Nuclear Engineering and Design 265 (2013) 918931

Contents lists available at ScienceDirect

Nuclear Engineering and Design

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

Finite element limit loads for non-idealized through-wall cracks in

thick-walled pipe
Do-Jun Shim a , Tae-Song Han b , Nam-Su Huh b,
a
Engineering Mechanics Corporation of Columbus, 3518 Riverside Drive, Suite 202, Columbus, OH 43221, USA
b
Department of Mechanical System Design Engineering, Seoul National University of Science and Technology, 172 Gongreung 2-dong, Nowon-gu, Seoul
139-743, Republic of Korea

h i g h l i g h t s

The lower bound bulging factor of thin-walled pipe can be used for thick-walled pipe.
The limit loads are proposed for thick-walled, transition through-wall cracked pipe.
The correction factors are proposed for estimating limit loads of transition cracks.
The limit loads of short transition cracks are similar to those of idealized cracks.

a b s t r a c t
a r t i c l e i n f o
The present paper provides plastic limit loads for non-idealized through-wall cracks in thick-walled pipe.
Article history:
These solutions are based on detailed 3-dimensional nite element (FE) analyses which can be used for
Received 2 July 2013
Received in revised form structural integrity assessment of nuclear piping. To cover a practical range of interest, the geometric
16 September 2013 variables and loading conditions affecting the plastic limit loads of thick-walled pipe with non-idealized
Accepted 26 September 2013 through-wall cracks were systematically varied. In terms of crack orientation, both circumferential and
axial through-wall cracks were considered. As for loading conditions, axial tension, global bending, and
internal pressure were considered for circumferential cracks, whereas only internal pressure was con-
sidered for axial cracks. Furthermore, the values of geometric factor representing shape characteristics
of non-idealized through-wall cracks were also systematically varied. In order to provide condence in
the present FE analyses results, plastic limit loads of un-cracked, thick-walled pipe resulting from the
present FE analyses were compared with the theoretical solutions. Finally, correction factors to the ide-
alized through-wall crack solutions were developed to determine the plastic limit loads of non-idealized
through-wall cracks in thick-walled pipe.
2013 Elsevier B.V. All rights reserved.

1. Introduction
In the design of nuclear piping, if it is demonstrated that the
rupture probabilities of nuclear piping are extremely low under
conditions consistent with the design basis for the nuclear pip-
ing system, a local dynamic effect associated with the postulated
pipe ruptures in the nuclear piping system can be excluded from
the design basis (USNRC, 2013). Accordingly, to ensure and sat-
isfy this criterion, the Leak-Before-Break (LBB) concept has been
widely applied to nuclear piping design as a deterministic approach
(USNRC, 1987). In addition, conservative crack tolerance assess-
ment procedure to satisfy this extremely low rupture condition
has been developed (USNRC, 1984). According to existing screening
Corresponding author. Tel.: +82 0 2 970 6317; fax: +82 0 2 974 8270.
E-mail address: nam-su.huh@seoultech.ac.kr (N.-S. Huh).
criteria for LBB assessment, it is specied that a nuclear piping sys-
tem that is exposed to active degradation mechanism (e.g., primary
water stress corrosion cracking; PWSCC) should not be designed
based on the LBB concept (USNRC, 1987, 1984).
However, during the last two decades, several cracking incidents
due to PWSCC have been observed in the nuclear components,
even in piping systems that have been approved for LBB prior to
operational PWSCC experiences. Based on the current LBB speci-
cations, this means that these piping systems no longer satisfy the
existing deterministic LBB assessment procedure (Electric Power
Research Institute, 2010; Nana and Yoon, 2006). To resolve this
issue regarding LBB design of piping related to PWSCC, there are
on-going efforts that are attempting to directly demonstrate full
compliance with the current design criteria (USNRC, 2013), i.e.
the condition of extremely low probabilities of rupture of nuclear
piping. These efforts have been made by using the probabilistic
assessment approach for piping system with active degradation

0029-5493/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.nucengdes.2013.09.033
 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931 919

mechanism. One of the well-known work is the xLPR (eXtremely

Low Probabilities of Rupture) program (Rudland and Harrington,
2012). Related to this work, it was demonstrated that a subcritical
surface crack can transition to a through-wall crack with signicant
differences between the inner diameter (ID) and outer diameter
(OD) crack lengths (Rudland et al., 2010). In the present work, such
a crack is referred to as a non-idealized through-wall crack, whereas
an idealized through-wall crack has a crack front that is parallel to
the radial direction. Moreover, it has been shown that more accu-
rate predictions (e.g., leak rate) can be made when non-idealized
through-wall cracks are used in the LBB assessment (Shim et al.,
2011). Thus, in order to evaluate piping rupture probabilities more
accurately, it is necessary to characterize the fracture behavior of
non-idealized through-wall cracks.
For typical fracture mechanics assessment, relevant fracture
mechanics parameters are needed. In terms of fracture mechanics
parameters, the elastic stress intensity factor is needed to pre-
dict subcritical crack growth behavior and elastic-plastic J-integral
(Rice, 1968) and plastic limit load are required for prediction of
unstable fracture (or instability). For leak-rate estimates, solutions
for crack opening displacement (or crack opening area) are needed.
For both surface and idealized through-wall cracks, the engineering
estimates of these fracture mechanics parameters have been well
established during the last three decades (Zahoor, 1991; Kumar and
German, 1988; Rahman et al., 1998a,b; France et al., 1997; Kim et al.,
2001; Miller, 1988; Huh et al., 2007). However, fracture mechanics
parameters for non-idealized through-wall cracks are still limited
(Huh et al., 2008a,b, 2010). Thus, there is a need to develop frac-
ture mechanics parameters for non-idealized through-wall cracks
(both in axial and circumferential directions) to support the rupture
probability calculations of nuclear piping. In addition, since several
crack indications have been found in a thick-walled piping (e.g.,
pressurizer nozzle of a light water reactor (Materials Reliability
Program, 2004)), the development of fracture mechanics param-
eters for non-idealized through-wall crack should be extended to
thick-walled components as well.
In the present work, plastic limit loads for thick-walled pip-
ing with non-idealized through-wall cracks are proposed based
on detailed 3-dimensional (3-D) nite element (FE) analyses. In
terms of crack orientation, both axial and circumferential cracks are
considered. As for loading conditions, internal pressure is consid-
ered for axial crack, whereas axial tension, global bending moment
and internal pressure are considered for circumferential crack. For
estimating plastic limit loads, the geometric variables affecting the
plastic limit loads were systematically varied. Based on the FE plas-
tic limit loads, a correction factor to quantify the effect of the crack
shape of non-idealized through-wall cracks on plastic limit loads is
newly proposed.

