Engineering

Fracture
Mechanics
Engineering Fracture Mechanics 73 (2006) 829–854
www.elsevier.com/locate/engfracmech

New plastic collapse moment equations of defect-free

and throughwall circumferentially cracked elbows subjected
to combined internal pressure and in-plane bending moment
J. Chattopadhyay *, A.K.S. Tomar
Reactor Safety Division, Hall-7, Bhabha Atomic Research Center, Trombay, Mumbai 400 085, India

Received 25 August 2005; received in revised form 24 November 2005; accepted 2 December 2005

Abstract

Closed-form plastic collapse moments (PCM) equations were earlier proposed for throughwall circumferentially cracked
(TCC) elbow subjected to pure in-plane bending moment. However, an elbow is often subjected to combined internal pres-
sure and bending moment in actual service condition. Therefore, the present study investigates the effect of internal pressure
on the in-plane PCM of a TCC elbow. The PCM of a cracked elbow is usually expressed as a product of two parameters:
PCM of a defect-free elbow multiplied by a weakening factor due to the crack. Therefore, the present study also includes
analysis of defect-free elbows. Elastic–plastic finite element analysis is employed for the present analysis. A total of 396 cases
of elbows with various sizes of circumferential cracks (2h = 0–150
), different wall thickness (R/t = 5–20), different levels of
normalized internal pressure (p = PR/(try) = 0–1), different elbow bend radii (Rb/R = 2,3) and two different bending
modes, namely closing and opening are considered in the analysis. Elastic–perfectly plastic stress–strain response of material
is assumed. The load in the elbows is split in two components: a constant internal pressure applied initially followed by in-
plane bending moment monotonically increasing in definite steps. PCM are evaluated from moment—end rotation curves
by twice-elastic slope method. From these results, closed-form equations are proposed to evaluate PCM of TCC and defect-
free elbows subjected to combined internal pressure and in-plane closing/opening bending moment. Attempt has been made
to compare the predictions of the proposed equations with the available experimental/numerical results and to rationally
explain the behaviour where no experimental/numerical data is available for comparison.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Elbow; Pipe bend; Throughwall crack; Plastic collapse moment; Internal pressure; Finite element analysis; Closed-form
equation

1. Introduction

Pipe bends or elbows are commonly used components in a piping system. It is important to know its limit
load for the safe operation of the plant. The term ‘limit load’ is used in this paper in a generic sense to

*
Corresponding author. Tel.: +91 22 25591522; fax: +91 22 25505151.
E-mail address: [email protected] (J. Chattopadhyay).

0013-7944/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfracmech.2005.12.002
830 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

Nomenclature

a semi-crack length
D, Dm outer, mean diameter of elbow cross section
E Young’s modulus
h = tRb/R2 elbow factor or pipe bend characteristics
M0 plastic collapse moment of defect-free elbow
m0 = M0/(4R2try) normalized plastic collapse moment of defect-free elbow
ML plastic collapse moment of cracked elbow
mL = ML/(4R2try) normalized plastic collapse moment of cracked elbow
P internal pressure
p = PR/(try) normalized internal pressure
R mean radius of elbow cross section
Rb mean bend radius at elbow crown
t wall thickness of elbow
X = ML/M0 weakening factor of throughwall circumferentially cracked elbow plastic collapse moment
due to the presence of crack
/ circumferential position at elbow mid-plane (/ = 0
at intrados, 90
at crown and 180
at extra-
dos)
h semi-circumferential crack angle
ry yield stress of elbow material

Abbreviations
PCM plastic collapse moment
TCC throughwall circumferentially cracked
TES twice-elastic slope

collectively indicate either plastic instability or plastic collapse load. As per the definition of Gerdeen [1], the
plastic instability load is characterized by the zero slope of the load–deflection curve, which means the maxi-
mum load in the monotonic load–deflection curve (see Fig. 1). Plastic (collapse) load indicates a load where
significant plastic deformation occurs (not necessarily physical ‘collapse’ of the structure), determined by
applying a criterion of plastic collapse (e.g. twice-elastic slope (TES) as recommended by ASME [2]) on
the load–deflection curve. TES criterion is shown in Fig. 1. In this paper, plastic collapse load has always

tan φ1 = 2tan φ Plastic Instability load

φ1
φ
Plastic (collapse) load
(TES criterion)
Load

Deflection

Fig. 1. Definitions of limit load.

 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 831

been evaluated by TES criterion. At the limit load, the deformation of the elbow increases without signif-
icant increase in the load. Elbows may potentially contain cracks due to manufacturing defects or service
related degradation mechanisms. It is very important to know the effect of cracks on the limit moments
of elbows for integrity assessment of the piping system, especially when the piping material is very ductile
and the assessment is made through the CEGB R6 [3] approach. It has become all the more important in
nuclear industry because of the application of Leak-Before-Break concept in the design of primary heat
transport system piping. Recently, new plastic collapse moment (PCM) equations of throughwall circumfer-
entially cracked (TCC) elbows subjected to in-plane bending moments (both closing and opening mode)
were proposed by Chattopadhyay et al. [4,5]. These equations were an improvement over the existing solu-
tions of Miller [6] and Zahoor [7]. However, all these equations were valid for pure bending moment load-
ing. In actual service condition, an elbow is often subjected to combined internal pressure and bending
moment. It is, therefore, important to know the effect of internal pressure on the load carrying capacity of
elbows. Limit load of any cracked component is generally expressed as product of limit load of defect-free
component and a weakening factor due to the presence of crack. Therefore, before studying the limit load of
any cracked component, one should know the limit load of a defect-free component. PCM equations of
defect-free elbow subjected to combined internal pressure and bending moment were proposed by Chatto-
padhyay et al. [8,9]. However, those equations were developed assuming strain-hardening material proper-
ties. In limit analysis, the advantage of material strain hardening is not usually taken and elastic–perfectly
plastic material response is assumed.

2. Scope of the present work

From the above discussion it is clear that there is no closed-form equations available in the open literature
to evaluate limit moments of throughwall circumferentially cracked elbows subjected to combined internal
pressure and in-plane bending moment. Against this backdrop, the present study has been undertaken to eval-
uate the collapse load of throughwall circumferentially cracked elbows subjected to combined internal pres-
sure and in-plane bending moments through elastic–plastic finite element analysis. Opening and closing
mode of bending moment are analyzed separately as deformation characteristics of elbow under these two dif-
ferent modes are distinctly different, which have been observed by several researchers [4,5,8–12]. A total of 324
long radius (Rb/R = 3) elbows with various sizes of circumferential cracks including no-crack (2h = 0–150
),
different wall thickness (R/t = 5–20), different normalized internal pressure values (p = PR/(try) = 0–1) and
two different bending modes, namely closing and opening are considered in the analysis. Few short radius
(Rb/R = 2) elbows numbering 72 have also been analyzed to investigate the effect of bend radius on PCM.
Elastic–perfectly plastic stress–strain response of material is assumed. PCM are evaluated from moment—
end rotation curves by twice-elastic slope method. From these results, new closed-form expressions are pro-
posed to evaluate PCM of defect-free and TCC elbows under combined internal pressure and in-plane bending
moment. Separate equations are proposed for closing and opening modes. The consistency of these newly pro-
posed equations have been checked against the existing experimental and analytical results. The present work
concerns only the throughwall circumferential crack because, Griffiths [13], Yahiaoui et al. [14] and Chatto-
padhyay et al. [4,5] had concluded that throughwall circumferential crack could significantly weaken an elbow
by reducing its PCM.

