Anda di halaman 1dari 11

Proceedings of the 2012 9th International Pipeline Conference IPC2012 September 24-28, 2012, Calgary, Alberta, Canada

IPC2012-90289

FRACTURE TOUGHNESS OF X100 PIPE GIRTH WELDS USING SE(T) AND SE(B) TESTS
Dong-Yeob Park CanmetMATERIALS Natural Resources Canada 3303 33 Street NW, Calgary, AB, Canada William R. Tyson CanmetMATERIALS Natural Resources Canada 555 Booth Street, Ottawa, ON, Canada

James A. Gianetto CanmetMATERIALS Natural Resources Canada 183 Longwood Road South, Hamilton, ON, Canada

Guowu Shen CanmetMATERIALS Natural Resources Canada 183 Longwood Road South, Hamilton, ON, Canada

Robert S. Eagleson CANMET Mining and Mineral Sciences Laboratories Natural Resources Canada 1 Haanel Drive - Building 10, Ottawa, ON, Canada

ABSTRACT Extensive single-edge notched tension (SE(T)) and singleedge notched bend (SE(B)) tests were performed to apply the SE(T) procedure developed at CANMET and the standard SE(B) procedure of ASTM E1820 to pipeline girth welds as a contribution to a broader project on strain-based design (SBD) for pipeline girth weld integrity. Specimens were precracked from the pipe inner surface to target lengths of 3 and 6 mm to represent surface-breaking weld flaws of single- and doublepass height, respectively. It was found that resistance curves for weld metal (WM) were much lower than those for base metal (BM) or heat affected zone (HAZ) specimens even though the WM strength overmatched the BM, owing to effects from the WM microstructure. The small crack-tip opening angle observed in post-test WM specimens was consistent with the low crack propagation resistance, which in turn results in small tearing resistance, TR. In general, toughness decreases with increase in initial crack length. J values and tearing resistance are found to be influenced by loading mode; toughness is higher in tension, SE(T), than in bending, SE(B) for a given crack length and test temperature. Key Words: toughness, facture resistance, low constraint, SE(T), SE(B), X100 pipe, girth welds, weld metal, heataffected zone

INTRODUCTION Onshore pipelines subject to geotechnical loads often contain flaws subject to low-constraint loading, e.g., crack-like circumferential defects in a pipe under tension. A highly constrained toughness test, such as ASTM E1820 (1), using single-edge notched bend, SE(B), and compact tension, C(T), specimens, will produce conservative toughness measurements. This can lead to costly material selection and pipeline design. As part of a large consolidated project on strain-based design for pipeline girth weld integrity, CanmetMATERIALS (formerly, CANMET Materials Technology Laboratory) developed a single-edged-notched-tension, SE(T), procedure and equations for J-integral calculation for toughness measurement applicable to pipe flaws under low-constraint conditions (2-4). The proposed CANMET SE(T) procedure uses the crack-mouth-opening displacement (CMOD) unloading-compliance technique from a clamped SE(T) specimen. The procedure has been successfully applied to measure toughness of X100 pipe steel, reported elsewhere (3, 5). The present work extends the procedure to girth welds and provides data for pipe girth weld integrity assessment and strain-based design. Extensive SE(T) and SE(B) tests of X100 pipe single-and dual-torch girth welds were performed according to the recommended SE(T) procedure and to ASTM E1820 (1), respectively. This paper reports resistance curves obtained from these tests.

