JID: JPSE

ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

Journal of Pipeline Science and Engineering xxx (xxxx) xxx

Contents lists available at ScienceDirect

Journal of Pipeline Science and Engineering

journal homepage:
http://www.keaipublishing.com/en/journals/journal-of-pipeline-science-and-engineering/

Analysis of fracture energy in drop weight tear testing of API X65 gas
pipeline steel
E. Fathi, S.H. Hashemi∗
Department of Mechanical Engineering, University of Birjand, POBOX 97175-376, Birjand, Iran

a r t i c l e i n f o a b s t r a c t

Keywords: In this research, the effect of low velocity impact (LVI) on different components of fracture energy of API X65
Dynamic pre-loading gas pipeline steel is investigated using drop weight tear test (DWTT) specimens. For this purpose, twenty one
Low velocity impact (LVI) test specimens (arranged in seven similar test groups) were used. The LVI tests were carried out first on six test
Partitioning of fracture energy
groups by releasing the machine hammer from different heights from 0.05 m to 0.3 m. Then, full fracture DWTTs
API X65 steel
were conducted on all seven test groups using 2 m drop height. The fracture energy of tested specimens and its
Drop weight tear test (DWTT)
different components were calculated from test data using an instrumented hammer. These were compared to
similar energies of virgin test specimens. It was found that by increasing LVI energy, yield energy increased up
to 26% due to work hardening. However, crack initiation energy decreased significantly (down to 48%) due to
plastic deformation at crack tip and crack tip blunting. By considering energy consumption, a novel test technique
was presented for exact determination of crack initiation point.

1. Introduction Xu et al., 2010; Zhen et al., 2020). The high dimensions of DWTT spec-
imen and its relatively long ligament (71 mm) as compared to smaller
The safety of high-pressure long-distance gas transportation test samples such as Charpy impact specimen, provide the conditions of
pipelines has been the subject of extensive study for more than stable crack growth (Hwang et al., 2004; Fang et al., 2014; Zheng et al.,
seven decades (Maxey, 1974; Eiber et al., 1993; Leis et al., 1998; 2008; Kim et al., 2021; Hong et al., 2011) during fracture experiment.
O’Donoghue et al., 1997). These pipelines are often designed to op- Moreover, test samples are cut from full thickness steel pipe which al-
erate at low ambient temperatures and elevated working pressures up lows realistic fracture test conditions.
to 15 MPa. This is correspondent to high hoop stresses, equivalent to The second author and his co-workers have been working for more
80% (and even more) of the specified minimum yield strength (SMYS) than a decade on different aspects of this invaluable test specimen us-
of pipeline materials. Different elastic and elastic-plastic fracture cri- ing a home-made testing machine which has led to some interesting re-
teria, including stress intensity factor, energy release rate, crack tip sults (Majidi-Jirandehi and Hashemi, 2019, 2017; Fathi-Asgarabad and
opening displacement (CTOD), crack tip opening angle (CTOA), and J- Hashemi, 2020a,b,c; Tazimi et al., 2020a,b; Khosravi Khor et al., 2020).
integral have been proposed to characterise the ductile fracture in these In this research, the effect of dynamic pre-loading on fracture behaviour
structures (Zhu and Joyce, 2012; Cotterell, 2002; Newman et al., 2003; of DWTT specimen is studied experimentally. It should be noticed that
Horsley, 2003; Zhu and Leis, 2006). Similarly, different fracture test static and dynamic pre-loads from impact of pipeline structure during
specimens have been suggested and used in laboratory scale to measure transportation, welding and installation of pipe segments or from earth-
such fracture criteria. The drop weight tear test (DWTT) specimen is quake, landslide or excavators can cause plastic deformation and perma-
widely used in oil and gas pipeline industries. Different test standards nent pre-strains. This in turn can change material properties such as its
(API, 1996; BSI, 1999; ASTM, 2014; ASTM, 2016) have been introduced yield strength, fracture energy and fatigue life. In current research, the
for this test sample from which ductile and brittle fracture appearances dynamic pre-loads were introduced to stationary test samples (placed
can be easily determined. Using various instrumentations such as load- on machine anvils) by means of LVI using a heavy drop hammer of 700
cell, accelerometer and laser velocity sensor, full fracture data can be kg. Using different drop heights between 0.05 m to 0.3 m, plasticity and
obtained from this important experiment (Cosham et al., 2010; Rudland crack tip blunting occurred at the vicinity of crack tip. Following this,
et al., 2003; Salvini et al., 2003; Yu and Ru, 2016; Demofonti et al., 1995; the pre-loaded test samples were subjected to second blow from full
Fang et al., 2014; Xu et al., 2016; Strnadel et al., 2013; Xu et al., 2007; height drop of 2 m. The total fracture energy of tested specimens was
measured from test data using an instrumented hammer. This fracture

∗
Corresponding author.
E-mail address: [email protected] (S.H. Hashemi).

