7/21/2019 Welding Underwater http://slidepdf.com/reader/full/welding-underwater 1/15 doi: 10.1111/j.1460-2695.2006.01049.x Crack resistance of austenitic stainless steel pipe and pipe welds with a circumferential crack under monotonic loading P. K. SINGH 1 , K. K. VAZE 1 , A. K. GHOSH 1 , H. S. KUSHWAHA 1 , D. M. PUKAZHENDI 2 and D. S. R. MURTHY 2 1 Health Safety & Environmental Group, Bhabha Atomic Research Centre, Mumbai-400085, India, 2 Fatigue Testing Laboratory, Structural Engineering Research Centre, Chennai-600113, India Received in final form 8 October 2005 ABSTRACT This paper describes the experimental studies carried out on cracked austenitic stainless steel pipe and pipe welds under bending loads. Pipe welds were produced by gas tungsten arc welding (GTAW) and shielded metal arc welding (SMAW). Fracture resistance curves forpipe andpipeweldswere compared.Results indicatethat thefractureresistance ofpipe and pipe weld (GTAW) is comparable but that of pipe weld (GTAW +SMAW) is inferior. Cracksdo notdeviatefromtheiroriginalplaneduringpropagationasobservedinthecases of carbon steel pipe and pipe welds. The fracture resistance of pipe welds does not depend on the loading histories to which it has been subjected prior to fracture test. Initiation and crack propagation were observed prior to the maximum moment. An existing limit load expressionisapplicableforthepipebasematerialbutgivesnon-conservativeresultsforthe pipe welds. Multiplication factors have been suggested for the pipe welds for evaluation of limit loads using the existing expression. Fracture resistance for the pipe and compact tension specimens have also been compared for base material and welds. Keywords austenitic stainless steel; gas tungsten arc welding (GTAW) and shielded metal arc welding (SMAW); fracture resistance; limit load; pipe; pipe welds. INTRODUCTION Modern design and safety considerations require satisfac- tion of leak-before-break (LBB) criteria for high-energy piping system of nuclear power plants. Evaluation of the fracture the resistance of components and its point of in- stability is necessary for the demonstration of the safety case. High-energy piping systems in these plants use duc- tile and strain hardening materials. Fracture resistance of such materials can be characterized by the J -integral pa- rameter, which takes into account extensive plastic defor- mation ahead of the defect under loading condition. A component integrity test program was started in Bhabha Atomic Research Centre (BARC), Mumbai to evaluate the application of LBB design criteria to the main heat trans- port system piping of the advanced heavy water reactors. Generally the fracture resistance of the material is eval- uated according to the guidelines given in ASTM E1820 1 for semi ductile materials. But use of test results obtained using compact tension specimen, as in the ASTM E1820 standard, in component analysis remains subjective for Correspondence : P.K. Singh. E-mail: [email protected]highly ductile materials because of the following reasons. Firstly, minimum thickness requirement for the determi- nation of valid J IC is very high whereas pipes are usu- ally much thinner. Secondly, the range of stable crack growthobtainedfrom thestandardcompact tension (CT) orthree-point bend (TPB)specimensisa fewmillimetres, which is small compared to that observed in components. Thirdly, there is a considerable range of scatter in the frac- ture toughness data of welds in the standard specimen test becauseoflimitedstablecrackgrowth.Austeniticstainless steel material is one of the examples that have consider- able scatter. 2 Therefore, there is a need to evaluate the crack resistance behaviour of cracked austenitic stainless steel pipes and pipe welds. In this program, a few large size pipe tests were performed to obtain the data for the representative of the actual pipes used. A large number of tests were conducted on small size pipes for validating the available analytical methods. The fracture resistance studies performed by various in- vestigators 3–5 in the past were also referred to and were taken into account while performing the tests and analyz- ing the data. In the past several methods have been devel- oped for the estimation of the J parameter for cracked c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915 901
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Crack resistance of austenitic stainless steel pipe and pipe welds witha circumferential crack under monotonic loading
P . K . SIN G H 1, K . K . V A Z E 1, A . K . G H O S H 1, H . S . K U S H W AH A1, D . M . P U K A Z H E ND I 2 an d D. S. R. MURTHY 2
1 Health Safety & Environmental Group, Bhabha Atomic Research Centre, Mumbai-400085, India, 2 Fatigue Testing Laboratory, Structural
Engineering Research Centre, Chennai-600113, India
Received in final form 8 October 2005
A B S T R A C T This paper describes the experimental studies carried out on cracked austenitic stainlesssteel pipe and pipe welds under bending loads. Pipe welds were produced by gas tungstenarc welding (GTAW) and shielded metal arc welding (SMAW). Fracture resistance curvesfor pipe and pipe welds were compared. Results indicate that the fracture resistance of pipeand pipe weld (GTAW) is comparable but that of pipe weld (GTAW +SMAW) is inferior.Cracksdo not deviate from their original plane during propagation as observed in the casesof carbon steel pipe and pipe welds. The fracture resistance of pipe welds does not depend
on the loading histories to which it has been subjected prior to fracture test. Initiation andcrack propagation were observed prior to the maximum moment. An existing limit loadexpression is applicable forthe pipe base material butgives non-conservative results forthepipe welds. Multiplication factors have been suggested for the pipe welds for evaluationof limit loads using the existing expression. Fracture resistance for the pipe and compact tension specimens have also been compared for base material and welds.