3. Background

3.1. Defect-free elbows

Spence and Findlay [15] expressed the lower bound in-plane limit moment of a defect-free elbow as
(
0:8h0:6 ðD2 try Þ; for h < 1:45
M0 ¼ 2
ð1Þ
ðD try Þ; for h > 1:45
832 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

where, ry is the material yield stress, h is the bend characteristics equal to tRb/R2, Rb is the mean bend radius
of elbow and R, D and t are the mean radius, outer diameter and wall thickness of the elbow cross section
respectively.
Calladine [16] expressed the lower bound in-plane limit moment as

M 0 ¼ 0:935h2=3 ð4R2 try Þ; for h < 0:5 ð2Þ

Both the above expressions were based on small displacement analysis and assume ideal plastic material
behaviour. Based on large displacement analysis Goodall [17] proposed the maximum load carrying capacity
of the defect-free elbow subjected to closing bending moment as

1:04h2=3 2
M0 ¼ ðD try Þ ð3Þ
1þb
where
! pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi !
2=3
ð3hÞ 4 3ð1
m2 Þ ry R
b¼ 2þ
3 p E t

Touboul et al. [12] proposed the following equations of PCM of elbows based on the experimental study at
CEA DEMT
Closing collapse: M 0 ¼ 0:715h2=3 ð4R2 try Þ ð4Þ
1=3 2
Opening collapse: M 0 ¼ 0:722h ð4R try Þ ð5Þ
All the above equations are applicable only for the pure in-plane bending moment. The effect of internal pres-
sure on PCM was studied by Goodall [18], Rodabaugh [19], Hilsenkopf et al. [20], Touboul et al. [12], Shalaby
and Younan [10,11] and Chattopadhyay et al. [8,9].
Chattopadhyay et al. [8,9] proposed closed-form equations of PCM of defect-free elbows subjected to com-
bined internal pressure and in-plane (closing/opening) bending moment. The proposed equations were as
follows:
M0 p
2
¼ 1:122h2=3 þ 0:175
0:508p2 ðfor closing caseÞ ð6Þ
4R try h
M0 p
¼ 1:047h1=3 þ 0:124 1:2
0:568p2 ðfor opening caseÞ ð7Þ
4R2 try h
Applicability : 0:24 6 h 6 0:6 and 0:0 6 p 6 1:0
where, p = PR/(try) is the normalized internal pressure.
However, the above equations were generated assuming strain-hardening material properties. In limit ana-
lysis, the advantage of material strain hardening is not usually taken.

3.2. Throughwall circumferentially cracked elbow

Miller [6] was the first to propose the closed-form PCM equation of elbow with throughwall cracks based
on Griffiths’ experimental data [13]. For elbows with throughwall circumferential cracks extending from
crown towards extrados subjected to in-plane bending moment, Miller [6] proposed the following equation
to evaluate collapse moment (ML):
ML 3h
¼1
ð8Þ
M0 2p
M 0 ¼ 0:935ð4R2 trf Þh2=3 ð9Þ
where, 2h is the total crack angle and ‘rf’ is the material flow stress usually taken as average of yield and ulti-
mate strength.
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 833

In general, a lower bound equation was proposed by Miller [6] as follows:

ML
¼ cosðh=2Þ
0:5 sin h ð10Þ
M0
Zahoor [7] also proposed closed-form PCM equations of throughwall cracked elbows under in-plane bend-
ing moment based on Griffiths’ experimental data [13]. The equation proposed for TCC elbow was as follows
[7]:
"    2  3 #
a a a
M L ¼ M 0 1
0:2137
0:0485
1:0559
Dm Dm Dm ð11Þ
Applicability: a=Dm 6 0:8; h 6 0:5 and Dm =t P 15

where, M0 is as defined in Eq. (9), ‘a’ is the half crack length, ‘Dm’ is the mean diameter of the elbow cross
section, ‘t’ is the elbow wall thickness, ‘h’ is the elbow factor or bend characteristics.
However, these equations do not differentiate between opening and closing modes of bending moment,
although elbow deformation characteristics are distinctly different for these two in-plane bending modes.
Yahiaoui et al. [14] while comparing their test results of through wall cracked elbows under opening moment
with the predictions of Miller [6] and Zahoor [7] found that the existing solutions were excessively conservative
and on occasions, non-applicable to the cases for which they were intended. Yahiaoui et al. [14], therefore,
advocated the need to revise the existing solutions. Accordingly, a comprehensive numerical exercise [4,5]
was undertaken to propose new PCM equations of TCC elbow. The proposed equations were as follows:
For closing mode

ML ¼ M0  X ð12Þ
2=3 2
M 0 ¼ 1:075h ð4R try Þ ð13Þ

where, ML and M0 are the limit moments of cracked and defect-free elbows respectively and X is the weak-
ening factor which is a function of crack size (2h) and the R/t of the elbow cross section. The function X is
shown in Table 1.
For Opening mode
The basic form of the equation is as given in Eq. (12) with M0 and X defined as follows:
M 0 ¼ ð1:0485h1=3
0:0617Þð4R2 try Þ ð14Þ
For 5 6 R/t 6 20
( h
1:127
1:8108 p
for 45 6 2h 6 150
X ¼  ð15Þ
1
0:8 ph for 0 6 2h 6 45

Figs. 2 and 3 show these equations in graphical form. The present work investigates the effect of internal pres-
sure on the in-plane (closing/opening) PCM of defect-free and TCC elbows.

Table 1
A0, A1 and A2 values for function X = A0 + A1(h/p) + A2(h/p)2
R/t A0 A1 A2 h limits
5 1.1194
0.7236
2.0806 for 45
6 2h 6 150
and X=1 for 2h < 45

7.5 1.1185
0.3420
2.5200 for 60
6 2h 6 150
and X=1 for 2h < 60

10 0.9655 1.0152
4.6800 for 60
6 2h 6 150
and X=1 for 2h < 60

15 1.1400 0.3000
3.6000 for 90
6 2h 6 150
and X=1 for 2h < 90

20 0.6400 3.4200
7.920 for 90
6 2h 6 150
and X=1 for 2h < 90

834 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

1.0
Eq.(1
1 ), Zah
0.9 oor [7
]
Eq
0.8 .(8
), Mi

X = ML/Mo
lle
0.7 r [6
] 20
15
0.6
10
0.5 7.5

R/t =5
0.4

0.3
0 20 40 60 80 100 120 140 160
Cicumferential crack angle, 2θ (deg.)

Fig. 2. Variation of weakening factor (X) with crack angle for TCC elbows under pure closing moment, no pressure (symbols show the FE
results and solid lines show predictions of closed-form equations).

1.0

0.8

0.6
X = ML/M0

0.4 R/t = 5
R/t = 10
R/t = 15
0.2 R/t = 20
Bi-Linear fit of R/t = 5 results

0.0
0 20 40 60 80 100 120 140 160
Circumferential crack angle, 2θ (deg.)

Fig. 3. Variation of weakening factor (X) with crack angle for TCC elbows under pure opening moment, no pressure (symbols show the
FE results and solid lines show predictions of closed-form equations).

4. Finite element analysis

The finite element method is used to conduct the investigation on the effect of internal pressure on in-plane
PCM of elbows. Finite element program WARP3D [21] is used for this study. Non-linear finite element anal-
ysis is carried out to determine the PCM of elbows for various geometric and loading combinations. Both geo-
metric and material non-linearity are considered in the analysis. Consideration of geometric nonlinearity is
very important to capture the ovalisation of elbow cross section during bending. Moment versus end rotation
curves are generated through finite element analysis where end rotation (w) is defined as: w = tan
1[(u1
u2)/
2R], u1 and u2 are the axial displacements of two diametrically opposite points at mean radial position, situ-
ated at the end plane of connecting straight pipe where moment is applied. PCM are obtained by twice-elastic
slope (TES) method from these curves. The following sections briefly describe the different aspects of the finite
element analysis.