2. Finite element analyses

Fig. 1. Schematics of pipes with (a) a non-idealized circumferential through-wall
2.1. Geometry crack under internal pressure, axial tension and global bending moment and (b) a
non-idealized axial through-wall crack under internal pressure.

In order to accurately calculate the probabilities of rupture of

nuclear piping exposed to PWSCC, the fracture behavior of a non-
idealized through-wall crack should be considered. In this context,
in the present work, thick-walled pipes with non-idealized axial
and circumferential through-wall cracks were considered. Fig. 1
depicts the geometric variables of pipes with non-idealized cir-
cumferential and axial through-wall cracks, where Ri , Rm , Ro , and
t represent the inner, mean, outer radius, and thickness of a pipe,
respectively. Fig. 1(a) depicts a pipe with a non-idealized circum-
ferential through-wall crack under axial tension, global bending,
and internal pressure. The non-idealized circumferential through-
wall crack is characterized by the half crack angle on the inner and
the outer surface of the pipe (dened as  1 and  2 ). In this study,
two parameters, i.e.,  1 / 2 (= 1, 2 and 3) and  1 / (= 0.125, 0.25, 0.3,
0.4 and 0.5) were systematically considered to cover various ranges
of crack sizes. The parameter  1 / 2 , i.e., the ratio of crack angle on
the inner surface of pipe to crack angle on the outer surface of pipe,
represents the shape of the non-idealized circumferential through-
wall crack, where  1  2 in the present study. Note that  1 / 2 = 1
represents an idealized circumferential through-wall crack. The
reference crack angle is dened as  1 / based on the crack angle
on the inner surface of a pipe.
920 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931

Table 1
List of analysis cases for the present FE analysis.

Geometry Loading condition Rm /t  1 /  1 / 2

Circumferential through-wall cracked pipe Axial tension, Global bending, Internal pressure 2, 3, 5 0.125, 0.25, 0.3, 0.4, 0.5 1, 2, 3

Geometry Loading condition Rm /t 1 1 /2

Axial through-wall cracked pipe Internal pressure 2, 3, 5 0.5, 1, 2, 3 1, 1.5, 2, 3, 4

Fig. 1(b) depicts a pipe with a non-idealized axial through-wall
crack subjected to an internal pressure. The shape of the non-
idealized axial through-wall crack is characterized by the half crack
length on the inner and outer pipe surfaces (dened as a1 and a2 ).
The crack front is assumed as a straight line. In this study, the crack
length is normalized as:
a a
1 =  1 , 2 =  2
Rm t Rm t
where, 1 and 2 represent the normalized crack length at inner
surface and outer surface, respectively. In this study, to cover
practical ranges of geometric variables for non-idealized axial
through-wall cracks, ve different values of 1 /2 (= 1, 1.5, 2, 3
and 4) were considered, where 1 2 in the present study. Note
that 1 /2 = 1 represents a pipe with idealized axial through-wall
crack. Similarly, the reference crack length is dened as crack length
at inner surface of pipe, then four different values of reference
crack length 1 (= 0.5, 1, 2 and 3) were systematically considered.
In addition, in order to provide solutions for practical ranges of
thick-walled pipe, three different values of Rm /t (= 2, 3 and 5) were
considered. Note that, in general, pipes with smaller Rm /t values
than 5 is regarded as thick-walled pipes (Zahoor, 1991). Thus, a
total of 195 cases were considered for various thick-walled pipes
with circumferential and axial through-wall cracks, which is sum-
marized in Table 1.
distributed load to the inner surface of cracked pipe. Considering
the effect of an end-closure of pipe, the equivalent axial tension due
to internal pressure was also applied to pipe-end as a distributed
load. Moreover, the effect of crack face pressure was also consid-
ered by applying 50% of the internal pressure to the crack face
(Kim et al., 2002). In addition, due to the convergence problem in
elastic-perfectly plastic calculations for calculating unstable load-
(1) ing, the modied RIKS option within ABAQUS was invoked in the FE
analyses. The plastic limit pressures of non-idealized through-wall
cracks were calculated by using the RIKS factor obtained from the
FE analyses. In the case of axial tension and global bending, these
loads were applied either by displacement or by rotational angle
using the MPC (Multi-Point Constraint) option within ABAQUS.
Then, the displacement and rotational angle were applied to the end
constrained node, so that resultant axial load and global bending
moment can be obtained from the nodal reaction force and reaction
moment, respectively. Since, a material was assumed to be elastic-
perfectly plastic, the reaction force and moment converge into
certain values, then these converged reaction force and moment
were dened as plastic limit load and plastic limit moment, respec-
tively.