Modern design and safety considerations require satisfac-tion of leak-before-break (LBB) criteria for high-energy piping system of nuclear power plants. Evaluation of thefracture the resistance of components and its point of in-stability is necessary for the demonstration of the safety case. High-energy piping systems in these plants use duc-tile and strain hardening materials. Fracture resistance of such materials can be characterized by the J -integral pa-rameter, which takes into account extensive plastic defor-mation ahead of the defect under loading condition. A component integrity test program was started in Bhabha
Atomic Research Centre (BARC), Mumbai to evaluate theapplication of LBB design criteria to the main heat trans-port system piping of the advanced heavy water reactors.
Generally the fracture resistance of the material is eval-uated according to the guidelines given in ASTM E18201
for semi ductile materials. But use of test results obtainedusing compact tension specimen, as in the ASTM E1820standard, in component analysis remains subjective for
highly ductile materials because of the following reasons.Firstly, minimum thickness requirement for the determi-nation of valid J IC is very high whereas pipes are usu-ally much thinner. Secondly, the range of stable crack growth obtained from the standard compact tension (CT)or three-point bend (TPB) specimens is a few millimetres,
which is small compared to that observed in components. Thirdly, there is a considerable range of scatter in the frac-ture toughness data of welds in the standard specimen test because of limited stable crack growth. Austenitic stainlesssteel material is one of the examples that have consider-able scatter.2 Therefore, there is a need to evaluate thecrack resistance behaviour of cracked austenitic stainless
steel pipes and pipe welds. In this program, a few largesize pipe tests were performed to obtain the data for therepresentative of the actual pipes used. A large number of tests were conducted on small size pipes for validating theavailable analytical methods. The fracture resistance studies performed by various in-
vestigators3–5 in the past were also referred to and weretaken into account while performing the tests and analyz-ing the data. In the past several methods have been devel-oped for the estimation of the J parameter for cracked
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915 901
components such as graphical estimation methods, R6procedure,3 direct analytical estimation methods by J Smethod4 and finite-element method for calculation of J -parameter. All these methods give good estimation of theapplied J and load till crack initiation and thereafter it starts deviating depending upon the material and geom-
etry of the components. All the earlier studies are mostly on pipes with homogeneous properties (i.e. base material) whereas studies on welded pipes having heterogeneousproperties (i.e. base, weld and HAZ) with cracks are rare.
Apart from the issues mentioned above, existing estima-tion methods for the calculation of J may not be suitablefor welded pipes, which have heterogeneous properties.
The tests on welded pipes with cracks were also to ver-ify the existing estimation and calculation procedure for J values.
The main objective of the present paper is to determinethe condition of initiation of crack growth and its subse-quent propagation in austenitic stainless steel pipes and
pipe welds subjected to monotonic loading. In additionto this, the dependence of fracture resistance on loadinghistories to which pipe was subjected prior to fracture test is also brought out.
M A T ER I A L S A N D S P E C I M E N S
The pipe material was austenitic stainless steel of SA312type 304LN. Tests were carried out on seamless pipe andgirth-welded pipe of nominal outer diameter 324 mm and168 mm having thickness of 27 mm and 14.3 mm, re-spectively. The pipes were in solution annealed condition
and conforming to the specifications of ASME Section IIand Section III. The chemical composition of the pipeand pipe weld materials are given in Table 1. Most of the tests were carried out on pipe having a girth weld at the centre, and some tests were also carried out on actualpipe. Welding of the pipe was carried out as per ASMESection IX6 of the Boiler and Pressure Vessel Code. Gastungsten arc welding (GTAW) was used for welding of 6inch NB pipe. GTAW (for root pass and few passes) andshielded metal arc welding (SMAW) (filling passes) wereused for welding of 12 inch NB pipe. Both sizes of pipe
were welded using a conventional V-groove preparation
Table 1 Chemical composition of material in wt%
Pipe size Material C Mn Si P S Cr Ni N
ASME Base 0.03 max 2.0 max 1.0 max 0.045 max 0.03 max 18–20 8–12 0.1–0.16
12 inch NB Base 0.024 1.73 0.55 0.022 0.001 18.8 9.25 0.15
6 inch NB Base 0.013 1.57 0.36 0.025 0.001 18.6 8.46 0.11
12 inch NB Weld 0.03 1.1 0.56 0.021 0.01 19.8 11.06 0.1
6 inch NB Weld 0.03 1.66 0.39 0.017 0.01 19.98 9.97 0.08
with an included angle of approximately 75◦. The detailsof the welding consumables, process and parameters aregiven in Table 2. The tensile properties of the pipe andpipe weld for each size of pipe are given in Table 3. The pipes and pipe welds were through-wall cracked in
a circumferential direction. Initial through wall crack an-
gle varied from 40 to 125◦
. Some tests were also carriedout on part-through cracked pipe and pipe welds. Thedetailed test matrix with crack dimensions and their lo-cations is given in Table 4. A typical pipe specimen withcrack location is shown in Fig. 1.