4.1. Geometry

Fig. 4a and 4b show the geometry of a throughwall circumferentially cracked (TCC) elbow. The crack is
centered at extrados or intrados depending on the mode of bending moment applied. The extrados crack is
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 835

6R
Extrados
crack
crack
2θ

Rb D t

Intrados

Cracked Section

Fig. 4a. Geometry of a throughwall circumferentially cracked elbow under closing moment.

6R
Extrados

Rb crack
D t
2θ

crack
Intrados

Cracked Section

Fig. 4b. Geometry of a throughwall circumferentially cracked elbow under opening moment.

Table 2
Various geometric and loading parameters of analyzed elbows
Bending mode: Closing Bending mode: Opening
R: 250 mm R: 250 mm
R/t: 5, 7.5, 10, 15, 20 R/t: 5, 10, 15, 20
Rb/R: 3 Rb/R: 3
2h: 0
, 45
, 60
, 90
, 120
, 150
2h: 0
, 45
, 60
, 90
, 120
, 150

p: 0, 0.2, 0.4, 0.6, 0.8, 0.9 p: 0, 0.2, 0.4, 0.6, 0.8, 0.9
No. of cases: 180 No. of cases: 144
Bending mode: Closing Bending mode: Opening
R: 250 mm R: 250 mm
R/t: 5, 20 R/t: 10
Rb/R: 2 Rb/R: 2
2h: 0
, 45
, 60
, 90
, 120
, 150
2h: 0
, 45
, 60
, 90
, 120
, 150

p: 0, 0.2, 0.4, 0.6 p: 0, 0.2, 0.4, 0.6
No. of cases: 48 No. of cases: 24
Total no. of cases: 396

assumed for closing mode of bending moment and intrados crack is assumed for opening mode of bending
moment. Geometrically, a 90
TCC elbow is characterized by three parameters, namely, Rb/R, R/t and 2h.
Table 2 shows different combinations of these parameters taken in the study. In general, Rb/R = 3 is assumed
in the present study indicating a long radius elbow. However, few short radius elbows with Rb/R = 2 have also
been analyzed to study the effect of bend radius on collapse moment of elbows. The R/t is varied from 5 to 20
covering a wide range for engineering use. Crack angles (2h) are varied from 0
to 150
. Crack angle 2h = 0

836 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

indicates a defect-free elbow. In the present analyses, the elbow is connected with straight pipes of length equal
to the six times the mean cross sectional radius. It is important to note that this straight pipe length allows free
ovalisation propagation from mean elbow section.

4.2. Material

Elastic–perfectly plastic material model is used in this study. The other material properties chosen are as
follows:

Young’s modulus: 200 GPa,

Yield stress: 300 MPa,
Poisson’s ratio: 0.3.

4.3. Finite element model

Twenty-noded solid elements with reduced integration order 2 · 2 · 2 are used to model the elbow. Because
of symmetry, only one fourth of the elbow is modeled. There are total 560 elements and 3405 nodes for elbows
with intrados crack (opening bending) and 528 elements and 3221 nodes for elbows with extrados crack (clos-
ing bending). Spider web type mesh is employed near the crack tip for mesh economy. Eight numbers of radial
and circumferential divisions are employed at the spider web with two elements across the thickness of elbow.
A very small hole of radius 0.5% of crack length is introduced at crack tip in the finite element model for faster
convergence without much change of results. This was recommended by Kumar et al. [22]. Fig. 5 shows a typ-
ical finite element mesh. The same mesh pattern is used for all the cases. A customized pre-processor of NISA
[23] is used to generate the FE mesh from the given geometry parameters. A mesh convergence study with four
elements across the thickness and more number of elements at the spider web has been performed to check the
adequacy of this mesh.

4.4. Loading

The load in the elbows is split in two components: a constant internal pressure and varying in-plane bend-
ing moment monotonically increasing in definite steps. The pressure is applied in initial 100 load steps and

Fig. 5. Typical finite element mesh of an elbow.

 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 837

subsequently held constant. The rationale behind keeping pressure constant is that internal pressure generally
does not increase during service. Whereas, bending moment may increase significantly in an accidental con-
dition (e.g. earthquake). Thus it is of interest to predict PCM of an elbow for a constant internal pressure.
Internal pressure is normalized as, p = PR/(try), where P is the applied internal pressure and ry is the material
yield stress. The various normalized pressures considered in the analysis are p = 0 (i.e. pure bending moment),
0.2, 0.4, 0.6, 0.8 and 0.9. Few additional cases with p = 0.1, 0.3, 0.5, 0.7 have also been analyzed for defect-free
elbows under closing mode of bending moment. The p = 1.0 case is analyzed for defect-free elbows and it has
been observed that numerical divergence occurs even before the full pressure is applied prior to application of
bending moment, which indicates plastic collapse at zero bending moment. Consequently, p = 1.0 cases are
not analyzed for cracked elbows. Closed end condition is simulated by applying axial pressure of intensity
‘PR/(2t)’ at the end of the connecting straight pipe. There are two modes of in-plane bending moment: closing
and opening. Several researchers [4,5,8–12] reported a significant difference in the elbow deformation charac-
teristics for these two modes of bending moment. Consequently, in the present study, closing and opening
bending moments are considered separately. Bending moment is simulated as triangularly varying face
pressure. However, within an element, face pressure has been kept constant. The face pressure value is
obtained as ‘M Æ c/I’ where ‘M’ is the applied bending moment, ‘I’ is the area moment of inertia of the elbow
cross section and ‘c’ is the vertical distance of the element face center from the neutral axis. The application of
bending moment in this way avoids the unwanted plastic deformation at the point of concentrated load
application.

5. Results and discussion

For each case, moment-end rotation curves are generated through non-linear finite element analysis. PCM
is then evaluated from the moment rotation curve by TES method. In few cases, numerical divergence
occurred indicating plastic instability prior to the attainment of TES PCM. In those cases, the instability point
has been taken as plastic collapse point. In some cases of deeply cracked elbows (2h P 120
) subjected to very
high pressure (p P 0.8), non-linearity sets in moment–rotation curves at very early stage. This makes it diffi-
cult to get a clear elastic slope and hence in those cases PCM based on TES criterion are not evaluated. Before
analyzing cracked elbows, defect-free elbows subjected to combined internal pressure and bending moment are
analyzed. Subsequently, the weakening factor because of the presence of crack has been quantified by evalu-
ating the ratio of collapse moments of cracked and defect-free elbows (X = ML/M0). Finally, closed-form
equations of PCM are proposed by curve fitting the FE data. The results of defect-free elbows are first
presented and then the effect of cracks and internal pressure on the PCM is shown. All the newly proposed
equations are also shown in Appendix A for viewing at a glance.

5.1. Defect-free elbow

Fig. 6 shows typical moment–rotation data for a defect-free long radius (Rb/R = 3) elbow having R/t = 10
and being subjected to various levels of internal pressure and closing bending moment. The figure also shows
the TES plastic collapse points. Fig. 7 shows the variation of normalized PCM (m0 = M0/(4R2try)) with nor-
malized internal pressure (p = PR/(try)) for various elbow thicknesses (R/t). The basic trend of these curves is
same as observed earlier by Chattopadhyay et al. [8,9] for strain-hardening material. It may be seen from Figs.
6 and 7 that internal pressure stiffens the elbow compared to when it is subjected to pure bending moment. The
stiffening effect is more significant for thinner elbow. It is observed that PCM increase gradually with appli-
cation of internal pressure, reach a peak and then start falling with further increase in internal pressure finally
heading towards zero at p = 1.0. This is in agreement with the observations of Robertson et al. [24]. The ovali-
sation of the elbow cross section plays an important role in its collapse. The application of uniform internal
pressure opposes the ovalisation of the elbow cross section, thus delaying the collapse phenomenon. Ovalisa-
tion is more prominent in case of thin walled elbow. That is why internal pressure enhances the limit moments
significantly in thin walled elbow. However, if the internal pressure is increased beyond a limit, the hoop stress
due to internal pressure nullifies the beneficial effect on the limit moments and finally the limit moment starts
reducing with further increase in internal pressure. The FE data of these normalized parameters (m0, p, h) have
838 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

1600 Defect-free Elbow, R/t = 10, Rb/R = 3, Closing mode

0.6 0.4
1400 0.3
0.2
1200

Moment (kN-m)
1000 0.9
p=0
800

600
TES plastic collapse moments
400

200

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
End rotation (deg.)

