omae83056 girthweld final

1 Copyright2012 by ASME

CRACK GROWTH TESTING OF AN X80 PIPELINE GIRTH WELDUSING SE(T) AND SE(B) FRACTURE SPECIMENS

Proceedings of the ASME 31st International Conference on Offshore Mechanics and Arctic EngineeringOMAE2012

July 1-6, 2012, Rio de Janeiro, Brazil

OMAE2012-83056

Diego F. S. BurgosDept. of Naval Arch. and Ocean Eng.

University of São PauloSão Paulo, SP 05508-900, Brazil

Leonardo L. S. MathiasDept. of Naval Arch. and Ocean Eng.

University of São PauloSão Paulo, SP 05508-900, Brazil

Marcelo ParedesDept. of Naval Arch. and Ocean Eng.

University of São Paulo,São Paulo, SP 05508-900, Brazil

Gustavo H. B. DonatoDept. of Mech. Eng.,FEI University Center

S. B. Campo, SP 09850-901, Brazil

Claudio RuggieriDept. of Naval Arch. and Ocean Eng.

University of São Paulo,São Paulo, SP 05508-900, Brazil

ABSTRACT

Accurate measurements of fracture resistance properties, in-cluding crack growth resistance curves for pipeline girth welds,become essential in defect assessment procedures of the weld-ment region and the heat affected zone, where undetected crack-like defects (such as lack of penetration, deep undercuts, rootcracks, etc.) may further extend due to to high tension stressesand strains. This work presents an investigation of the ductiletearing properties for a girth weld made of an API 5L X80 pipe-line steel using experimentally measured crack growth resist-ance curves ((J-Δa curves). Use of these materials is motivatedby the increasing demand in the number of applications formanufacturing high strength pipes for the oil and gas industryincluding marine applications and steel catenary risers. Testingof the pipeline girth welds utilized side-grooved, clamped single

edge notched tensiles (SE (T)) specimens and three-point (3P)bend single edge bend (SE(B)) specimens with a weld centerlinenotch to determine the crack growth resistance curves basedupon the unloading compliance (UC) method using a singlespecimen technique. Recently developed compliance func-tions and η-factors applicable for SE (T ) and SE(B) fracturespecimens with homogeneous material and overmatch weldsare introduced to determine crack growth resistance datafrom laboratory measurements of load-displacement records.This experimental characterization provides additional tough-ness data which serve to evaluate crack growth resistance prop-erties of pipeline girth welds using SE (T) and SE(B) specimenswith weld centerline cracks.


INTRODUCTION

Structural integrity assessments of pipe girth welds play a keyrole in design and safe operation of piping systems, includingdeep water steel catenary risers. More efficient and fasterinstallation methods now employ the pipe reeling processwhich allows welding and inspection to be conducted at on-shore facilities. The welded pipe is coiled around a large di-ameter reel on a vessel and then unreeled, straightened andfinally deployed to the sea floor. However, the reeling processsubjects the pipe to large bending loads and plastic deforma-tion (2~3%) well beyond the material’s elastic limits with apotentially strong impact on flaw acceptance criteria for thegirth weld. Consequently, accurate measurements of fractureresistance properties, including crack growth resistance curvesof the girth weld material, become essential in defect assess-ment procedures of the weldment region and the heat affectedzone, where undetected crack-like defects (such as lack ofpenetration, deep undercuts, root cracks, etc.) may further ex-tend due to to the high tension stresses and strains.

Fracture mechanics based approaches to describe ductilefracture behavior in structural components, including weldedstructures, rely upon crack growth resistance (J-Δa) curves (alsotermed R-curves) to characterize crack extension followed bycrack instability of the material [1,2]. These approaches allowthe specification of critical crack sizes based on the predictedgrowth of crack-like defects under service conditions. Currentstandardization efforts now underway [3-5] advocate the use ofsingle edge notch tension specimens (often termed SE(T) orSENT crack configurations) to measure experimental R-curvesmore applicable to high pressure piping systems, including girthwelds of marine steel risers.