D E S C R I P T I O N O F T E S T S
The test set up shown in Fig. 2 consists of a servo hy-draulic loading system, support for the pipes and variousinstruments for the measurement of the data during thetest. A servo-hydraulic-controlled actuator was used for
Table 2 Welding consumables, process and parameters
12 inch NB 6 inch NB
Welding parameters pipe weld pipe weld
Welding process GTAW & SMAW GTAW
Filler wire type ER 308L for GTAW ER 308L
Electrode type E308L for SMAW –
Filler wire diameter (mm) 3.2, 1.6 1.6, 3.2 mm
Electrode diameter (mm) 3.15, 4 –
Welding current (A) 100–120 DC 120–140 DC
Welding voltage (V) 15 & 24 18–20
Welding speed (mm min−1) 40–50 50–60
Inert gas Argon in GTAW Argon
Table 3 Tensile properties of materials
Pipe size Material σ y (MPa) σ u (MPa) % El
12 inch NB Base 324 660 63
6 inch NB Base 318 650 67
12 inch NB Weld 450 593 38
6 inch NB Weld 400 586 33
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
the loading of the pipe during the test. The support sys-tem for the pipe tests consists of two pedestals with pairsof rollers at the outer span and inner span, which cre-ates a four-point bending condition. The test specimen
was gripped between rollers. This type of loading ensuresthat the notched section of the pipe is subjected to pure
bending stress.Pipes with part through and through-wall notches werefatigue pre-cracked before the fracture tests to ensuresharpness of the crack tip. Some pipes with a surfacenotch were subjected to fatigue loading till the crack hadgrown through the wall thickness. Thereafter fracturetests were carried out on through-wall cracked pipes. Thefinal through-wall crack size after the fatigue precracking
was taken as the initial crack size for the fracture tests.Pipes were subjected to static loading with a loading dis-
placement rate of 0.036 mm s−1. The sequence of loadingand unloading followed for the pipe test is given in Fig. 3.
Fig. 2 Actual experimental set up for a fracture test.
0 1000 2000 3000 4000 5000
Time (sec)
10
30
50
0
20
40
60
D i s p l a c e m e n t ( m m )
Loading : 4 mmUnloading : 1.6 mm
Fig. 3 Typical loading and unloading
sequence followed for the test.
The loading of the pipe was continued till the load drop of 25% was observed with respect to maximum load carryingcapacity of the pipe.
During the test on through-wall cracked pipes, load linedisplacement (LLD), load and crack mouth opening dis-placement (CMOD) were recorded. The crack growth
at either tip of the through-wall crack and the load wasrecorded in the form of images captured simultaneously using cameras. Angular rotation of the pipe about the sup-port at either end was measured using an inclinometer.
Vertical deflection of the pipe at four locations along thelength of the pipe was also measured.
Duringthe tests on part-throughcracked pipes,the crack depth along the notch length was measured using a micro-gauge based on the principle of alternating current poten-tial difference (ACPD), until the crack reached throughthickness. Thereafter all the parameters measurement re-quired for through-wall cracked pipe test was carried out.
Final diameters parallel and perpendicular to the loading
direction were measured to characterize the ovality due tothe static loading by comparing with the initial diameter. The pipe and pipe weld was characterized by tensile
and fracture resistance ( J – R) properties. CT specimensof thickness 8 mm and 20 mm were machined from thesame pipe for evaluating fracture resistance. The test wascarried out as per the guidelines of ASTM E1820. A roundtensile specimen was used for determination of tensileproperties.
D E T E R M I N AT I O N O F F R A C T U R E
R E S I S T A N C E C U R V E
The J – R curves for the pipes were evaluated using theexpression given by Zahoor and Kanninen.7 This expres-sion uses load versus load line displacement and load
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
C R A CK R E S IS TA N C E O F A US T E N IT I C ST AI N L ES S S T EE L P IP E A N D P I P E W E L D S 905
versus crack extension obtained during the experiment. The J -integral for through-wall cracked pipes underbending can be given by expression
J = J el + J pl. (1)
The elastic solution for through-wall cracked pipe sub-
jected to four-point bending can be given by
J el = K 2/ E (2)
K = σ ( F G)
(πa), (3)
where
( F G)2 = (0.7631 − 1.7602 x + 1.3511 x2
− 0.3822 x3)/(1 − x)3,
x is the ratio of cracked area to the cross-sectionalarea of pipe,
a is the half the circumferential crack length, and
σ is the outer fibre tensile bending stress.
J pl = β
(2P ) d δ +
γ · J pl · dφ,[7] (4)
where δ varies from 0 to final δ and φ varies from φ 0 tofinal φ,
β = −h(φ)/ Rt h(φ),γ = h (φ)/h(φ),
2P = total bending load,δ = load line displacement ,φ = total crack angle
h(φ) = [cos(φ/4) − 1/2sin(φ/4)],
h(φ) = dh(φ)/dφ.
The second term in Eq. (4) is due to crack growth. Thisterm vanishes when the plastic J is computed at the point of crack initiation. For stable crack growth, the first termin Eq.(4) is calculated using area under theload versus loadline displacement curve to obtain an approximate value of J pl. Thereafter this J pl is substituted in the second termto consider the crack growth in addition to the plasticdeformation.