Fig. 6. Moment–rotation curves of a typical defect-free elbow subjected to various levels of internal pressure and in-plane closing bending
moment.
Normalized collapse moment (m0 = M0 /4R2tσy )

1.0

0.8 5

0.6 7.5

10

0.4 15 Closing bending mode

20
t = Symbols: FE data
R/ Solid lines: Proposed fit (Eq.16)
0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0
Normalized internal pressure (p = PR/t σy)

Fig. 7. Variation of normalized closing PCM (m0) with normalized internal pressures (p) for various wall thickness (R/t) of defect-free
elbows.

been best fitted and an equation is proposed to evaluate the PCM of defect-free elbow under combined loading
of in-plane closing bending moment and internal pressure:
 
M0 2:071p1:418
m0 ¼ 2 ¼ 1:075h2=3 þ þ 8:41p 12:129
½1
p ð16Þ
4R try h0:223
The above equation reduces to Eq. (13) in case of pure closing bending moment without any internal pressure
(p = 0).
In case of opening mode of bending moment, the basic trend is same as that observed in case of closing
moment. However, in case of opening mode, the fall of normalized collapse moment starts at lower value
of normalized internal pressure. Fig. 8 shows the variation of normalized PCM (m0 = M0/(4R2try)) with nor-
malized internal pressure (p = PR/(try)) for various elbow thickness (R/t). Similar to closing mode, the FE
data of these normalized parameters (m0, p, h) have been best fitted and the following equation is proposed
to evaluate the PCM of defect-free elbow under combined loading of in-plane opening bending moment
and internal pressure:
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 839

1.0

Normalized collapse moment (m0 = M0 /4R tσy)

2
5
0.8
10
15
0.6
R/t = 20

0.4 Opening bending moment

Symbols: FE data
Solid lines: Proposed fit (Eq.17)
0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0
Normalized internal pressure (p = PR/tσy)

Fig. 8. Variation of normalized opening PCM (m0) with normalized internal pressures (p) for various wall thickness (R/t) of defect-free
elbows.

 
M0 1=3 1:2182p 9:6431
m0 ¼ ¼
0:0617 þ 1:0485h þ þ 7:8509p ½1
p ð17Þ
4R2 try h1=3
The above equation reduces to Eq. (14) in case of pure opening bending moment without any internal pressure
(p = 0).

5.2. Cracked elbow

5.2.1. Closing mode

Fig. 9 shows a typical moment rotation data of various TCC elbows subjected to combined internal pres-
sure and closing bending moment for R/t = 15 and p = 0.4. The curves also show the TES plastic collapse
points. Almost similar trend is observed for all other cases. It may be seen from Fig. 9 that elbows with smaller
crack angles subjected to internal pressure require certain amount of closing moment to initiate closing rota-
tion. In contrast, elbows (closed end) with larger crack angles subjected to internal pressure will have initial
closing rotation even before closing moment is applied. This may be explained as follows. An elbow when sub-
jected to internal pressure has larger pressure boundary at the extrados (ab in Fig. 10) compared to intrados

1200

1000

800
Moment (kN-m)

R/t = 15, p = 0.4, Closing mode

600 2θ = 0°
2θ = 45°
2θ = 60°
400
2θ = 90°
2θ = 120°
200 2θ = 150°
TES plastic collapse

0
0 1 2 3 4 5 6
End rotation (deg.)

Fig. 9. Moment–rotation curves of typical TCC elbows subjected to internal pressure and in-plane closing bending moment.
840 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

a
PR/2t P
crack
c

d b

PR/2t

Fig. 10. Loadings on an internally pressurized elbow.

(cd in Fig. 10). This generates an unbalance opening moment (the so called Bourdon effect). The magnitude of
this unbalance opening moment is proportional to the difference of pressure boundary between extrados and
intrados of elbow and is independent of crack size. Another associated effect of internal pressure when applied
on an elbow with closed end is the axial stress (equal to PR/(2t)), which generates a closing rotation that is pro-
portional to the crack size. For smaller crack sizes, the opening rotation because of Bourdon effect dominates
and hence certain amount of closing bending moment is required to initiate closing rotation. For higher crack
sizes, the closing rotation because of axial stress PR/(2t) dominates and hence elbows will have initial closing
rotation even before closing moment is applied.
Figs. 11–15 show the effect of internal pressure on the PCM of TCC elbows subjected to combined internal
pressure and closing bending moment. All the data points are generated for long radius elbows (Rb/R = 3). To
express the PCM of cracked elbows, a weakening factor (X) is defined as the ratio of PCM of cracked elbows
to that of the defect-free elbows (X = ML/M0). It was earlier shown by Chattopadhyay et al. [4,5] and also
shown in Fig. 2 that there is a threshold crack angle below which a through wall circumferential crack at elbow
extrados does not weaken an elbow. This threshold crack angle increases with increasing R/t. An elbow when
subjected to pure closing bending moment generates tensile bending stress along the extrados and compressive
stress along the intrados (see Fig. 16). Because of the curved profile of elbow, this, in turn, generates compo-
nent of forces, which ovalises the circular cross section of the elbow in such a way that reduces the diameter
across the intrados–extrados (across ‘a–b’ in Fig. 16). This pattern of ovalisation stretches the part ‘c–d’ of the

1.0

0.8
X = ML/M0

0.6

R/t = 5, Closing mode p = 0.0

0.4 Symbols: FE data
Solid Lines: Proposed fit (Eq.18) 0.2

0.2 0.4
0.6
0.9 0.8
0.0
0 20 40 60 80 100 120 140 160
2θ (deg.)

Fig. 11. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows under combined internal pressure
and closing bending moment: R/t = 5.
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 841

1.0

0.8

X = ML/M0
0.6
p = 0.0
R/t = 7.5, Closing mode
0.4 Symbols: FE data 0.2
Solid Lines: Proposed fit (Eq.19)
0.4
0.2
0.6
0.9 0.8
0.0
0 20 40 60 80 100 120 140 160
2θ (deg.)

Fig. 12. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows under combined internal pressure
and closing bending moment: R/t = 7.5.

1.0

0.8
X = ML/M0

0.6 p = 0.0
R/t = 10, Closing mode
0.4 Symbols: FE data 0.2
Solid Lines: Proposed fit (Eq.20)
0.4
0.2
0.6
0.9 0.8
0.0
0 20 40 60 80 100 120 140 160
2θ (deg.)

Fig. 13. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows under combined internal pressure
and closing bending moment: R/t = 10.

1.0

0.8

p = 0.0
X = ML/M0

0.6

0.4 R/t = 15, Closing mode 0.2

Symbols: FE data
Solid Lines: Proposed fit (Eq.21)
0.2 0.4

0.6
0.9
0.8
0.0
0 20 40 60 80 100 120 140 160
2θ (deg.)

Fig. 14. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows under combined internal pressure
and closing bending moment: R/t = 15.
842 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

1.0

0.8
p = 0.0

X = ML/M0
0.6

0.2
0.4
R/t = 20, Closing mode
Symbols: FE data
0.2 Solid lines: Proposed fit (Eq.22) 0.4

0.9
0.8 0.6
0.0
0 20 40 60 80 100 120 140 160
2θ (deg.)

Fig. 15. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows under combined internal pressure
and closing bending moment: R/t = 20.