2.2. Finite element limit analysis

Elastic-perfectly plastic FE analyses for two geometries,

depicted in Fig. 1, were performed using the general-purpose
FE program, ABAQUS (Dassault Systemes, 2012). Material was
assumed to be elastic-perfectly plastic with a Poissons ratio of
0.3, and non-hardening J2 ow theory was used using a small
geometry change continuum FE model. The yield strength and
Youngs modulus were assumed to be 165 MPa and 190 GPa,
respectively. Considering symmetric conditions of geometries, only
a quarter of a pipe was modeled to reduce computing time.
Fig. 2 shows the typical 3-D FE meshes employed in the present
FE works, where reduced integration 20-nodes brick elements
(C3D20R in ABAQUS element library) were used to avoid prob-
lem associated with incompressibility. Fig. 2(a) depicts a 3-D FE
model of a thick-walled pipe with non-idealized circumferential
through-wall crack and Fig. 2(b) for a thick-walled pipe with
non-idealized axial through-wall crack. The crack-tip region was
modeled using focused element, and wedge-shaped elements to
create sharp crack were used as shown in Fig. 2. Five elements
were used through the wall thickness of pipe. The number of
elements in the present FE mesh is 2270 and 3600 elements
for circumferential and axial through-wall cracked pipes, respec-
tively.
As for the loading condition, axial tension, global bending, and
internal pressure were considered for non-idealized circumfer-
ential through-wall cracks, whereas only internal pressure was
considered for non-idealized axial through-wall cracks (in consid-
eration of dominant stress component, i.e., hoop stress component,
to crack opening or crack instability). When calculating plastic limit Fig. 2. Typical FE meshes employed in the present FE analyses: (a) a non-idealized
pressures from the FE analyses, an internal pressure was applied as circumferential through-wall crack and (b) a non-idealized axial through-wall crack.
 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931 921

Fig. 3. Comparison of the present FE limit loads for un-cracked, thick-walled pipes with existing solutions (Hill, 1950): (a) axial tension (b) global bending moment, and (c)
internal pressure.

3. Plastic limit loads of thick-walled pipe with
non-idealized through-wall cracks
3.1. Plastic limit load solutions
In the present paper, a comprehensive plastic limit load solu-
tions for non-idealized through-wall cracks in thick-walled pipes
were developed based on detailed 3-D FE analyses. Prior to devel-
opment of FE based plastic limit loads of thick-walled pipes with
non-idealized through-wall cracks, FE analyses for thick-walled
pipes, respectively.  y is the limiting (yield) strength of the elastic-
perfectly plastic material.
Tables 24 and Fig. 3 compare the plastic limit loads from
the present FE analyses with the existing analytical ones for un-
cracked, thick-walled pipe (Rm /t = 2, 3, 5) under axial tension, global
bending, and internal pressure. The FE analysis method employed
Table 2
Ratios of the present FE limit tension to existing solution for the thick-walled, un-
cracked pipe under axial tension.

pipes are performed either for 3-D un-cracked pipes or for idealized Rm /t Normalized plastic limit load
through-wall cracked pipes to validate the FE based limit analysis
NLFE, Un-Cracked /NL Diff. (%)
method employed in the present study. The plastic limit loads of
thick-walled, un-cracked pipes under axial tension, global bending 2 1.0006 0.0560
3 1.0006 0.0590
moment and internal pressure are given by Hill (1950).
5 1.0006 0.0597
 
NL = y  Ro2 Ri2
Table 3
4   Ratios of the present FE limit moment to existing solution for the thick-walled,
ML = y Ro3 Ri3 (2) un-cracked pipe under global bending.
3
Rm /t Normalized plastic limit load
2
PL = y ln(Ro /Ri ) MLFE, Un-Cracked /ML Diff. (%)
3
2 1.0002 0.0168
where, NL , ML and PL denote the plastic limit load, plastic limit 3 1.0003 0.0252
5 1.0003 0.0323
moment and plastic limit pressure of un-cracked, thick-walled
922 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931

/ /

Fig. 4. Comparison of the present FE limit loads for thick-walled pipes with idealized circumferential through-wall cracks with lower bound bulging factors from thin-walled
pipes: (a) axial tension, (b) global bending and (c) internal pressure.

for un-cracked pipes was identical to that for pipes with non-
idealized through-wall cracks described in Section 2.2. As shown in
Tables 24 and Fig. 3, the present FE results for un-cracked, thick-
walled pipes agreed very well with the existing analytical solution,
Eq. (2), and the maximum difference was less than approximately
0.1% (see Tables 24). Thus, the present FE limit analysis method
was veried against existing analytical solutions for un-cracked,
thick-walled pipes.
The present FE limit loads of thick-walled pipe with through-
wall cracks (both non-idealized and idealized) were normalized
with respect to plastic limit loads of un-cracked, thick-walled pipes,
Eq. (2), as:
non-idealized through-wall cracks and idealized through-wall
cracks (for both circumferential and axial), respectively, and the
subscription Tension, Bending, and Pressure denote each load-
ing conditions.
Although the present FE limit analysis method was veried
against analytical solutions of un-cracked, thick-walled pipes, fur-
ther condence was gained by comparing the present FE limit loads
for thick-walled pipes with idealized through-wall cracks with the
existing FE limit loads for thin-walled pipes with idealized through-
wall cracks (Huh et al., 2007). Huh et al. (2007) provided plastic limit
loads (lower bound solutions for conservative estimates) based on

NLFE, Non-idealized NLFE, Idealized

nNon-idealized
L,Tension
= , nIdealized
L,Tension
=
NL NL
Table 4
MLFE, Non-idealized MLFE, Idealized Ratios of the present FE limit pressure to existing solution for the thick-walled,
nNon-idealized = , nIdealized = (3)
L,Bending ML L,Bending ML un-cracked pipe under internal pressure.

PLFE, Non-idealized PLFE, Idealized Rm /t Normalized plastic limit load

nNon-idealized
L,Pressure
= , nIdealized
L,Pressure =
PL PL PLFE, Un-Cracked /PL Diff. (%)

In Eq. (3), nL is normalized plastic limit loads of thick-walled pipe 2 0.9997 0.0260
with through-wall cracks. The superscriptions FE, Non-Idealized 3 0.9999 0.0103
5 0.9991 0.0861
and FE, Idealized denote the FE based plastic limit loads for
 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931 923

Table 5
Values of normalized plastic limit loads of idealized circumferential through-wall
cracks and correction factors (S) for non-idealized circumferential through-wall
cracks in pipe under axial tension.