R E S U L T S A N D D I S C U S S I O N S
In all the cases, load line displacement, rotation of thepipes about the support, crack mouth opening and crack extension were recorded with respect to applied load.Crack initiation load and maximum load carrying capacity
were also determined. The details of the results are givenbelow:
Fracture resistance of pipe material and pipe welds
The values of applied bending moment versus bending ro-tation for the 168 mm OD pipe and pipe welds are shown
0 2 4 6 8 10 12
0
20
40
60
80
100
120
140
M o m e n t , K
N - m
Rotation (deg.)
Initial crack angle 41-43o
Pipe diameter=168 mm
Thickness= 14.2 mm
Weld
Base
Fig. 4 Comparison of bending rotation for pipe and pipe welds.
0 10 20 30 40 500
20
40
60
80
100
120
140
M o m e n t , K N - m
Crack extension (mm)
Initial crack ang le 41-43o
Pipe diameter=168 mm
Thickness= 14.2 mm
Weld
Base
Fig. 5 Comparison of crack extension for pipe and pipe weld.
in Fig. 4. This figure shows that bending rotations for the weld and base material at maximum moment are 6.5 and8.5◦, respectively, although maximum moments in boththe cases are approximately the same. The drop in bend-ing moment after attaining maximum moment is faster inpipe welds compared to that in pipe material. This ob-servation can be supported by the applied moment versuscrack extension and fracture resistance ( J – R) curves forthe pipe and pipe weld, as shown in Figs. 5 and 6. Figure 5shows that the crack extension at maximum bending mo-ment in pipe and pipe weld is 4.5 and 8.5 mm, respectively.
Thereby showing that failure is not due to plastic collapsealone and that fracture mechanics has a role in the predic-tion of the instability condition in pipes. Figure 6 showsthat the fracture resistance of pipe material is superiorcompared to the pipe weld. The crack extension direc-tion is along the crack plane in a circumferential directionfor pipe as well as pipe welds. Moulin and Delliou2 haveshown that crack extension in pipes change direction after
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
Fig. 6 Comparison of J – R curve of through wall cracked pipe and
pipe weld.
0 2 4 6 8 10 12400
500
600
700
800
900
1000Pipe diameter=324 mm
Thickness=26 mmInitial Crack angle=45-47
o
M o m e n t ( K N m )
Rotation (o)
Weld
Base
Fig. 7 Comparison of moment and bending rotation of pipe and
pipe welds.
crack initiation whereas in the case of pipe welds crackspropagate in the same plane. Our findings show that crack extensions in pipe and pipe welds (GTAW) do not devi-ate from the original crack plane i.e. plane perpendicularto the pipe axis. This is due to the absence of toughnessanisotropy in the pipe material because of solution anneal-ing of pipes after manufacturing by hot working processes.Carrying out fracture toughness tests on CT specimens(machined from pipe) with cracks oriented in longitudi-nal and circumferential directions has substantiated thisobservation. Results indicate that there is no significant difference in fracture toughness property in longitudinaland circumferential directions. Therefore, there is no de-
viation of the cracks from their original plane. Applied bending moment versus bending rotation results
for the 324 mm OD pipe and pipe welds are shown inFig. 7. This figure shows that bending rotations for the
0 20 40 60 80 100 120400
500
600
700
800
Pipe diameter=324 mm
Thickness=26 mm
Initial Crack angle=45-47o
M o m e n t ( K N m )
Crcak Extension (∆a) ,mm
Weld
Base
Fig. 8 Comparison of crack extension in pipe and pipe welds.
0 20 40 60 80 100 1200
2000
4000
6000
8000
10000 Pipe diameter=324 mm
Thickness=28.5 mm
J R
( K J / m
2 )
∆ a (mm)
Crack angle 47o (Weld)
Crack angle 45o (Base)
Fig. 9 Comparison of J – R curve of through wall cracked pipe and
pipe welds.
pipe weldat maximum momentof 750 kN m is4◦ whereasthe pipe maximum moments could not be achieved evenfor bend rotation of 12◦. The drop in bending moment after attaining maximum moment is faster in the welds.
This observation can be supported with applied moment versus crack extension and fracture resistance ( J – R) curvesfor the pipe and pipe weld, as shown in Figs. 8 and 9.
These observations are similar to that in 168 mm ODpipes. Figure 8 shows the crack extension at maximumbending moment in weld is 12.5 mm. Figure 9 shows that fracture resistance of pipe is superior compared to that of the pipe welds. The crack extension direction is alongthe crack plane in a circumferential direction for pipe as
well as pipe weld. Moulin and Delliou2 have shown that fracture resistance of pipe weld (SMAW) is lower than that of pipe.
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
C R A CK R E S IS TA N C E O F A US T E N IT I C ST AI N L ES S S T EE L P IP E A N D P I P E W E L D S 907
0 2 4 6 80
40
80
120
160
M o m e n t , K N - m
Rotation (deg.)
Initial crack angle 54-56°
Pipe diameter=168 mm
Thickness= 14.2 mm
Fatigued pipe-200 thousand cycles of
max. load 250KN (240 MPa) and R=0.1
Virgin pipe with fatigue pre cracking
Fig. 10 Comparison of moment and bending rotation for pipe and
pipe weld.
Fracture resistance of pre-cracked
and fatigue-tested pipe weldsPre-cracked pipe specimens are those with through-wallmachined notch and then fatigue pre-cracked at very low load, i.e.10% of limit load, to generate a sharp crack tip.Fatigue tested pipes are part through notched pipe sub-
jected to constant amplitude cyclic loading at more than50% of limit load until the crack has grown through-wall.
