Proceedings of IPC2008

7th International Pipeline Conference

Proceedings of IPC 2008
September 29-October 3,72008,
th
International PipelineCanada
Calgary, Alberta, Conference
September 29 - October 3, 2008, Calgary, Canada

IPC2008-64565
IPC2008 - 64565
STRAIN-BASED MODELS FOR DENT ASSESSMENT – A REVIEW
Ming Gao Rick McNealy Ravi Krishnamurthy
Blade Energy Partners Blade Energy Partners Blade Energy Partners
Houston, TX 77084 Houston, TX 77084 Houston, TX 77084

Iain Colquhoun
PII Pipeline Solutions, GE Infra, Oil & Gas, Chonburi, Thailand

ABSTRACT in the present paper are specific to ASME B31.8 – 2003 as

previously noted. Examination of these other criteria appears
The depth-based dent assessment criteria as required by to support the conclusions of the present paper.
various government regulations and industry codes have
served the pipeline industry well by requiring pipeline A detailed review of ASME B31.8 strain based methods is
operators to investigate dents with potentially severe presented in the Dent Study report by Baker [3]. Papers have
deformational damage. However, the use of depth criterion also been published to further address strain-based dent
alone can result in both unnecessary excavations and the assessment approach [5, 8-11], simplification of strain-based
mischaracterizing of otherwise severe dents. ASME B31.8- assessment method has been proposed for in-ditch application
2003 and PHMSA IMP Rules for gas pipelines affecting high [12], and significant progress has been made in more precisely
consequence areas offer an option for using strain-based calculating membrane and bending strain components and
criteria and B31.8 provides non-mandatory formulas for strain total or effective strain [5, 8-11]. It should be noted, however,
calculation. Due to complications in dent profile, other that considerable discrepancies in values of membrane strain
formulas in the open literature or those derived by qualified components and total strain were also found [10, 11].
engineers are also allowed by the ASME code. The present
paper reviews dent strain calculation methods that are In the present paper, a review of the currently available dent
currently available. Comparisons in the calculated strains strain calculation methods (mainly ASME 31.8 and
between methods are made using several case studies. Lukasiewicz et al) is given. Comparisons in the calculated
Discrepancies and their effect on the equivalent (or effective) strains between these methods are made and illustrated with
strain are identified. Improvements and alternative strain limit examples. Discrepancies and their effect on the equivalent
approaches are explored and discussed. (or effective) strain are identified. Improvements and
alternative strain limit approaches are explored and discussed.
INTRODUCTION
It should be noted that the present paper does not address
The depth-based dent assessment criteria required by dents with metal loss or stress concentrators such as gouges,
government regulations and industry codes have served the grooves, arc burns, or cracks. The paper addresses static dent
pipeline industry well by requiring pipeline operators to evaluation and does not discuss fatigue life of dents.
investigate dents with potentially severe deformational
damage [1-3]. However, the use of depth alone can result in ASME B31.8 BASED STRAIN CALCULATION
both unnecessary excavations due to the repair of many deep METHODS
dents and ovalities that are not necessarily harmful, and also
potentially miss of moderate dents that are in fact severe due Because of complexity of dent geometry, no standard methods
to their overall size and sharpness. Since metal damage due to for dent strain calculation exist to date. ASME B31.8 (2003)
offers relatively simple non-mandatory equations to estimate
plastic deformation in principle is related to strain, strain
levels derived from the dent profile offer another measure of the circumferential bending strain, the longitudinal bending
the severity of a plain dent. ASME B31.8-2003 offers an strain, and the longitudinal extensional strain using dent
option for using strain-based criteria and provides non- geometry data and the total equivalent (or, effective) strain
mandatory formulas for strain calculation [4]. Because of based on these components [1-4]. In 2005, Noronha et. al. [5]
identified that ASME B31.8 (2003) overestimates the
possible complications in dent profile, other formulas in the
open literature or those derived by qualified engineers are also circumferential and longitudinal bending strains by a factor of
allowed by the ASME code [4]. The code comparisons made two [4]. After Noronha et. al. [5], a corrected circumferential
1 Copyright © 2008 by ASME
and longitudinal bending strain component has been widely Where
adopted by the pipeline industry. The following are the
ASME B31.8 (2003) based formulas incorporating the t = Wall thickness
corrected Equations 1 and 2 by Noronha et. al.. The Ro = Initial pipe surface radius
parameters used in the equations are defined in Figure 1. R1 = Radius of curvature in transverse plane, negative for
reentrant dents
Circumferential bending strain (ε1), R2 = Radius of curvature in longitudinal plane, negative for
reentrant dents
t⎛ 1 1⎞ L = Dent length
ε1 = ⎜⎜ − ⎟⎟ (1) d = Dent depth
2 ⎝ Ro R1 ⎠