Crack is
postulated
here a
a
c d
2θth

f e

Deformed
Undeformed b
c/s
c/s

Fig. 16. Sketch to explain the presence of threshold through wall circumferential crack angle at extrados to cause weakening of elbow
subjected to closing moment.

circular cross section of elbow (as shown in Fig. 16), which generates a Poisson’s compressive stress field per-
pendicular to the cross section i.e. along the axis of the elbow. If the crack size at extrados is less than the
circumferential length ‘cd’ (or 2hth as shown in Fig. 16), the crack will be under a compressive normal stress
field and will not open up and hence will not cause any weakening of the elbow. Fig. 17 depicts the typical

Inside surface Extrados

Mid surface
400 R/t = 20, p = 0
Outside surface
φ Crown
300
200
Intrados
Axial Stress (MPa)

100
0 30 60 90 120 150 180
0
Intrados Crown Extrados
-100 φ (deg.)

-200

-300

-400

Fig. 17. Variation of axial stress at mid-plane of elbow (R/t = 20 and Rb/R = 3, no crack) under closing bending moment at the TES
plastic collapse moment.
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 843

axial stress (which is responsible for opening a circumferential crack) variation of a thin (R/t = 20) elbow sub-
jected to closing bending moment. It is seen from Fig. 17 that in the region from extrados extended towards
crown, axial stress in the thin elbow (R/t = 20) is compressive in nature from / = 180
to almost / = 143

(/ indicates the circumferential position at elbow mid-plane). When crack tip falls in the tensile stress field,
it starts weakening the elbow. This explains the presence of a threshold crack angle to initiate the weakening.
It should also be noted that the more is the ovalisation, the higher is the stretch of compressive stress field and
hence higher is the threshold crack angle ‘2hth’. It is well known that ovalisation is more for thin elbow com-
pared to thick elbow. It explains smaller threshold crack angle for thicker elbow to initiate weakening of load
bearing capacity of elbow.
It may be seen from Figs. 11–15 that application of internal pressure reduces the threshold crack angle and
increases the weakening due to the crack (i.e. reduces the weakening factor X). This may be explained as fol-
lows. Pure in-plane closing bending moment ovalises an elbow cross section as shown in Fig. 16. Internal pres-
sure opposes this ovalisation and tries to make the cross section circular resulting in lower ovality. This, in
turn, reduces or completely removes the stretch of compressive axial stress at the elbow extrados and hence
increases the weakening due to the presence of crack. Fig. 18 shows the axial stress distribution in the same
elbow (as shown in Fig. 17 with R/t = 20) subjected to combined internal pressure (p = 0.4) and TES plastic
collapse moment. It may be noted that the compressive stress field near the extrados is significantly reduced at
p = 0.4.
The FE data of these normalized parameters (m0, p, 2h) have been best fitted and equations are proposed to
evaluate the PCM of TCC elbows under combined internal pressure and in-plane closing bending moment.
The basic form of the equation is same as Eq. (12) with M0 defined by Eq. (16) and X defined as a function
of (h/p) and normalized pressure (p) for various R/t. The proposed equations of X are as follows:

For R/t = 5
   2  2:1758
h h h
X ¼ 1:1194
0:7236
2:0806
3:4164p0:8408 ð18Þ
p p p
For R/t = 7.5
 0:6687  1:7345
h 0:7136 h
X ¼ 1:4423
1:495
2:9803p ð19Þ
p p
For R/t = 10
 0:4082  1:4381
h h
X ¼ 1:6039
1:0847
3:1773p0:4807 ð20Þ
p p

500 Inside surface R/t = 20, p = 0.4

400 Midsurface
Outside surface
300
200
Axial stress (MPa)

100
Intrados Crown Extrados
0
0 30 60 90 120 150 180
-100 φ (deg.)
-200

-300
-400
-500

Fig. 18. Variation of axial stress at mid-plane of elbow (R/t = 20 and Rb/R = 3, no crack) under combined internal pressure (p = 0.4) and
in-plane closing bending moment equal to TES plastic collapse moment.
844 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

For R/t = 15
 0:5435  0:9644
h h
X ¼ 1:4298
0:0789
3:3789p0:2336 ð21Þ
p p
For R/t = 20
 0:0231  1:2776
h 0:5464 h
X ¼ 7:7803
6:8959
4:1061p ð22Þ
p p

Applicability of Eqs. (18)–(22): 0 6X 6 1, 2h 6 150
, 1 > p > 0

For intermediate R/t, X can be linearly interpolated between the adjacent R/t values. However, for conser-
vative results, the equation applicable for next lower R/t may be chosen. Please note that the above Eqs. (18)–
(22) should not be used for p = 0 and 1.

5.2.2. Opening mode

Fig. 19 shows a typical moment rotation data of various TCC elbows subjected to combined internal pres-
sure and opening bending moment for R/t = 5 and p = 0.4. The curves also show the TES plastic collapse
points. Almost similar trend is observed for all other cases. Figs. 20–23 show the effect of internal pressure
on the PCM of TCC elbows subjected to combined internal pressure and opening bending moment. All the
data points have been generated for long radius elbows (Rb/R = 3). As in the case of closing mode, results
are expressed as variation of weakening factor (X) with normalized internal pressure (p) and crack angle
(2h) for various R/t. Comparing Figs. 20–23 with Figs. 11–15, it may be observed that for opening mode,
the weakening factor (X) is not a strong function of internal pressure as in the case of closing mode. It
may be explained as follows. An elbow when subjected to pure in-plane opening bending moment ovalises
as shown in Fig. 24. This generates a typical axial stress pattern as shown in Fig. 25. It may be observed from
Fig. 25 that axial stress is mostly tensile in the region from intrados towards crown, where the throughwall
circumferential crack is postulated. Therefore, internal pressure has much smaller role in additional crack
opening and hence enhancing the weakening due to the presence of a crack at intrados of an elbow subjected
to opening bending moment. This is in contrast to the closing mode where internal pressure converts a stretch
of compressive stress field at elbow extrados to tensile stress zone and thus creating additional crack opening
leading to enhanced weakening due to the crack.
Accordingly, on a conservative estimate, a lower bound curve of weakening factor (X) versus crack angle
(2h) is plotted for each R/t (see Figs. 20–23). Fig. 26 shows all these lower bound curves for various R/t. It may

R/t = 5, p = 0.4, Opening mode

3500
-fre e
De fect 2θ = 45
o

3000 o
60

2500 90
o
Moment (kN-m)

o
2000 120

1500 150
o

1000

TES plastic collapse points

500

0
0 1 2 3 4 5
End rotation (deg.)

Fig. 19. Moment–rotation curves of typical TCC elbows subjected to internal pressure and in-plane opening bending moment.
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 845

R/t = 5, Opening mode

1.0

0.8

X= ML/M0
0.6
p = 0.0
p = 0.2
0.4 p = 0.4
p = 0.6
p = 0.8
0.2 p = 0.9
Lower bound curve

0.0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 20. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows subjected to combined internal
pressure and opening bending moment: R/t = 5.

R/t = 10, Opening mode

1.0

0.8
X= ML/M0

0.6

p = 0.0
0.4 p = 0.2
p = 0.4
p = 0.6
p = 0.8
0.2
p = 0.9
Lower bound curve

0.0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 21. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows subjected to combined internal
pressure and opening bending moment: R/t = 10.

be observed that weakening factor is nearly identical for various R/t. Almost same observation was made by
Chattopadhyay et al. [4,5] while analysing TCC elbows under pure in-plane opening bending moment. Hence,
a lowest bound conservative curve of weakening factor (X) versus crack angle (2h) is plotted disregarding the
effect of internal pressure and R/t. A two-part curve-fitting is done on this lowest bound curve and the pro-
posed equations are as follows:
( 
1
1:2776 ph for 0 6 2h 6 45
X ¼ h  h 2 ð23Þ
1:3053
4:2133 p þ 3:8386 p for 45 < 2h 6 150

Fig. 27 compares the predictions of Eqs. (23) and (15) to show the effect of internal pressure on the weakening
factor (X) due to a throughwall circumferential crack at intrados of an elbow subjected to combined internal
pressure and opening bending moment. It may be noted from Fig. 27 that internal pressure moderately in-
creases the weakening effect (i.e. decreases the weakening factor X) of a throughwall circumferential crack
at elbow intrados, compared to the pure in-plane opening moment loading (p = 0).
846 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

R/t = 15, Opening mode

1.0

0.8

X= ML/M0
0.6

p = 0.0
0.4 p = 0.2
p = 0.4
p = 0.6
0.2 p = 0.8
p = 0.9
Lower bound curve
0.0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 22. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows subjected to combined internal
pressure and opening bending moment: R/t = 15.

