BARC/1993/E/016

t
3
I
i

A DATABASE TO EVALUATE STRESS INTENSITY FACTORS OF ELBOWS WITH

THROUGHWALL FLAWS UNDER COMBINED INTERNAL PRESSURE AND BENDING
MOMENT
by
J. Chattopadhyay, B. K. Dutta. H. S. Kushwaha, S. C. Mahajan and A. Kakodkar
Reactor Design & Development Group

1993
 BARC/1993/L/016

GOVERNMENT OF INDIA
ATOMIC ENERGY COMMISSION

CO

A DATABASE TO EVALUATE STRESS INTENSITY FACTORS

OF ELBOWS WITH THROUGHWALL FLAWS UNDER
COMBINED INTERNAL PRESSURE AND
BENDING MOMENT
by
J. Chattopadhyay, B.K. Dutta, H.S. Kushwaha,
S.C. Mahajan and A. Kakodkar
Reactor Design and Development Group

BHABHA ATOMIC RESEARCH CENTRE

BOMBAY, INDIA
1993
 BARC/1993/E/O16
BIBLIOGRAPHY DESCRIPTION SHEET FOR TECHNICAL REPORT
(as per IS x 94OO - 1980)

Ol Security classication x Unclassified

02 Distribution : External
O3 Report status t New
O4 Series J BARC External

05 Report type t Technical Report

06 Report No. : BARC/1993/E/O16

07 Part No. or Volume No. z

08 Contract No. :

10 Title and subtitle : A database to evaluate stress intensity

factors of elbows with throughwall
flaws under combined internal pressure
and bending moment

11 Collation : 24 p., 12 figs., 3 tabs.

13 Project No. :

20 Personal author (s) t J. Chattopadhyay; B.K. Outta;

H.S. Kushwaha; S.C. Mahajan;
A. Kakodkar

21 Affiliation of author <s) x Reactor Design and Development Group,

Bhabha Atomic Research Centre, Bombay

22 Corporate author<s> : Bhabha Atomic Research Centre,

Bombay - 4OO 085
23 Originating unit i Reactor Engineering Division, BARC,
Bombay

24 Sponsor(s) Name : Department of Atomic Energy

Type s Government
30 Date of submission : June 1993
31 Publication/Issue date July 1993

contd...l(A)
40 Publisher/Distributor s Head, Library and Information Division,
Bhabha Atomic Research Centre, Bombay

42 Form of distribution : Hard Copy

50 Language of text s English

51 Language of summary i English

52 No. of references : 8 refs.

53 Given data on :

60 Abstract :The advent of Leak-Before-Breafc (LBB) concept has replaced

the traditional design basis event of Double Ended Guillotine Break
(DEGB) in the design of primary heat transport <PHT) piping. The use
of LBB concept requires postulation of largest credible cracks at
highly stressed locations and demonstration of its stability under
the maximum credible loading conditions. Stress analysis of PHT
piping in nuclear power plants shows that the highly stressed piping
components are normally elbows and branch tees. This necessiates
detailed fracture mechanics evaluation of piping connections by
computing Stress Intensity Factor <SIF) and/or J-integral. Simple
analytical solutions for evaluation of SIF and J-integral for cracks
in straight pipes are readily available in literature. However, the
same type -of solutions for elbows and tees are limited in open
literature. In the present work, a database is generated to evaluate
SIF for throughwall circumferential and longitudinal cracks under
combined internal pressure and bending moment. Different parameters
to characterise a cracked elbow are pipe factor <h>, pipe bore radius
to thickness ratio (r/t> and crack length. Another parameter
(p) is used to consider the ralative magnitude of stresses due to
internal pressure and remote bending moment. The database has been
used to derive closed form expressions to evaluate SIF for elbow with
cracks in terms of the aforementioned parameters.