Longitudinal bending strain (ε2),

−t
ε2 = (2)
2R2
z
Extensional strain (ε3),

2
1⎛d ⎞ x
ε3 = ⎜ ⎟ (3) y x
2⎝ L⎠

Total effective strain of the dent in the pipe: Figure 1: Dent Geometry: Non-Reentrant versus
Reentrant Dents
ε eff = ε x2 − ε x ε y + ε y2 (4)
It is noted that Equations 1 to 3 were developed considering
strain in a thin plate and empirically established against a
where x and y refer to the longitudinal and circumferential limited number of finite element analysis [1-3, 6].
axes of the pipe, Figure 1.
However, there are no documented explanations in the public
From Equation 4, the total combined strain on the pipe surface domain regarding how Equations 4, i.e., the total effective
can be readily derived using ε x = ±ε 2 + ε 3 , ε y = ±ε 1 strain, was empirically established [1-3]. Baker [3] indicates
that Equation 4 can be seen to be either the effective strain for
where the positive and negative signs refer to the inner and
2 2
outer wall surface, respectively. The nomenclature in a plane strain condition eeffective = e1 − e1e2 + e22 where
equations (1), (2), and (3) is that used in ASME B31.8. The 3
subscripts in equation (4) refer to the strain directions shown e1=ε1, e2=ε2+ε3, e3 =0 without having a preceded constant 2/3,
in Figure 1. or can be thought of as analogous to von Mises stress for a

Total strain on inside pipe surface ( ε i )

plane stress condition, σ effective = σ 12 − σ 1σ 2 + σ 22
where σ3 = 0 .
ε i = [ε 12 − ε 1 (ε 2 + ε 3 ) + (ε 2 + ε 3 ) ]
1
2 2
(5) Since the calculation of membrane strain in the longitudinal
direction (i.e., Equation 3, ε3 = 1/2(d/L)2,) does not rely on the
complex geometry interpolation techniques but on the length
Total strain on outside pipe surface ( ε o ) and depth of the dent, efforts made for implementation of the
strain based assessment using ASME B31.8 based method
have been focused on how to accurately determine the radii of
ε o = [ε 12 + ε 1 (−ε 2 + ε 3 ) + (−ε 2 + ε 3 ) 2 ]2
1
curvature R1 and R2 [5,6,8].
(6)
All methods are based on a piece-wise interpolation technique.
The overall strain, ε max , is then defined by equations (5) and Rosenfeld et al. employed a piece-wise Bessel cubic
interpolation method to characterize the geometry of the dent
(6) and used to be compared with the acceptance and rejection
contour, and the osculating circle technique to estimate radii of
criteria recommended by ASME B31.8 (2003)
curvature [6]. Noronha et al. made use of fourth-order B-
ε max = Max[ε i , ε o ]
spline curves [5] to approximate the dent profile in both
(7) longitudinal and circumferential directions. Since fourth-order
B-splines have second-order continuity, radii of curvature can
2 Copyright © 2008 by ASME
be calculated at any location directly from a classical equation The bending strain becomes maximum at surface, i.e., z = ±t/2
of curvature [6]. Dawson et. al. [8] used piece-wise quadratic where t is the pipe wall thickness. The strain at the outer
curves fitted by least squares to simulate the dent profile. surface is:
More than three points were used for each quadratic equation
fit in order to increase the accuracy. The radii of curvature t ∂2w b t ∂2w
were also calculated using a classical equation [8]. ε xb = ,ε y = (9)
2 ∂x 2 2 ∂y 2
In addition to the foregoing, a simplified method for in-ditch
strain measurement using a profile gauge was developed [12] and has a opposite sign at the inner surface.
based on the ASME 31.8 strain criteria. The length, depth and
axial profile replicated by the profile gauge are transposed to These authors indicate that a 4th order B-splines approach may
graph paper. The maximum longitudinal strain value is be employed with FEM simulation to calculate curvature of
measured by matching the dent trace to a pre-plotted series of dents from caliper data, which was demonstrated by Noronha
strain curves calculated for the given wall thickness. This et. al. [5].
maximum longitudinal strain is then used to determine the
severity of the dents. It is unclear, however, if the The main difficulty in strain-based methods is in determining
circumferential depth profile is measured and the effective membrane strains in the dented region. Lukasiewicz et. al. [9]
strain critria as defined in ASME B31.8 are used for this in- indicated that at present, no exact (analytical) solutions are
ditch evaluation of dent severity. available for calculation of membrane strains of the dented
region in the pipe. The authors proposed to use the
METHODS PROPOSED BY LUKASIEWICZ AND CZYZ relationships between membrane strain and displacement for
ET. AL. [9, 11] large deformations of a cylindrical shell [9] to calculate the
membrane strains:
Lukasiewicz et. al. [9] and Czyz et al. [11] proposed an
alternative method that combines mathematical algorithms and
FEA tool for dent strain calculation. The authors indicate that Z
strains in a pipe wall consist of two main components:
longitudinal and circumferential. Each of which can be w
further separated into bending and membrane strains. The X
membrane strain is constant through the wall, while the u
bending component changes linearly from the inner to outer
surface. Figure 2 shows these strain components in a pipe v
wall.