The primary motivation to use SE(T) fracture specimens indefect assessment procedures for this category of structural com-ponents is the strong similarity in crack-tip stress and strainfields which drive the fracture process for both crack configura-tions [6-9]. Recent applications of SE(T) fracture specimens tocharacterize crack growth resistance properties in pipeline steels[10] have been effective in providing larger flaw tolerances and,at the same time, reducing the otherwise excessive conservatismwhich arises when measuring the material’s fracture toughnessbased on high constraint, deeply-cracked, single edge notchbend (SE(B)) or compact tension (C(T)) specimens. However,while now utilized effectively in fracture testing of pipeline girthwelds, some difficulties associated with SE(T) testing proce-dures, including fixture and gripping conditions, raise concernsabout the significance and qualification of measured crackgrowth resistance curves. Such uncertainties in measured frac-

ture toughness may potentially affect tolerable defect sizes ob-tained from engineering critical assessment (ECA) procedures.While slightly more conservative, testing of shallow−crack bendspecimens (which is a nonstandard SE(B) configuration) maybecome more attractive due to its simpler testing protocol, labo-ratory procedures and much smaller loads required to propagatethe crack. Consequently, use of smaller specimens which yetguarantee adequate levels of crack-tip constraint to measure thematerial’s fracture toughness becomes an attractive alternative.

This work presents an investigation of the ductile tearingproperties for a girth weld made of an API 5L X80 pipeline steelusing experimentally measured crack growth resistance curves(J-Δa curves). Use of these materials is motivated by the in-creasing demand in the number of applications for manufac-turing high strength pipes for the oil and gas industry includingmarine applications and steel catenary risers. Testing of thepipeline girth welds employed side-grooved, clamped SE (T)specimens and 3P bend SE(B) specimens with a weld centerlinenotch to determine the crack growth resistance curves basedupon the unloading compliance (UC) method using a singlespecimen technique. Recently developed compliance func-tions and η-factors applicable for SE (T ) and SE(B) fracturespecimens with homogeneous material and overmatch weldsare introduced to determine crack growth resistance datafrom laboratory measurements of load-displacement records.This experimental characterization provides additional tough-ness data which serve to evaluate crack growth resistance prop-erties of pipeline girth welds using SE (T) and SE(B) specimenswith weld centerline cracks.

EXPERIMENTAL DETAILS

Material Description

The material utilized in this study was a high strength, low alloy(HSLA), API grade X80 pipeline steel produced by Usiminas,Brazil, as a base plate using a control-rolled processing routewithout accelerated cooling. The mechanical properties andstrength/toughness combination for this material are mainly ob-tained by both grain size refinement and second-phase strength-ening due to the small-size precipitates in the matrix. The20-inch pipe with longitudinal seam weld from which the girthweld SE(T) and SE(B) specimens were extracted was fabricatedat Tenaris-Confab plant in Pindamonhangaba, Brazil, using theUOE process. Table 1 provides the mechanical properties of thebase plate material (measured values based on standard tensile


testing) and the weld metal (estimated values based on measuredhardness HV values [11]).

Table 1 Material properties of the base plate and weldmentfor the tested pipeline girth weld.

Base Plate Weld

σys (MPa) σuts (MPa) n σys (MPa) σuts (MPa) n

558 722 11 613 793 11

Using Annex F of API 579 [12], the true stress-true strain re-sponse for the tested materials can be described by

Á= σE+σH1∕N (1)

where σ and Á are the (total) true stress and logarithmic strain,H is a material constant and n = 1∕N defines the Ramberg-Os-good exponent which is often required in J (and CTOD) estima-tion procedures using the η-method (to be addressed in the fol-lowing sections). Now, using the estimation expressionsprovided in Annex F of API 579 [12], the Ramberg-Osgood ex-ponent for the base plate and weld metal is given by n≈ 11 andH≈984 MPa.

Welding Procedure

The tested weld joint wasmade from the API X80UOE pipe hav-ing thickness, tw=19 mm. Girth welding of the pipe was per-formed using the FCAW process in the 1G (flat) position with asingle V-groove configuration in which the root pass was madeby GMAW welding employing the Surface Tension Transfer(STT) process developed by Lincoln Electric. The main weld pa-rameters used for preparation of the test weld using the FCAWprocess are: i) Number of passes 12 (including the root passmadeby the STT process); ii) Welding current 165 A; iii) Weldingvoltage 23 V; iv) Average heat input 1.5 kJ/mm. Fernandes [11]provided additional details on the welding process and weldpreparation using the STT procedure for the root pass followedby the FCAW process to produce the test weld employed in thiswork.