Copyright 2012 by Her Majesty the Queen in Right of Canada

MATERIAL, SPECIMEN PREPARATION AND TESTING Material Properties and Specimen Preparation Relevant material property data for the X100 pipe steel and a description of the welding procedures employed to produce the experimental X100 mechanized rolled welds and their corresponding mechanical properties are available elsewhere (6). Tensile properties are presented in Table 1. Low temperature tensile properties were estimated using the conversion equation in BS 7448: Part 2, and weld metal properties were used for the HAZ (7). The following information relates to the preparation of fracture test specimens from the X100 pipe steel and the experimental single and dual torch welds. Pipe material: pipes of X100 (Grade 690), 36 (914 mm) diameter by 0.750 (19.1 mm) wall thickness. Experimental pipeline welds: single torch (round 1 (R1)) and dual torch (round 2 (R2)) rolled girth welds. Figure 1 shows base metal (BM), heat-affected zone (HAZ) and weld metal (WM). Specimen cross-section: 17.5 mm 17.5 mm for base metal and dual torch welds, weld metal (WM) centerline and heat-affected zone (HAZ). SE(T) specimen length: 345 mm with length between grips (daylight) of H=10W=172 mm. Specimen length direction: longitudinal to pipe axis (transverse to girth weld). Side-groove details: 45 included angle with 0.5 mm root radius to total depth 15% (7.5% on each side) of thickness B. Targeted fatigue precrack lengths: 3 and 6 mm, chosen to represent surface-breaking weld flaws of single- and double-pass heights, respectively (Fig. 2). Notch and fatigue precrack: circumferential crack from inner diameter surface of the pipe. Test temperatures: room temperature (RT), -20C and -40C. Local compression: None.
Table 1. Yield and ultimate tensile strengths of X100 pipe steel and weld metal, measured at room temperature Material Base metal (Rounds 1 (R1) and 2 (R2)) Single-torch weld metal (R1) Dual-torch weld metal (R2) Notes: YS (0.2 %) MPa 720 825 827 UTS MPa 830 915 889

X100 pipe steel tested using strap tensile specimen parallel to pipe axis All-weld-metal strip specimen cut from girth weld

5 mm

5 mm

(a)

(b)

Figure 1. X100 rolled welds: (a) single torch and (b) dual torch.

(a)

(b)

Figure 2. Precrack locations: (a) weld metal centerline and (b) HAZ region (solid line: 3 mm precrack and dotted line: 6 mm precrack).

YS: yield strength, UTS: ultimate tensile strength

SE(T) and SE(B) Tests For a given material and initial crack length (3 and 6 mm), testing was repeated at -20oC and room temperature to examine the reproducibility, except for base metal tested at 20oC for which only one specimen was tested at each crack length. Two specimens each of single torch WM and HAZ were tested at -40oC. A total of 49 SE(T) tests were conducted. For low-temperature testing, a customized liquid-nitrogen environmental chamber was placed between the grips. Figure 3 shows SE(T) test setups at room temperature and low temperatures. The temperature within the distance of twice the specimen thickness from the notch was stable to 2oC before testing. However, temperature fluctuated up to 5oC, especially at -40oC, during testing. A total of 24 SE(B) tests were performed, following the procedure of ASTM E1820 (1). Low-temperature SE(B) testing was conducted in an environmental chamber, and the temperature was maintained within 1oC.

Copyright 2012 by Her Majesty the Queen in Right of Canada

regression curves, using data between 0.15 mm and 1.5 mm exclusion lines according to the procedure of ASTM E1820 (e.g., Fig. 5): (1) J or CTOD C 1 a C 2 where J and CTOD are the J-integral and CTOD converted from the J-integral, a is crack extension and C1 and C2 are regression constants.
2400

aoq = 4 mm, RT
2000

1600

J (kJ/m )

1200

(a)

(b)

Figure 3. Test setup: (a) room temperature and (b) low temperature

800

400

EXPERIMENTAL RESULTS AND DISCUSSION


0 0 0.5 1 1.5 2 2.5

J-Resistance Curves Resistance curves were evaluated according to the standard J-integral calculation procedure of ASTM E1820-11 (1) for SE(B) specimens and to the CANMET procedure for SE(T) specimens (2-4), using CMOD unloading compliance to measure crack length in both cases. Figure 4 shows examples of load (P)-CMOD curves with periodic unloadings recorded during SE(T) testing of BM, WM and HAZ specimens. The slope of the unloading line and peak load both decrease with increasing crack size (6.0, 6.4, and 7.2 mm respectively for BM, HAZ and WM).
180

Da (mm)

(a)
2400

aoq = 6.5 mm, RT


2000

1600

J (kJ/m )

1200

800

Base metal
160

400

HAZ
140 120
0 0 0.5 1 1.5 2 2.5

Da (mm)

Load (kN)

100 80

(b)

60 40

Weld metal centerline

Figure 5. Data points and fitted resistance curves: SE(T) single torch (R1) (a) weld metal and (b) HAZ. The solid red symbol (near a = 0.5 mm) corresponds to the maximum force attained during the test.
Base metal (ao=6mm) HAZ (ao=6.4mm) Weld metal (ao=7.2mm)

20 0 0.0 0.2 0.4 0.6 0.8

1.0

1.2

1.4

1.6

CMOD (mm)