Figure 3: Coordinate system of the pipe and displacement

components [9]

∂u 1 ⎛ ∂w ⎞
2

ε m
= + ⎜ ⎟ + ε xo
∂x 2 ⎝ ∂x ⎠
x

2
∂u w 1 ⎛ ∂w ⎞
ε m
= + + ⎜⎜ ⎟⎟ + ε yo (10)
∂y R 2 ⎝ ∂y ⎠
y

Figure 2: Strain components in a pipe wall [9, 11]. ∂u ∂v ⎡ ∂w ⎤ ⎡ ∂w ⎤

γ yx = + +
The authors indicated that calculation of the bending ∂y ∂x ⎢⎣ ∂x ⎥⎦ ⎢⎣ ∂y ⎥⎦
component is fairly straightforward, directly from the dent
displacement measured by in-line high-resolution caliper tool. where
The in-line caliper measures the pipe wall deflection w (see
Figure 3) in the Z(x,y) direction. The respective longitudinal
ε xm and ε ym are the membrane strains in the axial (x) direction,
and circumferential bending strain, i.e., ε xb and ε yb , can then
and the circumferential (y) direction
be calculated from the curvature of the radial displacement w
γ xy is the shear strain in the plane x, y
in the axial x and circumferential y directions:
R is the mean radius of the pipe
∂ w 2
∂ w 2 ε xo and ε yo are the initial strains due to the pressure in the
ε = Z 2 , ε yb = Z 2
b
(8)
∂x ∂y
x pipe, thermal expansion, etc.

3 Copyright © 2008 by ASME

In the above equations, the displacements u and v can be 1. Effective Strain: ASME B31.8 (2003) recommends
solved using a two dimensional FEM model [9] and the
normal displacement w(x,y) that is the only measured ε eff = ε x2 − ε x ε y + ε y2 while Lukasiewicz and Czyz et.
parameter on the dent geometry with the high-resolution 2
geometry tool. Figure 3 shows the displacement components al. [9, 11] suggest ε eq = ε x2 + ε x ε y + ε y2 and, later
u, v, and w in a pipe coordinate system. 3
2
These strains can be added to the bending strain components more precisely, ε eq = ε x2 + ε x ε y + ε y2 + γ xy2 / 2 .
producing the maximum values of strain in the axial and 3
circumferential directions: The differences with ASME are; (a) a preceding constant
(2/ 3 ), (b) a sign difference in the mid-term εxεy, and (c) an
ε x = ε xm ± ε xb additional shear strain γ xy . These differences could result in a
ε y = ε ym ± ε yb significant discrepancy in the calculated effective strain value.
Czyz et. al. [11], for example, demonstrated that the difference
The positive and negative signs refer to the outer and inner in effective strain can be more than 50% with the assumption
wall surface, respectively. that γ xy2 / 2 = 0. Figure 4 shows the difference in effective
strain as a function of the ratio of the hoop to axial strain – the
The equivalent strain on the inner and outer surface in the difference being the largest when the hoop to axial strain ratio
dented area of the pipe is then given by is close to one.