Specimen Geometry

Unloading compliance (UC) tests at room temperature were per-formed on center notch weld, clamped SE(T) specimens with afixed overall geometry and crack length to width ratios, a∕W ,

to measure tearing resistance curves in terms of J-Δa. Theclamped SE(T) specimens have a∕W=0.4 and H∕W=10 withthickness B=14.8 mm, width W=14.8 mm and clamp distanceH=148 mm (refer to Fig. 1(a)). Here, a is the crack depth andW is the specimen width which is slightly smaller than the pipethickness, tw. UC tests at room temperature were also conductedon center notched welded SE(B) specimens with a∕W=0.25with thickness B=14.8 mm, width W=14.8 mm and spanS=4W (refer to Fig. 1(b)). Conducted as part of a collaborativeprogram between the University of São Paulo, Tenaris-Confab,Lincoln Electric Brasil and Petrobrás, testing of these specimensfocused on the development of accurate procedures to evaluatecrack growth resistance data for pipeline girth welds.

All specimens, including the SE(T) configuration, wereprecracked in bending using a three-point bend apparatus verysimilar to a conventional three-point bend test. After fatigue pre-cracking, the specimens were side-grooved to a net thickness of~85% the overall thickness (7.5% side-groove on each side) topromote uniform crack growth and tested following some gener-al guidelines described in ASTM E1820 standard [13]. Recordsof load vs. crack mouth opening displacements (CMOD) wereobtained for the specimens using a clip gauge mounted on knifeedges attached to the specimen surface.

ESTIMATION PROCEDURE OF J-R CURVES

Laboratory measurements of fracture toughness resistance datafocus primarily on the unloading compliance method based upontesting of a single specimen. The methodology essentially relieson determining the instantaneous value of J and specimen com-pliance at each partial unloading during the measurement of theload-displacement curve as illustrated in Fig. 2(a). This tech-nique enables accurate estimations of J and Δa at several loca-tions on the load-displacement records from which the J-R curvecan be developed. Here, we provide an overview on the method-ology which builds upon previous work of Cravero and Ruggieri[5,14], Ruggieri [15] and Paredes and Ruggieri [16].

The procedure to estimate crack growth resistance data con-siders the elastic and plastic contributions to the strain energy fora cracked body under Mode I deformation [2] so that J can beconveniently defined in terms of its elastic component, Je , andplastic component, Jp, as

J = Je+ Jp =K2IE′ +

ηJ ApBN b0

(2)

where KI is the elastic stress intensity factor for the cracked con-figuration, Ap is the plastic area under the load-displacementcurve, BN is the net specimen thickness at the side groove roots


(a) (b)

Figure 1 Geometry of tested fracture specimens with weld centerline notch. (a) Clamped SE(T) specimen with aW=0.4 andHW=10; (b) 3P SE(B) specimens with aW=0.25 and SW=4 (BxB configuration).

H

B

W<tw

a

S

B

W<tw

Longitudinal axisof the pipe

(BN=B if the specimen has no side grooves where B is the speci-men gross thickness), b0 is the initial uncracked ligament(b0 =W−a0 where W is the width of the cracked configurationand a0 is the initial crack length). In writing the first term of Eq.(2), plane-strain conditions are adopted such thatE′ = E∕(1 −ν2 ) where E and ν are the (longitudinal) elastic mo-dulus and Poisson’s ratio, respectively. Factor ηJ introduced bySumpter and Turner [17] represents a nondimensional parameterwhich relates the plastic contribution to the strain energy for thecracked body and J. Fig. 2(b) illustrates the essential features ofthe estimation procedure for Jp. Here, we note that Ap (and con-sequently ηJ ) can be defined in terms of load-load line displace-ment (LLD or Δ) data or load-crack mouth opening displace-ment (CMOD or V ) data. For definiteness, these quantities aredenoted ηCMODJ and ηLLDJ . While both definitions serve essentiallyas a means to quantify the effect of plastic work on the J-integral,ηJ -values based on LLD have a different character than the cor-responding ηJ -values based on CMOD. Specifically, factor ηLLDJ

enters into the expression to correct the J-value due to crack ex-tension as discussed next.