Under axial tension nIdealized

L,Tension
Correction factor (S)

 1 / 2

Rm /t  1 / 1 2 3

2 0.125 0.889 1.000 1.077 1.095

0.25 0.627 1.000 1.274 1.389
0.3 0.531 1.000 1.354 1.525
0.4 0.365 1.000 1.589 1.842
0.5 0.230 1.000 2.041 2.438
3 0.125 0.864 1.000 1.094 1.106
0.25 0.601 1.000 1.278 1.404
0.3 0.508 1.000 1.355 1.532
0.4 0.344 1.000 1.617 1.875

1 5
0.5
0.125
0.212
0.838
1.000
1.000
2.124
1.117
2.562
1.132
0.25 0.578 1.000 1.300 1.433
Fig. 5. Comparison of the present FE limit pressures for thick-walled pipes with 0.3 0.486 1.000 1.383 1.571
idealized axial through-wall cracks with lower bound bulging factors from thin- 0.4 0.327 1.000 1.686 1.959
walled pipes. 0.5 0.199 1.000 2.269 2.744

the 3-D FE analysis for thin-walled pipe with idealized circumfer-

ential through-wall cracks as:
  2   
FE,Idealized(thin)
NL 2.19  6.27 
= +1 for axial tension (4)
2Rm ty    

  2   
FE,Idealized(thin)
ML  
= 0.35 1.74 +1 for global bending (5)
2 t
4Rm y  

 2   

  

0.92 0.28 +1 for internal pressure (0 0.263)

  
FE, Idealized(thin)

 
3 PL Rm
sin 
=  + 2 sin1 (6)
2 y t 2

3 1 for internal pressure (0.263  0.5)

 

where, superscript FE,Idealized(thin) represents the FE limit loads

of thin-walled pipes with idealized through-wall cracks.
The lower bound FE limit pressure of thin-walled pipes with
idealized axial through-wall cracks is given as (Huh et al., 2007)
FE, Idealized(thin)
PL Rm 3 1 The values of normalized plastic limit loads of thick-walled pipe
=  (7)
y t 2 1 + 0.34 + 1.342
As shown in Eqs. (4)(7), the lower bound FE limit loads of
thin-walled pipes with idealized through-wall cracks are normal-
ized with those of un-cracked, thin-walled pipes (the left-hand
side of Eqs. (4)(7)). Hence, the right-hand side of the Eqs. (4)(7)
could be considered as the lower bound bulging factor reecting
the effect of through-wall cracks. Thus, in the present work, these
lower bound bulging factors, i.e., the right-hand side of Eqs. (4)(7),
obtained from the thin-walled pipes with idealized through-wall
cracks are compared with the present FE plastic limit loads of thick-
walled pipes with idealized through-wall cracks normalized with
Eq. (2), i.e., nIdealized
L
in Eq. (3). Fig. 4 shows comparison results for
thick-walled pipes (Rm /t = 2, 3, 5) with idealized circumferential
through-wall cracks, and the comparison results for thick-walled
pipes (Rm /t = 2, 3, 5) with axial through-wall cracks are given in
Fig. 5. As shown in Figs. 4 and 5, it is revealed that the lower bound
plastic limit solutions, i.e. the right-hand side of the Eqs. (4)(7),
could also be considered as lower bound plastic limit solutions of
thick-walled pipes with idealized through-wall cracks if the plastic
limit loads of thick-walled pipes with idealized through-wall cracks
is normalized with those of un-cracked, thick-walled pipes. Thus,
the condence of the present FE limit analysis was gained again.
with idealized through-wall cracks are also given in Tables 57 for
circumferential cracks and in Table 8 for axial cracks.
The values of the normalized plastic limit loads of thick-walled
pipe with non-idealized circumferential and axial through-wall
cracks dened in Eq. (3) are provided in Figs. 68 together with
those of thick-walled pipe with idealized through-wall cracks for
three different values of Rm /t (= 2, 3, 5). As shown in Figs. 68, the
normalized plastic limit loads, nL , decrease as the reference crack
angle or length in the inner surface, i.e.,  1 / and 1 , increase. Fur-
thermore, nL values increase when crack angle or length on the
outer surface decrease, i.e., values of  1 / 2 and 1 /2 increase (see
Fig. 1 for crack shape denitions). In particular, in the small range of
crack length ratio, the plastic limit loads increase more rapidly. On
the other hands, as crack length ratio increase, the plastic limit loads
of thick-walled pipe with non-idealized through-wall cracks tend
to be constant. The differences of plastic limit loads of idealized
through-wall cracks between non-idealized through-wall cracks
increase as the reference crack length increase for both circumfer-
ential and axial cracks. These trends of nL values were independent
on the Rm /t and loading condition.
924 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931

= 0.125 = 0.125
= 0.25 = 0.25
= 0.3 = 0.3
= 0.4 = 0.4
= 0.5 = 0.5

1/2 1/2

= 0.125
= 0.5
= 0.25
=1
= 0.3
=2
= 0.4
=3
= 0.5

1/2 1/2

Fig. 6. Normalized plastic limit loads of thick-walled pipe with non-idealized through-wall cracks (Rm /t = 2); (a)(c) circumferential cracks and (d) axial cracks (The values
of idealized through-wall cracks are also given as solid symbols).

Table 6 Table 7
Values of normalized plastic limit loads of idealized circumferential through-wall Values of normalized plastic limit loads of idealized circumferential through-wall
cracks and correction factors (S) for non-idealized circumferential through-wall cracks and correction factors (S) for non-idealized circumferential through-wall
cracks in pipe under global bending moment. cracks in pipe under internal pressure.