https://doi.org/10.1016/j.jpse.2021.03.001
Received 27 December 2020; Received in revised form 15 February 2021; Accepted 6 March 2021
Available online xxx
2667-1433/© 2021 The Authors. Published by Elsevier B.V. on behalf of PetroChina Pipeline Company. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Please cite this article as: E. Fathi and S.H. Hashemi, Analysis of fracture energy in drop weight tear testing of API X65 gas pipeline steel, Journal
of Pipeline Science and Engineering, https://doi.org/10.1016/j.jpse.2021.03.001
JID: JPSE
ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

E. Fathi and S.H. Hashemi Journal of Pipeline Science and Engineering xxx (xxxx) xxx

Table 1
Geometry and design characteristics of API X65 pipe (Hashemi and Mohammadyani, 2012).

Manufacturing Welding a. Seam Angle Dimensions

API Grade Process Technique (°) Outside Wall Thickness
Diameter (mm) (mm) D/t

X65 TMCR Spiral SAW 35 1,219 14.3 85

Table 2
Measured mechanical properties of API X65 pipe in transverse direction and target values speci-
fied by API 5L.

API grade X65 Tensile properties

Yield strength (MPa) Tensile strength (MPa) Y/T Elongation (in 2 inches)

490 552 0.89 21

minimum 448 531 – –
maximum 600 758 0.93 –

Fig. 2. Micrographs demonstrating microstructure of API X65 steel.

Fig. 1. Nominal and true stress-strain plots of API X65 steel.

energy then was partitioned into different components such as yield en-
ergy, crack initiation energy and crack propagation energy. The compar-
ison of these energy components with those from virgin test specimens
are discussed in details in this paper. Moreover, a novel test technique
is presented for the determination of crack initiation point in specimen
ligament.

2. Characteristics of test material Fig. 3. Dimensions of DWTT specimen (API, 2007).

The API X65 steel, together with API X70 are the most com-
monly used pipeline materials in Iran high-pressure gas transportation The excellent mechanical properties of API X65 steel can be at-
pipelines and networks. The typical working pressure of these pipelines tribiuted to fine grianed microstrcuture (bainite-acicular ferrite), pro-
is around 750–1,000 psi (equivalent to 5–7 MPa), resulting in 220– duced via TMCR process (Hashemi and Mohammadyani, 2012), shown
400 MPa hoop stress in API X65 pipeline body. in Fig. 2.
The API X65 pipe under consideration was formed by spiral weld- The average ferrite grain size of this steel was around 3–7 𝜇m. The
ing technique. The original coil used for pipe manufacture was pro- improved cooling rate and reduced cooling interrupt temperature led to
duced by thermo-mechanical control-rolled (TMCR) process. The API the complete change of the final microstructure of this steel from tradi-
grade X65 gas pipe had 1,219 mm outside diameter (OD) and 14.3 mm tional ferrite-pearlite structure to bainite-acicular ferrite microstructure
wall thickness (WT), as given in Table 1. Its average chemical compo- (Hashemi and Mohammadyani, 2012; Hashemi, 2011, 2009).
sition (wt.%) was 0.071 C, 0.209 Si, 1.515 Mn, 0.018 Cu, 0.011 Ni,
0.003Mo, 0.044 Nb, 0.042 V, 0.017 Ti, and 0.157 Cr (Hashemi and Mo- 3. DWTT specimen and experimental set up
hammadyani, 2012).
In Fig. 1, nominal (engineering) and true stress-strain plots for tested The geometry and design dimensions of DWTT specimen are shown
steel are shown. in Fig. 3 (API, 2007).
From test data, the average mechanical properties of API X65 steel To prepare test samples, a test plate was first marked and cut from
were extracted and set out in Table 2 together with target values speci- the pipe body as shown in Fig. 4.
fied by API 5L. Second, the test samples were cut from this test plate by water jet
As can be seen from this data, the material tensile properties fulfilled and flattened to remove their initial curvature as recommended by test
the API specifications (yield strength > 448 MPa, tensile strength > 531 standard (API, 2007). It should be noted that all test specimens con-
MPa) for API X65 pipeline steel (API, 2007). tained only base metal and had no seam weld. A Chevron notch then

2
JID: JPSE
ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

E. Fathi and S.H. Hashemi Journal of Pipeline Science and Engineering xxx (xxxx) xxx

Fig. 4. Photograph of original API X65 pipe for specimen preparation.

Fig. 6. Photograph of DWTT machine and experimental set up.