The previous expression (2) defines the key quantities driv-ing the evaluation procedure for J as a function of applied (re-mote) loading and crack size. However, the area under the actualload-displacement curve for a growing crack differs significant-ly from the corresponding area for a stationary crack (which thedeformation definition of J is based on) [2]. Consequently, themeasured load-displacement records must be corrected for crackextension to obtain an accurate estimate of J-values with in-creased crack growth. A widely used approach (which forms thebasis of current standards such as ASTME1820 [13]) to evaluateJ with crack extension follows from an incremental procedurewhich updates Je and Jp at each partial unloading point, denotedk, during the measurement of the load vs. displacement curve inthe form


Figure 2 (a) Partial unloading during the evolution of loadwith crack mouth opening displacement; (b) Definition of theplastic area under the load-displacement curve.

CMOD

1Ck

P

(a)

P

A p

V(b)

k-th unloading

V

P

V

P

Jk = Jke+ Jkp . (3)

Within this approach, the k-th elastic term of J is directlycalculated from the corresponding k-th value of KI using the firstterm of previous Eq. (2) which yields

JKe =KkI

2

E′ . (4)

Similarly, the k-th plastic term of J follows from the secondterm of Eq. (2) using the current plastic area, Akp . Given an esti-mated value for Jp at k −1, the k-th value of J is given by

Jkp = Jk−1p +ηCMODk−1

bk−1BNAkp− Ak−1p × Γ (5)

in which Γ is defined as

Γ = 1−γLLDk−1bk−1(ak− ak−1) (6)

where factor γLLD is evaluated from

γLLDk−1 =− 1+ ηLLDk−1− bk−1W η^

LLD

k−1

ηLLDk−1 (7)

with

η^LLD

k−1 = WdηLLDk−1dak−1

(8)

Another key step in the experimental evaluation of crackgrowth resistance response involves the accurate estimation ofthe instantaneous crack length as testing progresses. The unload-ing compliance technique provides a convenient and yet simpleprocedure to correlate crack extension to the specimen com-pliance with increased crack growth. Fig. 2(a) illustrates the es-sential features of the method. The slope of the load-displace-ment curve during the k-th unloading defines the instantaneousspecimen compliance, denoted Ck , which depends on specimengeometry and crack length. The following section provides im-proved expressions for factors ηCMODJ and ηLLDJ as well as the non-dimensional compliance, μ , for the tested clamped SE(T) speci-men and the 3P bend SE(B) specimen which lead to accuratemeasurements of crack growth resistance data.

The previous development based upon the η -factor conceptretains strong contact with current standards to determine exper-imental J-values using common fracture specimens for homoge-neous materials. Computation of η -factors for fracture speci-mens made of heterogeneous materials, such as welded crackconfigurations, is relatively straightforward and depends uponthe mismatch ratio, My , defined by

My=σWMysσBMys

(9)

where σBMys and σWMys denote the yield stress for the base plate met-al and weld metal.

J-R CURVE TESTING

Stress Intensity Factors

Following the procedure described before, parameter KI is eval-

uated at the current load, Pk , using


KkI =PkBN W

F ak∕W (10)

where F (ak∕W ) defines a nondimensional stress intensity factordependent upon specimen geometry, crack size and loading con-dition. For the tested SE(T) specimen with clamped conditionswith H∕W=10 analyzed here, Cravero and Ruggieri [5] provideanalytical expressions for the nondimensional stress intensityfactors F (ak∕W ) in the form

F (ak∕W) =5

r=0

ξr ak∕W r

(11)

where it is understood that a 5-th order polynomial fitting isemployed with the polynomial coefficients, ξr , given byξ0=0.2832, ξ1=3.8497, ξ2=−1.4885, ξ3=4.1716,ξ4=9.9094 and ξ5=−7.4188. For the SE(B) specimen, analyt-ical expressions for F (ak∕W ) are given in Tada et al. [18].