R/t = 20, Opening mode

1.0

0.8

0.6
X=ML/M0

p = 0.0
0.4 p = 0.2
p = 0.4
p = 0.6
0.2 p = 0.8
p = 0.9
Lower bound curve
0.0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 23. Variation of weakening factor (X) with internal pressure and crack angle (2h) for TCC elbows subjected to combined internal
pressure and opening bending moment: R/t = 20.

Deformed
Undeformed
c/s
c/s a
c d
a

f e
b

Crack is postulated here

Fig. 24. Sketch to explain pattern of ovalisation of elbow cross section when subjected to opening bending moment.
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 847

Inside surface Extrados

600 Mid surface
Outside surface
φ Crown

400
Intrados

Axial stress (MPa)

200

0 30 60 90 120 150 180

0 Intrados Crown Extrados
φ (deg.)

-200

-400

Fig. 25. Variation of axial stress at mid-plane of elbow (R/t = 10 and Rb/R = 3, no crack) under opening bending at load beyond the TES
collapse moment.

R/t = 05 Lower bound curve

1.0 R/t = 10 Lower bound curve
R/t = 15 Lower bound curve
R/t = 20 Lower bound curve
0.8 Lowest bound curve
X = ML/M0

0.6

0.4

0.2

0.0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 26. Lowest bound curve showing the variation of weakening factor (X) with crack angle (2h) disregarding the effect of internal
pressure and R/t for TCC elbows subjected to combined internal pressure and opening bending moment.

Combined Internal Pressure and Opening Moment

1.0

Eq.(15) by Chattopadhyay et al [4,5] for p = 0

0.8
X = ML/M0

0.6

0.4
Proposed fit (Eq.23) for 1 > p > 0
0.2

0.0
0 30 60 90 120 150 180
Crack angle, 2θ (deg.)

Fig. 27. Proposed fit for weakening factor (X) versus crack angle (2h) curve for TCC elbows subjected to combined internal pressure and
opening bending moment and its comparison with un-pressurized (p = 0) case.
848 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

5.3. Effect of bend radius on weakening factor (X)

Effect of bend radius on in-plane collapse moment of cracked elbows has been investigated by comparing
weakening factors (X) of few short radius (Rb/R = 2) elbows with those of long radius (Rb/R = 3) elbows for
various crack sizes (2h), normalized internal pressure (p) and R/t ratios and mode of bending of elbows. Figs.
28–30 show the comparison of weakening factors. These figures show that there is insignificant effect of bend
radius on the weakening factors (X) for thin (R/t = 20) elbows under closing mode and for other elbows at
lower level of internal pressure (p 6 0.2). However, for thicker elbows under closing mode and elbows under
opening mode subjected to higher internal pressure (p P 0.4), short radius (Rb/R = 2) elbows show higher
weakening (i.e. lower X) compared to long radius (Rb/R = 3) elbows for the same size of a throughwall cir-
cumferential crack. The difference is more for opening mode of bending moment, compared to the closing
mode. Therefore, the proposed equations of weakening factors due to presence of a throughwall circumferen-
tial crack (Eqs. (18)–(23)) are mainly applicable for long radius elbows and may be used for short radius
elbows at lower level of internal pressure (p 6 0.2).

R/t = 20, Closing mode

1.0

0.8
p=0

0.6
X = ML/M0

Rb/R = 2
Rb/R = 3 p = 0.2
0.4

p = 0.4
0.2 p = 0.6

0.0
0 30 60 90 120 150 180
Crack angle, 2θ (deg.)

Fig. 28. Effect of elbow bend radius on weakening factors (X) for various levels of internal pressure: R/t = 20, Closing mode.

R/t = 5, Closing mode

1.0

0.8
X = ML/M0

0.6 Rb/R = 2, p = 0
Rb/R = 3, p = 0
Rb/R = 2, p = 0.2
0.4 Rb/R = 3, p = 0.2
Rb/R = 2, p = 0.4
Rb/R = 3, p = 0.4
0.2 Rb/R = 2, p = 0.6
Rb/R = 3, p = 0.6

0.0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 29. Effect of elbow bend radius on weakening factors (X) for various levels of internal pressure: R/t = 5, Closing mode.
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 849

R/t = 10, Opening mode

p = 0.2, Rb/R = 2
1.0 p = 0.2, Rb/R = 3
p = 0.4, Rb/R = 2
p = 0.4, Rb/R = 3
0.8
p = 0.6, Rb/R = 2
p = 0.6, Rb/R = 3

X = ML/M0
Lower bound
0.6 values, Rb/R = 3

0.4

0.2

0.0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 30. Effect of elbow bend radius on weakening factors (X) for various levels of internal pressure: R/t = 10, Opening mode.

6. Validation

Any new equations predicting a physical phenomenon (e.g. plastic collapse here) must be validated, pref-
erably comparing its predictions with the experimental results or in the absence of experimental results with
the available numerical results. It should also be ensured that the new equations are consistent with the earlier
observations related to the physical phenomenon and the predicted behaviour can be explained rationally.
Therefore, attempt has been made to compare the predictions of the proposed equations with the available
experimental/numerical results and to rationally explain the behaviour where no experimental/numerical data
is available for comparison. The discussion is split in two parts: first un-pressurized and subsequently pressur-
ized cases are discussed.
In case of un-pressurized (i.e. pure in-plane bending moment loading without any internal pressure) cases,
elbow test data of Griffiths [13], Moulin et al. [25], Yahiaoui et al. [14] and Chattopadhyay et al. [26] are
available. Fig. 31 shows the comparison of experimental maximum moment observed by Moulin et al. [25]
and reported by Wilkowski et al. [27] with the predictions of the proposed equations (Eqs. (12), (13) and
Table 1). A very good matching of the variation trend of moment with throughwall circumferential crack
angle (2h) may be observed. The predictions of Eqs. (12), (13) and Table 1 are seen to be consistently lower

45 Circumferential through wall cracked elbow

under in-plane closing bending
40
OD = 114.3 mm, t = 8.56 mm, No pressure,
35 Bend characteristics (h) = 0.465
σy = 289 MPa, σu = 599 MPa
30
Moment (kN-m)

25

20

15

10 Proposed Elbow TES plastic

collapse moment (Eqs.12,13 & Table 1)
5 Moulin elbow expt. maxm. moment [25,27]
Pipe limit moment (Eq.24)
0
0 30 60 90 120 150
Crack angle, 2θ (deg.)

Fig. 31. Comparison of maximum moments observed in elbow fracture experiments [25,27] with pipe and elbow limit moment predictions.
850 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

1.0

Moment normalized with

0.8

defect-free component
0.6

0.4

Chattopadhyay et al [4,5],Table 1
0.2
Moulin [25,27], maximum expt. moment
Pipe limit moment
0.0
0 30 60 90 120 150
Crack angle, 2 θ (deg.)

Fig. 32. Comparison of maximum moments normalized with defect-free component [25,27] with pipe and elbow limit moment predictions.