Fig. 5. Photograph of DWTT specimens (ready for experiment).

was introduced to test samples by wire cutting. Although both types

of press notch and Chevron notch are acceptable according to current
DWTT test standards, pipeline industries often use press notch due to
Fig. 7. Schematic of load–displacement data demonstrating the contribution of
its simplicity, low cost and quick specimen preparation. The purpose different energy components, (A) initiation energy, and (B) propagation energy
of this research was to control test conditions in laboratory in the best (Hashemi, 2009).
way. As Chevron notch is placed in test specimen by (CNC) wire cutting,
this process is quite precise and gives similar test specimens from which
consistent results can be obtained. 2) Initiation energy, is that part of total fracture energy which causes
In total, twenty test samples were prepared and labeled into seven crack initiation. Its value is the area under load-displacement record
similar groups from A to G as shown in Fig. 5. from zero load to peak load.
A home-made DWTT machine whose details and instrumenta- 3) Propagation energy, is that part of total fracture energy which cause
tion technique can be found in references (Majidi-Jirandehi and crack growth through specimen ligament. Its value is the area under
Hashemi, 2019, 2017; Fathi-Asgarabad and Hashemi, 2020a,b,c) was load-displacement record from peak load to end load (minimum load
used for experiments, see Fig. 6. at the end of fracture test).
4) Total fracture energy is the area under load-displacement record
from zero load to end load (minimum load at the end of fracture
3.1. Definition of fracture energy components test).

Before measuring different energy components, the following defi- 3.2. Full fracture experiments
nitions are given for each energy term (see Fig. 7):
Full fracture tests (with 2 m drop height equal to 13,734 J energy)
1) Yield energy, is that part of total fracture energy which causes yield- were conducted on three similar DWTT specimens (marked as group G).
ing of test material ahead of the crack tip. Its value is the area under Fig. 8 demonstrates DWTT specimens after fracture test.
load-displacement record (triangle region) from zero load to yield From the instrumented hammer, load history of test specimens was
load. captured using a high-frequency digital oscilloscope. The plot of load

3
JID: JPSE
ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

E. Fathi and S.H. Hashemi Journal of Pipeline Science and Engineering xxx (xxxx) xxx

7000

6500
Total Fracture Energy of Virgin Sample (6187J)
6000

Fr a ct ur e Ene r gy ( J)
y = -0.7581x + 5807.6
5500 R2 = 0.6003
5000

4500

4000 Test Data

3500 Linear Regression

3000
0 500 1000 1500 2000 2500 3000
Fig. 8. Photograph of DWTT specimens after fracture tests. Test Series

250 Fig. 11. Variation of fracture energy versus hammer displacement for all 21
y = -6E-08x6 - 1E-07x5 + 0.0008x4 - 0.0431x3 + 0.3461x2 + 14.362x - 8.9116 test data.
R2 = 0.9952
200
J) and crack initiation energy (2,060 J) from full fracture tests on virgin
specimens, test conditions (including height drop and impact velocity)
Loa d ( k N )

150
for six test groups A to F, were determined to conduct LVI experiments
according to Table 3.
100 As can be seen, drop height was increased from 0.05 m to 0.3 m to
provide 343 J to 2,060 J energy for LVI (near 353 J yield energy and
Load-displacement 2,493 J crack initiation energy of virgin G specimens). For example,
50
all three specimens of group A were hit by releasing the hammer from
0.05 m height so that 343 J energy was introduced to test samples. It
0 was assumed that this energy was absorbed totally by specimen and the
0 10 20 30 40 50 effect of probable rebounding of machine hammer was ignored.
Displacement ( mm)
4. Results and discussion
Fig. 9. Variation of load versus hammer displacement (mean values for group
G test specimens). Using test data, a summary of all test results is given in Table 4.
As it is well known, the geometry of DWTT specimen and its loading
7000 condition was three point bending, so the crack tip location was below
y = -7E-07x6 + 0.0001x5 - 0.0079x4 + 0.0028x3 + 8.9406x2 - 14.401x + 12.393 the neutral axis. Therefore, tensile stresses act on opposite sides of crack
R2 = 1
6000 tip and yielding occurs due to tensile normal loads from dynamic pre-
loading (first LVI test). When the test specimen was subjected to the
5000 second blow in full fracture experiment, yield energy ahead of the crack
Ene r gy ( J)

tip was higher than the yield energy of virgin specimens (353 J) due to
4000
work hardening effect from dynamic pre-loading. As can be seen from
3000 test data in Table 4, around 500% increase of LVI energy from 343 J
(group A) to 2,060 J (group F), caused the increase of yield energy from
2000 Energy-displacement 324 J to 446 J (up to 38%). In the same way, comparison of yield energy
of group F (446 J) with same data of group G (353 J) showed 26%
1000 increase in yield energy.
The variation of fracture energy versus LVI energy is shown in
0
Figs. 11 and 12.
0 10 20 30 40 50 Linear regression of all test data (Fig. 11) and mean test values
Displacement ( mm) (Fig. 12) gave almost similar results. However, R2 (mean squared error)
was closer to unity for the latter one. From this, the following approxi-
Fig. 10. Variation of energy versus hammer displacement (mean values for mating equation was obtained:
group G test specimens).
𝐸f = −0.7585𝐸LVI + 5807.9 (1)