The basic aim of this comparison is to bring out the effect of fatigue loading applied prior to the fracture test on thefracture resistance of material. There is no such reported
work on austenitic stainless steel pipes that fracture resis-tance depends on the prior loading histories, or fatigue
loading before conducting fracture tests. Similar studiesconducted on carbon steel pipes showed that fracture re-sistance of pipe depends on the prior loading histories to
which pipe has been subjected.13 One such comparisonin terms of bending moment versus rotation and fractureresistance ( J – R) curve for pre-cracked and fatigue-testedpipe weld is shown in Figs. 10 and 11. The comparisonactually shows higher fracture resistance for fatigued pipe.
This may be within the experimental scatter and one canonly say that a large reduction in fracture resistance doesnot take place as observed in the case of carbon steelpipes.13 Vickers Hardness (VHN) measured on the vir-gin and fatigued pipe and pipe welds are given in table 5.
There is strain hardening in pipe and pipe welds due tothe fatigue loading, similar to carbon steel pipe.13
Fracture resistance of pipe welds with different initial crack angle
Bending rotation and crack extension with respect tobending moment are shown in Figs. 12–13 for 6 and 12inch pipe welds having different initial crack angle. Corre-sponding fracture resistance curves for the pipe welds are
0 10 20 30 40 500
1000
2000
3000
4000
5000
pipe weld (GTAW)
Pipe diameter=168 mm
Thickness=14.2 mm
J R
( K J / m 2 )
∆a (mm)
Crack angle 54° (Fatigued)
Crack angle 56° (Virgin)
Fig. 11 Comparison of J – R cruve of through wall cracked fatigued
and virgin pipe weld.
0 1 2 3 4 5 6 7 80
100
200
300
400
500
600
700
800
S P W T W 1 2 - 1 3
S P W T W 1 2 -1 2
S P W T W 1 2 - 1 1
S P W T W 6 - 6S P W T W 6 - 5
S P W T W 6 - 4
M o m e n t , K N - m
Rotation (deg.)
Fig. 12 Moment vs rotation curves for through wall cracked pipe
welds.
0 20 40 60 80 100 1200
100
200
300
400
500
600
700
800
S P W T W 1 2 -1 3
S P W T W 1 2 - 1 2
S P W T W 1 2 -1 1
S P W T W 6 - 6S P W T W 6 -5
S P W T W 6 -4 M o m e n t , K N - m
Crack extension (mm)
Fig. 13 Moment vs crack extension for through wall cracked pipe
welds.
shown in Fig. 14. Extensive studies carried out by Moulinand Delliou,2 Shibata et al .,10 Delliou and Crouzet,11
Forster et al .,12 Stadtmuller et al .14 and Moulin et al .15 onaustenitic stainless steel pipe having different initial crack
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
Fig. 14 J – R curves for through wall cracked pipe welds.
angle shows that there is considerable crack extension af-ter initiation and before attainment of maximum moment.
Moulin and Delliou have also found that there is no ini-tiation of initial crack growth in the case of pipe with acrack angle less than 45◦. Results described in this papershow that initiation of crack growth has been observedeven for a crack angle of 41◦. The Wilkowski et al .8 work on fracture instability on circumferential cracked pipe of austenitic stainless steel (SS304) has shown dependenceof geometry (pipe size) on the fracture resistance curve.
They have also shown that fracture resistance curves donot depend on the initial crack angle. The Moulin andDelliou studies show a dependence of fracture resistanceon the initial crack angle. Our findings show that fracture
resistance depends on the pipe size and the initial crack angle. Fracture resistance of pipes with different initialcrack angles will differ if the constraint encountered by the cracked pipe is different. In the present study, all thetests have been carried out on pipe welds having different crack angles and show a variation in the J – R curve. The
variation in the J – R curve is more for 324 mm outer di-ameter pipe welds compared to 168 mm diameter pipes.
The difference in J – R curves may be attributed to the dif-ferences in toughness and tensile strength of the base and
weld, which leads to variations in constraint. This differ-ence is enhanced for the larger sizes of pipe.
Initiation fracture toughness
Usually initiation fracture toughness is determined by fol-lowing the guidelines given in ASTM standard E1820 us-ing J – R curve and blunting line. In this paper, initiationfracture toughness for the pipes was evaluated from thecrack initiation load. Crack initiation was assumed to oc-cur corresponding to crack growth of 0.2 mm, which ispossible to measure, by imaging of the crack tip. Initia-tion fracture toughness is the intersection of the line cor-
Table 5 Vickers Hardness of virgin and fatigued pipe
Precracked Fatigued
Material condition (VHN) (VHN)
168 mm OD pipe (base) 114 143
168 mm OD pipe weld (GTAW) 80–95 120–150
324 mm OD pipe (base) 140 173324 mm OD pipe weld (GTAW +SMAW) 120 140–165
responding to 0.2 mm of crack growth and the J – R curve. Values for all the cases are given in Table 6.
Table 6 shows that there is reduction in the initiationfracture toughness of the pipe weld compared to pipe.
Variation in the values of initiation fracture toughness forpipe weld is due to the scatter in the measured data in theinitial region of crack growth.