Under global bending nIdealized

L,Bending
Correction factor (S) Under internal pressure nIdealized
L,Pressure
Correction factor (S)

 1 / 2  1 / 2

Rm /t  1 / 1 2 3 Rm /t  1 / 1 2 3

2 0.125 0.907 1.000 1.062 1.072 2 0.125 0.981 1.000 1.006 1.008
0.25 0.667 1.000 1.237 1.339 0.25 0.899 1.000 1.063 1.081
0.3 0.572 1.000 1.310 1.461 0.3 0.827 1.000 1.130 1.160
0.4 0.407 1.000 1.528 1.744 0.4 0.581 1.000 1.488 1.588
0.5 0.268 1.000 1.943 2.273 0.5 0.361 1.000 2.057 2.364
3 0.125 0.877 1.000 1.089 1.107 3 0.125 0.975 1.000 1.009 1.016
0.25 0.646 1.000 1.237 1.343 0.25 0.893 1.000 1.065 1.081
0.3 0.556 1.000 1.305 1.454 0.3 0.818 1.000 1.141 1.164
0.4 0.391 1.000 1.547 1.761 0.4 0.575 1.000 1.508 1.587
0.5 0.253 1.000 2.000 2.356 0.5 0.350 1.000 2.145 2.328
5 0.125 0.857 1.000 1.103 1.127 5 0.125 0.969 1.000 1.009 1.014
0.25 0.628 1.000 1.250 1.361 0.25 0.887 1.000 1.061 1.076
0.3 0.540 1.000 1.323 1.481 0.3 0.811 1.000 1.145 1.165
0.4 0.378 1.000 1.600 1.825 0.4 0.566 1.000 1.549 1.613
0.5 0.242 1.000 2.125 2.516 0.5 0.340 1.000 2.280 2.565
 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931 925

= 0.125 = 0.125
= 0.25 = 0.25
= 0.3 = 0.3
= 0.4 = 0.4
= 0.5 = 0.5

1/2 1/2

= 0.125
= 0.25 = 0.5
= 0.3 =1
= 0.4 =2
= 0.5 =3

1/2 1/2

Fig. 7. Normalized plastic limit loads of thick-walled pipe with non-idealized through-wall cracks (Rm /t = 3); (a)(c) circumferential cracks and (d) axial cracks (The values
of idealized through-wall cracks are also given as solid symbols).

3.2. Correction factors for non-idealized through-wall cracks was proposed. These correction factors of non-idealized through-
wall cracks in thick-walled pipes under each loading condition is
In order to quantify the effect of the crack shape of non-idealized dened as:
through-wall cracks on plastic limit loads, a correction factor (S)
nNon-idealized
L,i
= nIdealized
L,i
Si (8)

Table 8 where, subscription i denotes the each loading condition, i.e., axial
Values of normalized plastic limit loads of idealized axial through-wall cracks and tension, global bending moment, and internal pressure.
correction factors (S) for non-idealized axial through-wall cracks in pipe under
Using these correction factors (S) together with the plastic limit
internal pressure.
loads of un-cracked pipe, the plastic limit loads of thick-walled
Under internal nIdealized
L,Pressure
Correction factor (S) pipe with non-idealized through-wall cracks can be determined.
pressure
Figs. 911 show the correction factor, S, for non-idealized through-
1 /2 wall cracks in thick-walled pipes, that is, ratio of plastic limit loads
Rm /t 1 1 1.5 2 3 4
of non-idealized through-wall cracks to those of idealized through-
wall cracks for three different values of Rm /t (= 2, 3, 5). As shown in
2 0.5 0.852 1.000 1.049 1.074 1.101 1.111
these gures, the values of S increase as values of crack length ratio
1 0.637 1.000 1.113 1.182 1.259 1.301
2 0.397 1.000 1.174 1.310 1.490 1.602 increase. Moreover, as values of reference crack angle or length
3 0.288 1.000 1.202 1.369 1.621 1.781 in inner surface of pipe increase, the ratio of plastic limit loads of
3 0.5 0.853 1.000 1.049 1.072 1.095 1.107 non-idealized through-wall cracks to those of idealized through-
1 0.642 1.000 1.113 1.182 1.260 1.302 wall cracks increase. In particular, correction factors of axial tension
2 0.393 1.000 1.184 1.335 1.531 1.650
3 0.280 1.000 1.216 1.396 1.692 1.867
show higher values than other loading conditions, and those of
5 0.5 0.856 1.000 1.047 1.069 1.090 1.105 internal pressure show the lowest values of correction factors.
1 0.646 1.000 1.117 1.189 1.269 1.312 When the reference crack length or angle (i.e., crack length or angle
2 0.389 1.000 1.198 1.385 1.615 1.741 at the inner surface of pipe) is relatively large, the values of correc-
3 0.272 1.000 1.233 1.436 1.853 2.051
tion factor are around 22.5, which means that the plastic limit
926 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931

= 0.125 = 0.125
= 0.25 = 0.25
= 0.3 = 0.3
= 0.4 = 0.4
= 0.5 = 0.5

1/2 1/2

= 0.125
= 0.25 = 0.5
= 0.3 =1
= 0.4 =2
= 0.5 =3

1/2 1/2

Fig. 8. Normalized plastic limit loads of thick-walled pipe with non-idealized through-wall cracks (Rm /t = 5); (a)(c) circumferential cracks and (d) axial cracks (The values
of idealized through-wall cracks are also given as solid symbols).