(and hence conservative) with respect to the maximum experimental moments. This is expected since Eqs. (12),
(13) and Table 1 predict the PCM based on TES criterion, which is normally lower than the maximum
moments. However, moments when normalized with the defect-free component show (Fig. 32) much better
matching with the predictions. Fig. 31 also compares the elbow test data with pipe limit moments predicted
by the following equation [28]:
M L ¼ 4R2 trf ½cosðh=2Þ
0:5 sinðhÞ ð24Þ
where, R is the mean radius of the pipe cross section, t is the wall thickness, rf is the pipe material flow stress
taken as the average of yield and ultimate stress i.e. rf = 0.5(ry + ru) and h is the semi-crack angle. The ratio-
nale of comparing the predicted pipe limit moments with elbow test data is that some literature (e.g. API 579
[29]) recommends to use pipe equations to assess the integrity of elbows. It may be observed that Eq. (24)
highly over-predicts (and hence non-conservative) the load bearing capacity of elbows, specially for lower
crack angles.
Additional comparison of elbow test data of Griffiths [13], Yahiaoui et al. [14] and Chattopadhyay et al.
[26] with those predicted by proposed equations are shown in Table 3. A very close matching may be observed.
It provides confidence to the general accuracy of the proposed equations.

Table 3
Comparison of elbow PCM test data with the analytical predictions
Reference Test ref. no. 2h (deg) Bending mode Expt. TES collapse Predicted by Eqs. (12), (13) % diff.a
moment (kN m) and Table 1 (kN m)
Chattopadhyay ELTWIN8-1 94.96 Opening 98.4 101.1
2.7
et al. [4,5,26] ELTWIN8-2 125.16 Opening 80.0 76.0 5
ELTWEX8-4 98.24 Closing 108.7 100.2 7.8
ELTWIN16-1 95.89 Opening 857.0 847.1 1.2
ELTWIN16-2 122.79 Opening 699.1 678.8 2.9
ELTWEX16-3 64.85 Closing 1161.2 1092.3 5.9
ELTWEX16-4 94.11 Closing 985.6 962.1 2.4
ELTWEX16-5 124.0 Closing 792.3 819.4
3.4
Griffiths [13] 1 0 Closing 0.1932 0.1714 11
5 90 Closing 0.1705 0.1707
0.1
Yahiaoui et al. [14] E0 0 Opening 9.01 8.68 3.7
E6 46 Opening 7.46 8.076
8.2
E11 120 Opening 5.89 4.923 16.4
a
% difference = [(expt.
predicted)/expt.] · 100%.
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 851

h = 0.18
0.8

m0 = M0 /4R tσy
0.6

2
0.4

Proposed Eq.(16)
0.2 Chattopadhyay et al [8,9], Eq.(6)
Robertson et al [24], Plastic Collapse
Robertson et al [24], Instability

0.0
0.0 0.2 0.4 0.6 0.8 1.0
p = PR/tσy

Fig. 33. Comparison of normalized limit moment versus normalized pressure variation for defect-free elbows subjected to closing
moment.

In case of pressurized cases, Fig. 33 shows the comparison between the PCM predicted by Eq. (16) and the
numerical results of Robertson et al. [24] for defect-free elbows. It may be seen that the basic trend of variation
predicted by Eq. (16) matches well with Robertson et al. [24]. The basic trend is that normalized PCM increases
with normalized internal pressure up to a certain limit, reaches a peak and then falls to zero when normalized
internal pressure (p) reaches unity. The present results are also compared with the predictions of Eq. (6) pro-
posed by Chattopadhyay et al. [8,9]. Again the basic trend is matching up to a large extent except near p = 1.0.
Eq. (6) cannot predict the drop of PCM to zero at p = 1.0, because strain hardening was assumed in the analysis
of [8,9] in contrast to the assumption of elastic–perfectly plastic material response in the present analysis.
In case of throughwall circumferentially cracked elbow under combined loading of internal pressure and
bending moment, no test/numerical data is available in the open literature for comparison. Therefore, rational
explanations are attempted in Sections 5.2.1 and 5.2.2 in support of the observations made in the present study.
Finally, it may be concluded that the present results are consistent with the available test/numerical data,
which provide confidence to the newly proposed equations in the paper.

7. Conclusions

The effect of internal pressure on in-plane plastic collapse moments (PCM) of defect-free and throughwall
circumferentially cracked (TCC) elbows are investigated by elastic–plastic finite element analysis. A total
of 396 cases of elbows with various sizes of circumferential cracks (2h = 0–150
), different wall thickness
(R/t = 5–20), different levels of normalized internal pressure (p = PR/(try) = 0–1), different elbow bend radii
(Rb/R = 2,3 and h = 0.1–0.6) and two different bending modes, namely closing and opening are analyzed. The
following conclusions are drawn:

• For defect-free elbow, the normalized PCM increases with normalized internal pressure, reach a peak and
then drop to zero at normalized pressure p = PR/(try) = 1.0. Almost same behaviour is observed for both
opening and closing mode of bending moment.
• In case of TCC elbow subjected to pure in-plane closing moment, there is a threshold crack angle below
which it has no weakening effect on PCM. Application of internal pressure reduces the threshold crack
angle and also significantly increases the weakening (i.e. reduces X) due to the presence of the throughwall
circumferential crack at extrados. The variations of weakening factor (X) with crack angle (2h) also show
strong dependence on R/t of elbow cross section.
• In case of TCC elbow subjected to in-plane opening moment, the effect of internal pressure on PCM is
much less compared to closing mode. Internal pressure moderately reduces the PCM in the opening mode.
The weakening factor (X) depends mainly on crack angle (2h) and is almost independent of internal pres-
sure (p) and R/t.
• Based on the above results, new closed-form equations are proposed to evaluate PCM of defect-free and
TCC elbows subjected to combined internal pressure and in-plane closing/opening bending moments.
852 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

• The predictions of the proposed equations are compared with the available experimental/numerical results
and rational explanations are attempted to describe the behaviour where no experimental/numerical data is
available for comparison.

Acknowledgements

The authors acknowledge the contribution of Mr. Avinash D.Patil, Trainee, 48th Batch BARC Training
School for doing some calculations for elbows under opening moment.

Appendix A

Proposed plastic collapse moment equations (Tables A.1–A.3) of elbows at a glance.

Note: See the nomenclature for the meaning of symbols.

Table A.1
Defect-free elbow under combined internal pressure and in-plane bending moment
Bending mode Proposed equations Applicability
 
Closing moment (Fig. 7) M0 2:071p1:418 0.1 6 h 6 0.6
¼ 1:075h2=3 þ þ 8:41p12:129 ½1
p
2
4R try h0:223 0 6 p 6 1.0
5 6 R/t 6 20

 
Opening moment (Fig. 8) M0 1=3 1:2182p 9:6431 0.1 6 h 6 0.6
¼
0:0617 þ 1:0485h þ þ 7:8509p ½1
p
4R2 try h1=3 0 6 p 6 1.0
5 6 R/t 6 20

Table A.2
Throughwall circumferentially cracked elbow under pure in-plane bending moment (p = 0) (ML = M0X with M0 to be evaluated using
Table A.1 putting p = 0)
Bending mode R/t Proposed equations
(   2
Closing moment (Fig. 2) 5 1:1194
0:7236 ph
2:0806 ph for 45 6 2h 6 150
X ¼
1 for 2h < 45

( h  2
7.5 1:1185
0:342
2:52 ph for 60 6 2h 6 150
X ¼ p
1 for 2h < 60

(   2
10 0:9655 þ 1:0152 ph
4:68 ph for 60 6 2h 6 150
X ¼
1 for 2h < 60

(   2
15 1:14 þ 0:3 ph
3:6 ph for 90 6 2h 6 150
X ¼
1 for 2h < 90

(   2
20 0:64 þ 3:42 ph
7:92 ph for 90 6 2h 6 150
X ¼
1 for 2h < 90

( 
Opening moment (Fig. 3) 5–20 1
0:8 ph for 2h < 45
X ¼ h
1:127
1:8108 p for 45 6 2h 6 150
 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854 853