Elastic Compliance

Standard elastic analyses under plane-strain conditions typicallydefine the (linear) dependence of applied load on displacementfor a given specimen geometry with different crack length. Suchdependence provides the required data from which the specimencompliance for varying crack size is extracted. For a fixed dis-placement, V or CMOD, increasing the crack size decreases thespecimen stiffness which, consequently, reduces the appliedload, P. The slope of each curve on the P-CMOD relationshipthen defines the inverse of the specimen compliance, denoted asC , which enables building the functional dependence of C anda∕W . Figures 3-4 provide the variation of normalized com-pliance, μSE(T)CMOD, with crack length to specimen width ratio, a∕W ,for the clamped SE(T) specimens with varying H∕W -ratios, andthe variation of μSE(B)CMOD, with a∕W for the SE(B) specimen. In thisplot, the normalized compliance μCMOD is defined by [13,19]

μSE(T)CMOD =1

1+ EBe C (12)

and

μSE(B)=CMOD BeWEC S∕4 + 1−1 (13)

where

Be = B−B− BN

2

B. (14)

Figure 3 Variation of normalized compliance in termsof CMOD with increased aW-ratio for the clampedSE(T) fracture specimen.

0.1

0.2

0.3

00.40.30.20.10

a∕W

0.4

μCMOD

0.5 0.6 0.7 0.8

0.5

0.6

HPW=10

HPW=6

HPW=20

Figure 4 Variation of normalized compliance in termsof CMOD with increased aW-ratio for the 3P SE(B)fracture specimen.

0.2

0.3

0.1

0.40.30.20.10

a∕W

0.4

μCMOD

0.5 0.6 0.7 0.8

0.5

0

Ruggieri [20]

Tada et al [18]

ISO 12135 [22]

ASTM 1820 [13]

SPW=4

Using now the previous results for guidance, Cravero andRuggieri [5] and Ruggieri [20] provide the functional depen-dence of crack length and specimen compliance in the form

a∕W =5m=0

βm μ m (15)


where it is understood that a 5-th order polynomial fitting isemployed. For the clamped SE(T) specimen with H∕W=10, thepolynomial coefficients, βm, are given by β0=1.6485,β1=−9.1005 β2=33.0250, β3=−78.4670, β4=97.3440and β5=−47.2270. For the 3P SE(B) specimen with S∕W=4,Cravero [23] gives the polynomial coefficients, βm, given byβ0=0.7590, β1=1.2961 β2=−37.9076, β3=154.9671,β4=−270.3815 and β5=182.6632. The previous equationsdefine the key steps in the evaluation procedure of the crackgrowth resistance curve. By measuring the instantaneous com-pliance during unloading of the specimen (see Fig. 2), the currentcrack length follows directly from solving the above expressionfor μ.

Nondimensional Plastic Eta Factors

SE(T) Fracture Specimens

Based upon the plastic component of the area under the load vs.displacement curve (see Fig. 2), Ruggieri [15] provided im-proved nondimensional η-factors for the analyzed clampedSE(T) specimens with varying H∕W -ratios. Further, Paredesand Ruggieri [16] examined the effects of weld strength mis-match on fracture toughness measurements defined by J andCTOD fracture parameters using SE(T) specimens and arrivedat a set of nondimensional η-factors for this crack configurationas a function of the mismatch ratio, My . Figure 5 displays pro-

vide factors ηJ derived from CMOD (ηCMODJ ) and LLD (ηLLDJ )for the clamped SE(T) specimens with varying a∕W-ratios anddifferent mismatch levels for a groove size, 2h=15mm. In theseplots, the solid symbols correspond to the compute η-valueswhereas the lines represent fitting curves to the numerical data.

The results given by the plots of Fig. 5 allow expressingthe η-factors for the analyzed clamped SE(T) specimens withH∕W=10 in the form

η(a∕W)= C0+ C1(a∕W )+ C2 a∕W 2+ C3 a∕W

3(16)

where the coefficients Ck are function of the mismatch level, i.e.,Ck= f My, valid in the range 0.2 ≤ a∕W ≤ 0.7 and1.0 ≤ My≤ 1.5 as follows

S Coefficients Ck for ηCMODJ :

C0= 0.6960− 0.2411My+ 0.2888M2y

Figure 5 Variation of η-factor to determine J with increasedcrack size (aW-ratio) for the clamped SENT specimens withdifferent levels of strength mismatch: (a) CMOD; (b) LLD.