70 Keywords/Descriptors i STRESS INTENSITY FACTORS; BENDING; EVALUATION;

THEORETICAL DATA; PRIMARY COOLANT CIRCUITS; HEAT TRANSPORT; PIPES;
PIPE JOINTS; ANALYTICAL SOLUTION; DEFECTS; LEAKS; F CODES; DYNAMIC
LOADS; CRACKS; FRACTURE MECHANICS; PIPELINES; STRESSES;
FRACTURES; RUPTURES; NUCLEAR POWER PLANTS; REACTORS;
FINITE ELEMENT METHOD

71 Class No. : INIS Subject Category : E1300; E3000; E2200

99 Supplementary elements c
A DATABASE TO EVALUATE STRESS INTENSITY FACTORS OF ELBOWS WITH
THROUGHWALL FLAWS UNDER COMBINED INTERNAL PRESSURE AND BENDING
MOMENT

J. Chattopadhyay, B. K. Dutta, H. S. Kushwaha,S. C. MahaJan & A.Kakodkar

Reactor Design and Development Group
Bhabha Atomic Research Centre
Trombay, Bombay - 4OOO85, India

ABSTRACT

The advent of Leak-Before-Break (LBB) concept has replaced

the traditional design basis event of Double Ended Guillotine
Break (DEGE) in the design of primary heat transport (PUT)
piping. The use of LBB concept requires postulation of largest
credible cracks at highly stressed locations and demonstration of
its stability under the maximum credible loading conditions.
Stress analysis of PUT piping in nuclear poner plants shows that
the highly stressed piping components are normally elbows and
branch tees. This necessiates detailed fracture mechanics
evaluation of piping connections by computing Stress Intensity
Fcator (SIF) and/or J-integral. Simple analytical solutions for
evaluation of SIF and J-integral for cracks in straight
pipes are readily available in literature. However r the same
type of solutions for elbows and tees are limited in open
literature. In the present work, a database is generated to
evaluate SIF for throughwall circumferential and longitudinal
cracks under combined internal pressure and bending moment.
Different parameters to characterise a cracked elbow are pipe
factor (h), pipe bore radius to thickness ratio (r/t) and crack
length. Another parameter (p) is used to consider the ralative
magnitude of stresses due to internal pressure and remote bending
moment. The database has been used to derive closed form
expressions to evaluate SIF for elbow with cracks in terms of the
aforementioned parameters.
NOMENCLATURE

K Stress Intensity Factor (SIF)

K.mld SIF at mid-layer

Kt SIF computed at radial distance rt

(SIF) C SIF extrapolated to the crack tip

r Mean pipe radius

rt Radial distance of the sampling point from crack tip to

evaluate <SIF) C

R Mean pipe bend radius

t Pipe thickness

a Semi crack length

aT Reference stress in defining AA

Aw = K mid / or V n.a , a non—dimensional parameter to define SIF

n Number of sampling points considered to evaluate (SIF) C

J)< J—integral in local x,y and z direction ( k = x,y,z)

f Integration path to evaluate J-integral

nk Normal to the integration path V

nk' Normal to the evaluation layer

o^j Stress tensor along the integration path r

ut Displacements

W Strain energy density

A Area enclosed by the integration path V
 Young's modulus
Poisson's ratio

= a/y rt , longitudinal crack length in non-dimensional form

"Ft , shell parameter to define crack

length in cylinder
9 Angle subtended by the elbow axis at the bend centre
9C Angle subtended by the longitudinal semi crack on elbow
at the bend centre