and energy (mean values) versus hammer displacement for Group G where Ef is total fracture energy and ELVI is low velocity impact energy.
specimens are shown in Figs. 9 and 10. Using this equation, the estimated total fracture energy was calcu-
lated for all test specimens as reported in Table 5.
3.3. LVI experiments As can be seen, the maximum error between test and estimated frac-
ture energy was around 6%. Moreover, total 5,808 J fracture energy for
Using the definitions given in Section 3.1, the mean values (for group virgin specimen (group G for which LVI energy was zero) was in good
G test specimens) of yield energy and crack initiation energy (assumed agreement with 6,187 J test result (with only 6.13% error).
at the point of peak load on load-displacement plot) were measured Similar approximating linear equations were obtained by curve fit-
from Fig. 9 as 353 J and 2,493 J, respectively. The mean value of total ting on the mean six test data to estimate the yield energy, initiation
fracture energy for this steel was 6,187 ± 644 J. Using yield energy (343 energy and propagation energy from known LVI as shown in Fig. 13.

4
JID: JPSE
ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

E. Fathi and S.H. Hashemi Journal of Pipeline Science and Engineering xxx (xxxx) xxx

Table 3
Test conditions and experimental results for API X65 steel.

Mean
Specimen Group Drop height Height ratio Velocity LVI energy Fracture fracture
No. symbol (m) (%) V (m/s) ratio (%) (J) energy (J) energy (J) SD (J)

1 0.05 2.5 0.99 16 343 5,590 5,608 ±355

2 A 5,183
3 6,051

4 0.1 5 1.4 22 686 5,306 5,260 ±178

5 B 5,023
6 5,451

7 0.15 7.5 1.71 27 1030 5,167 5,022 ±399

8 C 4,478
9 5,422

10 D 0.2 10 1.98 32 1373 5,107 4,699 ±297

11 4,407
12 4,582

13 E 0.25 12.5 2.21 35 1716 5,069 4,465 ±439

14 4,283
15 4,042

16 F 0.3 15 2.42 39 2060 4,510 4,328 ±418

17 3,749
18 4,724

19 G 2 100 6.26 100 – 7,085 6,187 ±644

20 5,605
21 5,872

Table 4
Mean experimental results from DWT testing on seven test groups.

Group symbol LVI energy (J) Yield point energy (J) Initiation energy (J) Propagation energy (J) Total fracture energy (J)

A 343 324 2,298 3,310 5,608

B 686 365 1,949 3,311 5,260
C 1,030 426 1,978 3,044 5,022
D 1,373 429 1,597 3,102 4,699
E 1,716 368 1,458 3,006 4,464
F 2,060 446 1,300 3,027 4,327
G – 353 2,493 3,694 6,187

Table 5
Comparison of test and estimated fracture energy for seven test groups.

Group symbol LVI energy (J) Test fracture energy (J) Estimated fracture energy (J) from Eq.1. Energy difference (J) Error (%)

A 343 5,608 5,740 +132 +2.35

B 686 5,260 5,406 +146 +2.78
C 1,030 5,022 5,072 +50 +1.00
D 1,373 4,699 4,739 +40 +0.85
E 1,716 4,464 4,405 −59 −1.32
F 2,060 4,327 4,071 −256 −5.92
G 0 6,187 5,808 −379 −6.13

7000 The approximating linear equations are given below:

6500 Test Fracture Energy (6187J) 𝐸y = 0.0518𝐸LVI + 330.8 (2)

6000
Fr a ct ur e Ene r gy ( J)

5500 y = -0.7585x + 5807.9

R2 = 0.9855 𝐸i = −0.5694ELVI + 2447.4 (3)
5000

4500

4000 Test Data 𝐸p = −0.1891𝐸LVI + 3360.5 (4)

3500 Linear Regression
where Ey , Ei , and Ep are yield energy, initiation energy and propagation
3000 energy, respectively.
0 500 1000 1500 2000 2500 3000 Using Eqs. (1) to (4), variation of all energy components versus LVI
Low Velocity I mpact Energy ( J) energy was estimated from the point of yield load to peak load as can
be seen in Tables 6–9.
Fig. 12. Variation of fracture energy versus hammer displacement for mean 7 As can be seen in Fig. 14, LVI energy can be partitioned in two parts.
test data. The first part was assumed to be used for yielding of the region ahead of

5
JID: JPSE
ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

E. Fathi and S.H. Hashemi Journal of Pipeline Science and Engineering xxx (xxxx) xxx

Table 7
7000 Estimation of initiation energy for API X65 steel using Eq.3.
Total Fracture Energy (6187J)
6000 LVI energy (J) Estimated Decrease of Ratio of initiation
Yield Energy initiation initiation energy to initiation
Initiation Energy
energy (J) energy (J) energy of virgin
5000 Propagation Energy
Total Fracture Energy from Eq.3. sample (2,493 J)
y = -0.7578x + 5807.4
Ene r gy ( J)