(b)

(a)

0.25

0.50

0.75

1.00

1.25

a∕W

0.40.30.20.1 0.5 0.70.6 0.80

H∕W= 10

ηCMODJ,C

My = 1.2My = 1.3My = 1.5

a∕W

0.40.30.20.1 0.5 0.70.6 0.80

ηLLDJ,C

0.50

0.75

1.00

1.25

1.50

0

0.25

My = 1.1

H∕W= 10

My = 1.2My = 1.3My = 1.5

My = 1.1

Base Plate

Base Plate

C1= 2.2141+ 2.7164My− 3.0284M2y

C2=− 3.3160− 12.1180My+ 9.1392M2y


C3= 0.8558+ 11.2970My− 7.6189M2y

S Coefficients Ck for ηLLDJ :

C0=− 0.8380+ 1.1698My− 0.8782M2y

C1= 10.4110− 2.4038My+ 2.7719M2y

C2=− 12.5970− 9.0348My+ 0.0200M2y

C3= 2.4767+ 13.2410My− 3.1169M2y

SE(B) Fracture Specimens

For the tested SE(B) fracture specimens, Donato and Ruggieri[21] provide a set plastic of η-factors for homogeneous materialswhich are displayed in Figs. 6. To facilitate manipulation of theprevious results to estimate J in fracture testing, a polynomial fit-ting of the functional dependence of the η -factors and on a∕Wfor varying hardening properties based upon a least squarescheme is given as follows.

ηLLDJ =− 17.260(a∕W )2+ 9.946(a∕W )+ 0.383

all n ; 0.05≤ a∕W≤ 0.3

(17)

ηLLDJ = 0.127(a∕W )+ 1.816

all n ; 0.3< a∕W≤ 0.7

(18)

ηCMODJ = 0.341(a∕W )2− 2.111(a∕W )+ 3.6496

all n ; 0.15≤ a∕W≤ 0.7

(19)

Crack Growth Resistance Results

This section now describes the evaluation of ductile tearingproperties for the tested X80 pipeline girth weld from laboratorymeasurements of load and CMOD for the clamped SE(T) speci-mens and the 3P bend SE(B) specimens with center notchedwelds. The clamped SE(T) specimens have a∕W=0.4 andH∕W=10 with thickness B=14.8 mm, widthW=14.8 mm andclamp distanceH=148 mm (refer to Fig. 1(a)). The SE(B) speci-mens have a∕W=0.25 with thickness B=14.8 mm, widthW=14.8 mm and span S=4W (refer to Fig. 1(b)). Figure 7shows the measured load-displacement curve (as described byCMOD) for both test specimens.

Figure 6 Variation of plastic ηJ derived from CMOD andLLD with increased aW-ratio and varying hardeningproperties for the 3P SE(B) fracture specimen.

2.5

3.0

3.5

2.00.40.30.20.10

a∕W

4.0ηCMODJ

0.5 0.6 0.7 0.8

n=5n=10n=20

n=5n=10n=200.75

1.00

1.25

0.500.40.30.20.10

a∕W

1.50

ηLLDJ

0.5 0.6 0.7 0.8

1.75

2.00

(b)

(a)

Evaluation of the crack growth resistance curve followsfrom determining J and Δa at each unloading point of the mea-sured load-displacement data. Based upon the previous resultsfor the η -factors displayed in the plots of Fig. 5 with My=1.1,the present analysis employs ηCMODJ and ηLLDJ to estimate theplastic component of the J-integral, Jp. Figures 8-NO TAGpresent the measured crack growth resistance curves for thetested pipelines girth weld. Figure 8 displays the J vs. acurves for for tested clamped SE(T) and 3P SE(B) specimensbased upon η-factors for homogeneous materials. Figure 9


Figure 7Measured load-CMOD curve for the tested X80 pipe-line girth weld using clamped SE(T) specimens with aW=0.4and 3P SE(B) specimens with aW=0.25.

0

20

40

60

80P (kN)

3.02.01.00

V (mm)

2.51.50.5

API X80 Girth Weld

shows the resistance curves for one of the tested clamped SE(T)specimens based upon η-factors for homogeneous materials andovermatched welds. The significant features associated withthese plots include: (1) The shallow-crack SE(B) specimen pro-vides an R-curve which, albeit slightly more conservative, ex-hibits levels of J-values which are somewhat comparable to theJ-values corresponding to the deeply-cracked SE(T) specimenat a fixed amount of crack growth, a; (2) Use of the η-factorsspecifically derived for overmatched materials in the estimationprocedure for Jp does provide a better description of the ductiletearing behavior for the material; here, use of the η-factors forhomogeneous materials (i.e., not taking into account the degreeof weld strength overmatch) leads to slightly nonconservative(higher) estimates of the R-curve (we should emphasize that thelarger the levels of weld strength mismatch the larger the degreeof nonconservativeness).