4> Angle subtended by the elbow circumference at cross-

sectional centre at a particular &
<pc Semi circumferential crack angle on elbow

p Internal pressure

M Remote bending moment

p = (2/n). (M/pr 9 ), lioad ratio to define relative magnitude of

internal pressure and bending moment

h = tR/r , pipe factor

1. INTRODUCTION
The safety of nuclear reactor has always engaged attention of
designers. Piping system, which enables cooling of reactor core
has been one of the important items where considerable research
has been carried out. A failure in the piping would lead to a
loss of coolant accident and may lead to release of radioactive
materials to public domain. Hence one of the current active
research area is the fracture mechanics evaluation of piping
components with the presence of flaws.
One of the hypothetical design basis events traditionally
considered in the design of Primary Heat Transport <PHT) piping
of Pressurised Heavy Water Reactor (PHWR) is instantaneous Double
Ended Guillotine Break (DEGB) of the largest heat transport pipe.
This concept was originally initiated for sizing the containment
and Emergency Core Cooling Systems (ECCS). Regulatory philosophy
for design of piping system, however, tended to shift the
postulate of DEGB to design basis for making provisions of
protection against DEGB. A natural consequence of an accepted
pipe break postulate would require provision of massive pipe whip
restraints to minimise pipe deflection.
However, over the years, it has been found that these
restraints have resulted in many problems, namely, restricted
access for in-service inspection and hence additional man-rem
expenditure, complication in pipe system design, incresased heat
loss to the surrounding environment , unanticipated thermal
 expansion stress. Against this backdrop, nuclear community has
started giving second thought to the real efficacy of the pipe
rupture protection devices in enhancing the safety of the plant.
The question being asked is whether these devices really increase
the plant safety or not. With this background, United States
Nuclear Regulatory Commasion (USNRC) formed a pipe break task
group to investigate the probability of occurance of catastrophic
pipe rupture. The task group findings show that there is
vanishingly low probability of pipe rupture under all
circumstances1 provided one adopts well directed design and
inservice maintenance procedures. Consequently, USNRC recommended
the elimination of requirements of mechanical pipe rupture
protection against the arbitary intermediate breaks and
henceforth to replace the traditional design basis event of DEGB
by the concept of Leak—Before—Break <LBB> in design of piping
system. The LBB approach means the application of fracture
mehanics technology to demonstrate that the piping are very
unlikely to experience double—ended rupture or their equivalent
as longitudinal or diagonal splits. The methodology consists of
demonstrating three levels of confidence. These three levels of
confidence may be viewed as defence in depth strategy. Level 1
confidence is inherent in the design philosophy of ASME Section
III which is normally followed in piping design. The design is
done with a well defined factor of safety. However, it does not
consider the presence of any flaw in the pipe. Level 2 consists
of postulating a part-through crack at the inside surface of the
PHT piping and then to demonstrate that it will not grow
throughwall during the period between two successive inservice
inspection/repair or, possibly, during the entire life period of
the reactor. It should be also shown that the final flaw size at
the end of evaluation period is sufficiently smaller than the
critical flaw. Level 3 postulates a throughwall crack that will
ensure detectable leakage and then demonstrates that the crack
will be stable under the maximum credible loading conditions.

2. SCOPE OF THE PRESENT WORK

The LBB approach requires the extensive application of fracture

mechanics for the piping system with the presence of flaw. It
postulates the largest credible flaw at the highly stressed
points in the piping sytem and demonstrates its stability under
the most severe loading conditons. Stress analysis of PHT piping
in nuclear power plant shows that highly stressed piping
components are normally elbows and tees. This necessiates
detailed fracture mechanics evaluation of piping components by
computing stress intensity factor (SIF) and/or J-integral. Simple
analytical solutions for SIF and J-integral are available for
straight pipes. 2 ' 3 However, the same type of solutions are not
available for elbows and tees in open literature. Hence, a
computational technique ,e.g., finite element method is adopted
to evaluate the solutions for elbows and tees. However, the
application of finite element method requires high level of
expertise and large computation time. In the present work, a
database is generated to evaluate SIF for various sizes of elbows
with various sizes of flaws under different types of loadings. We
have also tried to derive closed form expressions by using this
database to evaluate stress intensity factors for throughwall
flaws in elbows as a function of different geometrical parameters
and loadings.
3. DEVELOPMENT OF A FINITE ELEMENT CODE 'FABS'

To carry out the above mentioned task of analysing elbows with

throughwill flaws, a finite element code 'FABS' (Fracture
mechanics Analysis of Bending Structure) has been developed. It
uses nine noded heterosis/ eight noded degenerate shell bending
elements. Stresses along the thickness is considered using
layered approach. Each layer contains stress points and stresses
are assumed to be constant over the thickness of each layer.
Thus,the actual stress distribution is modelled by a piecewise
constant approximation. The capability of the code has further
been enhanced by adding subroutines to calculate J—integral using
contour approach at different layers of thickness in a three
dimension body. The database of 5IF has been generated using this
code. Considering a large number of case studies to be done to
generate the database, a versatile pre—processor is developed
which generates element topology, nodal coordinates, boundary
conditions and loading data. The output of this pre-processor is
made compatible with the input data of the code 'FABS'. The
output of the code 'FABS' is then further processed to get the
SIF expressed in a non-dimensional form 'Ae' where Ae is
expressed as,
Ae = KmiA / a, iT^a" (1)