Initiation of crack growth occurred at a load lower thanthe maximum load bearing capacity of the pipe. The con-dition of maximum moment can be used for characteriz-ing the stability of the cracked pipe if the crack growthat the maximum moment is known. The crack growth at maximum moment for the 6 and 12 inch pipe weld wasobserved in the range of 4 to 12 mm.
In all the cases crack growth is in the direction of initialcrack plane. There is no deviation of growing crack fromthe initial crack plane in pipe and pipe weld. Crack growthin the circumferential plane for pipe and pipe welds forboth sizes are shown in Fig. 15 (a–d). It can be also seenthat the crack extension in the case of pipe of 324 mm
outer diameter is less compared to the pipe weld with thesame initial crack angle, whereas in the case of pipe weldand pipe of 168 mm outer diameter, the crack extensionis comparable. This also indicates that fracture resistanceof the pipe and pipe weld of 324 mm outer diameter dif-fers significantly, whereas in the case of 168 mm outerdiameter this difference is not very significant. Thinning of pipe during fracture test has also substan-
tiated comparable fracture resistance in the case of pipeand pipe weld of 168 mm outer diameter. The extent of thinning is comparable in the case of pipe and pipe weldand is shown in Fig. 16(a and b).
A N A LY T I C A L M E T H O D S F O R P R E D I C T I O N
O F L I M I T L O A D
Limit load prediction
Limit load analysis is one of the simplest methods to pre-dict the instability of the piping system. This is also knownas net section collapse. Existing limit load expressions donot account for the crack extension or growth. Existingexpressions give reasonable prediction for pipes as long
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
C R A CK R E S IS TA N C E O F A US T E N IT I C ST AI N L ES S S T EE L P IP E A N D P I P E W E L D S 909
(a) (b)
(c) (d)
Fig. 15 (a) Pipe weld of 168 mm OD. (b) Pipe based of 168 mm OD. (c) Pipe weld of 324 mm OD. (d) Pipe base of 324 mm OD.
(a)
(b)Fig. 16 Thinning effect in cracked pipe (a)
and pipe weld, GTAW (b).
as there is no crack extension. In pipes of higher diam-eter and thickness where crack extension occurs prior tonet section collapse, the expressions give non-conservativeprediction. In the case of pipe welds, there are two ambi-guities. Oneis useof flow strength (base or weld) andotherinfluence of base material (higher toughness) on crack ex-tension in weld. Because of these factors, prediction of limit load gives non-conservative results. In the present investigation, the flow strength of the base material has
been used for prediction because it is lower compared tothe weld properties. The multiplication factor suggestedin the present paper takes care of crack extension priorto net section collapse and the influence of base materialon crack extension. Keeping the above points in consid-eration, the limit load expression can be used for pipe
welds.Experimental maximum moment and the calculatedlimit
moment for all the cases are given in Table 7. Analytical
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
limit load (P o) were calculated based on initial crack angleby expression,9
P o = (16σ f R2t )/(Z − L)[cos{(a/2t )(θ/2)} − (a/2t )sin θ ],
(5)
where σ f is the flow stress i.e. averageof yield and ultimatetensile strength of material in MPa, R is the mean radiusof pipe in mm, Z is the outer span and L is the inner spanin mm, θ is half the crack angle in degree, t is the thicknessin mm.
Limit loads were calculated based on the flow stress and yield strength of the material. Tensile properties (yieldand ultimate tensile strength) were used for the evalu-ation of limit load for pipe welds. Comparison showedthat the predicted limit load based on flow stress is higherthan the experimental limit load by 2–12% and 5–32%for 168 mm (GTAW) and 324 mm (GTAW + SMAW)outer diameter pipe welds, respectively. The lower limit load obtained experimentally in comparison to that pre-dicted is due to the substantial crack growth in pipe be-fore attaining maximum moment. The expression givenfor calculation of limit load is based on the initial crack length or crack angle. Calculated limit load gives non-conservative results can also be shown as variation of theratio of [ M max /(4 R2t σ f )] with respect to initial crack an-gle. The plot in Fig. 17 shows that all the experimen-tal points lie under the curve given by [cos{(a/2t )(θ/2)}− (a/2t )sinθ ] which is indicative of non-conservatism.
0 50 100 150 200 250 300 3500.0
0.2
0.4
0.6
0.8
1.0
Crack Angle (°)
M / 4 R
m 2 t σ
f Cos(θ /4)-1/2Sin(θ /2)
Theoretical
ϕ 324 mm pipe weld (SMAW)
ϕ168 mm pipe weld (GTAW)
Flow stress (σf)= (σu+σy)/ 2
Fig. 17 Comparison of analytical and experimental results.
Delliou and Crouzet 11 have shown that limit load anal-
ysis is not always conservative for large diameter pipes(406 mm outer diameter).