Table 2 shows a comparison of the predicted and estimatedcrack extension for the tested fracture specimens. For the SE(T)specimens, the accuracy is within 1,4 − 17,8% while for theSE(B) specimen, the accuracy is within 5,3%. These results indi-cate that the UC procedure provides reasonable estimates of thefinal crack length. Unfortunately, two of the three tested SE(B)specimens presented unexpected crack front geometry and werenot considered for the predictions, which motivates further ex-perimental investigation in order to better support these conclu-sions.

0 2.0 3.0 4.01.0

Δa (mm)

J (kJ∕m2)

400

1000

1200

0

API X80 Girth Weld

Figure 8 Experimental R−curvesfor tested clamped SE(T)and 3P SE(B) specimens based upon η-factors for homo-geneous materials (20_C).

200

600

800

−0.5 0.5 1.5 2.5 3.5

SE(T) Data

SE(B) Data

0 2.0 3.0 4.01.0

Δa (mm)

J (kJ∕m2)

400

1000

1200

0

API X80 Girth Weld

Figure 9 Experimental R−curvesfor tested clamped SE(T)specimen based upon η-factors for homogeneous materi-als and overmatched welds (20_C).

200

600

800

−0.5 0.5 1.5 2.5 3.5

SE(T) Base Plate

SE(T) 10% Overmatch


Table 2 Final crack length estimation based on UC proce-dure. Specimens whose crack front measurements were not vali-dated according to ASTM E1820 requirements were not consid-ered for predictions.

DISCUSSION AND CONCLUDING REMARKS

This work presents an investigation of the ductile tearing proper-ties for a girth weld made of an API 5L X80 pipeline steel usingexperimentally measured crack growth resistance curves((J-Δa curves). Testing of the pipeline girth welds utilized side-grooved, clamped SE (T) specimens and 3P bend SE(B) speci-mens with a weld centerline notch to determine the crack growthresistance curves based upon the unloading compliance (UC)method using a single specimen technique. This experimentalcharacterization provides additional toughness data which serveto evaluate crack growth resistance properties of pipeline girthwelds using SE (T) and SE(B) specimens with weld centerlinecracks. Additional work is in progress to further validate the useof shallow-crack SE(B) specimens as an alternative fracturespecimen to measure crack growth properties for pipeline girthwelds. Ongoing investigation also focuses on establishing robustcorrelations between J and CTOD for stationary and growingcracks in SE(T) and SE(B) fracture specimens.

Acknowledgments

This investigation is supported by the Brazilian Scientific Coun-cil for Research and Technology (CNPq) (Grant 304132/2009-8)and by Fundação de Amparo à Pesquisa do Estado de São Pau-

lo (FAPESP Grant 2009/54229-3). The authors acknowledgeTenaris-Confab and Lincoln Electric Brasil for providing sup-port for the experiments described in this work. The authorsalso acknowledge the encouragement and useful discussionsprovided by Sergio Kojima (Tenaris-Confab) and Antonio C.Souza (Lincoln Electric Brasil).

References

[1] Hutchinson, J. W., ‘‘Fundamentals of the Phenomenolog-ical Theory of Nonlinear Fracture Mechanics,’’ Journal ofApplied Mechanics, Vol. 50, pp. 1042-1051, 1983.

[2] Anderson, T. L. Fracture Mechanics: Fundaments andApplications - 3rd Edition, CRC Press, New York, 2005.

[3] Det Norsk Veritas. “Fracture Control for Pipeline Installa-tion Methods Introducing Cyclic Plastic Strain”. Recom-mended Practice DNV-RP-F108, 2006.

[4] Shen, G. and W.R. Tyson, ”Crack Size Evaluation UsingUnloading Compliance in Single−Specimen Single-Edge-Notched Tension Fracture Toughness Testing”, Journal ofTesting and Evaluation, Vol. 37 (4), paper ID JTE102368,2009.

[5] Cravero, S. and Ruggieri, C., “Estimation Procedure of J-Resistance Curves for SE(T) Fracture Specimens UsingUnloading Compliance”, Engineering Fracture Mechan-ics, Vol. 74, pp. 2735-2757, 2007.