2
ε eq = ε x2 + ε x ε y + ε y2 (11) 60%
3 40%

Or, more precisely [11] Strain Didderence % 20%

0%
2
ε eq = ε + ε xε y + ε + γ / 2
2
x
2
y
2
xy (12) -20%
3
-40%

Equations 11 and 12 are derived directly on basis of the plastic -60%

strain theory [13], with an incompressibility condition in the -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

plastic regime εx+εy+εz=0. They can be used to compare Ratio of Hoop to Axial Strain

against the ASME B3.8 based strain acceptance criteria. Note

Figure 4: Difference between Eq. 11 (Lukasiewicz et al10 )
that εz cannot be assumed to be identically zero [11]. The and Eq. 7 ( ASME B31.8).
pipe wall behaves as a thin shell and therefore it is a plane-
stress, not a plane strain, condition. The through-thickness Further comparison between Eq. 11 and Eq 4 is made using
strain is in fact the sum of the two in-plane components. the 3-case data provided by Baker [3], with corrections made
to ε1 and ε 2 in accordance with Noronha et. al. [5]. Again, the
This mathematical algorithm combined with a specially
comparison assumes γxy=0. The results are presented in Table
developed FEA tool allows the calculation of all the strain
1. It is seen for all the three cases studied that the ASME
components based on the measurement of the radial
B31.8 method underestimates the effective strain by a factor of
deformation of a pipe obtained from a high-resolution
up to a factor of 1.8 and 2 for external and inner surfaces,
geometry in-line tool. Case studies have been given by the
respectively, which is essentially consistent with the
authors and compared with three-dimensional shell model
comparison made by Czyz et. al. [11], Figure 4.
results. Their proposed method is likely to provide a practical
tool for assessment of strain in dents. For the best accuracy,
Table 1: Difference in the calculated effective strain
the authors suggested that caliper data collected by high-
between Eq. 11 using the 3 cases data in the Dent Study
resolution in-line inspection tools with narrow sensors not
report by Baker [3]
exceeding one inch (25.4 mm) is required, and errors due to
surface irregularities and measurement error must be Case 1 Case 2 Case 3
minimized. Strain
ε1, Circumferetial Bending 2.3% 2.3% 2.3%
ε2, Longitudinal Beding 1.1% 1.5% 2.2%
Component
ε3, Longitudinal membrane 0.59% 0.8% 1.0%
COMPARISON OF ASME B31.8 AND LUKASIEWICZ ASME 31.8 2.10% 2.29% 2.92%
and CZYZ ET. AL. METHODS Inner
Lukasiewicz
10
4.06% 4.58% 5.64%
Surface
Effective Strain Difference 193.8% 199.9% 193.1%
From the above review, two major differences from the External
ASME B31.8 2.13% 2.08% 2.03%
10
ASME B31.8 based calculations are noted: Lukasiewicz 3.05% 3.18% 3.59%
Surface
Difference 143.4% 153.1% 177.2%

4 Copyright © 2008 by ASME

Further review showed that the difference in effective strain
formula between ASME B31.8 (2003) and Lukasiewicz and
Czyz et. al. [9, 11] may be fundamental.