loads of non-idealized through-wall cracks are 22.5 times higher
than those of idealized through-wall cracks with same reference
crack angle or length. Furthermore, when the reference crack angle
or length is small, the shape of non-idealized through-wall cracks,
i.e. the ratio of crack angle on the inner surface to that on the outer
surface, has little or no effect on the plastic limit loads of non-
idealized through-wall cracks, thus for shorter crack, the plastic
limit loads of thick-walled pipe with idealized through-wall cracks
could be used for thick-walled pipe with non-idealized through-
wall cracks within the accuracy of approximately 10%. The values
of these correction factors together with normalized plastic limit
loads of thick-walled pipe with idealized through-wall cracks are
summarized in Tables 58.
pressure induced axial tension and global bending moment with
arbitrary-shaped circumferential crack, where thin-wall assump-
tion together with force and moment equilibrium conditions
was utilized. Using analytical solutions proposed by Rahman
and Wilkowski (1998) incorporating the 1-dimensional numeri-
cal integration and geometric variables modeling arbitrary-shaped
circumferential cracks, the plastic limit loads of the present non-
idealized circumferential through-wall cracks can be predicted
based on the thin-wall assumption.
Tables 9 and 10 compare the present FE plastic limit moments
of non-idealized circumferential through-wall cracks with the pre-
dictions based on Rahman and Wilkowski (1998) for the case

Table 9
Comparison of the present plastic limit moments of non-idealized through-wall
4. Discussion
crack with the predictions from Rahman and Wilkowski (1998) (Rm /t = 5,  1 / 2 = 2).

Although the present FE limit analysis method was veried  1 / Normalized plastic limit moment, nNon-idealized
L,Bending

either against analytical solutions of un-cracked, thick-walled pipes Rahman and The present FE Diff. (%)
or against lower bound bulging factors of idealized through-wall Wilkowski (1998) results
cracked, thin-walled pipes based on 3-D FE analyses in Section 0.125 0.840 0.946 11.2
3.1, more direct comparisons were made for the present plastic 0.25 0.673 0.785 14.3
limit loads of pipes with non-idealized circumferential through- 0.3 0.606 0.715 15.2
wall cracks. Rahman and Wilkowski (1998) provided analytical 0.4 0.484 0.605 19.9
0.5 0.392 0.515 24.0
plastic limit load solutions of pipes under axial tension including
 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931 927

1/2 1/2

1/2 1/2

Fig. 9. Values of correction factors for thick-walled pipe (Rm /t = 2); (a)(c) a non-idealized circumferential through-wall cracks and (d) a non-idealized axial through-wall
cracks.

of Rm /t = 5 ( 1 / 2 = 2 in Table 9 and  1 / 2 = 3 in Table 10). As (1998). The differences between the two results are expected to
mentioned above, both plastic limit moments of non-idealized increase for thicker pipes (Rm /t = 2, 3) since the predictions from
circumferential through-wall cracks are normalized with respect Rahman and Wilkowski (1998) were derived based on the thin-wall
to plastic limit loads of un-cracked, thick-walled pipes (see Eqs. (2) assumption.
and (3)). As given in these tables, the predictions based on Rahman Plastic limit loads of cracked pipes are related to the remaining
and Wilkowski (1998) provided lower values (conservative) than ligament area (or the cracked area). Thus, in the present paper,
the present FE results, and the differences between two results work was carried out to investigate if equivalent idealized through-
increased as the reference crack angle ( 1 /) increased. This is due wall cracks can be used to predict the plastic limit loads for
to the fact that the analytical solutions from Rahman and Wilkowski non-idealized through-wall cracks. The plastic limit loads of
(1998) were derived based on the thin-wall assumption. However, non-idealized through-wall cracks were compared with those of
the differences in two results are approximately 10% for shorter idealized through-wall cracks with average crack angle or length
crack. In the present paper, only the results for the case of Rm /t = 5 as shown in Fig. 12, where crack angle or length in the inner
are compared with the predictions from Rahman and Wilkowski and outer pipe surfaces were averaged in circumferential and
axial non-idealized through-wall cracks. Tables 1114 compare the
present FE plastic limit loads of non-idealized through-wall cracks
Table 10 in thick-walled pipes with those of idealized through-wall cracks in
Comparison of the present plastic limit moments of non-idealized through-wall
thick-walled pipes with average crack angle or length. For brevity,
crack with the predictions from Rahman and Wilkowski (1998) (Rm /t = 5,  1 / 2 = 3).
only results for  1 / 2 = 3 for each loading condition are provided
 1 / Normalized plastic limit moment, nNon-idealized
L,Bending for circumferential through-wall cracks, while results for 1 /2 = 3
Rahman and The present FE Diff. (%) are given for axial through-wall cracks. In these tables, similar to
Wilkowski (1998) results other results, the plastic limit loads are normalized with respect
0.125 0.856 0.966 11.4
to plastic limit loads of un-cracked, thick-walled pipes (see Eqs.
0.25 0.723 0.855 15.5 (2) and (3)). As shown in these tables, the estimations using the
0.3 0.660 0.799 17.5 idealized through-wall cracks with average crack angle or length
0.4 0.547 0.690 20.7 give lower values than those using the non-idealized through-
0.5 0.429 0.610 29.7
wall cracks. The differences between the plastic limit loads of
928 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931

1/2 1/2

1/2 1/2

Fig. 10. Values of correction factors for thick-walled pipe (Rm /t = 3); (a)(c) a non-idealized circumferential through-wall cracks and (d) a non-idealized axial through-wall
cracks.

non-idealized through-wall cracks and those of idealized through-
wall cracks with average crack angle or crack length increased
as values of Rm /t and  1 / (or 1 ) increased, where the
maximum differences were around 23% for relatively longer ref-
erence cracks. However, the plastic limit loads of non-idealized
through-wall cracks agreed well with those of idealized through-
wall cracks with average crack angle or length when the reference
crack angle or length ( 1 / or 1 ) was relatively small. Although
only the results for  1 / 2 = 3 and 1 /2 = 3 are given here, the dif-
ferences between the two results are expected to increase as the

Table 11
Comparison of the present plastic limit loads of non-idealized through-wall crack with the those of idealized through-wall crack with average crack angle (Circumferential
through-wall cracks, Axial tension,  1 / 2 = 3).