Table A.3
Throughwall circumferentially cracked elbow under combined internal pressure and in-plane bending moment (ML = M0X with M0 to be
evaluated using Table A.1 putting 1 > p > 0)
Bending mode R/t Proposed equations Applicability
   2  2:1758
Closing moment 5 (Fig. 11) h h h 2h 6 150

X ¼ 1:1194
0:7236
2:0806
3:4164p0:8408
p p p 06X61
0<p<1
 0:6687  1:7345
h h
7.5 (Fig. 12) X ¼ 1:4423
1:495
2:9803p0:7136 2h 6 150

p p 06X61
0<p<1
 0:4082  1:4381
h h
10 (Fig. 13) X ¼ 1:6039
1:0847
3:1773p0:4807 2h 6 150

p p 0 6 X61
0<p<1
 0:5435  0:9644
h h
15 (Fig. 14) X ¼ 1:4298
0:0789
3:3789p0:2336 2h 6 150

p p 06X61
0<p<1
 0:0231  1:2776
h h
20 (Fig. 15) X ¼ 7:7803
6:8959
4:1061p0:5464 2h 6 150

p p
06X61
0<p<1
( 
1
1:2776 ph for 2h < 45
X ¼   2 0<p<1
Opening moment (Fig. 27) 5–20 1:3053
4:2133 ph þ 3:8386 ph for 45 6 2h 6 150
Note: In case of closing mode of bending moment, for intermediate R/t, X can be linearly interpolated between the adjacent R/t values.
However, for conservative results, the equation applicable for next lower R/t may be chosen.

References

[1] Gerdeen JC. A critical evaluation of plastic behavior data and a united definition of plastic loads for pressure components. WRC Bull
1979;254.
[2] ASME Boiler and Pressure Vessel Code, Sec. III, American Society of Mechanical Engineers, 2000.
[3] R6 (2001). Assessment of the integrity of structures containing defects, Revision 4, British Energy.
[4] Chattopadhyay J, Tomar AKS, Dutta BK, Kushwaha HS. Closed form collapse moment equation of throughwall circumferentially
cracked elbows subjected to in-plane bending moment. J Press Ves Technol, ASME Trans 2004;126:307–17.
[5] Chattopadhyay J, Tomar AKS, Dutta BK, Kushwaha HS. Limit load of throughwall cracked elbows: comparison of test results with
theoretical predictions. Fatigue Fracture Engng Mater Struct 2004;27:1091–103.
[6] Miller AG. Review of limit loads of structures containing defects. Int J Press Ves Piping 1988;32:197–327.
[7] Zahoor A. Ductile fracture handbook, Vol. 1–3, EPRI-NP-6301-D, N14–1. Research Project 1757–69, Electric Power Research
Institute, 1989–1991.
[8] Chattopadhyay J, Nathani DK, Dutta BK, Kushwaha HS. Closed-form collapse moment equations of elbows under combined
internal pressure and in-plane bending moment. J Press Ves Technol, ASME Trans 2000;122:431–6.
[9] Chattopadhyay J. The effect of internal pressure on in-plane collapse moment of elbows. Nucl Engng Des 2002;212:133–44.
[10] Shalaby MA, Younan MYA. Limit loads for pipe elbows with internal pressure under in-plane closing bending moment. J Press Ves
Technol ASME Trans 1998;120(1):35–42.
[11] Shalaby MA, Younan, MYA. Limit loads for pipe elbows subjected to in-plane opening moment and internal pressure. Paper
presented at the 1998 ASME/JSME joint pressure vessels and piping conference held at San Diego, California, 26–30 July 1998, PVP-
Vol. 368, 1998. p. 163–70.
[12] Touboul F, Ben Djedidia M, Acker D. Design criteria for piping components against plastic collapse: application to pipe bend
experiments. In: Liu C, Nichols RW, editors. Pressure vessel technology, proceedings of 6th international conference held in Beijing,
11–15th September 1988. p. 73–84.
[13] Griffiths JE. The effect of cracks on the limit load of pipe bends under in-plane bending: experimental study. Int J Mech Sci
1979;21:119–30.
[14] Yahiaoui K, Moreton DN, Moffat DG. Evaluation of limit load data for cracked pipe bends under opening bending and comparisons
with existing solutions. Int J Press Ves Piping 2002;79:27–36.
[15] Spence J, Findlay GE. Limit load for pipe bends under in-plane bending. In: Proc 2nd int conf on pressure vessel technology, San
Antonio, 1–28, 1973. p. 393–9.
854 J. Chattopadhyay, A.K.S. Tomar / Engineering Fracture Mechanics 73 (2006) 829–854

[16] Calladine CR. Limit analysis of curved tubes. J Mech Engng Sci, Inst Mech Engs 1974;16(2):85–7.
[17] Goodall IW. Large deformations in plastically deforming curved tubes subjected to in-plane bending. Research Division Report RD/
B/N4312, Central Electricity Generating Board, UK, 1978.
[18] Goodall IW. Lower bound limit analysis of curved tubes loaded by combined internal pressure and in-plane bending moment.
Research Division Report RD/B/N4360, Central Electricity Generating Board, UK, 1978.
[19] Rodabaugh EC. Interpretive report on limit load analysis and plastic deformations of piping products. Weld Res Council Bull
1979;254:65–82.
[20] Hilsenkopf P, Boneh B, Sollogoub P. Experimental study of behavior and functional capability of ferritic steel elbows and austenitic
stainless steel thin-walled elbows. Int J Press Ves Piping 1988;33(2):111–28.
[21] Gullerud K, Koppenhoefer A, Roy Chowdhury S, Walters M, Dodds RH Jr. WARP3D-Release 14.0. 3-D Dynamic nonlinear
fracture analysis of solids using parallel computers and workstations, University of Illinois, USA, 2002.
[22] Kumar V, German MD, Shih CF. Elastic–plastic and fully plastic analysis of crack initiation, stable growth and instability in flawed
cylinders. In: Shih CF, Gudas JP, editors. Elastic–plastic fracture: second symposium. Vol. I—Inelastic crack analysis. ASTM STP,
vol. 803. American Society for Testing and Materials; 1983. p. I-306–53.
[23] Customized pre-processor of NISA. A general purpose finite element program, Engineering Mechanics Research Corporation,
Michigan, USA, 2002.
[24] Robertson A, Li H, Mackenzie D. Plastic collapse of pipe bends under combined internal pressure and in-plane bending. Int J Press
Ves Piping 2004;33(2):111–28.
[25] Moulin D et al. Experimental evaluation of J in cracked straight and curved pipes under bending. In: Tenth international conference
on structural mechanics in reactor technology, 10th SMiRT, vol. G, 1989. p. 323–6.
[26] Chattopadhyay J, Pavankumar TV, Dutta BK, Kushwaha HS. Fracture experiments on throughwall cracked elbows under in-plane
bending moment: test results and theoretical/numerical analyses. Engng Fract Mech 2005;72:1461–97.
[27] Wilkowski GM, Olson RJ, Scott PM. State-of-the-art report on piping fracture mechanics. NUREG/CR-6540. BMI-2196. United
States Nuclear Regulatory Commission, 1998.
[28] Kanninen MF, Zahoor A, Wilkowski GM, Abousayed I, Marschall C, Broek D, et al. Instability predictions for circumferentially
cracked type 304 stainless steel pipes under dynamic loading, EPRI-NP-2347. Palo Alto, CA: Electric Power Research Institute;
1982. p. 1–2.
[29] API. Recommended practice for fitness-for-service, API 579. Washington, DC: American Petroleum Institute; 2001.