As indicated previously, no details were given regarding how

the ASME B31.8 effective strain was empirically established
[1-3]. However, in accordance with Dent Study by Baker [3],
the formula is analogous to the effective stress (i.e., von Mises
stress) for a plane stress condition,
σ effective = σ 12 − σ 1σ 2 + σ 22 where, σ3 = 0 , or, similar
Figure 5: Comparison of equivalent strain along the pipe
to the effective strain for a plane strain condition without axis obtained by Czyz et al [11] using Dent Strain Analysis
having a preceded constant 2/3. DSA program and ASME B31.8 based method [1-3].
As noted by Czyz et. al. [11], there is no justification for using IMPROVEMENTS & ALTERNATIVES
the effective stress formula for calculation of the effective
strain by simply replacing the stress with strain symbol. And In the above sections, we have reviewed the currently
also there is no justification for making the comparison available strain-based dent assessment methods in pipeline
between the effective strain formula derived for a plane strain industry. It is shown that the discrepancy in the calculated total
condition as per the basis of ASME B31.8 (2003) since the strain between ASME B31.8 and Lukasieicz and Czyz et. al.
thin wall pipe is under plane stress condition and εz=0 should [9,11] can be significant.
not be used.
Principally, the algorithm and FEA methods, as described and
The effective strain formula derived by Lukasiewicz and Czyz demonstrated by Lukasieicz and Czyz et. al. [9,11], provides
et al [9,11] includes all the non-negligible strain components better solutions for strain components and can incorporate the
in a generalized format and is consistent with well established radial displacements measured by high resolution in-line
plastic strain theory [13]. The only condition used for the inspection caliper tools. FEA tools such as 3D shell finite
derivation is εx + εy + εz = 0, which is one of the basic element software for the non-linear elastic-plastic material and
characteristics of plastic deformation of a metal, i.e., the large deformation models (for example ANSYS and
volume of a metal remains constant during plastic ABAQUS) are commercially available. A specialized but
deformation. simpler FEM model with only two degrees of freedom per
node was developed by Lukasieicz and Czyz et. al. [9,11].
2. Membrane Strains. Rosenfeld et. al. [1-3, 6] present an However, despite the availability of FEA tools,
equation which is an empirical approximation to the implementation of strain calculations is as yet not routine and
longitudinal membrane strain and was quoted as a validated requires expertise both in areas of FEA and mechanical
formula using finite element analysis [3]. However, as damage. For dents with complex geometry, critically located
described by Lukasiewicz and Czyz et. al. [9, 11], this formula in high consequence areas, there is no doubt that FEA methods
is based on an analogy to the radial strain in a circular plate should be used. For other cases, a simple and easy to use
and not on rigorous shell theory, and is therefore simplistic method for strain calculation is still desired, and the ASME
and inaccurate [9]. B31.8 based strain method provide a useful tool for pipeline
industry in this case
Lukasiewicz et. al. further indicate that the existing ASME
31.8 (2003) techniques are limited to the longitudinal Since it has been shown that the current ASME B31.8 2003
membrane strain, which neglects the circumferential based strain methods are under-predicting the total strain
membrane and shear strains. The FEM analysis of actual (potentially by a factor of three), dents that may actually fail
dents has shown that these strain components can have a the current ASME B31.8 strain limit critria could remain un-
similar magnitude to the longitudinal strain [9]. For a more repaired in pipelines. Therefore, update of the B31.8 strain
accurate calculation, the shear strain γxy must also be included calculation should be considered to ensure that the strain
as shown in Equation 12 [11]. Czyz et. al. have demonstrated critria is being appropriately applied to dented pipelines.
that the difference between these two methods can be as high
as three times. Figure 5 is an example provided by Czyz et. 1. Improvement of ASME B31.8 Based Method: In terms
al.[12] About half of the 200% increase was found to be due of the current ASME 31.8 2003 method, three improvements
to the difference in the calculated membrane strains, and half may be considered to reduce its non-conservatism in strain
due to the different formulas for effective strain calculations. prediction:

a) Adopt the strain formula suggested by Lukasiewcz and

Czyz et. al. [9, 11] for total strain calculation

2
ε eq = ε x2 + ε x ε y + ε y2 (11)
3
5 Copyright © 2008 by ASME
With this improvement, ASME B31.8 could significantly 2
⎛L⎞ L
reduce the degree of its underestimate of stain by about 50% d +⎜ ⎟ −
2
[11]. ⎝2⎠ 2 4d 2
ε xm = = +1 −1 (14)
The more precise formulation for the equivalent strain given L L2
by equation (12) contains the membrane shear strain γ xy . 2
Czyz et. Al. [11] did not provide values for the shear strain in 0.5
their calculation, but did note that this term cannot be ⎛ 4d 2 ⎞
neglected. This of course is true for a FEA model since it will By using Taylor series expansion of ⎜⎜ 2 + 1⎟⎟ , Eq. 14
have a significant influence on the non-radial displacement ⎝ L ⎠
field and therefore the membrane stresses constraining the becomes
shape of the dent. However, if we assume that the mid point
of the dent as a point of symmetry, there would be no shear 2 2
⎛d ⎞ ⎛d ⎞
deformation here and the shear strain can/should therefore be ε xm = 1 + 2⎜ ⎟ + ... − 1 = 2⎜ ⎟ + ...... , and
removed from the expression for calculating the equivalent ⎝ L⎠ ⎝ L⎠
strain for comparison to the limit strain. That is to say it is the
equivalent strain calculated from equation (11) that should be 2
⎛d ⎞
compared with the limit strain criteria at the center of the dent. ε xm ≈ 2⎜ ⎟ (15)
⎝ L⎠
b) Improve the formula for longitudinal strain calculation
when d/L <<1.
Furthermore, examination of the longitudinal membrane
formula developed by Rosenfeld et. al. [1-3], ε3 = ½ (d/L)2 , By comparing Eq. (4), i.e., εxm =1/2(d/L)2 and Eq. (15), it is
appears to also underestimate the strain by a factor of up to 4. seen that ASME B31.8 can underestimate the longitudinal
This may be shown with a simplified geometrical model using membrane strain a factor of 4.
triangular approximation, Figure 6.
c) Circumferential strain consideration
Ls