Figure 4. Load-CMOD curves for SE(T) specimens

The J- and CTOD-resistances computed from these experimental measurements were curve-fitted to power-law

Target initial crack lengths were 3 and 6 mm, but actual lengths were a little larger. To extrapolate/interpolate a small amount to the target initial crack values, the curve-fit parameters for the actual crack sizes (aoq) were in turn fitted to an assumed linear dependence on initial crack length (e.g., Fig. 6). Tables 2 and 3 report power-law fitting parameters (C1 and C2) extrapolated/interpolated to initial crack lengths ao = 3 and 6 mm for SE(T) and SE(B) J-resistance curves, respectively. Figures 7 and 8 show J-resistance curves

Copyright 2012 by Her Majesty the Queen in Right of Canada

calculated using these power-law parameters for initial crack lengths of 3 and 6 mm. (Owing to lack of data, fits of the
parameters to aoq were not attempted for base metal and HAZ (R2)). R1 and R2 represent single- and dual-torch welds, respectively.

In general, the J-resistance curves of the WM (single and dual torch) lie well below those of base metal and HAZ regions for both SE(T) and SE(B) geometries. For instance, J of the BM is higher than single torch (R1) WM by a factor of two at crack growth (a) of 1 mm for SE(T) specimens of ao = 6 mm at room temperature. This lower toughness of the WM is in part addressed by the microstructure observation near the crack tip in the subsequent section. The R curves of the HAZ are also lower than the BM curves (except for the SE(T) tests with ao = 3 mm tested at room temperature) as a result of microstructural change caused by thermal cycling.

Table 2. Power-law fitting parameters (C1 and C2) extrapolated/interpolated to the target initial crack lengths (ao = 3 and 6 mm) by linear fits to the compliance-estimated initial crack lengths (aoq) for SE(T) J-resistance curves
Target ao (mm) Temp. (oC) Material BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM 3 512 0.930 WM (R1) 616 0.740 from an individual specimen (a\oq = 3.3 mm) from an individual specimen (aoq = 3.5 mm) from an individual specimen (aoq = 3.6 mm) from an individual specimen (aoq = 2.6 mm) C1 (kJ/m2) 1304 674 1424 641 1157 1500 622 1308 505 1147 C2 0.830 0.590 0.671 0.653 0.862 0.857 0.745 0.741 0.886 0.821 Remarks

RT

1350 SE(T), BM, Room temp

-20

1300 Individual C1 Linear (Individual C1) 1250

C1 (kJ/m )

1200

1150 y = -46.743x + 1444.2 R2 = 0.8342 1100

-40 1400 HAZ (R1) 1310 0.878 0.668

1050 0 1 2 3 4 5 6 7 8

aoq (mm)

(a)
0.9 SE(T), BM, Room temp 0.8

RT

0.7

6
y = -0.07x + 1.0404 R2 = 0.916

0.6

0.5

-20

WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2)

1164 564 1132 647 948 1353 505 1006 648 985

0.620 0.710 0.560 0.784 0.604 0.520 0.763 0.712 0.793 0.830

C2
0.4 0.3 Individual C2 Linear (Individual C2) 0.2 0.1 0.0 0 1 2 3 4 5 6 7 8

aoq (mm)

(b)

Figure 6. Power-law fitting parameters (C1 and C2), fitted to an assumed linear dependence on the initial crack length aoq to enable small interpolation/extrapolation to the target crack size.

Copyright 2012 by Her Majesty the Queen in Right of Canada

4000 SE(T), -20C, a=3mm 3500 BM WM (R1) HAZ (R1) WM (R2) HAZ (R2)

Table 3. Power-law fitting parameters (C1 and C2) extrapolated/interpolated to the target initial crack lengths (ao = 3 and 6 mm) by linear fits to aoq for SE(B) J-resistance curves
Target ao (mm) Temp. (oC) Material BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM -20 WM (R1) HAZ (R1) WM (R2) HAZ (R2) 486 1208 626 0.532 0.691 0.325 C1 (kJ/m2) 555 979 416
J, kJ/m2

3000

2500

C2

Remarks

2000

1500

RT

0.632 0.685 0.605

1000

500

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(b)
4000 SE(T), RT, a=6mm BM WM (R1) HAZ (R1) WM (R2) HAZ (R2)

1121 BM 1124 RT WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM WM (R1) -20 HAZ (R1) WM (R2) HAZ (R2) 406 936 591