Rm /t  1 / Normalized plastic limit loads, Eq. (3)

Non-idealized Idealized through-wall crack Diff. (%)

through-wall crack with average crack angle

2 0.125 0.973 0.958 1.5

0.25 0.872 0.807 7.4
0.3 0.810 0.735 9.6
0.4 0.672 0.595 11.4
0.5 0.562 0.473 15.7
3 0.125 0.956 0.944 1.3
0.25 0.844 0.777 7.9
0.3 0.778 0.706 9.2
0.4 0.644 0.570 11.5
0.5 0.542 0.451 16.9
5 0.125 0.948 0.927 2.3
0.25 0.828 0.750 9.4
0.3 0.764 0.679 11.1
0.4 0.641 0.547 14.6
0.5 0.546 0.431 21.1
 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931 929

1/2 1/2

1/2 1/2

Fig. 11. Values of correction factors for thick-walled pipe (Rm /t = 5); (a)(c) a non-idealized circumferential through-wall cracks and (d) a non-idealized axial through-wall
cracks.

Table 12
Comparison of the present plastic limit loads of non-idealized through-wall crack with the those of idealized through-wall crack with average crack angle (Circumferential
through-wall cracks, Global bending,  1 / 2 = 3).

Rm /t  1 / Normalized plastic limit loads, Eq. (3)

Non-idealized Idealized through-wall crack Diff. (%)

through-wall crack with average crack angle

2 0.125 0.973 0.965 0.8

0.25 0.893 0.832 6.8
0.3 0.835 0.762 8.7
0.4 0.709 0.634 10.6
0.5 0.610 0.515 15.6
3 0.125 0.971 0.957 1.5
0.25 0.867 0.810 6.6
0.3 0.808 0.744 8.0
0.4 0.689 0.617 10.5
0.5 0.597 0.499 16.4
5 0.125 0.966 0.944 2.3
0.25 0.855 0.785 8.2
0.3 0.800 0.724 9.5
0.4 0.690 0.600 13.1
0.5 0.610 0.484 20.6
930 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931

+
=

+
=

Fig. 12. Simplication of non-idealized through-wall crack into idealized through-wall crack by analogy with the ligament area: (a) a circumferential through-wall crack and
(b) an axial through-wall crack.

Table 13
Comparison of the present plastic limit loads of non-idealized through-wall crack with the those of idealized through-wall crack with average crack angle (Circumferential
through-wall cracks, Internal pressure,  1 / 2 = 3).

Rm /t  1 / Normalized plastic limit loads, Eq. (3)

Non-idealized through-wall crack Idealized through-wall crack with average crack angle Diff. (%)

2 0.125 0.989 0.985 0.4

0.25 0.972 0.958 1.5
0.3 0.960 0.937 2.4
0.4 0.922 0.870 5.6
0.5 0.854 0.745 12.8
3 0.125 0.991 0.983 0.8
0.25 0.966 0.956 1.0
0.3 0.952 0.936 1.6
0.4 0.913 0.858 6.0
0.5 0.816 0.711 12.8
5 0.125 0.983 0.978 0.5
0.25 0.955 0.952 0.3
0.3 0.944 0.931 1.4
0.4 0.913 0.866 5.2
0.5 0.871 0.738 15.3

Table 14
Comparison of the present plastic limit loads of non-idealized through-wall crack with the those of idealized through-wall crack with average crack length (Axial through-wall
cracks, Internal pressure, 1 /2 = 3).

Rm /t 1 Normalized plastic limit loads, Eq. (3)

Non-idealized through-wall crack Idealized through-wall crack with average crack length Diff. (%)