In principal the same approach used to calculate longitudinal

x d x extensional strain could be used to calculate circumferential
L/2 L/2
extensional strain. However the distortion in the cross-section
makes it difficult to relate the width of the dent to the original
undeformed state. The ASME approach in fact assumes this
component to be negligible. The FEA based approach takes it
R into account intrinsically. During deformation, the membrane
stress will be initially compressive as the cross section resists
load acting essentially as an arch (see Roark’s Formulas for
Stress and Strain, Table 30, Case 8, [14]). As the profile
becomes re-entrant, the stress will progressively relax to zero,
and then on further loading, become tensile due to elastic
restraint on the deformation of the cross section at the dent
Figure 6: A simplified model used for estimate the location from the adjacent, non-dented pipe. No simplified
longitudinal membrane strain model for this behavior is presented, but the foregoing
consideration of the loading sequence suggests the ASME
Following the Rosenfeld et. al. [6] assumption that all surfaces assumption of zero for circumferential strain presents a
deflect only radially (i.e., the circular plate assumption), the practical approach.
membrane unit elongation in the longitudinal axis is εxm = ε3 =
(Ls – Lo)/Lo, where Ls is the arc length of the deformed With this proposed improvement, the effective strains are
longitudinal cross section and Lo is its initial straight length. recalculated and compared with the respective results of
ASME 31.8 method using the same 3-cases data as used
Assuming L x ≈ 2 x , i.e., the arc length may be approximated previously. The results are summarized in Table 2. It is seen
for all the three cases studied that the ASME B31.8 method
with two straight lines (arc string) and each straight line forms underestimates the effective strain by a factor of up to 3 on the
a right triangle with d and L/2, and yields the following inner surface and up to 1.2 on the external surface, indicating
relationship: two of the three cases that are acceptable based on ASME
B31.8 strain limit 6% criterion now become un-acceptable
L 2 L2 using the improved method.
x = d + ( ) and x = d +
2 2 2
(13)
2 4
6 Copyright © 2008 by ASME
Table 2: The calculated effective strain using the proposed Once the maximum equivalent strain is determined, a
method as compared with the ASME B31.8 comparison between the determined maximum strain against
the ASME B31.8 acceptance criteria can be made to determine
ASME B31.8 2.10% 2.29% 2.92%
Inner necessary repair actions.
Proposed Method 5.85% 7.01% 7.83%
Surface
Effective Difference 279.2% 305.8% 268.2%
Strain ASME B31.8 2.1% 2.1% 2.0% It should be emphasized that the maximum "effective strain" is
External
Surface Poposed method 2.3% 2.4% 2.5% only a measure of mechanical damage introduced by denting
Difference 110.1% 115.1% 123.3%
process. As long as the value is within ASME B31.8 (2003)
acceptance criteria, and it is assumed that the dent does not
Moreover, the 3-times high in the calculated maximum contain micro and/or macro cracks, there is no immediate
effective strains is consistent with the comparison made by harm to the pipeline [17]. However, fatigue and SCC life
Czyz et al [11] using their Dent Strain Analysis DSA program, assessment should be applied to evaluate lifecycle and
which is shown in Figure 4 in the present paper. A direct determine if repairs are needed based on a worst-case
comparison in the effective strain prediction between the scenario, such as assuming the dent is unconstrained and
proposed method and Lukasiewicz and Czyz et. al. [9, 11] formed at the time of construction of the pipeline. Stress hot
FEM model cannot be made at this time, however, with the spots where the membrane and bending stresses and/or
improved method, the non-conservatism of ASME 31.8 (2003) residual stresses are the highest which may or may not overlap
strain-based model is expected to be greatly reduced. with the maximum effective strain in the dent, should also be
identified [18-20] and incorporated into the assessment.
2. Alternative Approach: Although the concept of
equivalent strain is a convenient way to account for the SUMMARY
combined effect of imposed biaxial strains, and equivalent
stresses have clear physical significance in force-controlled By not considering the sharpness of the dent profile leading to
loadings, no rigorous failure criterion appears to exist linking high local plastic strains in the dent that are conducive to
the equivalent strain to damage in the plastically deformed cracking or like damage, existing criteria that use only the
pipe where the loading is intrinsically displacement- depth of dents to evaluate their severity may underestimate the
controlled. It is recommended, therefore, as a measure of failure potential. To remedy this, ASME B31.8 proposed
conservatism, to apply a limit strain criterion to the individual criteria in which the local strains were estimated as a function
tensile strain components in addition to the calculated of the dent depth and length measured along the axis of the
equivalent strain in the event that one of the individual strain pipe. Recognizing short-comings in the ASME method,
components is negative. Lukasiewicz and Czyz et. al. developed an approach using a
combination of improved algorithms and post-elastic, large-
Furthermore, other strain-based damage models in the deflection FEA which provides a useful tool for the rigorous
literature may be also considered. One of the models assessment of dents. The reliance on advanced FEA however
proposed by Mazars [15] for example suggests that since requires significant skill and judgment. In light of this, the
damage in many materials propagates much more easily under present paper presented an alternative evaluation method
tension than under compression, the influence of tension based on improvements in the ASME methodology which
should be emphasized. The applicability of this damage allows an initial, rapid strain-based assessment of dents
consideration is obvious in the case of brittle materials, but we without relying on FEA to estimated strain components.
believe it is also applicable in the limit to steel at the failure
strain where ductile fracture takes place on a plane REFERENCES
perpendicular to the maximum tensile strain. The fracture
occurs as a consequence of the coalescence of voids resulting 1. Rosenfeld, M. J.: ”Proposed New Guidelines for ASME
from the presences of secondary phase particles and/or slip B13.8 on Assessment of Dents and mechanical damage”,
displacements [16]. As a first approximation then, one can GRI Topical Report No. GRI-01/0084, May, 2001.
assume that compressive strains do not indicate damage in the 2. Rosenfeld, M., Pepper, J. and Leewis, K.: “Basis of the
as-found, deformed state. Therefore, the equivalent, or New Criteria in ASME B31.8 for Prioritization and
effective, strain εeq is related to positive strain only, and may Repair of Mechanical Damage”, Proceedings of IPC’02,
be defined as 4th International Pipeline Conference, Paper No. 27122,
Sept. 29 to Oct. 3, 2003. Calgary, Canada.
3 3. Baker, M., Dent Study, Final Report, TTO Number 10,
ε effective = ∑ εi
2
(16) Integrity Management Program, Delivery Order DTRS56-
i =1 02-D-70036, DOT Research and Special Programs
where εI, i=1, 2, 3 are the principal strains, and the brackets Administration, Office of Pipeline Safety, November
2004
... denotes the positive part, i.e., ε = ε for ε >0 and 4. Anon: “Gas Transmission and Distribution Piping
ε =0 for ε ≤ 0. This method could provide a reasonable System”, ASME B31.8 2003, Paragraph 851.4, 2003.
5. Noronha, D. B., Martins, R., Jacob, B. and Souza E.: “The
equivalent strain for strain based dent assessment. Use of B-Splines in the Assessment of Strain Levels
Associated with Plain Dents”, Rio Pipeline Conference &