4000 R2 = 0.9854 353 2,246 247 0.90

y = -0.1891x + 3360.5 400 2,220 273 0.89
3000 R2 = 0.7431
500 2,163 330 0.87
600 2,106 387 0.84
y = -0.5694x + 2447.4 700 2,049 444 0.82
2000 R2 = 0.9531 800 1,992 501 0.80
900 1,935 558 0.78
1000 y = 0.0518x + 330.8 1,000 1,878 615 0.75
R2 = 0.4867 1,100 1,821 672 0.73
0 1,200 1,764 729 0.71
1,300 1,707 786 0.68
0 1000 2000 3000
1,400 1,650 843 0.66
Low Velocity I mpact Enerngy ( J) 1,500 1,593 900 0.64
1,600 1,536 957 0.62
Fig. 13. Variation of energy components versus LVI energy. 1,700 1,479 1,014 0.59
1,800 1,422 1,071 0.57
1,900 1,366 1,127 0.55
Table 6 2,000 1,309 1,184 0.52
Estimation of yield energy for API X65 steel using Eq.2. 2,100 1,252 1,241 0.50
2,200 1,195 1,298 0.48
Estimated Ratio of yield 2,300 1,138 1,355 0.46
yield point energy to yield 2,400 1,081 1,412 0.43
energy (J) Dissipated Increase of energy of virgin 2,493 1,028 1,465 0.41
LVI energy (J) from Eq.2. energy (J) yield energy (J) sample (353 J) 2,500 1,024 1,469 0.41

353 349 4 – 0.99

400 352 48 – 1.00
500 357 143 4 1.01 Table 8
600 362 238 9 1.03
Estimation of propagation energy for API X65 steel using Eq.4.
700 367 333 14 1.04
800 372 428 19 1.05 Ratio of propagation
900 377 523 24 1.07 Estimated energy to
1,000 383 617 30 1.08 propagation Decrease of propagation energy
1,100 388 712 35 1.10 energy (J) from propagation of virgin sample
1,200 393 807 40 1.11 LVI energy (J) Eq.4. energy (J) (3,694 J)
1,300 398 902 45 1.13
1,400 403 997 50 1.14 353 3,294 400 0.89
1,500 409 1,092 56 1.16 400 3,285 409 0.89
1,600 414 1,186 61 1.17 500 3,266 428 0.88
1,700 419 1,281 66 1.19 600 3,247 447 0.88
1,800 424 1,376 71 1.20 700 3,228 466 0.87
1,900 429 1,471 76 1.22 800 3,209 485 0.87
2,000 434 1,566 81 1.23 900 3,190 504 0.86
2,100 440 1,660 87 1.25 1000 3,171 523 0.86
2,200 445 1,755 92 1.26 1,100 3,152 542 0.85
2,300 450 1,850 97 1.27 1,200 3,134 560 0.85
2,400 455 1,945 102 1.29 1,300 3,115 579 0.84
2,493 460 2,033 107 1.30 1,400 3,096 598 0.84
2,500 460 2,040 107 1.30 1,500 3,077 617 0.83
1,600 3,058 636 0.83
1,700 3,039 655 0.82
1,800 3,020 674 0.82
2400 1,900 3,001 693 0.81
LVI Energy 2,000 2,982 712 0.81
2000 LVI Energy = 2,100 2,963 731 0.80
Yield Energy 2,200 2,944 750 0.80
+ Dissipated Energy 2,300 2,926 768 0.79
1600
2,400 2,907 787 0.79
Ene r gy ( J)

Dissipated 2,493 2,889 805 0.78

1200 Energy 2,500 2,888 806 0.78

800
Yield Energy
400 the crack tip whereas the remaining part (dissipated energy) was used
for crack tip blunting.
0
In the first and second regions in Fig. 15, the remaining part (dissi-
400 800 1200 1600 2000 2400
pated energy) of input LVI energy was less than crack initiation energy
Low Velocity I mpact Energy ( J)
(the straight line of dissipated energy was below the line of crack initia-
tion energy). In the third region, the remaining part (dissipated energy)
Fig. 14. Partitioning of LVI energy into yield energy and remaining (dissipated)
energy. of input LVI energy was more than crack initiation energy (the straight
line of dissipated energy was above the line of crack initiation energy),

6
JID: JPSE
ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

E. Fathi and S.H. Hashemi Journal of Pipeline Science and Engineering xxx (xxxx) xxx

Table 9 2400
Estimation of total fracture energy for API X65 steel using Eq.1. Initiation Dissipated
2000 Energy Energy
Ratio of fracture
energy to
Estimated total Decrease of total fracture energy 1600