4. EVALUATION OF STRESS INTENSITY FACTOR

There are number of methods to evaluate SIF. However, only

two, namely, displacement correlation and J—integral methods have
been adopted here.
In case of displacement correlation method, SIFs have been
evaluated by considering the crack face nodes only. However,
normal procedure for evaluating SIF by extrapolating a straight
line on a graph paper is more of an art where uniqueness is
lost 8 . Hence, in the present work, a best fit straight line
equation is used for number of nodes and then extrapolated
to the crack tip to get a unique solution of SIF. The equation
used for this extrapolation is
(SIF) C = S( Kv - ft rt)/n (2)

where ft = C n SKkrt - SKi 2 ^ 3 / Cn Sr^2 - (Er^ 2 1 (3)

As mentioned above, we have also implemented the J-integral
procedure to evaluate SIFs. The expression which have been used
in the present study far J-integral evaluation is given below. 4

- J(W n k -
r
5. ANALYSIS OF STRAIGHT PIPES WITH CRACKS
A number of case studies have been done to test the different
capabilities of the present code to compute stress intensity
factor (SIF) and J—integral of the structures containing flaws.
These are described in Ref.5. The details of the analysis done to
compute SIF for straight pipes with circumferential and
longitudinal cracks under internal pressure are given below.
Fig.l shows the details of circumferential and longitudinal
cracks on a straight pipe. The pipe is subjected to internal
pressure. A parametric study is carried out to evaluate SIF
expressed in terms of 'Aw' for different r/t and crack length.
Crack length is expressed in terms of a shell parameter (\c>
which is expressed as:
o. 23 .
\ c = C12(l-t>2)3 . a/V <rt> (4)
We have computed SIFs without considering the crack face loading
and also with considering the crack face loading equal to
internal pressure. To calculate A e from eqn.Cll the reference
stress crr is taken as nominal hoop stress (pr/t) in case of
longitudinal flaw and nominal axial stress (pr/2t) in case of
circumferential flaw.
Figs.2 and 3 show the variation of A e with shell parameter
for circumferentially cracked pipe. Figs.4 and 5 show the similar
results for longitudinally cracked pipe. The figures also compare
these results with those quoted in reference 6.

6. FINITE ELEMENT ANALYSIS OF A NON-WEAKENED ELBOW AND

COMPARISON OF RESULTS WITH ANALYTICAL VALUES

Prior to analysing an elbow with a crack a non—weakened elbow

has been analysed to compute the stresses by the present computer
code and the results have been compared with the analytical
values. This exercise is done to test the capability of the
present element in modelling an elbow under various types of
loadings. Fig.6 shows the geometry of an elbow. Fig.7 shows the
computed stresses for an elbow under internal pressure on 0 =
45° plane. The same figure also shows the analytical values 7 .
The variation shows that longitudinal stress is constant along
the circumference and hoop stress varies from maximum at intrados
to minimum at extrados. The figs. 8 and 9 show the variation of
computed hoop and longitudinal stresses for the elbow subjected
to remote bending moment. The stress distribution shows that
maximum hoop stress occurs at 4> ~ 90°. Hence, in the subsequent
analysis the longitudinal crack has been postulated at <p ~ 90°
with its centre on & = 45° plane and the circumferential crack
has been postulated on & = 45° plane with its centre at 4> -
180°.

7. PARAMETRIC STUDY OF ELBOW WITH CIRCUMFERENTIAL CRACK

A detailed parametric study of elbows with circumferential

throughwall cracks under combined loading of closing bending
moment and internal pressure has been carried out. Fig.6 shows
the loaction of circumferential crack on the elbow. The elbow has
been modelled along with the connecting straight pipe. The length
of the straight pipe is chosen as six times the radius of the
pipe to eliminate the boundary effect. One fourth of the elbow is
modelled due to symmetry. The finite element model consists of
552 eight noded thick shell bending elements and 1751 nodes. The
crack tip singularity is modelled by using quarter point
crack-tip elements. The structure is assumed to be in elastic
condition. Three parametrs are chosen to characterise an elbow
with a circumferential crack, namely,r/t, <pc and pipe factor (h).
Another parameter is chosen to consider the relative magnitude of
stresses due to internal pressure and bending moment. It is
expressed as,