7 Copyright © 2008 by ASME

 Exposition 2005, Paper No. IBP 1245_05, October 2005, rupture du beton de structure”, These de Doctorat d’Etat,
Rio de Janeiro, Brazil. 1984, Universite Paris VI, France.
6. Rosenfeld, M. J.., Porter, P. C., Cox, J. A., “Strain 16. Broek, D. “Elementary Engineering Fracture Mechanics”,
Estimation Using Vetco Deformation Tool Data”, ASME Third Revised Edition, Martinus Nijhoff Publishers, 1984.
2nd International Pipeline Conference, Calgary, 1998. Library of Congress Catalogue Card Number 82-45135.
7. Weisstein, E. W.: “Curvature.” from MathWorld – A 17. PDAM (2003): Pipeline Defect Assessment Manual,
Wolfram Web Resource, 2003.
http://mathworld.wolfram.com/curvature.html 18. Rinehart A. J. and Keating P. B.: “Length Effects on the
8. Dawson, S. J., Russell A. and Patterson, A.: ”Emerging Fatigue Behavior of Longitudinal Pipeline Dents”, 4th
Techniques For Enhanced Assessment and Analysis of International Pipeline Conference, PAPER Number
Dents”, Proceedings of IPC 2006, Paper No. 10264, 6th 27244, September 2002, Calgary, Alberta, Canada.
International Pipeline Conference, September 25-29, 19. Dinovitzer A., Lazor, R., Carroll, L. B., Zhou, J.,
2006, Calgary, Canada. McCarver, F., Ironside, S. and Raghu, D.: “Geometric
9. Lukasiewicz, S. A., Czyz, J. A., Sun, C., Adeeb, S. Dent Characterization”, 4th International Pipeline
“Calculation od=f Strains in Dents Based on High Conference, Paper Number 27076, September 2002,
Resolution In-Line Caliper Survey”, IPC2006, Paper Calgary, Alberta, Canada.
10101, 6th International Pipeline Conference, September 20. Been, J., Carroll, B, Dinovitzer, A., and Suthherby, R.:
25-29, 2006, Calgary, Canada. “Stress Intensification and Crack Growth in the Presence
10. Lukasiewicz, S.A.: “Local Loads on Plates and Shells”, of Dent on Pipelines, IPC 2006, Paper Number 10415,
Fridhoff-Noordhoff, The Netherlands, Leyden, 1979. September 25-29, 2006, Calgary, Alberta, Canada.
11. Czyz, J. A., Lukasiewicz, S. A., and Addeeb, S.:
“Calculating Dent Strain”, Pipeline and Gas Technology,
January/February, 2008, pp. 38-45.
12. Warman, D.J, Johnston, D., Mackenzie, J. D, Rapp, S.
and Travers, B. “Management of Pipeline Dents and
mechanical Damage in Gas Pipelines”, IPC 2006, Paper
10407, 6th International Pipeline Conference, September
25-29, 2006, Calgary, Canada.
13. Dieter, G.E., Mechanical Metallurgy, McGraw-Hill, 1976.
14. Young, W.C., Roark’s Formulas for Stress & Strain, 6th
Edition, McGraw-Hill, 1989
15. Marzrs J., “Application de la mecanique de
I’endommagement au comportement non lineaire et a la

8 Copyright © 2008 by ASME