2 0.5 0.938 0.927 1.2

1 0.802 0.772 3.8
2 0.592 0.531 10.3
3 0.466 0.396 15.2
3 0.5 0.934 0.923 1.2
1 0.809 0.778 3.7
2 0.602 0.532 11.7
3 0.474 0.392 17.3
5 0.5 0.933 0.926 0.8
1 0.820 0.790 3.6
2 0.628 0.534 15.1
3 0.504 0.387 23.2
 D.-J. Shim et al. / Nuclear Engineering and Design 265 (2013) 918931
values of  1 / 2 and 1 /2 increase. Thus, it can be concluded that
the estimations using idealized through-wall cracks with average
crack angle or length provides conservative values of plastic limit
loads of thick-walled pipes with non-idealized through-wall cracks,
particularly for longer reference crack angle and length.
5. Concluding remarks
In the present paper, plastic limit loads were provided for
thick-walled pipes with non-idealized circumferential and axial
through-wall cracks based on detailed elastic-perfectly plastic 3-D
FE analyses. Regarding loading conditions, only internal pressure
was considered for the non-idealized axial through-wall cracks,
whereas axial tension, global bending, and internal pressure were
considered for the non-idealized circumferential through-wall
cracks. In order to cover a practical range of geometries of thick-
walled pipes with idealized and non-idealized through-wall cracks,
the variables affecting the plastic limit loads were systematically
varied.
Prior to development of the plastic limit loads of thick-walled
pipe with non-idealized through-wall cracks, the present FE anal-
ysis was veried against the existing analytical limit solutions of
un-cracked pipe and FE limit loads of thin-walled pipes with ide-
alized through-wall cracks. It was demonstrated that the lower
bound bulging factors of thin-walled pipe with idealized through-
wall cracks can be used for thick-walled pipe with idealized
through-wall cracks, provided the plastic limit loads of thick-
walled pipe with idealized through-wall cracks are normalized
appropriately with respect to those of un-cracked, thick-walled
pipe.
Based on the present FE results on thick-walled pipe with non-
idealized through-wall cracks, correction factors were provided to
quantify the effect of non-idealized through-wall crack shapes on
the plastic limit loads. The present results can be used to predict
load carrying capacity of thick-walled pipes with circumferential
and axial through-wall cracks (both idealized and non-idealized).
Based on the present FE results, the plastic limit loads of thick-
walled pipes with non-idealized through-wall cracks can easily
be obtained, from which the elastic-plastic J-integral of a thick-
walled pipe with non-idealized through-wall cracks can also be
calculated based on the reference stress concept. Moreover, pos-
sibility of simplication of non-idealized through-wall crack into
idealized through-wall crack using average crack angle or length
was also investigated in the present paper. The estimates from
using the idealized through-wall cracks with average crack angle
or length provided conservative results, particularly for longer ref-
erence crack angle and length.
Acknowledgements
This work was supported by the Nuclear Research and Devel-
opment Program of the Korea Institute of Energy Technology
931
Evaluation and Planning (KETEP) grant funded by the Korea gov-
ernment Ministry of Knowledge Economy (No. 2011520100010).
References
Dassault Systemes, 2012. Users Manual ABAQUS Version 6. Dassault Systemes, pp.
1121.
Electric Power Research Institute, 2010. Materials Reliability Program: Technical
Basis for Preemptive Weld Overlay for Alloy 82/182 Butt Welds in Pressurized
Water Reactors (PWRs) (MRP-169). Electric Power Research Institute, Revision
1-A.
France, C.C., Green, D., Sharple, J.K., Chivers, T.C., 1997. New stress intensity factor
and crack opening area solutions for through-wall cracks in pipes and cylinders.
Int. Con. Pres. Ves. Pip. 350, Fatigue and Fracture.
Hill, R., 1950. The Mathematical Theory of Plasticity. Oxford University Press, Oxford.
Huh, N.S., Kim, Y.J., Kim, Y.J., 2007. Limit load solutions for pipes with through-wall
crack under single and combined loading based on nite element analyses. J.
Press. Vess.-T ASME 129, 468473.
Huh, N.S., Shim, D.J., Choi, S., Wilkowski, G.M., Yang, J.S., 2008a. Stress intensity
factors for slanted through-wall cracks based on elastic nite element analyses.
Fatigue Fract. Eng. M. 31, 197208.
Huh, N.S., Shim, D.J., Choi, S., Park, K.B., 2008b. Stress intensity factors and crack
opening displacements for slanted axial through-wall cracks in pressurized
pipes. Fatigue Fract. Eng. M. 31 (6), 428440.
Huh, N.S., Shim, D.J., Yoo, Y.S., Choi, S., Park, K.B., 2010. Estimates of elastic crack
opening displacements of slanted through-wall cracks in plate and cylinder. J.
Press. Vess.-T ASME 132, 110.
Kim, Y.J., Huh, N.S., Kim, Y.J., 2001. Enhanced reference stress-based J and crack
opening displacement estimation method for leak-before-break analysis and
comparison with GE/EPRI method. Fatigue Fract. Eng. M. 24, 243254.
Kim, Y.J., Huh, N.S., Kim, Y.J., 2002. Quantication of pressure-induced hoop stress
effect on fracture analysis of circumferential through-wall cracked pipe. Eng.
Fract. Mech. 69, 12491267.
Kumar, V., German, M.D., 1988. Elastic-Plastic Fracture Analysis of Through-Wall
and Surface Flaws in Cylinders. EPRI, EPRI Report, NP-5596.
Materials Reliability Program, 2004. Welding Residual and Operating Stresses in
PWR Alloy 182 Butt Welds (MRP-106). EPRI.
Miller, A.G., 1988. Review of limit loads of structures containing defects. Int. J. Pres.
Ves. Pip. 32, 191327.
Nana, A.D., Yoon, K.K., 2006. Comparison of leak rates from Alloy 82/182 butt weld
cracks for leak-before-break applications. Int. Con. Pres. Ves. Pip., 2006-ICPVT11-
93767.
Rahman, S., Brust, F., Ghadiali, N., Wilkowski, G., 1998a. Crack-opening-area analyses
for circumferential through-wall cracks in pipes part I: analytical models. Int.
J. Pres. Ves. Pip. 75, 357373.
Rahman, S., Brust, F., Ghadiali, N., Wilkowski, G., 1998b. Crack-opening-area analyses
for circumferential through-wall cracks in pipes part II: model validation. Int.
J. Pres. Ves. Pip. 75, 375396.
Rahman, S., Wilkowski, G., 1998. Net-section-collapse analysis of circumferentially
cracked cylinders part I: arbitrary-shaped cracks and generalized equations.
Eng. Fract. Mech. 61, 191211.
Rice, J.R., 1968. A path independent integral and the approximate analysis of strain
concentration by notches and cracks. J. Appl. Mech. 35, 379386.
Rudland, D., Csontos, A., Shim, D.J., 2010. Stress corrosion crack shape development
using AFEA. J. Press. Ves.-T ASME 132 (1), 011406.
Rudland, D., Harrington, C., 2012. xLPR Pilot Study Report. U.S. Nuclear Regulatory
Commission, NUREG-2110.
Shim, D.J., Rudland, D., Harris, D., 2011. Modeling of subcritical crack growth due to
stress corrosion cracking: transition from surface crack to through-wall crack.
Int. Con. Pres. Ves. Pip., 57267.
USNRC, 1984. Evaluation of Potential for Pipe Breaks, vol. 3. USNRC, NUREG-1061.
USNRC, 1987. Standard Review Plan for the Review of Safety Analysis Reports for
Nuclear Power Plants: LWR Edition (also Known as the SRP). USNRC, Section
3.6.3; NUREG-0800.
USNRC, 2013. General Design Criteria for Nuclear Power Plants (GDC) 4; Environ-
mental and Dynamic Effects Design Bases, 10 CFR Part 50, Appendix A.
Zahoor, A., 1991. Ductile Fracture Handbook. Novetech Corp.