mar. 11-15.
[39] M. Gao, R. Krishnamurthy, Mechanical damage in pipelines: a review of the [56] H.B. Xue, Y.F. Cheng, Electrochemical corrosion behavior of X80 pipeline steel in a
methods and improvements in characterization, evaluation, and mitigation, Oil near - neutral pH solution, Mater. Corros. 61 (2010) 756–761.
Gas Pipelines (2015) 289–326. [57] A. Valor, F. Caleyo, J.M. Hallen, J.C. Velázquez, Reliability assessment of buried
[40] K.F. Tee, A.H. Wordu, Burst strength analysis of pressurized steel pipelines with pipelines based on different corrosion rate models, Corrosion Sci. 66 (2013) 78–87.
corrosion and gouge defects, Eng. Fail. Anal. 108 (2020), 104347. [58] F. Caleyo, J.C. Velázquez, A. Valor, J.M. Hallen, Probability distribution of pitting
[41] M. Allouti, C. Schmitt, G. Pluvinage, Assessment of a gouge and dent defect in a corrosion depth and rate in underground pipelines: a Monte Carlo study, Corrosion
pipeline by a combined criterion, Eng. Fail. Anal. 36 (2014) 1–13. Sci. 51 (2009) 1925–1934.
[42] Z. He, W. Zhou, Fatigue reliability analysis of dented pipelines, J. Pipeline Sci. Eng. [59] J.C. Velázquez, F. Caleyo, A. Valor, J.M. Hallen, Predictive model for pitting
1 (2021) 290–297. corrosion in buried oil and gas pipelines, Corrosion 65 (2009) 332–342.
[43] P. Zhao, J. Shuai, M. Sun, Z. Lv, K. Xu, Y. Wang, Burst pressure of thin-walled pipes [60] American Society of Mechanical Engineers, Manual for Determining the Remaining
with dent and gouge defects, Thin-Walled Struct. 159 (2021), 107213. Strength of Corroded Pipelines, ASME B31G, Washington DC, US, 2012.
[44] M. Allouti, C. Schmitt, G. Pluvinage, J. Gilgert, S. Hariri, Study of the influence of [61] Det Norske Veritas, Corroded Pipelines, DNV-RP-F101, Norway, 2004.
dent depth on the critical pressure of pipeline, Eng. Fail. Anal. 21 (2012) 40–51.

10