0.678

0.729 0.575 0.712 0.629

from an individual specimen (aoq = 5.4 mm) from an individual specimen (aoq = 5.6 mm)
J, kJ/m
2

3500

3000

2500

2000

1500

1000

382 979 594

0.625 0.760 0.625


500

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(c)
4000 SE(T), -20C, a=6mm 3500 BM WM (R1) HAZ (R1) WM (R2) HAZ (R2)

4000 SE(T), RT, a=3mm 3500 BM WM (R1) HAZ (R1) HAZ (R2)

3000

2500

J, kJ/m

WM (R2) 3000

2000

2500

1500

J, kJ/m

2000

1000

1500

500

1000

0 0 0.5 1 1.5 2 2.5 3

500

Crack growth, mm

(d)
0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(a)

Figure 7. SE(T) J-resistance curves for crack lengths 3 and 6 mm calculated using the extrapolated/interpolated C1 and C2 power-law parameters fit to a linear dependence on aoq: (a) RT, ao= 3 mm, (b) -20oC, ao= 3 mm, (c) RT, ao= 6 mm and (d) -20oC, ao= 6 mm.

Copyright 2012 by Her Majesty the Queen in Right of Canada

4000 SE(B), -20C, a=6mm

4000 SE(B), RT, a=3mm WM (R1) 3500 HAZ (R1) WM (R2) 3000
2500 3000 3500

WM (R1) HAZ (R1) WM (R2)

2500

J, kJ/m2

J, kJ/m

2000

2000
1500

1500
1000

1000
500

500
0 0 0.5 1 1.5 2 2.5 3

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

Crack growth, mm

(d)

(a)
4000 SE(B), -20C, a=3mm WM (R1) 3500 HAZ (R1) WM (R2) 3000

Figure 8. SE(B) J-resistance curves for crack lengths 3 and 6 mm calculated using the extrapolated/interpolated C1 and C2 power-law parameters fit to a linear dependence on aoq: (a) RT, ao= 3 mm, (b) -20oC, ao= 3 mm, (c) RT, ao= 6 mm and (d) -20oC, ao= 6 mm.

2500

2000

1500

1000

500

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(b)
4000 SE(B), RT, a=6mm WM (R1) 3500 HAZ (R1) WM (R2) 3000

2500

J, kJ/m2

CTOD-Resistance Curves CTOD values were calculated from J and power-law fitting parameters (C1 and C2) were obtained for each specimen. Then, as for J resistance curve data, linear fits of the C1 and C2 values to aoq were obtained. The linear fits were extrapolated/interpolated to initial crack lengths ao= 3 and 6 mm; the results are summarized in Tables 4 and 5. CTODresistance curves calculated using these C1 and C2 parameters are presented in Figure 9. The general trends are similar to those of J-resistance curves. The major difference between CTOD- and J- resistance curves is that for CTOD curves, the relative position of the BM curve is much higher than HAZ and WM; for example, for SE(T) specimens with ao = 6 mm tested at room temperature the ratio of CTOD values at crack growth of 1 mm of BM to single torch WM is 50% higher than the corresponding ratio of J values. This is a result of differences in the yield strengths used for conversion from J to CTOD for both SE(T) and SE(B) (1, 2): J (2) CTOD
m Y

J, kJ/m

2000

1500

1000

500

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(c)

where m is a function of work-hardening coefficient (n) and ratio of crack length to width (a/W), and Y is the average of the 0.2 % offset yield strength and the ultimate tensile strength. The WM yield strength is higher than that of the BM (overmatching), and this causes the WM CTOD resistance curves to be depressed even further below those of the BM, relative to the J-R curves, according to Eq. (2). The WM yield strength is used in J-to-CTOD conversion for the HAZ so that the HAZ CTOD resistance curves are also relatively lowered below those of the BM for the same reason.