Limit load for pipe welds
The limit load expression takes into account flow prop-erties of material, pipe size and crack configuration. It has been seen that expression (5) for the limit loadgives reasonable prediction in the case of pipe and non-conservative prediction (i.e. higher value compared to
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
C R A CK R E S IS TA N C E O F A US T E N IT I C ST AI N L ES S S T EE L P IP E A N D P I P E W E L D S 911
experimental) for the pipe weld. This is due to the het-erogeneity in the pipe welds in terms of mechanical andmetallurgical properties. There is no available literatureon evaluation of the limit load for the pipe welds. It hasbeen reported that flow strength (average of yield and ul-timate tensile strength) of the base material can be used
for prediction of limit load for welds. The values of limit load considering flow strength of base material is given in Table 6. The results indicate that the limit load expressionrequires modification when using it for welds. Based onthe results of several pipe welds, expression (5) for limit
0 100 200 300 400 5000
1000
2000
3000
4000
J R
( K J / m 2 )
Limit load Stress (SR), MPa
168 mm OD Pipe
weld (GTAW)
56.27o
74.35o
102.46o
324 mm OD Pipe
weld (GTAW+SMAW)
65.96o
94.79o
125.38o
Fig. 18 Limit load stress variation with fracture resitance for pipe
welds.
Table 7 Ovality with respect to diameters and ovalization factor
Initial Crack Final crack Initial
Test no. angle (2θ ◦) angle (2θ ◦) diameter (mm) D V (mm) DH (mm)
load evaluation for pipe can be modified by incorporatinga multiplication factor for better prediction of limit loadin the case of pipe welds. The suggested factor is 0.85 forpipes welded using GTAW and 0.7 for pipes welded usingGTAW + SMAW. Modified expression can be written as
P o = 0.8516σ f R2t (Z − L)[cos{(a/2t )(θ/2)}
−a/2t
sin θ ] for GTAW welds (6)
P o = 0.716σ f R2t
(Z − L)[cos{(a/2t )(θ/2)}
− (a/2t )sin θ ] for GTAW + SMAW welds. (7)
Limit load stress is the stress corresponding to the maxi-mum bending moment evaluated by rearranging the limit load expression given in (6). Limit load stress evaluatedfrom the experimental maximum moment indicates lower
values compared to the flow strength used for evaluationof limit load. This is explained below.
Limit load stress
Limit load stress based on theexpression forlimit moment considering experimental bending moment has been eval-uated to show that limit load stress is not unique for theGTAW and GTAW + SMAW pipe welds. The formulaof limit load stress can be given as 12
S R = M exp/4 R2t (cos θ/4 − 0.5sin θ/2), (8)
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
where M exp is the experimental bending moment and θ
is the half-crack angle. Limit load stress variation withfracture resistance J values for pipe welds of 324 mm(GTAW + SMAW) and 168 mm (GTAW) outer diam-eter have been plotted and are shown in Fig. 18. It hasbeen observed that limit load stress curves do not con-
verge at a particular value of stress as reported by Moulin,Clement, Drubay and Goudet 12 for crack angle greaterthan 30◦. Figure 18 shows that limit load stress for GTAW pipe welds is in the range of 415–450 MPa where as forGTAW + SMAW Pipe welds is 325–375 MPa. Flow stressof the pipe, GTAW pipe weld and GTAW + SMAW pipe
welds are 488, 493 and 521 MPa, respectively. Compari-son of this limit load stress with flow stress (σ f = (σ u +
σ y )/2) indicates that flow stress is not suitable for the pre-diction of the limit load. Factors indicated in Eqs. (6) and(7) can be used for two different types of pipe welds. Ourfinding shows that for the given range of crack angles andsizes of the pipe, limit load criterion is sensitive to the tear-
ing taking place before the maximum bending moment isreached.
O VA L I T Y I N C R A C K E D P I P E A N D P I P E W E L D S
Ovalization of pipe is a geometrical effect and leads tolowering of the fracture resistance curve and the maxi-mum moment.11 In this study, ovality was not measuredcontinuously with increasing applied load during the tests;therefore, it is not possible to infer the effect of ovality onfracture resistance of pipe. After the fracture test, ovalitiesin pipe and pipe welds were measured. It was found that
pipe diameter to the loading direction has increased andthat in the perpendicular direction decreased with respect to initial diameter. Similar observation has been reportedby Delliou and Crouzet 11 and Forster etal .12 Ovality mea-surement for all the tests is given in Table 7.
F R A C T U R E R E S I S T A N C E U S I N G C O M P A C T
T E N S I O N S P E C I M E N S
Load against CMOD data obtained from testing of CTspecimens machined from pipe and pipe welds (GTAW and GTAW + SMAW) of 168 mm and 324 outer diam-eters are shown in Fig. 19(a and b). The correspondingfracture resistance ( J – R) curves obtained are shown inFig. 19c. The comparison of fracture resistance curvesobtained from CT specimens also shows the same be-haviour as in the actual pipe and pipe weld. This observa-tion has also been reported by Forster12 and Moullin andDelliou.2
Critical stretched zone width (SZW) was measured onthe fractured surface of compact tension specimens madeof pipe base and pipe weld material under scanning elec-tron microscope (SEM). Typical SZW of pipe and pipe
0 2 4 6 8 100
2
4
6
8
10
12(a)
L o a d K N
CMOD, MM
168 mm OD pipe
CT 304LN (Base)
CT 308L (Weld)
0 2 4 6 8 100
10
20
30
40
50
60(b)
L o a d K N
CMOD, MM
324 mm OD pipe
CT 304LN (Base)
CT 308L (Weld)
0 2 4 6 8 10 120
40 0
80 0
1 2 0 0
1 6 0 0
(c)
J R , K J / m
2
∆a, mm
304LN, 168mm O D p ipe ( Base)
E R 3 0 8 L ( G T A W ) , 1 6 8 m m O D w e l d E 3 0 8 L (G T A W + S M A W ) , 3 2 4 m m O D w e ld
Fig. 19 (a) Comparison of load and load line displacement. (b)
Comparison of load and CMOD. (c) Comparison of J – R curves.