Ene r gy ( J)
fracture energy fracture energy of virgin sample
LVI energy (J) (J) from Eq.1. (J) (6,187 J) 1200
Yield
353 5,540 647 0.9 Point
400 5,504 683 0.89
Crack
800
Initiation
500 5,429 759 0.88
Point
600 5,353 834 0.87
400
700 5,277 910 0.85 Yield Energy
800 5,201 986 0.84
900 5,125 1,062 0.83 0
1,000 5,050 1,137 0.82 400 800 1200 1600 2000 2400
1,100 4,974 1,213 0.80 Low Velocity I mpact Energy ( J)
1,200 4,898 1,289 0.79
1,300 4,822 1,365 0.78
Fig. 16. Variation of yield energy, remaining (dissipated) energy and crack ini-
1,400 4,746 1,441 0.77
tiation energy versus LVI energy.
1,500 4,671 1,516 0.75
1,600 4,595 1,592 0.74
1,700 4,519 1,668 0.73
1,800 4,443 1,744 0.72
1,900 4,368 1,819 0.71 5. Conclusions
2,000 4,292 1,895 0.69
2,100 4,216 1,971 0.68 In this paper, DWTTs were used to study the effects of dynamic pre-
2,200 4,140 2,047 0.67 loading on fracture behavior of API X65 gas pipeline steel. In doing
2,300 4,064 2,123 0.66
so, seven test groups consisting of three similar test specimens were
2,400 3,989 2,198 0.64
2,493 3,918 0.63 2,269 used. First, the LVI tests were carried out on six test group by releasing
2,500 3,913 2,274 0.63 the machine heavy hammer from different heights (ranged from 0.05
m to 0.3 m). Then, full fracture DWTTs were conducted on all seven
test groups at 2 m drop height. The fracture energy of tested specimens
and its different components (including yield energy, initiation energy
2400
and propagation energy) were calculated from test data using an instru-
Initiation Dissipated mented hammer. These were compared to similar test data of virgin test
2000 Energy Energy specimens. The obtained results are summarised below:

1600 (1) The virgin DWTT specimens (group G) had mean yield energy, crack
Ene r gy ( J)

initiation energy, crack propagation energy and total fracture energy

1200 of 353 J, 2,493 J, 3,694 J and 6,187 J, respectively.
1st 2nd 3rd
region region region (2) When DWTT specimens (six groups from A to F) were subjected to
800 dynamic pre-loading ranged from 343 J to 2,060 J (associated with
Yield Energy 97%–484% of yield energy of virgin specimens), it was observed
400 that the yield energy slightly increased from 353 J (group G) to 446
J (group F) corresponding to 26% increase in yield energy. This was
0
the effect of work hardening from LVI tests.
400 800 1200 1600 2000 2400
(3) Dynamic pre-loading had the most decreasing effect on crack initia-
Low Velocity I mpact Energy ( J)
tion energy. The initiation energy significantly decreased from 2,493
J (group G) to 1,300 J (group F) corresponding to 48% drop in ini-
Fig. 15. Demonstration of three separate regions for determination of yield
point and point of crack initiation. tiation energy. This was the effect of plastic deformation and crack
tip blunting from LVI tests.
(4) The propagation energy slightly decreased from 3,694 J (group G) to
3,027 J (group F) corresponding to 18% drop in propagation energy.
(5) The total fracture energy (area under load-displacement record) de-
so it can be assumed that part of remaining energy was used for crack creased from 6,187 J (group G) to 4,327 J (group F) corresponding
propagation through specimen ligament. to 31% drop in fracture energy.
From Fig. 16, the exact yield point can be determined (intersection of (6) An approximating linear equation was obtained by curve fitting on
two lines i.e. yield energy-LVI energy and dissipated energy-LVI energy). the mean six test data in the form of Ef = −0.7585ELVI +5807.9 from
Similarly, the exact crack initiation point can be obtained (intersection which fracture energy of API X65 steel can be derived from DWTT
of two lines, i.e. dissipated energy-LVI energy and initiation energy-LVI specimens in terms of known LVI energy. The maximum error of this
energy). equation was around 6% for tested specimens.
Using this technique, the exact yield point was found at 369 J which (7) By extrapolation of this equation, 5,808 J fracture energy was ob-
was slightly higher (4.5%) than mean yield energy point of group G tained for virgin specimens (group G with ELVI = 0) whose error was
(353 J). In the same way, the exact crack initiation point was found at only 6.13% as compared to 6,187 J test fracture energy.
1,405 J which was significantly lower (43.6%) than mean crack initi- (8) Similar approximating linear equations were obtained by curve fit-
ation energy point of group G (2,493 J). This indicted that crack initi- ting on the mean six test data to estimate the yield energy, initiation
ation occurred somewhere between yield point (353 J) and the point energy and propagation energy from known LVI tests.
of peak load (2,493 J) on load-displacement plot as reported elsewhere (9) A novel test technique was presented for exact determination of yield
(Kobayashi et al., 1993; Tanguy et al., 2005). and crack initiation points. From this, the exact yield point was found