Copyright 2012 by Her Majesty the Queen in Right of Canada

Table 4. Power-law fitting parameters (C1 and C2) extrapolated/interpolated to the target initial crack lengths (ao= 3 and 6 mm) using linear fits to aoq for SE(T) CTOD-resistance curves
Target ao (mm) Temp. (oC) Material BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM C1 (mm) 1.315 0.510 1.123 0.509 0.912 1.414 0.460 0.995 0.372 0.870 C2 0.819 0.590 0.660 0.661 0.853 0.847 0.756 0.750 0.912 0.818 from an individual specimen (aoq = 3.3 mm) from an individual specimen (aoq = 3.5 mm) from an individual specimen (aoq = 3.6 mm) from an individual specimen (aoq = 2.6 mm) Remarks

Table 5. Power-law fitting parameters (C1 and C2) extrapolated/interpolated to the target initial crack lengths (ao= 3 and 6 mm) by linear fits to aoq for SE(B) CTOD-resistance curves
Target ao (mm) Temp. (oC) Material BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) C1 (mm) 0.407 0.720 0.309 C2 Remarks

RT 3 -20

RT

0.611 0.659 0.584

0.355 0.891 0.459

0.509 0.673 0.298 from an individual specimen (aoq = 5.4 mm) from an individual specimen (aoq = 5.6 mm)

-20

0.855 BM

0.658

3 0.361 WM (R1) 0.441 -40 1.026 HAZ (R1) 0.971 WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) 0.904 0.673 0.758 0.960

RT 6 WM (R1) HAZ (R1) WM (R2) HAZ (R2) BM WM (R1) HAZ (R1) WM (R2) HAZ (R2)

0.850 0.280 0.641 0.404

0.708 0.556 0.688 0.606

-20

0.261 0.673 0.406

0.605 0.742 0.603

RT

-20

CTOD, mm

1.200 0.434 0.869 0.492 0.732 1.331 0.356 0.736 0.476 0.735

0.632 0.756 0.561 0.822 0.607 0.515 0.823 0.733 0.826 0.844

3.5 SE(T), RT, a=3mm 3 BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) 2.5

1.5

0.5

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(a) Fig. 9. (caption below)

Copyright 2012 by Her Majesty the Queen in Right of Canada

3.5 SE(T), -20C, a=3mm 3 BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) 2.5

Effects of Constraint and Microstructure Tearing resistance (or tearing modulus), TR, is used for a measure of the relative crack growth stability and determined by the slope of the J-R curve (8):
TR E dJ 2 da Y

CTOD, mm

(3)
a 1mm

1.5

0.5

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(b)
3.5 SE(T), RT, a=6mm 3 BM WM (R1) HAZ (R1) WM (R2) HAZ (R2) 2.5

1.5

0.5

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(c)
3.5 SE(T), -20C, a=6mm 3 BM WM (R1) HAZ (R1) WM (R2) HAZ (R2)

2.5

1.5

0.5

0 0 0.5 1 1.5 2 2.5 3

Crack growth, mm

(d)

Figure 9. SE(T) CTOD-resistance curves extrapolated/interpolated to crack lengths 3 and 6 mm using the C1 and C2 power-law parameters fit to a linear dependence on aoq: (a) RT, ao= 3 mm, (b) -20oC, ao= 3 mm, (c) RT, ao= 6 mm and (d) -20oC, ao= 6 mm. Owing to lack of data, fits of the parameters as a function of aoq were not attempted for base metal and HAZ (R2).

where E is Youngs modulus and Y is the average of the yield and ultimate strengths. J values and TR were calculated using the C1 and C2 power-law parameters fit to a linear dependence on aoq. It is seen in Fig. 10 that TR and small-growth J values (at a = 1.0 mm) generally depend on the initial crack length for both SE(B) and SE(T) specimens excluding dual torch (R2) weld metal. A similar tendency is observed for J values at a = 0.5 mm. Toughness increases with decrease in initial crack length as anticipated, owing to lower constraint for shallow cracks. However, for the dual torch weld metal, the initial J values and TR are apparently insensitive to the initial crack length. Joyce and Link (9) also reported negligible effect of constraint and initial crack length on JIc for HY80 steel. The cause of the insensitivity, however, is not clear. Because toughness is controlled by microstructure and inclusions, differences in these factors for the weld metal and base metal may cause the observed differences in sensitivity. Initial J values and tearing resistances are influenced by loading mode as well as initial crack length (Fig. 11). Initial J values and TR measured from the low constraint SE(T) test are higher than the high constraint SE(B) test by as much as 70%, except for TR of the single-torch (R1) HAZ. This indicates that the low constraint of tension testing increases resistance to both crack initiation and crack growth. WM resistance curves are much lower than BM and HAZ, as discussed earlier (Figs. 7-9). This can be ascribed to the weld metal microstructure (10-12). Figure 12 shows numerous large voids near the crack tip in WM but relatively fewer voids in the HAZ; large voids are marked by v in the figures. Voids can initiate in the process zone at (large) inclusions with relatively little plastic deformation. Weld metals generally contain a significant density of inclusions (10), but further examination of the WM microstructure is required to quantify the nature and density of the inclusions. The voids shown in Fig. 12 may contribute significantly to the observed decrease in WM toughness. Another characteristic observed in Fig. 12 is that the cracktip opening angle (CTOA) in the WM is only one-quarter of that in the HAZ. The CTOAs measured at 75 m behind the crack tip are 6o and 24o for WM and HAZ, respectively. This suggests a low crack propagation resistance of the WM, again possibly owing to the void content and the microstructure. The low crack propagation resistance in turn results in a low slope of the resistance curve as seen in Figs. 7-9. TR of the single torch HAZ (ao = 3 mm), for instance, is higher than that of the