[6] Nyhus, B., Polanco, M and Ørjasæter, O., ‘‘SENT Speci-mens as an Alternative to SENB Specimens for FractureMechanics Testing of Pipelines.” in ASME InternationalConference on Offshore Mechanics and Arctic Engineer-ing (OMAE 2003), American Society of Mechanical En-gineers, Cancun, Mexico, 2003.

[7] Cravero, S. and Ruggieri, C., “Correlation of Fracture Be-havior in High Pressure Pipelines with Axial Flaws UsingConstraint Designed Test Specimens - Part I: Plane-StrainAnalyses,” Engineering Fracture Mechanics, Vol. 72, pp.1344-1360, 2005.

[8] Silva, L. A. L., Cravero, S. and Ruggieri, C., “Correlationof Fracture Behavior in High Pressure Pipelines with Ax-ial Flaws Using Constraint Designed Test Specimens -Part II: 3-D Effects on Constraint,” Engineering FractureMechanics, Vol. 73, pp. 2123-2138, 2006.

[9] Shen, G., R. Bouchard, J. A. Gianetto and W. R. Tyson,”Fracture Toughness Evaluation of High-Strength SteelPipe”, inASME PVP 2008 Pressure Vessel and Piping Di-vision Conference (PVP 2008), American Society of Me-chanical Engineers, Chicago, IL, 2008.

[10] Park, D. Y., Tyson, W. R., Gianetto, J. A., Shen, G and Ea-gleson, R. S., “Evaluation of FRacture Toughness of X100Pipe Steel Using SE(B) and Clamped SE(T) Single Speci-mens,” in ASME International Pipeline Conference (IPC2010), American Society of Mechanical Engineers, Cal-gary, Canada, 2010.


[11] Fernandes, P. E. A., HAZ Notch Toughness of an API X80Steel Weldment Fabricated by SMAW and FCAWWeldingProcesses, Ph. D. Thesis, University of São Paulo, 2011(in Portuguese).

[12] American Petroleum Institute. “Fitness-for-Service”, APIRP-579-1 / ASME FFS-1, 2007.

[13] American Society for Testing and Materials, “StandardTest Method for Measurement of Fracture Toughness”,ASTM E1820, 2008.

[14] Cravero, S. and Ruggieri, C., “Further Developments inJ Evaluation Procedure for Growing Cracks Based onLLD and CMOD Data”, International Journal of Frac-ture, Vol. 148, pp. 387-400, 2008.

[15] Ruggieri, C., “Further Results in J and CTOD EstimationProcedures for SE(T) Fracture Specimens - Part I: Homo-geneous Materials,” Engineering Fracture Mechanics, (InPress), 2011.

[16] Paredes, M. and Ruggieri, C., “Further Results in J andCTOD Estimation Procedures for SE(T) Fracture Speci-mens - Part II: Weld Centerline Cracks,” EngineeringFracture Mechanics, (Submitted for Publication), 2012.

[17] Sumpter, J. D. G. and Turner, C. E., “Method for Laborato-ry Determination of Jc ,”Cracks and Fracture, ASTMSTP601, American Society for Testing and Materials, pp 3-18,1976.

[18] Tada, H, Paris, P. C. and Irwin, G. R., “The Stress Analysisof Cracks Handbook”, 3rd Edition, American Society ofMechanical Engineers, NY, 2000.

[19] Saxena, A., “Nonlinear Fracture Mechanics for Engi-neers”, CRC Press, Boca Raton, 1998.

[20] Ruggieri, C., “FRACTUS2D: Numerical Computation ofFracture Mechanics Parameters for 2-D Cracked Solids,”EPUSP, University of São Paulo, 2011.

[21] G. H. B. Donato and C. Ruggieri, “Estimation Procedurefor J and CTOD Fracture Parameters Using Three-PointBend Specimens” in 6th International Pipeline Confer-ence - IPC 2006, Anais CD-ROM, Calgary, Canadá, 2006.

[22] International Organization for Standardization, “MetallicMaterials − Unified Method of Test for the Determinationof Quasistatic Fracture Toughness”, ISO 12135, 2002.

[23] Cravero, S., Development of Evaluation Procedures forJ-R Curves of SE(T) Fracture Specimens Using the Un-loading Compliance Method, Ph. D. Thesis, University ofSão Paulo, 2007 (in Portuguese).

omae83056 girthweld final

Documents