p = <2/rc>.(M/pr9>

p = 0 indicates internal pressure only and p -oc indicates bending

moment only.
The parametric study involves the evaluation of 5IF for
different pipe factors, crack angles, r/t and load ratios ipi.
Table 1 shows the various combinations of r/t and pipe factors
considered for the present analysis. The range of *h' considered
is different for various r/t ratios, because the minimum possible
value of 'h' is restricted by t/r. The maximum value of pipe
factor is decided by considering the application of the database
to primary heat transport piping. A typical 'h' value for PHWR
PHT piping is O.5. Hence, the maximum value of *h* is restricted
to 0.65. For each pipe factor, computations are done for six
different crack angles (2*,. = 50°, 90°, 120°, 14O°, 160°and 180°)
and for five different load ratios <p = «, 2, 1, 0.5 and 0 ) . It
has been seen that pure closing bending moment cannot open a
small circumferential crack at the extrados. Hence, small crack
angle has not been included in the present parametric study.

Table 1 Different Combinations of r/t and Pipe Factor

r/t Different Pipe Factors Considered

2O 0 .075 O.15 O.25 0 .35 0 .50 O.65

1O O.15 0.25 O .35 O .50 O.65

5 0 .35 0 .50 0.65

In this manner a database is generated by calculating A, for

various crack angles <2<0C>, pipe factors (h), r/t and load ratios
(p). There are 404 different cases in this database. The A 0 is
calculated by using equation C13 for each case.The reference
stress Ccr) is taken as

ar = (pr/2t) + (H/nrt)

For each case, f\m is calculated by computing SIF by using two

independent methods. These are,displacement correlation and
J-integral methods. Figs. 10a, lOb and lOc show the variation of
A # with pipe factor (h), crack angle C2^c) and r/t for different
load ratios (p) .
8. PARAMETRIC STUDY OF ELBOW WITH LONGITUDINAL CRACK

A parametric study of elbows with longitudinal crack has also

been carried out along the similar lines as explained above.
Fig.6 shows the location of the longitudinal crack at # = 9O°
of an elbow. Half of the elbow is modelled due to symmetry. There
are 484 eight noded thick shell elements and 1484 nodes. The
parameters chosen to characterise the elbow and loading are same
as before except tnat the crack length is expressed as

\ = a/V <rt)

In the present case five different values of \ ( \ = O.5, l.O,

1.5, 2.O and 2.5) and four different values of load ratio <p =
2.0, 1.0, O.5 and O.O) have been considered, p = oc could not be
considered in the present case as a pure closing bending moment
(where internal pressure is equal to zero) cannot open a
longitudinal crack fully.
Here also, a database is generated for SIF expressed in terms
of 'A.' for different parameters as mentioned above. Fig.11 shows
a typical deformed plot of a longitudinally cracked elbow under
internal pressure. Figs.12a and 12b show the variation of 'A.'
with pipe factors, crack length and r/t for different load
ratios.

9. CLOSED FORM EXPRESSIONS OF SIF FOR CRACKED ELBOW

A closed form expression is being proposed here to evaluate
stress intensity factor of elbow with longitudinal and
circumferential crack under combined loading of internal pressure
and remote bending moment. SIF is expressed in terms of a
non-dimensional number (Ae) as defined in equation E l ] . Reference
stress (o-r) is taken as defined in equation C53. The 'AA' is
expressed as a function of pipe factor (h), r/t and \ or 9c/n
(depending on whether the crack is longitudinal or
circumferential). In case of longitudinal crack, the general
expression used for A # is,
PI P2 P3
A 8 = <Ct + C2h ) + (Ca + C 4 h ) . \
P4 ps pa p?
+ C(C 5 + C,sh ) + (C7 + C8h ) . \ ].(t/r) <6)
'\' in the above expression is substituted by &c/n for
circumferential crack. The coefficients <Cn) and exponents (pn)
are evaluated by least square curve fitting method. Table 2 and 3
show the values of C n and p n for different load ratios <p) for
circumferential and longitudinal crack respectively.
 Tabls 2 Derived Coefficients (Cn> and Exponents <pn) for
Circumferential Crack
p « 2 1 O.5 O

c* -3.4628 -1.391B -0.2061 -O.4565 O.251O

c2 4.446 1.3331 -O.2688 -O.O928 O.2245
c» -52.429 -45.488 1.OO52 53.608 -7.8911
c« 52.445 41.612 3.1153 193.82 5.2822
c9 -2.2524 -0.8677 -0.1474 O.1784 0.0773
ctf 1.1102 0.8342 O.4177 0.2258 0.0599
c7 O.8634 2.5685 1.8792 2.2646 O.5119
c. 1.72B3 2.1310 1.3741 2.O149 3.0222