7
 JID: JPSE
ARTICLE IN PRESS [m5GeSdc;June 9, 2021;17:49]

E. Fathi and S.H. Hashemi Journal of Pipeline Science and Engineering xxx (xxxx) xxx

at 369 J which was slightly higher (4.5%) than mean yield energy Horsley, D.J., 2003. Background to the use of CTOA for prediction of dynamic ductile
point of group G (353 J). fracture arrest in pipelines. Eng. Fract. Mech. 70, 547–552.
Hwang, B., Lee, S., Kima, Y.M., Kima, N.J., Yoo, J.Y., Woo, C.S., 2004. Analysis of abnor-
(10) Similarly, the exact crack initiation point was found at 1,405 J which mal fracture occurring during drop-weight tear test of high toughness line-pipe steel.
was significantly lower (43.6%) than mean crack initiation energy Mater. Sci. Eng. A 368 (1-2), 18–27.
point of group G (2,493 J). Khosravi Khor, H., Hashemi, S.H., Raghebi, M., 2020. Experimental study of natural fre-
quencies of notched homogeneous and inhomogeneous specimens made of API X65
steel in low-blow drop weight test. Modares Mech. Eng.
Declaration of Competing Interest Kim, J.S., Kim, Y.J., Lee, M.W., Kim, K.S., Shibanuma, K., 2021. Finite element simulation
of drop-weight tear test of API X80 at ductile-brittle transition temperatures. Int. J.
Mech. Sci. 191, 106103.
The authors declare that they have no known competing financial
Kobayashi, T., Yamamoto, I., Niinomi, M., 1993. Introduction to a new dynamic fracture
interests or personal relationships that could have appeared to influence toughness evaluation system. J. Test. Eval. 21 (3), 145–153.
the work reported in this paper. Leis, B.N., Eiber, R.J., Carlson, L., Gilroy-Scott, A., 1998. Relationship between apparent
(total) Charpy Vee-notch toughness and the corresponding dynamic crack-propaga-
tion resistance. In: Proceedings of the 3rd International Pipeline Conference. ASME,
Acknowledgement pp. 723–731.
Majidi-Jirandehi, A.A., Hashemi, S.H., 2017. Investigation of macroscopic fracture surface
The supply of API X65 pipeline steel by Sadid Pipe and Equipment characteristics of spiral welded API X65 gas transportation pipeline steel. Modares
Mech. Eng. 17 (11), 219–228.
Company (IRAN) and financial support by University of Birjand are Majidi-Jirandehi, A.A., Hashemi, S.H., 2019. Weld metal fracture characteriza-
greatly acknowledged. The authors are grateful to National Iranian Gas tion of API X65 steel using drop weight tear test. Mater. Res. Express 6.
Company (NIGC) and South Khorasan Gas Company (SKGC) for their doi:10.1088/2053-1591/aae797.
Maxey, W.A., 1974. Fracture initiation, propagation and arrest. In: Proceedings of the 5th
support in manufacturing of DWTT machine. Symposium on Line Pipe Research, Houston, USA, pp. 20–22.
Newman, J.C., James, M.A., Zerbst, U.A., 2003. A review of the CTOA/CTOD fracture
References criterion. Eng. Fract. Mech. 70, 371–385.
O’Donoghue, P.E., Kanninen, M.F., Leung, C.P., Demofonti, G., Venzi, S., 1997. The devel-
API, 1996. Recommended practice for conducting Drop-Weight Tear Test on line pipe, API opment and validation of a dynamic fracture propagation model for gas transmission
RP 5L3. American Petroleum Institute, Washington DC. pipelines. Int. J. Pres. Ves. Pip. 70, 11–25.
API, 2007. Specification for line pipe, API Specification 5L/ISO 3183 (Modified), 44th ed. Rudland, D.L., Wilkowski, G.M., Feng, Z., Wang, Y.-Y, Horsley, D., Golver, A., 2003. Exper-
American Petroleum Institute, Washington DC. imental investigation of CTOA in linepipe steels. Eng. Fract. Mech. 70 (3–4), 567–577.
ASTM, 2014. Standard test method for Drop-Weight Tear Tests of ferritic steels, ASTM E Salvini, P., Fonzo, A., Mannucci, G., 2003. Identification of CTOA and fracture process
436-03. American Society of Testing Materials. parameters by drop weight test and finite element simulation. Eng. Fract. Mech. 70,
ASTM, 2016. Standard test method for determination of crack-tip-opening- angle of pipe 553–566.
steels using DWTT specimens, ASTM E3039-16. American Society of Testing Materi- Strnadel, B., Ferfecki, P., Židlík, P., 2013. Statistical characteristics of fracture surfaces in
als. high-strength steel drop weight tear test specimens. Eng. Fract. Mech. 112–113, 1–13.
Cosham, A., Jones, D.G., Eiber, R., Hopkins, P., 2010. Don’t drop the drop weight tear Tanguy, B., Besson, J., Piques, R., Pineau, A., 2005. Ductile to brittle transition of an