weld are shown in Fig. 20(a and b). Initiation fracturetoughness ( J i) for pipe was also calculated based on thecritical SZW measured from CT specimen for compari-son purposes. J i was determined from the intersection of the J – R
curve (polynomial fit for crack growth of 1.5 mm) andthe vertical line corresponding to critical SZW. This isshown in Fig. 21(a and b). The figures indicate that there is no difference in the J i of CT and pipe. The J – R curve for pipe weld and CT (from pipe weld) is
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
C R A CK R E S IS TA N C E O F A US T E N IT I C ST AI N L ES S S T EE L P IP E A N D P I P E W E L D S 913
(a) (b)Fig 20 SZW on fractured CT specimen of
(a) pipe and (b) pipe weld.
0.0 0.5 1.0 1.5 2.00
1000
2000
3000
4000
Ji
Critical SZW (0.14 mm) line
CT, J=20+1658 a-428 a2, J
i=261 KJ/m
2
Pipe, J=-86+2281 a+56 a2, J
i=235 KJ/m
2
(a)
J R ,
K J
/ m 2
∆a, mm
CT
Pipe
0.0 0.5 1.0 1.5
400
800
1200
Ji
Critical SZW (0.06 mm) line
CT, J=60+1608 a-938 a2+237 a
3, J
i=153 KJ/m
2
Pipe, J=139+1195 a-572 a2+182 a
3, J
i=143 KJ/m
2
(b)
J R ,
K J / m 2
∆a, mm
CT
Pipe weld
Fig. 21 (a) Comparison of J i of pipe and CT using SZW. (b)Comparison of Ji of pipe weld and CT using SZW.
comparable whereas there is a significant difference in thecase of pipe and CT (Pipe) specimens. This is explainedbelow.
Full fracture resistance ( J – R) curves for the 168 and324 mm outer diameters pipe and pipe welds have beencompared with those of the corresponding compact ten-sion specimens and are shown in Fig. 22(a–c). This figure
shows that J – R curves of the CT specimen lay on thelower side compared to the pipe except in the case of pipe
weld of 168 mm outer diameter. This is due to the higherconstraint in the case of the CT specimen compared topipe in general. In the case of weld material, there is alsoinfluence of base material on the fracture resistance of the
weld. This influence varies depending upon the toughnessproperties of the base and weld. In the present study, pipe
welds of GTAW and GTAW + SMAW have been used.Fracture toughness properties of GTAW are comparable(see Fig. 19c) and that of GTAW + SMAW is inferior(see Fig. 19c). Figure 22a shows that the J – R curve of CT(base) specimen is lower compared to pipe base material,
which can be explained based on the fact that CT spec-imen constraint is higher, compared to pipe. Figure 22bshows that J – R curves for CT (weld) and pipe weld are thesame. This can be explained based on the fact that thereis the same constraint because of the comparable tough-ness and lower strength of the weld metal compared tothe base material. Because of the lower strength, plastic
deformation (critical plastic zone) remains restricted tothe weld and leads to a reduction in constraint. Figure22c shows that the J – R curve of CT (weld) is significantly lower compared to pipe weld. In this case, weld fracturetoughness is significantly lower and has lower strength
which results in reduced critical plastic zone size. Lowerfracture resistance of this weld shows that the influenceof base material is not very significant and leads to higherconstraint.
Fracture surface examination was also carried out un-der SEM for pipe and pipe welds. The size of dimples
was large for GTAW welds compared to the SMAW. Thephotographs of the fracture surfaces are shown in Fig. 23
(a and b). This observation indicates that the fractureresistance of the pipe weld (GTAW) is superior comparedto the pipe weld (GTAW + SMAW).
C O N C L U S I O N S
Studies on crack resistance behaviour of the austeniticstainless steel pipe and pipe welds along with CT spec-imens have been carried out and following are theconclusions.
c 2006 Blackwell Publishing Ltd. Fatigue Fract Engng Mater Struct 29, 901–915
11 Delliou, P. L. and Crouzet, D. (1990) Experimental and
numerical study of circumferentially through-wall cracked pipe
under bending including ductile crack growth and ovalization.
ASME PVP 195, 85–89.
12 Forster, K., Griiter, L., Setz, W., Bhandari, S., Debaene, J. P.,
Faidy, C. and Schwalbe, K. H. (1996) Crack resistance of austenitic pipes with circumferential through wall crack. Int. J.Pres. Ves. Piping 65, 335–342.
13 Singh, P. K., Vaze, K. K., Bhasin, V., Kushwaha, H. S., Gandhi,
P. and Murthy, D. S. R. (2003) Crack initiation and growth
behaviour of circumferentially cracked pipes under cyclic and
monotonic loading. Int. J. Pres. Ves. Piping 80(Issue 9)
September, 629–640.
14 Stadtmuller, W. and Strum, D. (1997) Leak-before-break
behaviour of austenitic and ferritic pipes containing