Pi 0.1366 0.1574 3.4968 5.4819 -0.0405

Pz -0.1848 -0.2604 -0.3772 -O.4787 0.1994
Pa 2.6137 4.0769 2.6686 10.248 1.3462
P* 0.1216 O.137O O.37O5 -O.19B3 O.O166
Ps 0.0695 0.0044 -0.116 O.OO76 -0.3614
P<» 0.4587 1.1139 1.1723 1.1981 1.2298
P7 -0.5119 -0.2564 -O.2955 -0.2474 -O.3662

Table 3 Derived Coefficients (Cn> and Exponents (pn)

for Longitudinal Crack
p 2 1 O.5 O

c* -1.8531 0.6716 0.509 0.5904

c* 0.0050 O.i932 O.S84 0.7452
1.3618 O.6OO6 O.3375 -O.5249
1.1382 O.4195 O.5764 O.5366
cs -1.8117 -0.6402 -0.8963 O.1675
c* 0.0660 -O.OOO5 O.5O76 O.1727
c7 2.4713 O.19O9 -O.2O26 O.3639
c. O.2726 O.3577 O.74 O-6O92

Pi -1.2866 -O.7353 -0.0426 0.0974

Pz -O.O16 -O.O239 0.1096 0.027
Pa O.4O25 0.1440 O.2889 3.986
P« -1.2O83 -1-9603 -0.2266 O.O539
Ps -0.5277 0.1927 0.1229 0.0257
P<» -0.2421 O.79B2 0.9113 0.90B9
P7 0.3353 -0.2501 -0.2676 -0.2O39

10. CONCLUSIONS

a) A finite element code 'FABS' has been developed for fracture

mechanics analysis of structures containing defects using
degenerate shell bending elements. The code has been validated
against a number of sample problems which are described in Ref.5.

b) The code has been used to analyse straight pipes with cracks
under internal pressure. Two cases have been considered. One of
the cas-r accounts for crack face loading equal to internal
pressure and another does not consider crack face loading. The
results quoted in Ref.6 lie between the results obtained in the
above two cases.
c) In case of the analysis of a non-weakened elbow under internal
pressure the hoop and longitudinal stress variations at the elbow
mid-plane have excellenet agreement with those given in Ref.7
(fig.7).
d) From the parametric study of elbow with circumferential crack
at the extrades the following points may be noted.
i) Small circumferential throughwall crack at the extrados of
the elbow does not open up under closing bending moment.

ii) The 'Aa' increases with the crack angle <0C* *or
cases.
iii) For higher values of pipe factors <h), A 9 seems to be
independent of 'h' (fig.lO).
e) From the parametric study of elbow with longitudinal crack at
the crown, the foilwing points may be noted.

i) A longitudinal throughwall crack at elbow crown does not open

up fully under pure bending moment.

ii) The value of 'A.' increases with crack length (\).

iii) Here also, the values of 'Ao' at higher pipe factors are
almost independent of 'h' (fig.12).

f) The closed form expression derived in this work are found to

be capable of calculating stress intensity factor of an elbow
with circumferential and longitudinal crack under combined
internal pressure and bending moment for the range of different
geometrical parameters considered in this analysis.

REFERENCES

1) Report of the United States Nuclear Regulatory Commision

Piping Review Committee, Evaluation of Potential for Pipe Breaks,
NUREB-CR-1O61, Vol.3, 1984

2) V.Kumar, M.D.German, C.F.Shih, " An Engineering Approach for

Elastic Plastic Fracture Analysis " , EPRI-NP-1931, Electric
Power Research Centre, Palo Alto, CA, July, 1981

3) V.Kumar et al, " Advances in Elastic Plastic Fracture

Analysis ", EPRI-NP-36O7, Aug.,1984
4) T.Watanbe et al, "J-integral Analysis of Plate and Shell
Structures with Throughwall Cracks Using Thick Shell Elements",
Engg.Fracture Mechanics, Vol.19, 1984, PPicos-tcia.