CTOD, mm

CTOD, mm

Copyright 2012 by Her Majesty the Queen in Right of Canada

single torch WM by a factor of approximately two for both SE(T) and SE(B) specimens (Fig. 10 (a)).

TR
ao = 3 mm ao = 6 mm

Line of equality
SE(T)/SE(B)
WM(R1) - SE(T) WM(R1) - SE(B) HAZ(R1) - SE(T) HAZ(R1) - SE(B) WM(R2) - SE(T) WM(R2) - SE(B)

300

250

200

TR

150

WM (R1)

HAZ (R1)

WM (R2)

WM (R1)

HAZ (R1)

WM (R2)

BM

100
2

(a)
JDa=1mm
ao = 3 mm ao = 6 mm

50

0 0 1 2 3 4 5 6 7
SE(T)/SE(B)

Line of equality

ao (mm)

(a)
1600

1400

WM(R1) - SE(T)
1200

WM(R1) - SE(B) HAZ(R1) - SE(T)

1000

J (kJ/mm )

HAZ(R1) - SE(B)
800

WM (R1)

HAZ (R1)

WM (R2)

WM (R1)

HAZ (R1)

WM (R2)

BM

WM(R2) - SE(T) WM(R2) - SE(B)

(b)
Figure 11. Ratio of SE(T) and SE(B) results for (a) TR and (b) J values at a = 1mm

600

400

200

0 0 1 2 3 4 5 6 7

ao (mm)

(b)

Figure 10. Effects of initial crack length, ao, on (a) TR and (b) J values at a = 1 mm for SE(T) and SE(B) specimens tested at room temperature. R1 and R2 represent single- and dual-torch welds,

respectively.

In an overmatched weld notched in the HAZ, the crack grows toward the lower-strength material (softened HAZ and BM). This is partly dependent on notch location since shallow cracks near the hot pass bevel tend to follow the HAZ profile, whereas those higher up in the joint tend to move towards the base metal region for both SE(T) and SE(B). Figure 12(c) clearly shows the biased crack growth direction as well as the severe deformation on the base metal side. The measured crack extensions on the BM and WM sides are 1.96 mm and 2.06 mm respectively, reflecting the extensive plastic deformation in the BM. The crack tip approaches the boundary between HAZ and BM in the figure. If it were to grow further, into the BM, the toughness will reflect BM properties rather than those of the HAZ. However, the resistance curve is relevant for engineering critical assessment (ECA), since it reflects the actual growth of a crack initiating in the HAZ.

Copyright 2012 by Her Majesty the Queen in Right of Canada

Change in temperature over the range tested did not significantly affect toughness of the base metal, HAZ or weld metal. J and CTOD values at room temperature and -20C lie in the same scatter band for a given region. Note that resistance curves in Figs. 7-9 were constructed from power-law fitting parameters extrapolated/interpolated to the target initial crack lengths, based on multiple tests repeated at a given specimen geometry. The scatter in C1 and C2 values was not large (see Fig. 6).
(a)