test. Pipeline Eng. 9 (2), 69–84. A508 steel characterized by Charpy impact test – part I: experimental results. Eng.
BSI, 1999. Metallic materials, Drop Weight Tear Test, BS EN 10274: 1999. Frac. Mech. 72, 49–72.
Cotterell, B., 2002. The past, present, and future of fracture mechanics. Eng. Fract. Mech. Tazimi, M.M., Hashemi, S.H., Rahnama, S., 2020a. Experimental study of fracture surface
69, 533–553. thickness variations of inhomogeneous drop weight tear test specimen (with hori-
Demofonti, G., Buzzichelli, G., Venzi, S., Kanninen, M., 1995. Step by step procedure for zontal weld seam) made from API X65 steel. Shahrood J. Solid Fluid Mech. 20-10,
the two specimen CTOA test. In: Proc. 2nd International Pipeline Technology Con- 2583–2592 Accepted for publication.
ference, Pipeline Technology. Elsevier Science, Ostend, Belgium, pp. 503–512 Sept. Tazimi, M.M., Hashemi, S.H., Rahnama, S., 2020b. Experimental study of fracture surface
11-14. characteristics of inhomogeneous drop weight tear test specimen made from API X65
Eiber, R.J., Bubenik, T.A., Maxey, W.A., 1993. Fracture control technology for natural gas steel. Sharood J. Solids Fluids Mech. 10-1, 77–91.
pipelines. PRC of AGA NG18 -Report #208. Xu, S., Bouchard, R., Tyson, W.R., 2007. Simplified single-specimen method for evaluating
Fang, J., Zhang, J.W., Wang, L., 2014. An energy based regression method to evaluate CTOA. Eng. Fract. Mech. 74, 2459–2464.
critical CTOA of pipeline steels by instrumented drop weight tear tests. Int. J. Fract. Xu, S., Tyson, W.R., Eagleson, R., Mccowan, C.N., Drexler, E.S., Mccolskey, J.D., et al.,
187, 123–131. 2010. Measurement of CTOA of pipe steels using MDCB and DWTT specimens. In:
Fang, J., Zhang, J., Wang, L., 2014. Evaluation of cracking behavior and critical CTOA Proceedings of the 8th International Pipeline Conference (IPC 2010). American Soci-
values of pipeline steel from DWTT specimens. Eng. Fract. Mech. 124–125, 18–29. ety of Mechanical Engineers, pp. 269–278.
Fathi-Asgarabad, E., Hashemi, S.H., 2020a. Experimental study of low velocity impact Xu, S., Tyson, W.R., Simha, C.H.M., Gesing, M., Liang, J., 2016. CTOA test method sing
effect on fracture energy of API X65 steel using drop weight tear test. Shahrood J. drop-weight tear test (DWTT): background and development of standard. In: Proceed-
Solid Fluid Mech. Accepted for publication. ings of the 2016 11th International Pipeline Conference IPC2016, September 26-30.
Fathi-Asgarabad, E., Hashemi, S.H., 2020c. Experimental measurement and numerical Calgary, Alberta, Canada IPC2016-64058.
evaluation of fracture energy in drop weight tear test specimen with Chevron notch Yu, P.S., Ru, C.Q., 2016. Analysis of energy absorptions in drop-weight tear tests of
in API X65 steel. Modares Mech. Eng. 20-5, 1145–1156. pipeline steel of pipeline steel. Eng. Fract. Mech. 160, 138–146.
Fathi-Asgarabad1, E., Hashemi, S.H., 2020b. Experimental and numerical study of energy Zhen, Y., Li, X., Caoa, Y., Zhang, S., 2020. A novel method to determine critical CTOA
absorption in drop weight tear test specimen with Chevron notch on API X65 steel. directly by load displacement curve. Eng. Fract. Mech. 230, 107013.
Modares Mech. Eng. 9. doi:10.22044/jsfm.2020.9440.3129. Zheng, Y., Chang-Boo, K., Yaorong, F., Chongdu, C., 2008. Abnormal fracture appear-
Hashemi, S.H., 2009. Correction factors for safe performance of API X65 pipeline steel. ance in drop-weight tear test specimens of pipeline steel. Mater. Sci. Eng. A 83–484,
Int. J. Pres. Ves. Pip. 86, 533–540. 239–241.
Hashemi, S.H., 2011. Strength-hardness statistical correlation in API X65 steel. Mater. Sci. Zhu, X.K, Leis, B.N., 2006. Application of constraint-corrected J–R curve to fracture anal-
Eng. A. 528-3, 1648–1655. ysis of pipelines. J. Press. Ves. Technol. 128, 581–589.
Hashemi, S.H., Mohammadyani, D., 2012. Characterisation of weldment hardness, impact Zhu, X.-K., Joyce, J.A., 2012. Review of fracture toughness (G, K, J, CTOD, CTOA) testing
energy and microstructure in API X65 steel. Int. J. Pres. Ves. Pip. 98, 8–15. and standardization. Eng. Fract. Mech. 85, 1–46.
Hong, S., Shin, S.Y., Lee, S., Kim, N.J., 2011. Effects of specimen thickness and notch
shape on fracture modes in the drop weight tear test of API X70 and X80 linepipe
steels. Metall. Mater. Trans. A 42, 2619–2632.