5) J.Chattopadhyay, B.K.Dutta and H.S.Kushwaha, " Elastoplatic

Fracture Mechanics Analysis of Elbows Using Layered Shell Bending
Elements", Proceedings of Internal Symposium on Fatigue and
Fracture of Steel and Concrete Structures, ISSF-91, Madras,India,
1991, pp 79 - 89
6) Barsoum et al, "Analysis of Cylindrical Shell With Creaks
Using Quarter Point Crack Tip Elements" , International Journal
of Fracture, Vol.15, 1979, pp 259 -280
7) Design of Piping System, The M.W.Ktllog Company, Nov.1955
8) Computational Methods in the Mechanics of Fracture, edited by
S. N. Atluri, Elsevier Science Publishers, 1986

1 f
 Circunferervtial Crack

Longi-udinol Crock

Fig.l A STRAIGHT PIPE WITH LONGITUDINAL AND

CIRCUMFERENTIAL CRACK
 Present
Present i /
»»»»• Present ( r / t =
3.00 -i * * * * * Barsoum /
+ + +• • Barsoum ( r / t 1 0
««««« Barsoum ( r / t = 5 )

2.00 '-

1.00
0.00 2.00 4.00 6.00 8.00
Shell Parameter

Fig.2 VARIATION OF Ae WITH SHELL PARAMETER FOR CIRCUMFER-

ENTIALLY CRACKED STRAIGHT PIPE (WITH CRACK FACE LOADING)

00000
Present (r/t-20)
3.00 -2 «""""» Present (r/t-10)
Present (r/t«5)_
* * * * * Barsoum (r/t-20)
+ + • • • Barsoum (r/t»10)
Barsoum (r/t»5)
2.00 i

1.00 :

0.00
0.00 2.00 4.00 6.00 8.00
Shell Parameter

Fig.3 VARIATION OF Ae WITH SHELL PARAMETER FOR CIRCUM-

FERENTIALLY CRACKED STRAIGHT PIPE (NO CRACK FACE LOADING)
9.00 -
»°» Present Ifr/t=20)
Present I r/t=10)
Present I r/t=5)
6.00 -. Barsoum (Ref.d

3.00 -.

0.00 rn
0.00 3.60 7.20 10.80 14.40 $0
Shell Parameter

Fig.4 VARIATION OF Ae WITH SHELL PARAMETER FOR LONGI-

TUDINALLY CRACKED STRAIGHT PIPE (WITH CRACK FACE LOADING)

9.00 -
Present ( r / t - 2 0 )
Present Cr/t=10)
Present ( r / t = 5 )
6.00 : Barsoum (Ref.6)

3.00 -_

0.00 111111 111 nrrri

0.00 3.60 7.20 10.80 14.40 18.00
Shell Parameter

Hg.5 VARIATION OF Ae WITH SHELL PARAMETER FOR LONGITU-

DINALLY CRACKED STRAIGHT PIPE (NO CRACK FACE LOADING)
Rg.6 GEOMETRIC DETAILS OF AN ELBOW AND
LOCATIONS OF VARIOUS POSTULATED CRACKS
 50.00 -

40.00 ooooo Hoop Stress (Analytical)

Present Hoop Stress
Axial Stress (Analytical)
.** Present Axial Stress
:
30.00
(A
tt)
(0

I
20.00 -

10.00

0.00
0.00 36.00 72.00 108.00 144.00 180.00

Fig.7 HOOP AND AXIAL STRESS VARIATION ON MID-PLANE

OF ELBOW UNDER INTERNAL PRESSURE (h=0.075, r / t = 2 0 )
 8.00 -

Top Surface

Bottom Surface

J
-8.00

Fig.8 HOOP STRESS VARIATION ON MID-PLANE OF ELBOW

UNDER REMOTE BENDING MOMENT ( h = 0.25 )

n
 4.00 -i

Top Surface
o_aa_aa Bottom Surface

2.00 -

80.00

J
-4.00

Fig.9 AX'AL STRESS VARIATION ON MID-PWNE OF ELBOW

UNDER REMOTE BENDING MOMENT (h = 0.25)