CONCLUSIONS Extensive SE(T) and SE(B) tests were conducted to apply the SE(T) procedure developed at CANMET and the standard SE(B) procedure of ASTM E1820 to high-strength pipeline single- and dual-torch girth welds. Specimens were precracked on the pipe inner surface to target crack depths of 3 and 6 mm to represent surface-breaking circumferential weld flaws of single- or double-pass height. J- and CTOD-resistance curves were derived using the CANMET and ASTM standard procedures for SE(T) and SE(B) tests, respectively, and posttest examinations of specimens were conducted. The following results and conclusions were obtained: 1. J values and tearing resistance, TR, are generally dependent on the initial crack length. For most cases, toughness increases with decrease in initial crack length. 2. Initial J values and tearing resistance are also sensitive to loading mode, i.e., constraint. Both values are higher in tension, SE(T), than in bending, SE(B), at a given crack length and temperature. 3. The WM strength overmatched the BM, as required by pipeline codes. However, resistance curves for WM were lower than those for BM and HAZ specimens. This is probably a result of a higher density of void-forming particles in the WM microstructure, although this was not investigated in detail. The small crack-tip opening angle observed in post-test WM specimens was consistent with the low crack propagation resistance slope which is directly related to the tearing resistance.

(b)

(c)

(d) Figure 12. Optical microscope images of etched cross sections of SE(T) specimens: weld metal centerline specimen ((a) and (b)) and HAZ specimen ((c) and (d)). Figures (b) and (d) show enlarged images near the crack tip of (a) and (c), respectively, and large voids are marked by V in the figures.

ACKNOWLEDGMENTS This work is part of a larger consolidated program jointly funded by the Canadian federal government Program of Energy Research and Development (PERD), the Pipeline and Hazardous Materials Safety Administration of the U.S. Department of Transportation (DOT), and the Pipeline Research Council International, Inc. (PRCI).

10

Copyright 2012 by Her Majesty the Queen in Right of Canada

DISCLAIMER The views and conclusions in this paper are those of the authors and should not be interpreted as representing the official policies of any of the sponsor organizations.

433-458

REFERENCES 1. ASTM E 1820-11. Standard Test Method for Measurement of Fracture Toughness, ASTM International, West Conshohocken, PA, U.S.A, 2011 2. Shen, G., Gianetto J.A., and Tyson, W.R. Development of Procedure for Low-Constraint Toughness Testing Using a Single-Specimen Technique, MTL Report 2008-18(TR) 3. Shen, G., Gianetto, J.A. and Tyson, W.R. Measurement of J-R curves using single-specimen technique on clamped SE(T) specimens, Proceedings of the 19th ISOPE Conference, Osaka, Japan, June 21-26, 2009 4. Shen, G. and Tyson, W.R. Crack length evaluation for SE(T) testing using unloading compliance, Journal of Testing and Evaluation, 2009; 37: 347-357 5. Park, D.-Y., W.R. Tyson, G. Shen, J.A. Gianetto, and R.S. Eagleson. Evaluation of fracture toughness of X100 steel using SE(B) and clamped SE(T) single specimens, the 8th International Pipeline Conference, IPC2010, Calgary, Canada, Sept 2010. 6. Gianetto, J., Tyson, W.R., Park, D.Y., Shen, G., Lucon, E., Weeks, T.S., Quintana, M.A., Rajan, V.B. and Y.-Y. Wang, Small Scale Tensile, Charpy V-Notch, and Fracture Toughness Tests, Final Report 277-T-05 to PHMSA per Agreement # DTPH56-07-T-000005, September 2011 7. British Standard BS 7448, Fracture Mechanics Toughness Tests, British Standards Institution, London, England, 1997 8. Paris, P.C., Tada, H., Zahoor, A. and Ernst, H., The theory of instability of the tearing mode of elastic-plastic crack growth, Elastic-Plastic Fracture, ASTM STP 668, ed. J. D. Landes, J.A. Begley and G. A. Clarke. American Society for Testing and Materials, Philadelphia, 1979, 5-36. 9. Joyce, J. A. and Link, R. E. Application of two parameter elastic-plastic fracture mechanics to analysis of structures, Eng. Fract. Mech., 1997, 57, 431-446 10. Wei, S, Lu, S, Li, D and Li, Y. Impact property analysis of weld metal and heat-affected zone for low-alloy carbon steel multi-pass welded joint, China Welding (English Edition), 2010, 19, 21-25 11. Neves, J and Loureiro, A. Fracture toughness of welds effect of brittle zones and strength mismatch, Journal of Materials Processing Technology, 2004, 153154, 537 543 12. Boothman, D.P., Lee, M.M.K and Luxmoore, A.R. The effects of weld mismatch on J-integrals and Q-values for semi-elliptical surface flaws, Eng. Fract. Mech., 1999, 64,

11

Copyright 2012 by Her Majesty the Queen in Right of Canada

Anda mungkin juga menyukai