ff
 740
r/t-20
r/t-20 Ittw - 0.5
Rho - 0

MO

UO

1.00

ado' ate' aio'" a « ato aw aoo 6.(3 OJO all otfo are
pip* factor
pip* factor

too
r/t - 10
f/t • >O
Rho - 0
SflL"- 0.5

1«
120°
MO
a so*
O-—0 0 -GSO*

OM
6' dii' ai»""aJ» ate an o3» w ffi ais
Pip* Factor
Pip* Faeter

r/i - 5
Rho - 0.S
st
r/t-_1O
3J0
MO
180*
2.70
JMO - f 140°
-© 120°
- 0 90°
1M
ato -© 50*

OJO oio
T ^ Foctor Pip* Factor

Fig. 10a VARIATION OF A. WITH PIPE FACTOR CIO, rxt AND CRACK
ANGLE C20C3 FOR CIRCUMFERENTIALLY CRACKED ELBOW C p » O, 0.5 3
 r/t-20
Rh2

100

xoo
UOO -a oaO*

UX>
MO

OdOO (40
OJSO 6.W*"' oio ali" o3o att oi» oJ»
factor plp« faetor

r/t - 10 r/t • 10
Rho - 1 Rho - 2

roo -BOO*

,50"

i ii) w t« aa £n oa»«*
Pip* Factor Plp« Factor

r/t - S
r/t -S
Rho>2

SJO
-*180C
18(f

-14OC
H- led*
1J0 -0 140*
-B90P -a I2d»
1J0
OJO
-©so* - O 9CC

0X0
Pips Factor
oJo 0.40
r ofc
Pip* Faoetor

Fig. 10b VARIATION OF A. WITH PIPE FACTOR (h), r/'t AND CRACK
ANGLE C2^>c> FOR CIRCUMFERENTIALLY CRACKED ELBOW C p • 1, 2 3
 7XO
jA-20
I

oM oil*oio a* olio
pip* f -tor

SIB

MO

1X0

"VK bis oiȣtt ofi^

Pip* Foetor

r/t -5
o-MMty
RhO"

-O120*

o« <ia6"
Pip* Factor

Fig. 10c VARIATION OF A o WITH P I P E FACTOR CIO, r V t AND CRACK

ANGLE CZ4>ay FOR CIRCUNFERENTIALLY CRACKED ELBOW C p » « )
 FACTOR OP
DCFCRWtTION

ROTX - 120 0
ROTY = 120.0
ROTZ = 60.0
Pig.H DEFORMED PLOT OF A LONGITUDINALLY
CRACKED ELBOW UNDER INTERNAL PRESSURE
 7J»
- 20
- OS
«M
2A

2.0

O 1.5

1.0

"Hi ia»"""F:J3"""a*i"
Pip* Foster Pip* Facior

r/t - 10
r/t - 10
RhO • 0
•JO

uo

a*> ait us a» an o.4i Si* o.h

Pip* Factor
P(p« Faeter

Oho"- 0
SOS

X - 1.5
ISO

100

ISO -O0.3

uo *tt«a7
bio oii Pip* Factor
Stt 6JS7 Mi 63s
Pip* Foelor

F i g . 1 2 a VARIATION OF A. WITH PIPE FACTOR CW» r ^ t AND CRACK

LENGTH C>O FOR LONGITUDINALLY CRACKED ELBOW C p « O, 0 . 5 3
 r/t - 20
100 Rho » 2.0

•A-2.5 240'

--+X-2.S
—•2.0
-

0*" «§"""«*• HM' *»""T3i S i lo «i» SX itS Hi

Ptoa Factor
Pip* Factor

•JO

2.0

I JO

OJO
j3i Ui OJJ a«i an Pip* Factor
Pfp* Footer

r/t - 0
law - 1.0
- 2.0
-0X-1JS
1.71

0.5
140
u ui w £M 6JB
P*p# Footor

? ' J S ^ VARIATION OF A. WITH PIPE FACTOR (h), r^t AND CRACK

LENGTH C\> FOR LONGITUDINALLY CRACKED ELBOW C p - l, zT
Pubbshed by : M. R. Balakrishnan,Head, Library & Information Services Division,
Bhabha Atomic Research Centre.Bombay 400 085